THE CAMBRIDGE
HISTORY OF
JAPAN
Volume 2
Heian Japan
Edited by
DONALD H. SHIVELY
and
WILLIAM H. McCULLOUGH
I CAMBRIDGE
UNIVERSITY PRESS
Cambridge Histories Online © Cambridge University Press, 2008
PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
CAMBRIDGE UNIVERSITY PRESS
The Edinburgh Building, Cambridge CB2 2RU, UK
http: //www.cup.cam.ac.uk
40 West 20th Street, New York, NY 10011-4211, USA http: //www.cup.org
10 Stamford Road, Oakleigh, Melbourne 3166, Australia
© Cambridge University Press 1999
This book is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press
First published 1999
Printed in the United States of America
Typeface P\an\in 11/13 pt.
System Quark [HVG]
A catalog record for this book is available from
the British Library.
Library of Congress Cataloging-in-Publication Data
(Revised for volume 2)
The Cambridge history of Japan.
Includes bibliographical references and index.
Contents: v. 1. Ancient Japan / edited by Delmer M.
Brown - v. 2. Heian Japan / edited by Donald H. Shively and
William H. McCullough - v. 3. Medieval Japan / edited by Kozo Yamamura v. 4. Early modern Japan / edited by John Whitney Hall - [etc.]
1. Japan - History. I. Hall, John Whitney, 1916-1997
952
88-2877
DS835.C36 1998
ISBN 0-521-22353-9 (v. 2) hardback
0-521-22352-0 (v. 1) hardback
0-521-22354-7 (v. 3) hardback
0-521-22355-5 (v. 4) hardback
0-521-22356-3 (v. 5) hardback
0-521-22357-1 (v. 6) hardback
0-521-65728-8 hardback set
Cambridge Histories Online © Cambridge University Press, 2008
GENERAL EDITORS' PREFACE
Since the beginning of this century the Cambridge histories have set
a pattern in the English-reading world for multivolume series containing chapters written by specialists under the guidance of volume
editors. Plans for a Cambridge history of Japan were begun in the
1970s and completed in 1978. The task was not to be easy. The details of Japanese history are not matters of common knowledge
among Western historians. The cultural mode of Japan differs greatly
from that of the West, and above all there are the daunting problems
of terminology and language. In compensation, however, foreign
scholars have been assisted by the remarkable achievements of the
Japanese scholars during the last century in recasting their history in
modern conceptual and methodological terms.
History has played a major role in Japanese culture and thought,
and the Japanese record is long and full. Japan's rulers from ancient
times have found legitimacy in tradition, both mythic and historic,
and Japan's thinkers have probed for a national morality and system
of values in their country's past. The importance of history was also
emphasized in the continental cultural influences that entered Japan
from early times. Its expression changed as the Japanese consciousness turned to questions of dynastic origin, as it came to reflect Buddhist views of time and reality, and as it sought justification for rule
by the samurai estate. By the eighteenth century the successive need
to explain the divinity of the government, justify the ruler's place
through his virtue and compassion, and interpret the flux of political change had resulted in the fashioning of a highly subjective fusion of Shinto, Buddhist, and Confucian norms.
In the nineteenth century the Japanese became familiar with Western forms of historical expression and felt the need to fit their national history into patterns of a larger world history. As the modern
Japanese state took its place among other nations, Japanese history
faced the task of reconciling a parochial past with a more catholic
present. Historians familiarized themselves with European accounts
Cambridge Histories Online © Cambridge University Press, 2008
VI
GENERAL EDITORS
PREFACE
of the course of civilization and described Japan's nineteenth-century
turn from military to civilian bureaucratic rule under monarchical
guidance as part of a larger, worldwide pattern. Buckle, Guizot,
Spencer, and then Marx successively provided interpretative schema.
The twentieth-century ideology of the imperial nation-state, however, operated to inhibit full play of universalism in historical interpretation. The growth and ideology of the imperial realm required
caution on the part of historians, particularly with reference to Japanese origins.
Japan's defeat in World War II brought release from these inhibitions and for a time replaced them with compulsive denunciation of
the pretensions of the imperial state. Soon the expansion of higher
education brought changes in the size and variety of the Japanese
scholarly world. Historical inquiry was now free to range widely. A
new opening to the West brought lively interest in historical expressions in the West, and a historical profession that had become cautiously and expertly positivist began to rethink its material in terms
of larger patterns.
At just this juncture the serious study of Japanese history began in
the West. Before World War II the only distinguished general survey
of Japanese history in English was G. B. Sansom's Japan: A Short
Cultural History, first published in 1931 and still in print. English and
American students of Japan, many trained in wartime language programs, were soon able to travel to Japan for study and participation
with Japanese scholars in cooperative projects. International conferences and symposia produced volumes of essays that served as
benchmarks of intellectual focus and technical advance. Within
Japan itself an outpouring of historical scholarship, popular publishing, and historical romance heightened the historical consciousness
of a nation aware of the dramatic changes to which it was witness.
In 1978 plans were adopted to produce this series on Japanese history as a way of taking stock of what has been learned. The present
generation of Western historians can draw upon the solid foundations of the modern Japanese historical profession. The decision to
limit the enterprise to six volumes meant that topics such as the history of art and literature, aspects of economics and technology and
science, and the riches of local history would have to be left out.
They too have been the beneficiaries of vigorous study and publication in Japan and in the Western world.
Multivolume series have appeared many times in Japanese since
the beginning of the century, but until the 1960s the number of proCambridge Histories Online © Cambridge University Press, 2008
GENERAL EDITORS
PREFACE
Vll
fessionally trained historians of Japan in the Western world was too
small to sustain such an enterprise. Although that number has
grown, the general editors have thought it best to draw on Japanese
specialists for contributions in areas where they retain a clear authority. In such cases the act of translation itself involves a form of
editorial cooperation that requires the skills of a trained historian
whose name deserves acknowledgment.
The primary objective of the present series is to put before the
English-reading audience as complete a record of Japanese history
as possible. But the Japanese case attracts our attention for other
reasons as well. To some it has seemed that the more we have come
to know about Japan, the more we are drawn to the apparent similarities with Western history. The long continuous course of Japan's
historical record has tempted historians to look for resemblances between its patterns of political and social organization and those of
the West. The rapid emergence of Japan's modern nation-state has
occupied the attention of comparative historians, both Japanese and
Western. On the other hand, specialists are inclined to point out the
dangers of being misled by seeming parallels.
The striking advances in our knowledge of Japan's past will continue and accelerate. Western historians of this great and complex
subject will continue to grapple with it, and they must as Japan's
world role becomes more prominent. The need for greater and
deeper understanding of Japan will continue to be evident. Japanese
history belongs to the world, not only as a right and necessity but
also as a subject of compelling interest.
JOHN WHITNEY HALL
MARIUS B. JANSEN
MADOKA KANAI
DENIS TWITCHETT
Cambridge Histories Online © Cambridge University Press, 2008
This is the final volume of The Cambridge History of Japan,
of which the first to be published appeared in 1988. Professor JohnW. Hall, A.Whitney Griswold Professor Emeritus of
Yale University, died in October 1997 and, sadly, was unable
to see the completion of this project. As one of the general
editors and as editor of Volume 4, Early Modern Japan, he
played a central role in shaping and executing every facet of
this undertaking, and his loss is mourned by all historians of
Japan.
MARIUS B. JANSEN
MADOKA KANAI
DENIS TWITCHETT
Cambridge Histories Online © Cambridge University Press, 2008
CONTENTS
v
xiii
xv
xviii
General editors'preface
List of maps, figures, and tables
Preface to Volume 2
Chronology
page
1
Introduction
by D O N A L D H . SHIVELY and W I L L I A M
H.
Department of East Asian Languages,
University of California, Berkeley
MCCULLOUGH,
The Heian court, 794-1070
by WILLIAM H. MCCULLOUGH, Department of East
Asian Languages, University of California, Berkeley
20
Kammu to Nimmyo, 781-850
Evolution of the statutory government
The establishment of Fujiwara ascendancy, 850-969
The Fujiwara regency, 970-1070
Regency government
Foreign relations, 794-1070
20
37
45
64
74
80
The capital and its society
by WILLIAM H. MCCULLOUGH, Department of East
Asian Languages, University of California, Berkeley
97
Site of the new capital
Plan of the city
Greater Imperial Palace
Emperor's Residential Compound
Other public buildings and spaces
Residential districts and population
Imperial clan and court nobility
Cambridge Histories Online © Cambridge University Press, 2008
97
102
108
113
116
119
123
CONTENTS
The noble family, marriage, and the position of women
Life in the mansion of a great noble
Officialdom and its functions
The city's economy
City administration
Changes in the city plan
. New imperial and Fujiwara buildings
Ceremony and ritual
Land and society
by D A N A M O R R I S , Department of History, University of
California, Berkeley
Agrarian technology
Peasant community
Tax structure
Landholding
Shoen
Provincial administration and land tenure in early Heian
by CORNELIUS J. KILEY, Department of History,
Villanova University
134
142
159
161
170
172
173
180
183
184
194
199
215
224
236
Regional administration
The establishment of custodial governorship
Land and taxes
The surrender of central control to provincial
authorities
Discretionary taxation and elite wealth
Local elites as a political force
254
265
272
Chinese learning and intellectual life
by M A R I A N U R Y , Comparative Literature Program,
University of California, Davis
341
Introduction and assimilation of Chinese learning
Ideal of the sage-king
Six National Histories
Compilation of statutes
State Academy
Scholars and their accomplishments
Cambridge Histories Online © Cambridge University Press, 2008
283
298
326
341
355
359
364
367
375
CONTENTS
6 Aristocratic culture
by HELEN CRAIG MCCULLOUGH, Department of East
XI
390
Asian Languages, University of California, Berkeley
Domestic architecture and furnishings
Textiles and costumes
Diet
Aristocratic occupations and pastimes
Secular painting
Calligraphy and paper
Buddhist art
Music
Literature: Poetry
Literature: Narrative prose
390
394
398
400
409
415
418
424
431
441
7 Aristocratic Buddhism
by STANLEY WEINSTEIN, Department of East Asian
Languages and Literatures, Yale University
The prelude to Heian Buddhism
The assertion of government control over the
Buddhist church
Saicho
Kukai
The Tendai school after Saicho
The Shingon school after Kukai
The growth of Pure Land Buddhism
449
454
462
473
478
497
507
8 Religious practices
517
449
by ALLAN G. GRAPARD, Department of Religious Studies,
University of California, Santa Barbara
The association of Shinto shrines with Buddhist temples
Ritualized and ritualizing activities
Dealing with the forces of nature
The association of kami with buddhas
Late Heian developments
9 Insei
by G.
520
532
547
564
572
576
CAMERON HURST,
III, Department of Asian and
Middle Eastern Studies, University of Pennsylvania
Abdication, regency, and the Japanese throne,
645-1068
Cambridge Histories Online © Cambridge University Press, 2008
576
Xll
CONTENTS
Go-Sanjo and the prelude to insei, 1068-1073
Shirakawa and the normalization of insei, 1073-1129
The hegemony ofToba, 1129-1156
Go-Shirakawa and theTaira, 1156-1185
Foreign relations, 1070-1185
The insei in retrospect
10 The rise of the warriors
595
608
618
632
637
644
by T A K E U C H I R I Z O , Faculty of Literature, Waseda University
Origins of the warriors
Revolts of Masakado and Sumitomo
Revolt of Tadatsune
Earlier Nine Year s' War
Later Three Years'War
Conditions in the capital
Hogen Disturbance
Heiji Disturbance
Taira rise to power
Gempei War
Works cited
Glossary-index
Cambridge Histories Online © Cambridge University Press, 2008
644
653
664
670
675
679
688
691
695
700
711
741
MAPS, FIGURES, AND TABLES
MAPS
Japan in the Heian period
10.1 Battle sites in the northeast
10.2 The GempeiWar
page xxiv
672
701
FIGURES
1.1
1.2
1.3
2.1
2.2
2.3
2.4
2.5
2.6
3.1
3.2
3.3
3.4
3.5
3.6
9.1
9.2
9.3
9.4
Genealogy of Heian emperors
Genealogy of the four Fujiwara houses
Genealogy of the Northern House of the Fujiwara
The Heian capital (Heian-kyd)
Greater Imperial Palace (Daidairi)
Major Heian governmental organs
The Emperor's Residential Compound (Dairi)
Plan of Ononomiya
Plan of Hoj5ji
Eighth- and ninth-century house sites at the
Hiraide site
Houses in Wakatsuki-no-sho
Houses in Kohigashi-no-sho
Tax structure in the late seventh and eighth centuries
Tax structure in the ninth century
Tax structure in the tenth century
Genealogy of emperors during the Insei period
Genealogy of the Murakami Genji (Minamoto)
Family relations of Go-Sanjo
Family relations of Shirakawa
Cambridge Histories Online © Cambridge University Press, 2008
22
27
46
104
110
112
114
145
178
196
196
197
210
211
211
584
587
594
597
XIV
9.5
9.6
9.7
10.1
10.2
MAPS, FIGURES, AND TABLES
Structure of the in-no-cho
Family relations of Toba
Genealogy of the Kammu Heishi (Taira)
Genealogy of the Seiwa Genji
The Kiyohara and Oshvi Fujiwara
605
609
615
651
676
TABLES
1.1
3.1
3.2
4.1
4.2
4.3
8.1
9.1
Heian emperors
Distribution of iron tools in farm households
Land tax rates in the late tenth century
Provincial officials mandated by the ritsuryo
Gun officials mandated by the ritsuryo
Stipend grants from stored rice in Izumo
The Twenty-two Shrine-temples sponsored by the
imperial government
Reigning and retired emperors in the Inset period
Cambridge Histories Online © Cambridge University Press, 2008
52
190
213
256
258
315
527
610
PREFACE TO VOLUME 2
Heian (794-1185) is regarded as Japan's classical age. The imperial
court was at its height as a political power and patron of aristocratic
culture in its most brilliant time. The Heian period has received special attention from Japanese historians through the centuries, as
might be expected, and became an important subject of modern
scholarship following the restoration of the imperial government in
1868. Japanese historians have been thorough and tireless in their investigations of the era. All of the primary materials known to have
survived from Heian times have been published in modern editions.
Japanese scholars have shared their erudition in a daunting wealth of
detailed monographs and articles as well as interpretive studies. The
chapters of this volume, in their content and notes, give evidence of
our debt to them. None contributed more to research on Heian history than the late Professor Takeuchi Rizo, who wrote a chapter for
this volume.
In this volume, Japanese is romanized according to the Modified
Hepburn system, and Chinese according to Wade-Giles. Japanese
and Chinese personal names follow their native form, with family or
clan name preceding given name, except in citations of Japanese authors writing in English. Characters for Japanese and Chinese names
and terms appear in the Glossary-Index. References cited in the
footnotes are listed in alphabetical order by author in the list of
Works Cited.
In footnotes Japanese dates are abbreviated as, for example, Jowa
9 (842) 3/6, meaning the ninth year of the Jowa era (dated to 842 in
the Western calendar), the sixth day of the third lunar month. Years
of the Heian lunar calendar and the Julian calendar do not correspond exactly. When the date of an event occurring late in the lunarcalendar year is known to fall at the beginning of the next year in the
Julian calendar, conversion is made to the next year, following the
practice of the Kodansha Encyclopedia of Japan, 9 vols. (Tokyo: K6dansha, 1983). When a person's age is given, it is expressed accord-
Cambridge Histories Online © Cambridge University Press, 2008
XVI
PREFACE TO VOLUME 2
ing to the Western method of counting full years, rather than the
Japanese practice of counting the calendar years in which the individual lived.
The Japanese sovereign is usually referred to as "emperor," the
conventional translation of tennd, his official title. The generally
recognized "names" of Japanese emperors are actually titles or toponymic cognomens, sometimes bestowed posthumously, as in the
case of Kammu and Konin, and sometimes acquired during the
person's lifetime or reign. For ease of identification, such names
are employed in the present volume to refer to their holders both
before and after their accession to the throne, and also after their
retirement.
In the translation of official titles, we generally follow the translations descending from Sir George Sansom's pioneering study, "Early
Japanese Law and Administration," Transactions of the Asiatic Society
of Japan, 2nd series, 9 (1932), as modified and expanded by Helen
C. McCullough and William H. McCullough, translators, A Tale of
Flowering Fortunes, 2 vols. (Stanford, Calif.: Stanford University
Press, 1980), with further modification as deemed necessary. In another terminological matter, we sometimes refer to the system of law
and government known to Japanese historians as the ritsuryo sei by a
romanized form, the "ritsuryo system," and sometimes by a translated form, the "statutory system." For confirmation of dates and
readings of Heian names and offices, we consulted Kokushi daijiten,
15 vols. (Tokyo: Yoshikawa kobunkan, 1979-97).
I should like to express my particular appreciation of the contributors for their chapters and their remarkable patience during the
long delay in publication. I am grateful to Dr. Patricia Sippel for her
translation of Chapter 10, and to Dr. Regine Johnson for her care in
adapting and expanding the chapter. Among those who assisted in
the preparation of the volume I should like to thank Professor
Robert Borgen for his collegial assistance to Marian Ury in attending to the final revisions of her chapter when she fell ill. In 1985,
when other responsibilities left me inadequate time to devote to editing, William H. McCullough, whom I had recently joined on the
Berkeley faculty, generously consented to join me as coeditor. Author of the first two chapters of this volume, he made important contributions to several other chapters before he was unexpectedly
stricken by a debilitating illness that eventually took his life in April
1997. I am deeply indebted to William McCullough.
The costs of publishing this book have been supported in part by
Cambridge Histories Online © Cambridge University Press, 2008
PREFACE TO VOLUME 2
XV11
an award from the Hiromi Arisawa Memorial Fund (named in
honor of the renowned economist and the first chairman of the
Board of the University of Tokyo Press) and financed by the generosity of Japanese citizens and Japanese corporations to recognize
excellence in scholarship on Japan. On behalf of the contributors to
this volume, I would also like to express our gratitude to the United
States-Japan Friendship Commission for a grant that funded a
workshop for the authors when we were planning the volume and
that supported the translation of Takeuchi's chapter.
I join the editors of the other five volumes in thanking the Japan
Foundation for funds that facilitated the production of this series.
Donald H. Shively
Cambridge Histories Online © Cambridge University Press, 2008
CHRONOLOGY
794
796
797
798
799
801
804
805
806
807
809
810
811
812
813
Emperor Kammu (r. 781-806) transfers the capital to Heian-kyo. Otomo
no Otomaro, appointed the first seii taishogun ("Barbarian-subduing Generalissimo"), is commander of a campaign against the Emishi in Mutsu.
Resettlement of 9,000 people from the eastern and northern provinces to
Iji Fort in Mutsu.
Sakanoue no Tamuramaro appointed seii taishogun, commander of forces
to subjugate the Emishi.
Shoku Nihongi, second of the national histories, covering 697-791, completed.
Provincial administrators ordered to register Buddhist monks and lay
practitioners.
Appointment of an embassy to Silla.
Provincial governors and bishops ordered to purge the kokubunji (provincial branch temples) of corrupt monks.
Tamuramaro subjugates the Emishi, constructs Isawa Fort, and moves the
Pacification and Defense Headquarters (chinjufu) there. Four thousand
vagrants settled at the fort.
Embassy to the T'ang court accompanied by monks Saicho and Kukai.
Abolition of the Office of Palace Construction.
More than one hundred princes and princesses reduced from imperial to
noble status and given clan names.
Monopolization of the use of uncultivated land by princely and noble
families and by Buddhist temples prohibited.
Purge of officials of the Southern House of the Fujiwara.
Emperor Saga (r. 809-23) succeeds his brother Heizei.
Establishment of the kurododokoro (Chamberlains' Office).
Attempt by Heizei to regain the throne fails and the Ceremonials House
of the Fujiwara is discredited.
Kamo Shrine Vestal first appointed.
Victorious campaign against the Emishi ends thirty years of conflict.
Hereditary district magistrates (gunryo), previously abolished by Kammu,
are reinstated.
Buddhist monks and nuns cautioned by imperial decree against depravity.
Sillan attack on the island of Ochika in Hizen.
Cambridge Histories Online © Cambridge University Press, 2008
CHRONOLOGY
814
815
816
818
823
825
827
832
838
842
848
857
858
866
873
875
878
887
889
891
894
XIX
Princes and princesses given the clan name of Minamoto no ason.
Ryounshu, an imperial anthology of poems in Chinese, completed.
An order issued directing the planting of tea in Kinai and other provinces.
Kebiishi (Imperial Police) office established.
Saga approves Kukai's plan to build a Shingon monastery on Mount
Koya, the beginning of Kongobuji.
The court uniform for ordinary and ceremonial occasions changed to the
T'ang style.
Bunka shureishu, an imperial anthology of poems in Chinese completed.
Saga puts Kukai in charge of Toji in Kyoto as a Shingon temple.
Circulating inspectors (junsauushi) appointed to examine the performances of provincial and district administrators.
Mahayana ordination hall is completed at Enryakuji, the central Tendai
monastery founded by Saicho on Mount Hiei.
Keikokushu, an imperial anthology of poems and prose in Chinese, completed.
Kukai establishes a Shingon chapel within the imperial palace.
The monk Ennin travels with the embassy to T'ang; he returns in 847 with
esoteric scriptures and ritual implements and introduces Tendai and
Mikkyo practices at the court.
Jowa Incident, a plot resulting in the deposition of Crown Prince Tsunesada. He is replaced by a nephew of Fujiwara no Yoshifusa, head of the
Northern House of die Fujiwara.
Ennin begins to establish Amida worship on Mount Hiei.
Yoshifusa appointed Chancellor (daijo daijiri) and becomes de jure regent
for his nephew, Emperor Montoku.
Yoshifusa's grandson, Seiwa, becomes emperor, the first of many child
emperors.
The scandal of die burning of Otemmon discredits the Otomo and Ki
clans. Yoshifusa the first person not of the imperial family to receive die
tide of regent (sessho). Thereafter the Northern House monopolizes die
office.
Fujiwara no Mototsune appointed regent and continues for four reigns.
Reizeiin, a detached palace, destroyed by fire with loss of books and documents.
Emishi revolt in Dewa.
Mototsune appointed regent with the tide kampaku.
He embarrasses Emperor Uda in die Ak5 Controversy.
First Kamo Shrine Special Festival.
Upon Mototsune's deadi, Uda appoints die scholar-official Sugawara no
Michizane Head of die Chamberlains' Office to check die power of die
Fujiwara.
Compilation of Nihonkoku genzai shomokuroku, a bibliography of texts,
mostly Chinese, existing in Japan.
The plan to send an embassy to T'ang is canceled.
Cambridge Histories Online © Cambridge University Press, 2008
XX
899
901
903
905
920
923
926
927
931
935
936
938
939
940
941
949
953
967
969
CHRONOLOGY
Sillan "bandits" attack Tsushima.
Catapult experts deployed to Noto and their number increased in
Kyushu.
Uda instructs Fujiwara no Tokihira and Michizane to share the supervision of government as Ministers of the Left and Right, respectively.
Tokihira succeeds in plotting the demotion and exile of Michizane to
Kyushu, where he dies in 903.
Nihon sandai jitsuroku, the last of the Six National Histories {Rikkokushi),
covering 858-87, completed.
Private purchase of Chinese goods by princes and nobles is forbidden.
Kokin(waka)shu, the first imperial anthology of poems in Japanese, compiled.
Last Po-hai embassy at court. Cessation of official relations with the continent.
Michizane posthumously pardoned and returned to office to placate his
vengeful ghost.
Po-hai destroyed by the Khitan. Two centuries of diplomatic relations
end.
Engi shiki, a compilation of 3,300 statutes, completed, enacted 967.
Quarrel between Taira no Masakado and his uncle Taira no Yoshikane in
Shimosa.
Beginning of the Johei-Tengyo Disturbance: Masakado said to have killed
his uncle, Taira no Kunika, a Hitachi official.
Ki no Tsurayuki composes a travel journal, Tosa nikki.
Minamoto no Shitago completes Wamyo ruiju shd, a large dictionaryencyclopedia, about this date.
Reunification of Korea under Koryo.
Kuya preaches in the streets of Kyoto.
Fujiwara no Sumitomo, an official turned pirate, causes havoc in the Inland Sea.
Emishi revolt in Dewa.
(or 940) Masakado, joined by Prince Okiyo, seizes several eastern
province headquarters and styles himself the "New Emperor" (shinno).
Masakado is killed by his cousin, Kunika's son, Taira no Sadamori and
Fujiwara no Hidesato.
T h e pirate Sumitomo is hunted down and killed.
Fujiwara no Tadahira resigns the office of sesshd and is appointed katnpaku; hereafter the title kampaku is used for regent of an adult emperor.
First major violent demonstration in the capital by warrior monks (sdkei),
these fromTodaiji.
A Chinese merchant from Wu-yiieh takes the monk Nichien to China.
Fujiwara no Saneyori appointed kampaku, beginning the full regency period (to 1068), during which heads of the Northern House are regents almost continuously.
Anna Incident results in the exile of Minamoto no Takaakira.
Cambridge Histories Online © Cambridge University Press, 2008
CHRONOLOGY
985
986
988
995
997
1000
1002
1005
1019
1020
1022
1028
1030
1031
1O
35
1039
1050
1051
1052
1057
1062
1063
1066
XXI
Genshin writes Ojoydshu (Anthology on Rebirth in Pure Land).
Emperor Kazan abdicates, succeeded by Ichijo. Fujiwara no Kaneie's
high-handed rule as regent to 990.
Printed edition of the Tripitaka is brought from China.
Petition of district magistrates and farmers of Owari Province requesting
the removal of the governor for gross misconduct.
Michinaga receives nairan ("private inspection") regental powers; his
control of the court until his death in 1028 is the height of Fujiwara
power.
Pirates from Koryo and Amami Islands attack Tsushima, Iki, and
Kyushu.
Two daughters of Michinaga become empresses of Ichijo concurrently:
Teishi as kogo, Shoshi as chiigu.
Sei Shonagon completes Makura no sdshi (Pillow Book) by this year.
Arrival of Sung traders in Kyushu.
Michinaga falls ill and takes holy orders, but continues to dominate the
court.
Toi (Jurchen) pirates in fifty or more ships ravage Tsushima, Iki, and the
northern coast of Kyushu.
Gen]i monogatari (The Tale of Genji) completed by Murasaki Shikibu
about this date.
Completion of the Golden Hall at Michinaga's Hojoji.
Taira no Tadatsune of Kazusa and ShimSsa plunders tax receipts and revolts. Taira no Naokata appointed commander of a punitive force but
fails to capture him.
Cedar-bark shingles and earthen walls forbidden to those of Sixth Rank
or lower.
Tadatsune surrenders to Minamoto noYorinobu without a fight, raising
the prestige of the Seiwa Genji.
Onjoji warrior monks attack Enryakuji.
Enryakuji monks protest at the regent's residence and set it on fire.
Governor of Yamato and his son exiled for failure to curb the violence
of the Kofukuji monks.
Beginning of the Earlier Nine Years' War, Minamoto no Yoriyoshi's attempt, on imperial orders, to discipline Abe noYoritoki in Mutsu.
Regent Yorimichi converts his Uji villa into a Buddhist temple, the
Byodoin, and constructs the Hoodo (Phoenix Hall) in 1053.
Yoritoki is killed, but the Abe continue a dogged resistance.
Kiyohara noTakenori of Dewa, with a large force, joins Yoriyoshi and his
son Yoshiie and ensures the defeat of Abe no Sadato, ending the Earlier
NineYears'War.
Yoriyoshi, in gratitude for his victory, secretly builds a shrine dedicated
to Hachiman at Yui-no-go, Sagami. (His descendant, Yoritomo, moves
the shrine to Kamakura in 1191 as the Tsurugaoka Hachimangu.)
A Sung merchant presents rare medicines and a parrot to the court.
Cambridge Histories Online © Cambridge University Press, 2008
XX11
1068
1069
1073
1074
1075
1078
1081
1083
X087
1091
1095
1105
1108
1113
1115
1126
1129
1135
1155
CHRONOLOGY
Emperor Go-Sanjo exercises direct rule (to 1073), th e fifst emperor in
170 years whose mother is not a Fujiwara.
Go-Sanj5 establishes the Office for Investigation of Estate Documents
(Kiroku shoen kenkeijd) and confiscates shoen (estates) established since
1045 as well as earlier shoen with questionable deeds.
Forty-two Japanese merchants visit Koryo, presenting gifts to the king
and beginning an active, quasi-legal trade.
Sung court lifts the prohibition on exporting Sung coins, which become
widely used in Japan.
Monks of Enryakuji and Onjoji fight over Onjoji's petition to establish an
ordination platform.
Chinese merchants arrive in Kyushu with a message from the Sung court.
Enryakuji monks and laymen burn Onjoji temples.
Emperor Shirakawa visits the Iwashimizu and Kamo shrines, guarded by
Minamoto no Yoshiie and Yoshitsuna against attack by Onjoji monks.
Yoshiie intervenes in a quarrel among the Kiyohara and the Later Three
Years' War begins.
Shirakawa, after fourteen years of strong rule, abdicates and opens the
Senior Retired Emperor's Office (in-no-cho), through which he dominates the court until his death in 1129.
Yoshiie finally defeats Kiyohara no Iehira, ending the Later Three Years'
War. Mutsu and Dewa are united under Kiyohara no Kiyohira, who assumes Fujiwara, his father's clan name, at Hiraizumi (theOshu Fujiwara).
T h e court is alarmed by the threat of a clash between forces of Yoshiie
and his brother Yoshitsuna near the capital.
Provincial troops forbidden to come up to the capital.
Landholders forbidden to commend land to Yoshiie.
Shirakawa establishes a guard unit (in-no-hokumeri) for the Senior Retired Emperor's Office.
Fujiwara no Kiyohira begins building a temple in Hiraizumi later known
as Chusonji.
Taira no Masamori, favored by Shirakawa, successfully leads a punitive
mission against Yoshiie's sonYoshichika.The martial reputation of the Ise
Heishi begins to rival the Minamoto's.
A force of 2,000 Enryakuji warrior monks comes to Shirakawa's residence, where they are confronted by Imperial Police led by Masamori
and Minamoto noTameyoshi.
ShirabySshi female dancers are said to have made their first appearance.
T h e Chusonji in Hiraizumi is dedicated.
Upon Shirakawa's death, his grandson Toba follows him as the senior retired emperor and proves to be equally strong-willed. Toba relies on the
Ise Heishi for military support.
Masamori's son, Tadamori, captures pirates in the Inland Sea and parades them in the capital.
Because of the lawless conduct of Minamoto noTametomo, his father,
Tameyoshi, is dismissed from office.
Cambridge Histories Online © Cambridge University Press, 2008
CHRONOLOGY
1156
1158
1160
1167
1168
1172
1175
1177
1179
1180
1181
1183
1184
1185
1189
XX1U
Enthronement of Go-Shirakawa.
Toba, the senior retired emperor, dies.
Hogen Disturbance results from rivalries within both the imperial and
Fujiwara families that bring mounted warriors into Kyoto for battle.
The faction supporting Go-Shirakawa, including Taira no Kiyomori and
Minamoto no Yoshitomo, is victorious. Ex-emperor Sutoku is exiled to
Sanuki.
Go-Shirakawa abdicates and plays a strong role at court as senior retired
emperor through the reigns of five emperors, all children but one, until
his death in 1192.
In the Heiji Disturbance, Yoshitomo's coup fails and Kiyomori, who is
again victorious, decimates the Minamoto leaders.
Kiyomori is appointed to die Third Rank, the first warrior to become a
senior noble (kugyo).
Kiyomori appointed Chancellor, the first warrior to rise to the First
Rank. His kinsmen monopolize court offices.
Kiyomori falls ill, resigns, and takes holy orders, but continues to dominate the government.
A Chinese merchant arrives as an emissary from the Sung and gifts are
exchanged.
Honen preaches Pure Land teaching in Kyoto, leading to the formation
of die first sect of popular Buddhism, the Jodo Sect.
Shishigatani plot of Go-Shirakawa's supporters to overthrow Kiyomori is
exposed and crushed.
Kiyomori, with a show of military strength against Go-Shirakawa, seizes
full control of the government.
Kiyomori's two-year-old grandson, Antoku, is enthroned.
Prince Mochihito, a son of Go-Shirakawa, issues a call for warriors
everywhere to rise against the Taira.
Minamoto no Yoritomo raises an army and die Taira army flees from a
confrontation at Fujigawa.
A Taira force torches die Nara temples.
Kiyomori dies.
Minamoto no Yoshinaka (Kiso Yoshinaka) defeats a Taira army at
Kurikara in Etchu and marches on Kyoto.
The Taira with Antoku flee to Kyushu.
Go-Shirakawa has Go-Toba, his grandson, enthroned, even diough Antoku is emperor.
Minamoto no Yoshitsune, half brother of Yoritomo, defeats Yoshinaka.
Yoshitsune surprises and destroys the Taira force at Ichinotani.
Yoshitsune defeats the Taira atYashima and in the final sea battle at Dannoura, where Antoku drowns.
Yoritomo destroys theOshu Fujiwara and extends his military control to
all of Japan.
Cambridge Histories Online © Cambridge University Press, 2008
Cambridge Histories Online © Cambridge University Press, 2008
KYOTO
AREA
_
..
Ninnaji 6
v
A
•£{
Upper K a m o / '
^
S
6 (En "akliji)'
h
Greater'
Imperial ^
Lake
t
^
^
Shrine
Bitva
Onjoji ( M i i d e r a ) ^
ft^.Saga
Awataguchi
' • Rokuhara
• \
i s h i n o m i \ ,i
f
* Osaka
/
Barrier
£ Inari Shrine
•
SEA
• Mansion
} * Daigo
•Fushimi
• Hino
Mutsu
JAPAN
Dewa
/Sado
Minase/
Villa A
Byodoinn^
/ ft ^
/
OF
Temple
Shrine
Iwashimizu
Hachiman Shrine
) Oki
Mansion/Villa
Echigo ,
Tsushima
Izumo , ' - ,
. . . Hoki
¥ T
^Iwami
Nagato .
Ikip
3hikuzen"\
J^™
ilnaba ,
Tango
3
1
.Wakasa EchizenN
Bingo]"
Aki
.Mimasaka'^
BitchuS
,.
Su6
Tanba
J
—~
Tosa
Bungo\
. Higo
y
Musashi j
,-Yamato^-
;'Hitachi/
Kai
IShim6sa""\
Mikawa,-'
.. Awa
/Shimotsuke
Mino'-'Owari
AwaHj,
Iyo
Kozuke
Hida
Shinano
IBuzen^
Hizen
<Kaga,-
Etchu
Suruga^-
Sagarhi
Totomi;
fKazusa
^Shima
Kii
Satsumai
Hyuga
/
0
50
100
150kms
1. Settsu
2. Izumi
3. Kawachi
4.Yamashiro
K Y O T O area
(see inset above)
S\wa
PACIFIC
o
Japan in the Heian period. Adapted from Helen McCullough,
Classical Japanese Prose, 1990.
Cambridge Histories Online © Cambridge University Press, 2008
OCEAN
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
The Heian period opened in 794 with the building of a new capital,
Heian-kyo, later known as Kyoto. The grand plan of the new city, on
a larger scale than earlier capitals, expressed the ambitious vision of
Emperor Kammu. No other Japanese emperor had ever taken into
his own hands so decisively the absolute powers of the emperor as
conceived in Chinese theory. He and some of his immediate successors not only asserted the authority of the throne; they took positive
measures designed to improve the effectiveness of the central government in administering the country. Theirs was a dedicated attempt to revitalize the system of administration modeled on the governmental machinery of T'ang China and operate it effectively.
Throughout the four centuries of the Heian era the imperial court
continued as the only political center, but the effectiveness of its administration declined gradually. The title of emperor continued in
the imperial line without dynastic change, as it does to this day, but
many of the reigning emperors were reduced to figureheads, manipulated by noble families at court, notably the Fujiwara, and later
by senior retired emperors. The Heian period closed in 1185 when
the struggle for hegemony among the warrior families resulted in the
victory of Minamoto no Yoritomo and most political initiatives devolved into his hands at his headquarters at Kamakura. The imperial
court continued at Kyoto, playing a largely ceremonial and legitimizing role, while political power was exercised by military overlords
until the Meiji Restoration of 1868.
During Heian times, however, there was no challenge to the central position of the imperial court; rather, there was a gradual decline
in its ability to derive adequate income from the provinces to sustain
itself in the style it had designed. Similarly, although the principle of
monarchical rule was unquestioned, actual political power was usually exercised, after the first century and a half, by a Fujiwara or by
a senior imperial relative acting in the name of the emperor. In fact,
historians, both medieval and modern, of loyalist sympathies regard
Cambridge Histories Online © Cambridge University Press, 2008
2
INTRODUCTION
as the golden age of Japanese history those decades of direct rule by
enlightened, "virtuous" emperors - Kammu (781-806) and Saga
(809-23), and again Uda and Daigo (887-930). For loyalists, these
were the years when Japanese rulers most closely approached the
ideal reigns of the sage-kings of ancient China. Probably the direct
rule that these emperors exercised was praised, without close inquiry
into how wise their policies may have been.
The system of government at the beginning of the Heian period was
a remarkable copy, somewhat modified, of the Chinese institutions
of the Sui (581-618) and T'ang (618-907) dynasties. The emperor
was expected to rule with absolute authority. He was served by high
ministers and a council of state, overseeing an elaborate centralized
bureaucracy arranged in ministries (eight in the Japanese version)
and numerous bureaus and offices. Japan, with a population estimated at 6 or 7 million, was divided into sixty-eight provinces (kuni),
including two island provinces, as of 823, each with a provincial
headquarters overseen by a governor. Provinces were subdivided
into districts (gun or kori), eight or nine on the average, each with an
administrative office. In China the bureaucratic structure was staffed
by officials selected for appointment on the basis of qualifications
as determined by examinations. The examination system was not
properly instituted in Japan. Some students at the state Academy
(Daigaku-ryo) did take examinations, but appointment to office was
determined largely by the court rank of the candidate's family and
by family and marital connections rather than by qualifications or
ability. The Chinese prohibition against appointment in one's native
district was not observed in Japan.
Rice land was nationalized, in principle, and it was allotted to
families of cultivators according to the ages and sex of family members. Allotments were subject to revision and adjustment every six
years. Uniform taxes were levied on the basis of rice-field holdings,
payable in grain, and there were handicraft-produce taxes frequently
payable in textiles, a corvee, and a military service tax. In order to
operate the land system, it was essential to survey the ricefieldsand
prepare current registers of land allotted and a census at six-year intervals. The procurement of a literate and efficient staff, dedicated to
the government's interests, to operate this complex land and tax system may well have been beyond the country's human resources.
Provincial administration required the building and maintenance of
a network of highways and a post system. It is not known when and
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
3
how extensively the facilities were completed in areas distant from
the capital, especially in the eastern provinces.
The T'ang system of government is referred to by Japanese historians as the ritsuryo (statutory) system: ritsu is the penal code, specifying punishments for various offenses, and ryo is the administrative
code that provides detailed regulations and instructions for the operation of government and society. The Taiho code, completed in
701, and its revision, theYoro code, in 718 (but not put into effect
until 757), generally followed the T'ang code. The ryo of Ydro had
over nine hundred articles. It was later amplified by supplementary
rules. Kyaku were new orders or modifications of existing law, while
shiki were issued to provide additional detail for provisions of the ritsuryo and kyaku. These supplements, designed to make the ritsuryo
system operate more effectively by expanding on existing provisions
or sometimes overruling them, were compiled diligently in early
Heian into the tenth century.
When Chinese institutions were introduced from the beginning of
the seventh century, they were seriously compromised by native traditions of aristocratic privilege. The elite class had long been organized in uji (clans or lineage groups), bound together by descent
from a common ancestor, the clan deity (ujigami). Among the large
clans in Yamato in central Japan, one rose above the others and assumed a kingly role. Claiming descent from the Sun Goddess, its
chief, in his sacerdotal function, interceded with the deities on behalf of all of the clans. He also mediated relations among clans. In
the sixth century, by conquest and negotiation, this Sun line extended its authority over much of Japan in league with its supporting clan chiefs, who performed specialized services as warriors, ritualists, administrators, and fiscal agents, functioning increasingly like
ministers of a king.
The relative status level of the leading uji chiefs was recognized
by hereditary court titles {kabane) conferred by the king of the Sun
line to honor superior lineages. Competitiveness among the chiefs
for status was the source of considerable turbulence within the Yamato group. It is significant that one of the first Chinese institutions
to be adopted in 604 was the system of twelve cap-ranks, designating the colors of caps in official court dress. The scheme of court
ranks was elaborated several times during the seventh century, increasing the number of gradations to nineteen, then to twenty-six,
and in 685 to an eight-rank system with forty-eight steps, demonstrating the care with which gradations in hierarchy among the clan
Cambridge Histories Online © Cambridge University Press, 2008
4
INTRODUCTION
chiefs and their followers were adjusted to proclaim their relationship to the Yamato king.
This king came to be called tenno (literally, "heaven [descended]
luminance"), conventionally translated "emperor." Like the Chinese
Son of Heaven, he claimed absolute authority to all of the land. Private titles to rice fields were abolished in principle, as were agricultural and craft support groups that had served and supported the
clans. To gain the acquiescence of the affiliated clans, he appointed
their members to new Chinese-style official positions and assigned
lands and households for their support. Rice land was granted according to court rank, office held, and meritorious service. Powerholders of regional clans whose cooperation was needed were also
given grants and official positions. Court rank gave honored status
and assurance of eligibility for hereditary appointments. The awarding of lower grades of rank restricted the number of lineages that
could compete for high office. In effect, a new court aristocracy was
formed selectively from the uji at great expense to the public domain. In the eighth century this was a large group, for more than 120
clans can be recognized, but by mid-Heian only about 10 of the
clans were playing a significant role.
The first chapter of this volume, by William McCullough, describes
the politics of the Heian court, beginning with the strong rule of
Emperor Kammu and continuing until the reassertion of direct imperial rule by Go-Sanj6 in 1068. During the first half-century the
highest officials were drawn mostly from the imperial clan. Thereafter, in the competition for appointment to high office, the Northern Branch of the Fujiwara clan increasingly succeeded in excluding
from high office other clans and also the other branches of the Fujiwara. This predominance came in good part from their success in
providing emperors with Fujiwara daughters who produced heirs to
the throne. The practice of a Fujiwara head, as grandfather of a
young emperor, serving as his grandson's regent began in 866 and
became the regular pattern. (Hitherto only imperial princes had
served as regent.) Emperors were often persuaded to abdicate before
reaching manhood. Early abdication had been common since the
middle of the seventh century. It was often a welcome escape from
the ceremonial demands of the position, palace intrigue, and the
constant requests for favors from imperial relatives and families
which had supplied women for the harem. It became a regular practice for Fujiwara relatives to serve as regent for adult as well as child
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
5
emperors, and after 967 an unbroken succession of Fujiwara held
the office of regent. This was possible because the Fujiwara played
the game of marriage politics with consummate skill, aided by the
good fortune of usually having available a supply of eligible imperial
consorts.
The few strong emperors who ruled directly in early Heian proved
to be the exception in Japanese history. The actual power of political
initiative was delegated to (or usurped by) Fujiwara relatives. The
emperor's role again became largely ceremonial and sacerdotal as
he receded to a position above politics. The preference for indirect
rule that had been evident since the clan period prevailed again. The
domination of the court by the Fujiwara leader reached its height
with Michinaga, especially in the years from 1016 to 1028.
During this time when the aristocracy was ascendant, the number
favored with privilege was not large, perhaps only two or three hundred principal male members of the nobility at any one time. Those
who actually wielded power may have numbered only a dozen or two
and were almost all Fujiwara, except for a few Minamoto (of courtier, nonwarrior families). The five or six highest offices were reserved
for lineages that traditionally might reach the first three court ranks,
senior nobles known as kugyo, a group that also included men of the
Fourth Rank who served on the Council of State. Lower on the
scale, but also among the privileged nobles, were those of the Fourth
Rank (ministers, for example) and Fifth Rank (governors of large
provinces, among others).
Although the wealth of the aristocracy came from a variety of,
sources, well into the eleventh century its mainsray was in(V"..v attached to rank and office. In the eleventh century, however, the
lower ranks of the nobility seem often to have received little or nothing in the way of official income, depending for their livelihood on
service in the households of the great aristocrats, who had concentrated most of the government's resources in their own hands. There
was, in addition, some income obtained from landed proprietorships
called shoen (estates). It was this small, ancient, completely urbanized society of aristocratic civil officials, living mostly on appointive
incomes, that produced either directly or through patronage most of
what we think of as Heian aristocratic culture, which is discussed in
Chapter 6 by Helen McCullough.
The monopoly of power held by the Fujiwara continued until 1068,
when Go-Sanjo came to the throne, the first emperor in 170 years
Cambridge Histories Online © Cambridge University Press, 2008
6
INTRODUCTION
whose mother was not a Fujiwara. During his short reign he reasserted monarchical powers. For the next century until the end of
the Heian period, the Fujiwara, while continuously holding the post
of regent and other high offices, had little actual power, being kept
in check by three strong emperors (Shirakawa, Toba, and Go-Shirakawa). Each of these three, after a relatively short reign, abdicated,
took up residence in a "cloister" (in), and established the Administrative Office of the Senior Retired Emperor (in-no-chd). This office,
in its peculiar mixture of private and public functions, resembled
the Administrative office (mandokord) that had been created much
earlier by the Fujiwara and other noble families to manage their
family affairs, even to the formation of a guard unit and client relationships with warrior families. This organization enabled the retired
emperors to develop more effectively landholding and other resources for the support of the imperial family and to hold the upper
hand over the Fujiwara and dominate the court as few emperors
had. As father or grandfather of the reigning emperor, the senior retired emperor played a regental role, dominating not only the emperor but also any younger retired emperors. He displaced the Fujiwara regent as the acknowledged authority in national affairs. This
practice of political domination by the senior retired emperor,
known as insei, "cloister government," is the subject of Chapter 9 by
G. Cameron Hurst. The shift of power to the ex-emperors can be
viewed as another phase in the ebb and flow of political strength between the imperial line and noble families with which it intermarried, a pattern that was already familiar in the Yamato clan period
and in the Nara court.
Perhaps the greatest attribute of the Chinese state was its capital, an
enormous walled city laid out in a symmetrical grid pattern dominated by the huge buildings of the imperial palace compound. The
city plan was a most impressive symbolic representation of imperial
grandeur. Kammu's new capital, Heian-kyo, following this model,
was designed to make a powerful statement. True, it was overly ambitious and too costly, for it was never possible to fill out the complete grid. But while earlier capitals were all short-lived, Kyoto remained the imperial capital for more than a thousand years, from
794 to 1869. Until the seventeenth century it was the only real city
in Japan. Chapter 2, by William McCullough, is devoted to a description of the capital, its plan and architecture, its economy and
commerce, its population and the social world of noble households.
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
7
For the major policy and administrative problems engaged by the
early Heian emperors, we turn again to the first chapter. In addition
to building a new capital, Kammu may also have had in mind the
T'ang example of mounting military campaigns to subjugate and absorb "barbarian" border areas. Since early Nara there had been numerous expeditions to subdue the ethnic people in the northeast.
Kammu intensified the effort and by 804 finally met with success.
However, the great expense of the expeditions together with the
building of the new capital exhausted the treasury.
Among measures of fiscal retrenchment taken by Kammu was to
restrict the large number of imperial princes and princesses receiving government support under provisions of the statutory system.
In 805, more than one hundred were reduced to noble status, a
measure taken by several succeeding emperors. Some sons of highranking consorts were granted the clan name Minamoto, and a few
imperial grandchildren were given such clan names as Taira, Ariwara, and others. Some of these imperial descendants found careers
as court nobles and others joined the provincial gentry.
Kammu and his immediate successors attempted to make their
administrations more effective by introducing several new offices. To
check misappropriation of tax rice and other assets from the provincial account by an outgoing governor, a board of agents known as
kageyushi was appointed in Kammu's reign to audit a governor's accounts and the transfer of property to the incoming governor. Both
parties were held accountable for discrepancies.
In 810, Emperor Saga established the kurododokoro (Chamberlains' Office), staffed by trusted men, to ensure confidentiality in the
handling of important documents. Later it transmitted imperial
edicts and supervised the imperial archives. In time it also came to
handle the emperor's household affairs.
Also in Saga's time, in response to lawlessness in the capital and
the surrounding region, a new police organization known as kebiishi
("Offenses Investigation Agents") gradually evolved. The functions
of the Imperial Police included not only security matters and the arrest of miscreants but also the investigation, trial, sentencing, and
imprisonment of criminals, thus replacing some of the duties of several existing offices. Later, branches were placed in some provinces
and its kebiishi agents investigated land ownership, tax evasion, and
other matters.
These new offices were established, one may conjecture, because
the functions they served were not being performed satisfactorily by
Cambridge Histories Online © Cambridge University Press, 2008
8
INTRODUCTION
the existing ritsuryo offices. Or, in some instances, the emperor and
his circle may have aimed to bring the functions in question under
more immediate control. Many of the offices provided for by the
statutory code were languishing, either because they were not considered essential or because the government had insufficient revenues to keep them in operation. By the end of the ninth century
half of the central government's ritsuryo offices had been abandoned
and the number of officials was much reduced.
The primary mission of the provincial government office and its
subunits, the district offices, was to collect local products in the form
of taxes and forward them to the capital. These products, including
rice and other foodstuffs, were all of the goods and services needed
to support officials of the central government and supply the specific
needs of the capital and its elites: textiles, handicrafts, and local
products such as salt, iron, paper mulberry, and many other goods.
Rice was collected not only by direct taxation, but also as rental on
rice land lent out by the provincial government. Rice was also collected on seed rice lent to cultivators for planting in the spring. At
the beginning, corvee and military service were also part of tax
obligations. In Chapter 3, "Land and Society," Dana Morris discusses how this all-important tax structure of the beginning of Heian
underwent continual changes during the following centuries.
The provincial capitals were designed as small versions of the grid
plan of the Heian-kyo. Detailed regulations for the staffing and operation of the provincial government, described by Cornelius Kiley
in Chapter 4, "Provincial Administration and Land Tenure in Early
Heian," indicate the importance central officials attached to the
province's mission. A directive of 822 specifies a large and specialized staff for the provincial office to perform various administrative
functions and compile the required annual reports, tax-grain inventory, list of tribute, number of taxable households, acreage under
cultivation, percentage of crop damage, and so forth.
Responsibility for the administration of a province was entrusted
to a governor selected from the middle ranks of the Kyoto nobility
and appointed usually for four years. He was accompanied to his
post by a number of staff members, but most of the officeholders in
the provincial headquarters were members of the local elites in positions that, by and large, were permanent and hereditary. It was
usually difficult for the governor to prevail against the interests of the
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
9
locally based officials. Increasingly the governors were absentee, the
position a sinecure, and the executive function left to a deputy.
The subprovincial district offices were staffed entirely by local
gentry. As a consequence, the administration of land and the collection of taxes were carried out by locals who, backed by governmental authority, benefited greatly from their positions in income and
landholding. Although the mission of a provincial office was to marshal local resources for the benefit of the central government, it developed into a bargaining place for the division of resources between
capital and country. Some governors sought the post, even purchased it, with the expectation of enriching themselves. As a consequence of these competing interests, the share that went to Kyoto
declined steadily.
In the attempt to ensure its income, the government set a revenue
quota for each province, charging the governor with the responsibility of meeting the contracted amount. In effect this policy recognized the provincial government as a semiautonomous unit with tax
obligations to the central government. It was permitted to make certain changes in the tax system, adding new taxes or occasional levies,
to meet its quota. This was a significant departure from the principle of the statutory code of a national, uniform tax system. Morris
argues that the modifications in the tax system, while abandoning
provisions of the code, were changes that better met the capital's
needs and, at the same time, were more efficient and betterfittedthe
rural economy. The changes succeeded at length in stabilizing the
government's income. However, the quotas were set using the tax
base as it stood about the year 900. As a consequence, Morris points
out, the central government did not benefit from the increase in
agricultural output brought about by expansion of acreage and by
higher yields produced by improvements in agricultural methods.
Among the improvements were the introduction of an animal-drawn
plow with moldboard, the use of draft animals, better fertilizer, and
other innovations discussed by Morris.
The system of allotting rice fields on the basis of census registration operated reasonably well in the Nara period, but reallocations
came to a halt about 840. Scholars have suggested a variety of causes
for the suspension of reallocation. Morris demonstrates that the primary reason was the shortage of land available for distribution. Population had increased by more than a million during Nara, creating
a demand for allocations that the government could not meet.
Cambridge Histories Online © Cambridge University Press, 2008
10
INTRODUCTION
Since early in the Nara period the government had encouraged
the opening of new rice fields, and much land continued to be reclaimed. Because of the high cost of developing irrigated rice land,
the government was obliged in 743, as an incentive, to grant developers permanent possession of reclaimed land, the source of many
of the first shoen (estates). In the Heian period, however, shoen were
created, in effect, when the central government ordered the transfer
of tax payments on segments of land from the provincial government
to a religious institution or a noble family in the capital. Subsequently there was an increase in the number of shoen established by
commendation. Local magnates or land managers were often in conflict with provincial officials over land rights, management authority,
and tax immunities. They tried to prevent the interference of provincial authorities by commending rights (shiki) to the land under their
control to a Kyoto aristocrat or a major religious institution as "proprietor" (ryoke) while retaining hereditary rights of management and
control of the cultivators. In return for a fee or a share of the shoen's
income, the ryoke sought to protect the rights claimed by the local
manager. If the ryoke could not command enough influence at court
to accomplish this task, he might make a further commendation to
a member of the imperial family or one of the most powerful Fujiwara or to a great temple (honke). By late Heian, nearly half of the
agricultural land had become shoen in this way. This privatization of
land, or rights to land, was carried out, for the most part, within the
provisions of the statutory code and was usually well supported by
documentation.
The greater part of the agricultural land may have remained in the
public domain under the administration of the provincial headquarters - designated as kokugaryo (provincial domains) - but shoen probably provided most of the economic support of court nobles, religious institutions, and even the imperial family. Emperor Go-Sanjd
in 1069 ordered a major nationwide registration of shoen to examine
their legality and rule on their tax exemption claims. Such inquiries,
which had begun as early as the ninth century, continued periodically
until the end of Heian. Sorting out claims of land parcels to shoen or
kokugaryo status was a continuous process. Also in frequent dispute
was the question of which parcels of shoen land were liable for which
provincial levies. Issues such as these were usually present in the centuries-long struggle of the central government to control local officials and landholders. But, lacking effective means of coercion, the
authorities gradually lost ground. Thus, landholding and tax systems
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
11
changed slowly but significantly in the course of the Heian period.
Through these changes, as Kiley states, the provincial governments
"proved to be among the most durable of the ritsuryo institutions,
probably because from the beginning they served to integrate the interests of local elites, capital officials, and court nobility."
Japan was a preliterate society until the adoption of Chinese as a system of writing. This probably occurred several centuries before the
writing of the first complete books that have come down to us, two
histories: the Kojiki, traditionally dated 712; and the first of the official histories modeled on Chinese dynastic histories, the Nihon shoki
of 720. The latter work reveals a high degree of assimilation of Chinese civilization: the writing of a historical chronicle in literary Chinese, the citing of earlier sources, a Confucian worldview and ethics,
and, of course, the re-creation of the complex T'ang organization of
government already spelled out in the codes of 702 and 718.
By this time the study of Chinese civilization had certainly had a
long history. The first study of Chinese texts is conventionally associated with the arrival of Wani, a Korean scholar from the kingdom
of Paekche who is thought to have arrived about 400 to tutor the
crown prince. Early in the sixth century, Paekche began sending, in
rotation, scholars of the classics as well as specialists in music, medicine, divination, and the calendar. The introduction of Buddhist
statues and sutras in 552 (or 538) is also attributed to the king of
Paekche. Knowledge of Chinese-style governmental institutions as
adapted in Korea came from several Korean states and informed the
measures taken by the prince regent Shotoku Taishi in 604 to introduce cap ranks and other Chinese institutions.
Regular relations with China began in 607 and 608 when official
embassies were sent to the Sui court, followed by many embassies to
the T'ang, beginning in 630 and continuing until 834. These missions enabled the court to send students to China, some for many
years of specialized study. Upon their return they made an invaluable contribution to the political and cultural transformation of
Japan on the Chinese model.
The invasion of the Korean peninsula by T'ang armies and the
overthrow of Paekche in 663 and Koguryo in 668 precipitated the
flight to Japan of many Koreans. Among the refugees were officials
and scholars who had knowledge of such fields as administration,
law, court ceremonial, military tactics, Chinese literature, and other
subjects. That well over a hundred of the immigrants were appointed
Cambridge Histories Online © Cambridge University Press, 2008
12
INTRODUCTION
to court rank and absorbed into the nobility is evidence of the value
placed on their knowledge and talents.
The court followed the Chinese example of opening an Academy
(Daigaku-ryo) in the capital to train officials. The written language
of the Academy was Chinese, as was the written language of government, the law codes, the surviving records and inscriptions, and
the national histories. The ability to compose Chinese poems became an essential skill at court banquets, as it was at receptions for
visiting Chinese and Korean officials. Banquet poems in Chinese
were represented in the earliest known anthology, the Kaifuso of 751.
It was followed in the early decades of the Heian period by three anthologies of poems in Chinese compiled by imperial order. The only
written works in the Japanese language surviving from the Nara period are written with Chinese characters used phonetically, that is,
for their sound, ignoring their meaning. However, there were exceptions when Chinese constructions were used, pronounced in the
Japanese approximation of their meaning. This cumbersome mixture
of Chinese and Japanese is found in the Kojiki and the great anthol-'
ogy of Japanese poetry, the Man'yoshu (after 759), and in records of
imperial proclamations (semmyo) and Shinto prayers (norito).
The Japanese elite appears to have been eager to adopt all aspects
of Chinese civilization, not merely governmental institutions, law,
and the written language, but also Chinese thought and ethics, Buddhism and its sculpture, painting, and architecture, continental forms
of music and dance, and many branches of Chinese knowledge, arts,
and crafts.
These remarkable advances in acquiring and assimilating Chinese
culture in the two centuries leading up to the Heian period are described by Edwin Cranston in his chapter in Volume 1 of The Cambridge History of Japan. In the present volume, Marian Ury's Chapter 5 continues the account of the absorption of Chinese learning
and intellectual life. Sons of aristocrats were tutored using Chinese
primers that were compilations of quotations drawn from the classics and other edifying texts. While memorizing the characters, children also learned moral maxims to live by. The more talented or
better-connected boys went on to private schools or the Academy
to prepare for careers as officials. The curriculum included Chinese
classics, histories, and belles lettres.
Many noblemen recorded their activities in diaries, written in
Chinese or hybrid Chinese, which have come down to us. Parts of
the personal diaries of emperors Uda and Murakami of the eighth
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
13
and ninth centuries have survived. Anthologies of prose written in
Chinese began to appear. Ury traces the careers of professional
scholars of Chinese patronized by emperors and Fujiwara leaders.
Among the scholars were a number of men of experience and insight
who, when commissioned to submit reports detailing the ills of society and government and their recommendations for reform, came
forward with trenchant proposals, most of which were not adopted.
Some of the early Heian rulers, however, tried to legislate an orderly
and harmonious society by compiling detailed regulations supplementing the ritsuryo statutes, culminating in the voluminous Engi
shiki, completed in 927. Scholars continued to record the histories of
past reigns, completing the last of the Six National Histories {Rikkokushi) in 901 to serve as mirrors in which to read the successes and
failures of past administrations.
During the ascendancy of the Chinese cultural style at the beginning of Heian, poetry composed in Japanese may have lost temporarily its role on public occasions. But an important change came
about 900 with the development of the hiragana syllabary. This was
the practice of using a limited number of the cursive, simplified
forms of characters for their sound to write Japanese phonetically.
The graceful and fluid kana were an efficient and elegant way to
write Japanese, a major improvement over the cumbersome method
of writing characters in their formal, angular style. Men as well as
women used hiragana to write poetry and informal prose.
There was a resurgence of attention to Japanese poetry, especially
the waka (thirty-one-syllable) form. The first and most celebrated of
the imperial anthologies of poems in Japanese, the Kokinshii, was
compiled in 905. One of the compilers, Ki no Tsurayuki, added a
preface in Japanese in which he seemed to imply that Japanese poetry could stand its ground with Chinese poetry. From about the
same date came the earliest surviving work of narrative prose, Taketori monoqatari. In 935,Tsurayuki wrote a travel journal (Tosa nikki),
in which the Japanese prose serves as a setting for his poems; in it he
adopted the persona of a woman, a nod to the convention that
Japanese prose was woman's language. In these prose works, as in
the Kokinshii, there is a confident assertion of the Japanese style at
the same time that Chinese stylistic influences are evident. The
blend of Japanese and Chinese elements is characteristic of the best
of the arts of aristocratic culture. The importance of hiragana was
that both women and men could now write fluently in their native
language. It was especially liberating for the women, ladies of the
Cambridge Histories Online © Cambridge University Press, 2008
14
INTRODUCTION
court, to find voices in which they could express their feelings about
their experiences - relationships between women and between men
and women.
The aesthetics of court society received its most sensitive expression in The Tale of Genji (early eleventh century) of Murasaki Shikibu. This acclaimed masterpiece of Japanese literature, together
with the Pillow Book (completed about 1001), a miscellany by Sei
Shonagon, and contemporary historical accounts (Eiga monogatari
and Okagami) of the Fujiwara at their height, provide us with extraordinarily rich descriptions of aristocratic culture of the early
eleventh century. Helen McCullough's Chapter 6 draws on this
range of literary sources to detail the artistic accomplishments and
entertainments of this privileged society in its most brilliant period.
This was the culture of a small circle of the highest-ranking nobles
and their associates. And yet we are better informed about their lives
and careers, their etiquette and ceremonies, their ideals of beauty
and aesthetic sensibilities, their romantic affairs and yearnings, than
we are about any other group in Japanese history, at least until the
seventeenth century.
Of the many elements of Chinese and continental civilization introduced into Japan, one that was embraced with special enthusiasm
and had profound and lasting importance in Japanese culture is
Buddhism. Following the arrival of Buddhist texts and statues from
Paekche in the middle of the sixth century, the Nakatomi clan, with
its Shint5 ritualists, opposed the reception of Buddhism by the
court. In 587 the dying emperor Yomei was converted to Buddhism.
In the succession struggle that followed, Soga no Umako and the future Prince Shotoku defeated the Nakatomi. To commemorate divine assistance in their victory, Shotoku founded the Shintennoji
Temple (in present day Osaka).
Buddhist sutras and images were considered to have magical powers akin to but greater than those of the native religion. They were regarded as beneficial in preventing and curing disease and safeguarding against famine and natural disasters. Buddhism's gorgeous
vestments and mysterious rites had a strong appeal, as did the
grandeur of its temple buildings with their statues and wall paintings.
Buddhism was an important vehicle for the transmission of many aspects of continental culture. In addition to the literate, intellectual
tradition and the large corpus of texts, it also carried with it beneficial practical knowledge: medicine, architecture, bridge building, and
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
15
road construction. The imperial clan and great families were extravagant in supporting Buddhism, founding temples and endowing
them with large landholdings. At the same time, the government
tried to maintain control of temples by appointing the high clerics,
restricting the number of ordinations, and disciplining wayward
monks and nuns.
The government recognized the value of Buddhism as a means of
spreading the spiritual authority of the state, following the practice
in China and Korea. In 741, Emperor Shomu ordered the establishment of a temple (kokubunji) and nunnery in each province as
branches of the central Todaiji in Nara. They were staffed by clerics
from the capital who served as religious agents of the central government and performed rituals for the protection of the state. Buddhism became, in effect, a state religion. It was brought under the
administration of the central government in somewhat the same way
that Shinto shrines were in the Taihd code that established the
Jingikan (Department of Shrines) as a government office. The emperor presided over both temples and shrines while continuing to
perform his historical role as chief priest in the worship of his ancestors and the national deities.
The government's attempt to check the uncontrolled increase of
temples and monks in early Nara was undercut by the lavish support
of temples by Emperor Shomu (r. 724-49) and his daughter, who
succeeded him as Koken (r. 749-58) and reigned again later as
Shotoku from 764 to 770. The large temples in Nara appeared to surround and overwhelm the imperial palace. The scandal of the priest
Dokyo's influence over Shotoku, which threatened even the imperial
succession, brought a sharp reaction that profoundly affected the
course of Heian Buddhism. The danger of Buddhist interference in
government was a factor in Emperor Kammu's decision to move the
capital first to Nagaoka, and it certainly determined his policy to forbid the Nara sects from establishing temples in Heian-kyo.
Stanley Weinstein, in Chapter 7, "Aristocratic Buddhism," traces
the imperial patronage of the newly introduced esoteric sects, Tendai
and Shingon, which were permitted to establish temples in and near
the new capital. These sects were adapted to Japanese needs. Their
rituals soon came to have an integral role in the spiritual as well as
ceremonial life of the imperial family and court aristocrats. Tendai
monks introduced from China recent developments in Pure Land
teachings that assured believers that the repeated recitation of
Amida's name would lead to rebirth in the Western Paradise. BudCambridge Histories Online © Cambridge University Press, 2008
l6
INTRODUCTION
dhism emerged from the monasteries to give direction to the spiritual lives of the aristocrats, especially in their later years, with the
promise of immediate salvation. Nobles patronized the copying of illuminated sutras, and the great Fujiwara constructed private chapels
and commissioned large gilded statues of Amida Buddha. On his
deathbed the believer waited holding strings in his hands, stretched
to a statue or painting of Amida descending to welcome him (raigo),
to ensure that his soul would be conducted to paradise. At the same
time, itinerant monks preaching the faith attracted an ecstatic following among the common people. In the twelfth and thirteenth
centuries, Pure Land teachings emerged from Tendai to become independent sects: the Jodo and Jddo Shin.
In Heian times, Buddhism diverged from its sources on the continent, and became a naturalized Japanese religion. It has continued to this day to be the prominent religion in Japan, in contrast
to its history in China and Korea, where its following gradually
declined.
Allan Grapard's Chapter 8, on religious practices, introduces recent
findings by Japanese and Western scholars that have broadened our
conception of Heian religion. One theme the chapter traces is the relationship that evolved between Buddhism and Shinto, as the native
religious practices came to be called. Beginning in mid-Nara and
increasingly in early Heian, Buddhist temples were built on the
grounds of important Shinto shrines. Buddhist monks administered
the temples and performed Buddhist rites in front of the native
deities (kami). There was no incongruity seen in monks worshipping
both buddhas and kami. Not only were temples associated with
shrines in this way, but buddhas and bodhisattvas came to be associated with individual ancestral kami of the imperial and clan lineages. In many instances kami came to be regarded as local manifestations of the more universal buddhas and bodhisattvas. A
number of shrines to which temples were added were distant from
the capital and were dedicated to the worship of deities associated
with lineages other than the imperial clan. Not only were kami coopted in this way and linked to Buddhist deities, but measures were
taken to establish a uniformity in the rites that were performed. The
primary rituals were those for the protection of the state and for the
health and protection of the emperor as the embodiment of the
state. This was, like the network of provincial branch temples, yet another strategy for extending throughout the country the religious
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
17
representation of the emperor and the central government, and for
presenting rites that had a recognizable, consistent pattern.
Grapard opens his chapter with a quotation from the tenthcentury testament of Fujiwara no Morosuke, who, as Minister of the
Right, held one of the highest offices of state. His words of guidance
for his heirs include advice on how they should conduct themselves
in their daily activities. He provides a remarkable illustration of the
variety of religious, cosmological, and ethical beliefs that guided the
Heian noble's actions. Morosuke's daily observances included reciting the names of buddhas and worshipping various kami. He was
mindful each day of cosmological and geomantic constraints, according to Chinese calendrical lore, on his movements and activities.
The testament proceeds to advocate at length some of the Confucian principles of behavior - rectitude, moderation, self-control, frugality, and a single-minded concern for his family's physical and material welfare. Heian prose literature and nobles' diaries provide
numerous instances of a noble's appeal to rites to ensure protection
from disease, recovery from illness, safe childbirth, success in projects, and progress in his career. A noble gentleman was expected to
cultivate knowledge of and skill in the correct performance of rites
and ceremonies which were akin to religious practices.
The professional warrior class (busht), its evolution and rise to dominance at the end of Heian, is the subject of Chapter 10 byTakeuchi
Riz5. When the government ended military conscription as a national requirement in 792, it ceased to maintain a standing army.
To keep order in the countryside, it recruited kondei ("stalwart
youths"), mounted fighters drawn from the families of the regional
gentry. A prescribed number of kondei, ranging from twenty to two
hundred, was specified for each province. Perhaps because they were
so few, they do not appear frequently in the literature. Little is
known about their employment, and they are rarely mentioned by
mid-Heian. In the capital, for the defense of the Greater Imperial
Palace, there were six guard units provided by law. Led by aristocrats, they became largely ceremonial and decorative. For the security of the palace and to check crime in the capital and its environs,
the Imperial Police {kebiishi) were formed early in the ninth century.
A number of government units, like the new Chamberlains' Office,
and later the Senior Retired Emperor's Office (in-no-cho), found it
expedient to recruit private guard units for security.
Lawlessness was prevalent in the countryside, for the central govCambridge Histories Online © Cambridge University Press, 2008
l8
INTRODUCTION
ernment had in effect ceded to the landed classes responsibility for
the preservation of the public peace. By the late ninth century,
provincial governors, first in the eastern provinces, requested permission to recruit warriors to protect themselves from attack and
also to employ as needed to enforce their orders. Landowners and
shoen managers engaged armed men to help them defend their
claims. Predominant among the professional fighters were relatives
of district magistrates and other prominent provincial families. Only
those in this class had the resources to maintain the horses, saddles,
armor, and weapons that distinguished them as professional warriors. Descended from regional uji, these mounted archers, skilled in
hunting, had been the effective forces in the campaigns to the northeast. Members of this class held appointments as provincial or district officials, or involved themselves as self-appointed land managers or tax collection agents, ever ready and equipped to defend
their land claims and perhaps to trespass on others' holdings. Military preparedness was a necessary adjunct to land management.
When a local strongman defied a provincial governor by refusing
to forward tax revenues or by intruding on government land, the imperial court, lacking a military force of its own, deputized the governor or, if more strength was needed, the leader of a court-related warrior family - a Taira or Minamoto - with a military or police title and
commissioned him to mount a force against the offender. The revolt
of Taira no Masakado in the 930s was the first major disturbance of
this kind. The Taira or Minamoto chief who was victorious as commander of a punitive expedition attracted followers to his warrior
band from among the provincial elite. These noble families of imperial descent sought their fortunes by landholding and police actions
in the provinces while maintaining ties with the capital. They were
used as intimidators by Fujiwara regents and senior retired emperors.
When rival factions at court brought mounted warriors of the Minamoto and Taira chiefs into the capital to stage a coup in 1156, the
gates to political power were opened for the bushi. The issue still to
be resolved was which of the network of warrior bands would dominate. That was determined in the Gempei (i.e., Minamoto - Taira)
War of 1180-85, when armies organized by Minamoto noYoritomo
defeated the Taira forces.
At the outset of the war, Yoritomo attracted supporters by proposing a bold plan. On his own authority he confirmed rights to land
and office of warriors in the eastern region who would pledge allegiance to him. By usurping in this way the authority of the imperial
Cambridge Histories Online © Cambridge University Press, 2008
INTRODUCTION
ig
government over land, he guaranteed enduring tenures that the
provincial warrior-gentry class had long sought. As he rose to
military supremacy, he gradually put in place an administrative
organization that supplemented the existing system of provincial
government. From the independent headquarters he maintained in
Kamakura in the east, Yoritomo exercised, with imperial sanction,
broad military and police powers and, through his involvement in
land rights, civil authority as well.
This revolutionary transfer of power brought the Heian period to
an end. The court nobility, together with the imperial line, had monopolized political power and enjoyed all of its material benefits for
more than five centuries since the beginning of the Nara period.
With the end of Heian, the substance of the civil aristocracy's privilege was gone, although as a class it continued to survive in reduced
circumstances. From this time political initiatives were taken only by
bushi rulers. As one warrior regime followed another in the Kamakura (1185-1333), Muromachi (1333-1568), Azuchi-Momoyama
(1568-1600), and Edo (1600-1868) periods, the imperial family and
nobility were sustained on meager allowances by bushi overlords who
continued to relish appointment to court rank and empty ritsuryo titles. Some of the vestiges of court traditions were preserved, along
with the lineages of the Fujiwara and other noble families, to be resurrected following the Meiji Restoration of 1868.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 1
THE HEIAN COURT, 794-1070
KAMMU TO NIMMYO, 7 8 1 - 8 5 0
The man known to history as Emperor Kammu (737-806, r. 781806) was an obscure official in his mid-thirties when the exigencies
of Nara politics catapulted his almost equally obscure father, Konin
(702-82, r. 770-81), onto the imperial throne in 770 and elevated
Kammu himself to the position of crown prince three years later.1 In
781 his father abdicated, and the former director of the state Academy, who may have been passed over originally in the selection of
Konin's heir apparent because of the humble immigrant origins of
his mother's patriline (a Korean-descended lineage, the Yamato),
now found himself installed as Emperor of Japan - a learned peer of
his illiterate contemporary Charlemagne and of the famous Abassid
caliph Harun Al-Rashid at the new capital of Baghdad. Governmental reform and retrenchment, coupled with vigorous action
against Buddhist and secular opponents, had laid a sound basis for
imperial power during Konin's reign, and Kammu quickly demonstrated that he was capable of exercising and enhancing that power.
After efficiently suppressing a plot against the throne in 782 by
high-ranking adherents of the imperial lineage displaced by Konin's
accession (the line of Emperor Temmu), the new emperor and his
advisers apparently decided the time had come to leave the capital,
which was the handiwork of the old imperial line and the stronghold
of both that line's adherents and the Buddhist forces that seem to
have very nearly succeeded in usurping imperial authority in the
time of Konin's immediate predecessor, Empress Shotoku (718-70;
1 Among the surveys of Heian history consulted in the writing of this chapter, one of the most
useful is Inoue Mitsusada, Nagahara Keiji, Kodama Kota, andOkuboToshiaki, eds., Nihon
rekishi taikei, vol. 1: Genshi. Kodai (Tokyo: Yamakawa shuppansha, 1984). See also George
Sansom, A Short Cultural History of Japan, rev. ed. (New York: Appleton-Century, 1943);
Sansom, A History of Japan to 1334 (Stanford, Calif.: Stanford University Press, 1958);
James Murdoch, A History of Japan, vol. 1 (London: Routledge & Kegan Paul, 1949; first
published 1910), and John Whitney Hall, Japan from Prehistory to Modern Times (New York:
Delacorte Press, 1970).
2O
Cambridge Histories Online © Cambridge University Press, 2008
KAMMU TO NIMMYO, 7 8 1 - 8 5 O
21
r. 749-58, 764-70). In the early summer of 784, Kammu directed
that a site for a new capital be surveyed at Nagaoka, an area near the
Katsura River about twenty miles northwest of Nara (Heizei-kyo or
Heij5-ky6) and approximately a half mile northwest of the point
where the Meishin Highway now first crosses the Tokaido Shinkansen south of Kyoto. At the end of the same year, he moved into his
new palace there, his haste to be gone from Nara inspired perhaps
by fear of interference in, or resistance to, the move by forces in the
old capital facing economic loss or ruin because of the court's abandonment of the city.
Nagaoka was in an area associated politically and economically
with Kammu's family line ever since the time of his great-grandfather
Emperor Tenji (626-72, r. 661-72), but it was doubtless particularly
attractive to him as the home of his matriline. (His mother, as already
noted, was aYamato, and her mother was a Korean-descended Haji,
or Hanishi, based in the same general area.) The new imperial seat
was more conveniently located for land and water communication
than the capital at Nara had been, but its cramped and flood-prone
site near a large marsh may have made it unsuitable for long-term occupancy. In any case, a series of inauspicious political events that accompanied the city's founding probably further condemned it in the
eyes of a court that set great store by omens and spirits.2
Fujiwara no Tanetsugu (737-85), although not the senior minister
at court in 784, was Kammu's chief adviser and, apparently, the leading advocate of the transfer of the capital to Nagaoka. (His maternal
family, the Hata, was, like Kammu's, also of Korean lineage and also
based in the Nagaoka area.) He probably owed his influential position at court to his being a nephew of Fujiwara no Momokawa
(732-79), who had almost certainly been instrumental both in the
selection of Konin as successor to Empress Shotoku in 770, and also
in the appointment of Kammu as Konin's heir in 773. Moreover,
Tanetsugu's cousin Otomuro (760-90) was Kammu's consort and
the mother of his eldest son, the future Emperor Heizei (774-824,
r. 806-9). (See Figure 1.1: Genealogy of Heian emperors.)
On a night in the autumn of 785, while Emperor Kammu was temporarily absent from the city, Tanetsugu was killed at Nagaoka. Because of his position it seems likely that the murder was a political
2 On the removal of the imperial seat to Nagaoka, see Ronald P. Toby, "Why Leave Nara?
Kammu and the Transfer of the Capital," Monumenta Nipponica 40 (1985): 331-47. On
floods at Nagaoka, see Hirakawa Minami, "Zoto to seii," in Hashimoto Yoshihiko, Komonjo
no kataru Nihon shi, vol. 2: Heian (Tokyo: Chikuma shobo, 1991), pp. 27-32.
Cambridge Histories Online © Cambridge University Press, 2008
KONIN
[49]
F. Ryoshi ==
(Momokawa) I
JUNNA
SAGA
[53]
[52]
= = Takano Niigasa
(Ototsugu)
KAMMU
[50]
= = Tachibana Kachiko
| (Kiyotomo)
F. Junshi = = NIMMYO
[54]
(Fuyutsugu)
F. Meishi =,
(Yoshifusa)
F. Koshi = =
(Nagara)
== F. Otomuro
I (Yoshitsugu)
HEIZEI
[51]
.= F.Takushi
(Fusatsugu)
MONTOKU
KOKO
[55]
[58]
= = Ps. Hanshi
(P. Nakano)
U D A = = F. Inshi
[59]
(Takafuji)
SEIWA
[56]
F. Onshi = =
(Mototsune) I
YOZEI
[57]
DAIGO
[60]
r
F. Anshi = =
(Morosuke) I
MURAKAMI
SUZAKU
[62]
[61]
1
EN'YU :
[64]
= F. Senshi
(Kaneie)
F. Choshi = =
(Kaneie)
= = F. Shoshi
(Michinaga)
ICHIJO
[66]
REIZEI
[63]
SANJO
= = F. Kaishi
(Koretada)
KAZAN
[67]
[65]
1
F. Kishi = =
GOICHIJO
GOSUZAKU
(Michinaga)
[66]
I
P. Sukehito
(Sanjo)
GOREIZEI
GOSANJO
[70]
[71]
F. Kenshi = =
(Morozane)
F. Ishi = =
(Sanesue) I
Arihito
[Minamoto]
GOSHIRAKAWA
[77]
NUO •
[78]
Key:
= Iki
(Muneto)
SHIRAKAWA
[72]
[73]
==
= = T. Shigeko
Kenshummonin
(Tokinobu)
F. Shokushi = =
= = F. Moshi
(Yoshinobu)
HORIKAWA
F. Shoshi = = T O B A
[74]
Taikemmonin
(Kinzane)
F. Ishi = =
(Tsunezane)
= = Ps.Teishi
[69]
F. Tokushi
Bifukumonin
(Nagazane)
SUTOKU
KONOE
[75]
[76]
TAKAKURA
[80]
= = T . Tokushi
Kenreimonin
(Kiyomori)
ROKUJO
GOTOBA
ANTOKU
[79]
[82]
[81]
Emperor's names are in uppercase type.
Names in parentheses are fathers of the imperial consorts.
Numbers in brackets designate the order of emperors.
Equals signs show marriages or liaisons.
Figure 1.1. Genealogy of Heian emperors.
Cambridge Histories Online © Cambridge University Press, 2008
Abbreviations:
F. Fujiwara
M . Minamoto
P.
Prince
Ps. Princess
T.
Taira
24
TH
E HEIAN COURT, 794-IO7O
act, as the spare account in the chronicles suggests. It could have
been instigated by entrenched interests at Nara opposed to the move
to Nagaoka, by elements excluded from power after the ascendancy
of the Momokawa line of the Fujiwara clan at court (i.e., by members of the Otomo and Saeki clans), and by friction between Tanetsugu and the heir apparent, Kammu's younger full brother, Prince
Sawara (d. 785), who appears to have clashed with Tanetsugu earlier,
and who may have felt that his own position was endangered by
Tanetsugu's familial connection with Heizei. Tanetsugu may in fact
have been looking for an opportunity to depose Sawara and establish
Heizei in his place as crown prince, a step that might well have had
Kammu's backing. In any case, Sawara was soon implicated in the assassination, deposed, and condemned to exile - which he avoided,
the chronicle alleges, by starving himself to death. That outcome
served what may be presumed to have been Kammu's interests - so
well, indeed, that one might suspect the emperor himself of having
manipulated the affair from the outset. However, no evidence supports such a view, and Kammu does not appear otherwise to have
pursued his ends with so ruthless a disregard of his nobles' lives.
Tanetsugu's assassination, Sawara's suicide, and the execution of
the assassins were soon followed by famine, devastating floods, epidemic disease, and a series of deaths and illnesses in Kammu's family, which diviners had no difficulty in interpreting as the revenge of
Sawara's angry spirit. It may have been the inauspiciousness of all
those circumstances, together with a lively fear of what the prince's
spirit might do in the future, that helped Kammu decide in the early
spring of 793 to accept the advice of his longtime confidant Wake
no Kiyomaro (733-99), who, reportedly dismayed by the unfinished
state of Nagaoka after ten years of effort and untold expenditures,
had urged the emperor to seek a new location for the capital in the
Kadono area northeast of Nagaoka. The choice of a site in what is
now the city of Kyoto was soon made, and the fifty-seven-year-old
Kammu moved into his new palace at Heian-kyo in the tenth month
of 794.
The move to Heian marked the end of a temporally long peregrination that had taken Japanese rulers and their courts from one site
to another ever since the inception of the statutory (ritsuryo) system
of government in the seventh century. Like the earlier transfers, the
moves to Nagaoka and Heian doubtless served to strengthen the position of the political leaders at the imperial court, increasing the dependence of the nobility on the government because of the great
Cambridge Histories Online © Cambridge University Press, 2008
KAMMU TO NIMMYO, 7 8 1 - 8 5 0
25
personal cost of the moves; and we can perhaps assume that such
was one of Kammu's chief objectives.3 The move to Heian was both
a continuation of, and a conclusion to, a century and a half of homogeneous historical development - in that sense, an apt symbol for
the reign of the sovereign who presided over the construction of the
new capital and occupied the throne for the first dozen years of its
life. Kammu was the last of a line of puissant, capital-building monarchs who were able to mobilize the entire country's wealth and military power for national or dynastic purposes. After him, the limelight of central political history shifted steadily and rapidly away
from the person of the sovereign toward erstwhile holders of nominally subordinate court posts, roles men whose growing power at
length confined their suzerains to largely ritualistic and ceremonial
functions. By the time death ended the reign of Kammu's grandson
Nimmyo (810-50, r. 833-50), the Northern House of the Fujiwara
clan was well on its way to complete domination of both the emperor
and the organs of his statutory government.
Kammu may have been the most powerful ruler the imperial line
ever produced. He was a mature man and an experienced official
when he came to the throne in 781, trained in the Confucian pedagogic tradition to what appears to have been a sober passion for government. The official chronicler notes that he had no use for "literary floweriness," and he appears to have been equally uninterested
in the kind of extravagant Buddhistic devotion that had nearly bankrupted the state under Shomu and Koken/Shotoku. It is characteristic of him that his reign is remembered chiefly not because of grand
temples, magnificent art, or superlative literature, but for its accomplishments in city building, war, and governmental reform.
Kammu was able to impose his will on the imperial court not only
because of his personal qualities, but also because of circumstances,
partly fortuitous and partly of his own making, that left him relatively free of influence from the old-line high nobility. Political
history until the middle of the eighth century had been characterized by the interaction between, on the one hand, an emperor who
was thought of as possessing absolute authority, and, on the other,
a powerful body of noble clans, based in the Council of State
(Daijokan or Dajokan), who wielded the government's executive authority. But after the downfall of Fujiwara no Nakamaro (706-64) of
3 Sasayama Haruo, ed. (Kodai o kangaeru) Heian no miyako (Tokyo: Yoshikawa kobunkan,
PP- i-5Cambridge Histories Online © Cambridge University Press, 2008
26
THE HEIAN COURT, 794-IO7O
the Southern House (Nanke) in 764 and the reconstitution of the
government under Doky5 (d. 772), men of lesser, "bureaucratic"
clans (Kibi, Ishikawa, Isonokami, etc.) entered the central councils
of government, and the ability of the Council of State to control and
gainsay the emperor was markedly reduced.
Heir to that development, Kammu carried it to its furthest extent. Both he and his father owed their positions to the backing of
Fujiwara leaders - notably Nagate (714-71) of the Northern House
(Hokke) and Momokawa of the Ceremonials House (Shikike) who, in concert with other nobles, seem to have sought a change of
imperial line as a means of checking the decay associated with Empress Koken/Shotoku (in theTemmu line of emperors), a process that
was threatening both their political and their economic well-being.
(See Figure 1.2: Genealogy of the Four Fujiwara Houses.) But by the
early years of Kammu's reign, Momokawa and most of the other influential Fujiwara were dead, and Kammu subsequently saw to it
that a new generation of clan leaders was not given the opportunity
to establish its hegemony at court. He was aided by the enormous
costs to the nobility of the successive moves of the capital, and by
great military campaigns in the northeastern part of the country. His
success in maintaining ascendancy may also have been related to his
lack of blood-kinship ties with the Fujiwara or any of the other leading clans; and it must have been aided, too, by his skillful creation of
a small, privileged group of supporters tied to him by kinship, by
marital alliances, and by the large land grants he had begun making
to his favorites as early as 793. (He was particularly generous in his
treatment of relatives of his mother and grandmother.)
That Kammu enjoyed substantial freedom from clan control may
be inferred from the fact that leading governmental offices were either left vacant or entrusted to imperial family members during
much of his reign. The post of Minister of the Left, the highest
regularly filled office in the court government, remained vacant
throughout all but the first year of Kammu's rule; after the death in
796 of Tsugutada (the Southern House Fujiwara who may have
been the chief promoter of the move to Heian), the post of Minister
of the Right was similarly vacant until 798 and was then occupied
until the end of the reign by Kammu's cousin Prince Miwa
(737?-8o6); the post of Palace Minister was vacant at Kammu's accession and never thereafter filled; and the post of Major Counselor,
the fourth highest in the government, was held from 796 on by another of Kammu's cousins, who shared it with Ki no Kosami
Cambridge Histories Online © Cambridge University Press, 2008
Fujiwara no Kamatari
614-669
I
Fuhito
659-720
I
I
I
Takatoshi
761-808
I
I
Otomaro
Nakamaro
(also known as
Emi no Oshikatsu)
706-764
I
I
Kosemaro Sadatsugu
Torikai
Ioe no Iratsume
I
Fusasaki
681-737
Northern House
(Hokke)
Muchimaro
680-737
Southern House
(Nanke)
Ibyonari
704-765
I
Fsugutada
727-796
I
Higami no Iratsume
I
Umakai
d737
Ceremonials House
(Shikike)
Komyoshi
Maro
701-760
695-737
Capital House
(Kyoke)
Miyako
d754
I
T
Matate
Uona
Nagate
715-766 721-783
714-771
Hamanari
Hirotsugu
d740
1
Yoshitsugu Kiyonari
716-777
Momokawa
732-779
I
Kurajimaro
Toyohiko
Oguromaru
733-794
Kadonomaro
755-818
Uchimaro
756-812
I
Onatsu
I
Manatsu
Otomuro
760-790
Fuyutsugu
775-826
Arachi
787-843
Tanetsugu
737-785
Nakanari
774-810
Otsugu
774-843
Tabiko
759-788
Kusuko
d810
Figure 1.2. Genealogy of the four Fujiwara houses. Adapted from Kodansha Encyclopedia of Japan.
Cambridge Histories Online © Cambridge University Press, 2008
Fuyuo
28
THE HEIAN COURT, 794-IO7O
(733-97) until 797. There was not a single Fujiwara officeholder
above the level of Middle Counselor during the final decade of the
reign (796-806). Although influential lower officials restricted the
emperor's role to something considerably less than that of an absolute despot, it seems, nevertheless, to have been true that the final
decision in major matters was his, both in name and in fact.
Under Kammu's domination, the imperial court addressed itself
to the political, social, and economic problems of the day with a
vigor and dedication that did credit to the emperor's Confucian
training. Among those problems, none was more pressing than the
security of the country's northeastern frontier.
At the time of Kammu's accession to the throne in 781, the Japanese government had reached a crisis in its relations with a nonsubject people who lived in the general area of Honshu now known as
Fukushima and Niigata prefectures, and also farther north. Little is
known with certainty about those people: even their name is a matter of dispute. The logographic orthography commonly used in reference to them until around the beginning of the ninth century is
generally thought to have been read as both "Emishi" (see GlossaryIndex) and "Ebisu" (also "Ezo" after the mid-Heian period), but
other readings ("Kai" and "Ainu") are still advocated in scholarly
circles. After 800 the orthography itself was replaced by others, especially "Fushu" (literally, "war captives"), a term that presumably
reflected the Japanese view of the Emishi after their subjugation
early in the ninth century.4 Nor do we know anything more of Emishi physical characteristics than that the mighty T'ang emperor
Kao-tsung (r. 649-83) was surprised by the "oddity" of the bodies
and faces of two introduced to him by a Japanese embassy in 659
(but "odd" in what way?), that both women and men did up their
hair in the shape of a "small mallet," and that men's beards might be
four or more feet long. The Emishi may, in fact, have been quite hirsute, as the frequent occurrence of the epithet "hairy people" suggests. On the other hand, that usage, it is suspected by some, may
have been simply an acknowledgment of an ancient Chinese notion
that the realm of the "hairy people" was in the northeast, which in
Japan's case was the direction of the Emishi homeland. Alternatively,
"hairy" may have had reference to their fur clothing.
Culturally, some Emishi were clearly hunters and gatherers; but
4 For a survey of the problems surrounding the name of the Emishi, see ArakiYoichiro, "Emishi no kosho, hyoki o meguru shomondai," Kokushi kenkyu (Hirosaki Daigaku) 87 (1989).
Cambridge Histories Online © Cambridge University Press, 2008
KAMMU TO NIMMYO, 7 8 1 - 8 5 0
2Q
others were agriculturalists at least as early as the eighth century and
possibly from Yayoi times (roughly 300 B.C.E.-300 C.E.). 5 For the
Japanese, however, their most notable cultural characteristic may
have been their fierce, brutish nature. This attitude seems to underlie an imperial order, quoted in the Nihon shoki (the first official
history of 720), which dispatched the legendary Yamatotakeru no
Mikoto against the Emishi during the reign of Keiko (traditionally,
r. 71-130). In its description of Emishi character and custom the
order is an almost verbatim reproduction of Chinese descriptions of
barbarian people. But though it is doubtless anachronistic or wholly
apocryphal, its contents may be taken as an expression of a view
commonly accepted at court in the latter half of the seventh century - one perhaps discernible even today in histories accounting for
the distinctive "brutality" (zangyaku) of Japan's eastern warriors by
reference to their experiences in battles against these foes:6
I hear that the eastern outlanders are by nature fierce and wild, and that
their chief interest is violent assault. Their villages lack chiefs, their settlements lack heads. Coveting territory, they all rob each other. Further, there
are evil deities in the mountains and perverse devils on the plains. They obstruct passage at intersections and block the roads, causing great affliction
on people.
The fiercest of those eastern outlanders are the Emishi. Men and
women live mixed together, nor is there distinction between father and
son. In winter they lodge in holes; in summer they dwell in nests. They
wear furs and drink blood; eldest and younger brothers are distrustful of
each other. Climbing mountains they are like flying birds; running through
grass they are like furry beasts in flight. They forget benefits received but
always requite wrongs suffered and for that purpose conceal an arrow in
their topknots and wear a sword under their robes. Sometimes they band
together and invade the border regions; sometimes they spy out opportune
times in agriculture and sericulture and rob the people. If attacked, they
hide in the grass; if pursued, they go into the mountains. Consequently,
they have never since ancient times been subject to the transforming royal
influence.7
The Emishi lived for the most part in tribal groups, but Japanese
cultural and political influences, which had penetrated the Fuku5 Takahashi Takashi, Emishi: kodai tohokujin no rekishi, Chufeo shinsho 804 (Tokyo: Chuo
koronsha, 1986), pp. 7-8.
6 See, for example, Inoue et al., eds., Nihon rekishi taikei, p. 857.
7 Nihon shoki, ed. Sakamoto Taro, Ienaga Saburo, Inoue Mitsusada, and Ono Susumu, vol. 67
of Nihon koten bungaku taikei (Tokyo: Iwanami shoten, 1967), p. 302; passage cited in Takahashi, Emishi, p. 9; translated by W. G. Aston, Nihonqi: Chronicles of Japan from the Earliest
Times toA.D. 697, reprint of 1896 ed. (London: George Allen & Unwin, 1956), part 1, p. 203
(translation revised).
Cambridge Histories Online © Cambridge University Press, 2008
30
THE HEIAN COURT, 794-IO7O
shima area as early as the fourth century,8 impelled some of them to
become agriculturalists, with leaders who might hold appointments
as local officials of the statutory regime. As the Japanese pushed their
administration and settlers farther northward and eastward after the
Taika Reforms of the mid-seventh century, relations with the Emishi
deteriorated, and there were sporadic eruptions of violence - isolated
attacks of one upon the other, killings, and warfare. But it was not
until 780, the year before Kammu's accession to the throne, that matters came to a head. In that year, just as Japanese forces were preparing to attack the Emishi stronghold of Isawa (about fifteen miles due
north of modern Hiraizumi in Iwate Prefecture), they were confronted by the revolt of a powerful Emishi ally named Koreharu no
Azamaro, who held a Japanese provincial title. Their leaders were
slain, and in the ensuing turmoil the chief military seat of the region,
Taga Fort, was taken by Emishi and put to the torch. The Japanese
then found themselves confronted by a sudden dangerous threat to
their presence in the northeastern territories.9
Azamaro's revolt in 780 was the beginning of more than thirty
years of large-scale warfare between the Emishi and the Japanese.
Major Japanese military campaigns against the Emishi in 780-81,
788-89, and 794 met with either defeat or inconclusive success. Not
until 801-2 were Japanese forces under the command of the redvisaged, yellow-bearded warrior "giant" Sakanoue no Tamuramaro
(758-811) able to defeat the main Emishi leader, Aterui, and permanently occupy and garrison his base at Isawa.10 A relatively limited action under Fun'ya no Watamaro (765-823) against two Emishi areas, in what is now the eastern part of Iwate Prefecture and
along the Aomori-Iwate border, brought the Emishi Wars to a tentative conclusion in 811, and in the following year most of the Japanese armies in the northeast were demobilized.
Although the back of armed Emishi resistance to Japanese encroachment was broken during the three decades of fighting from
8 A century or two earlier than commonly believed until recently. See "Kokogaku: kono ichinen," Yomiuri shimbun, 22 December 1990.
9 On the vexing question of the reading of Azamaro's surname, which has commonly been
read "Iji," see Inoue, Nihon rekishi taikei, p. 692, n. 2.
10 At about five feet ten inches, Tamuramaro was six inches taller than the inferred average
height of Japanese men during the Ancient-Tomb (kofun) period - i. e., from the end of the
third or the beginning of the fourth to the mid-sixth century, - and he may well have
looked like a giant to his contemporaries. His ancestors are said to have come from China.
A short, perhaps nearly contemporaneous, biography of him has been preserved: Tamuramaro denki, in (Shinko) Gunsho ruiju, Hanawa Hokiichi, comp. (Tokyo: Meicho fukyiikai,
1977), kan 64, vol. 3, p. 699. On die heights of early Japanese, see Joseishi Sogo Kenkyukai,
Nihon josei shi, vol. 1: Genshi. Kodai (Tokyo: Tokyo daigaku shuppankai, 1982), p. 15.
Cambridge Histories Online © Cambridge University Press, 2008
KAMMU TO NIMMYO, 7 8 1 - 8 5 0
31
780 to 811, the conflict was by no means finally resolved. Revolts by
Emishi against Japanese authority in Dewa Province in 878-79 and
939, the threat of further revolt, and outbreaks of fighting among the
Emishi themselves kept the area in an almost continuous state of
alarm and tension during the following century and a half. Then, as
central government control weakened and the assimilation of the
Emishi proceeded, with Emishi leaders coming to play a growing
role in local government, the conflict was internalized in Japanese
society, erupting again in open warfare during the Earlier Nine
Years' War of 1051-62 and the Later Three Years' War of 1083-87,
when Emishi-descended leaders sought to assert their independence
of the central government (see Chapter 10). The area subsequently
developed into the private satrapy of a local Fujiwara line that had
intermarried with an Emishi family, remaining independent of central control until the time of Minamoto no Yoritomo's conquest in
1189, more than four centuries after Azamaro's revolt in 780. By that
time the unassimilated Emishi were confined generally to the northernmost tip of Honshu and to Hokkaido.
Japanese penetration of the northeastern part of Honshu was
marked by the establishment of forts (Jo) and palisades (saku)
throughout the area as the line of colonization and conquest moved
east and north. The aim was not merely to provide strongholds for
defense against the Emishi but also to create administrative centers
through which the central government could exercise control over
the Japanese colonists who were sent en masse to the region from
the Kanto and central Honshu. In 796, for example, 9,000 people
from Sagami, Musashi, and seven other provinces are said to have
been moved to the Iji Fort region (in what is now Miyagi Prefecture)
and established as "palisade households" {sakukd) to cultivate the
land there.
As both military and administrative centers, the forts and palisades at least sometimes rivaled the administrative seats of provinces (kokufu) in their size and complexity, as archaeological excavation has shown. The site of the famous Taga Fort, for instance, was
a square nearly 3,000 feet on each side surrounded by an earthen
wall over two miles long. Several administrative buildings on an elevation at the center of the site were enclosed within their own
earthen wall, which measured 330 feet east to west and 390 feet
north to south. Elsewhere within the site were other groups of buildings, including storehouses and what are thought to have been quarters for artisans and soldiers. Five distinct periods in the history of
Cambridge Histories Online © Cambridge University Press, 2008
32
THE HEIAN COURT, 794-IO7O
the site can be identified, from the founding of the fort in the first
half of the eighth century through the Heian period.
Although the Emishi Wars consolidated and extended Japanese
control over the northern part of Honshu, their enormous cost must
have also contributed significantly to the weakening of the statutory
regime. Even allowing for exaggeration by the chroniclers, the
armies mustered for the campaigns were huge in relation to the contemporary population and economy, numbering in one instance, it
is said, as many as 100,000 men. The creation and supply of such
forces placed what ultimately proved to be an intolerable burden on
the resources of the government, both in direct expenses and presumably also in the removal of men from productive employment,
and in 805 Kammu decided to terminate the campaigns specifically
because of their cost. Fujiwara no Otsugu (774-843), whose recommendation it was that brought the large-scale campaigns to an end,
asserted that the wars against the Emishi and construction work in
the new capital at Kyoto were the two chief causes of distress among
the peasantry.
The wars and their aftermath were also apparently a stimulus to
the growth of private warrior forces in the eastern provinces of Honshu, whose men had borne the chief burden of the fighting, and they
further affected the social fabric of Japan as a result of the forced resettlement of captured and subjugated Emishi in almost every area
of the main islands. Although it is not possible to estimate the number of Emishi involved in the resettlement, it is clear that a major
shift of population was involved, undertaken almost certainly to remove rebellious and unassimilated Emishi elements from their still
restless homelands. Some of the transported Emishi became slaves
attached to noble households or government offices, but most apparently were forced to live in their own special communities, supported at least in part by governmental subsidies and experiencing
the problems familiar to minority peoples everywhere submerged in
a hostile dominant culture. Some communities quickly died out,
others were involved in armed clashes with Japanese neighbors and
rebellions against the government, and still others seem to have
moved fairly expeditiously toward full assimilation to Japanese society. The warlike skills of the resettled Emishi were put to good use in
Kyushu, where they were employed in coastal defense, and it is possible that Emishi also became ancestors of warrior families elsewhere in Japan.
The Emishi wars and the construction of the capital were made
Cambridge Histories Online © Cambridge University Press, 2008
KAMMU TO NIMMYO, 7 8 1 - 8 5 0
33
economically possible by the close attention that Kammu gave to the
government's revenue base in the countryside, which is discussed
later in the chapter.
After the death of Kammu in the spring of 806, the leading noble
clans, and especially the Fujiwara, began almost immediately to
move again into the highest offices of government, but the balance
of power at court until about the middle of the century seems to
have consistently favored the emperors, who thus continued to
maintain at least the semblance of rule in their own right. Although
Kammu's three immediate successors on the throne, his sons Heizei,
Saga (786-842, r. 809-23), and Junna (786-840, r. 823-33), each had
a Fujiwara mother, they were able during their reigns to steer relatively independent courses, thanks at least in part apparently to rivalries within the Fujiwara clan itself. It was only after the death of
the retired emperor Saga in 842 that the exercise of court power
began to shift back and forth between emperors and Fujiwara leaders, initiating the process that ended with the firm establishment of
Fujiwara ascendancy in the first half of the tenth century.
Insofar as imperial rule was maintained during the three or so
decades following Kammu's death, the chief credit is usually given
to Saga, who appears to have inherited his father's erudition and also
his skill in government. The reign of Saga's predecessor, his sickly
and possibly neurotic full brother Heizei, does not seem to have
been promising from that point of view.
Even before Heizei's accession in 806, he is said to have established a scandalous liaison with his consort's mother, Fujiwara no
Kusuko (d. 810), a daughter of the Tanetsugu who was assassinated
at Nagaoka in 785, and after he became emperor he seems to have
been much guided by her and her brother Nakanari (774?-8io).The
official chronicle says, in fact, that Heizei was completely under
Kusuko's sway, although that may be no more than another case of
the misogynistic scapegoating common in Chinese and Chinesederived historiography. The unfavorable view often taken of Heizei
may be mistaken. His reign was actually notable for its attention to
government and finances, so attentive, in fact, that the emperor may
have provoked the hostility of his nobility. Saga's illustrious reputation, by contrast, seems based more on his literary and cultural accomplishments than on his actual effectiveness as a ruler.
Nakanari and Kusuko may have been behind the most dramatic
event of Heizei's brief reign, the forced suicides in 807 of the emperor's younger half brother, Prince Iyo, and the prince's SouthernCambridge Histories Online © Cambridge University Press, 2008
34
THE HEIAN COURT, 7 9 4 - 1 0 7 0
House mother following their imprisonment on the probably false
charge of plotting rebellion against the government. The incident became the occasion for a purge of the leading members of the Southern House of the Fujiwara clan from the influential positions in government they had held since the days of Emperor Kammu, the
pretext for the purge being that the Southern-House men were
closely related to Prince Iyo through his mother. Since the purge removed from the court important rivals of Nakanari and Kusuko who
belonged to the Ceremonials House of the clan, it is plausible to
suppose, as the official chronicle seems to suggest, that the brother
and sister were responsible for the charges against Prince Iyo and the
implication of his maternal relatives. However, the incident may
have also been inspired at least in part by personal animosities of the
violent-tempered Heizei. The emperor is said to have long been resentful of the marked favor his father, Kammu, had shown Prince
Iyo and his mother during the last half of his reign. Heizei is also
known to have borne a particular grudge against one of the purged
Southern-House officials, Takatoshi (761-808), who, while Heizei
was still crown prince, had treated him with less than the respect the
future emperor felt was his due.
If Nakanari and Kusuko were involved in the downfall of Iyo and
his kin, their maneuvering gained them very little in the end. Just
two years later Emperor Heizei was forced by ill health to abdicate
in favor of Saga, leaving Nakanari and Kusuko isolated in a government whose chief minister was Fujiwara no Uchimaro (756-812),
the head of the clan's Northern House. Heizei withdrew to the old
capital at Nara, but in the following year, 810, he began issuing governmental orders again in his own name, decreeing finally in the late
fall of the year that the capital was to be moved back to Nara. The
step was seen as a move on the part of the ex-emperor to reclaim
the throne, and Saga moved swiftly against the pair held responsible
for his actions. Kusuko was relieved of her court rank and ordered
expelled from Heizei's palace, and her brother Nakanari was arrested and demoted to a provincial office. Heizei attempted to make
his way with Kusuko to the eastern provinces to raise an army in his
defense there, but his route was blocked by government forces, and
he returned to Nara and took Buddhist holy orders. Nakanari was
executed, his sister Kusuko committed suicide, and the crown
prince, a son of Heizei, was replaced by Junna, a half brother of Saga
and Heizei. Removed from the influence of Nakanari and Kusuko,
Cambridge Histories Online © Cambridge University Press, 2008
KAMMU TO NIMMYO, 7 8 1 - 8 5 O
35
the chastened ex-emperor thereafter led a pious, docile life until his
death fourteen years later.
Following the deaths of Kusuko and Nakanari and the hasty ordination of Heizei in 810, the imperial court enjoyed a thirty-year period of relative peace and stability, undisturbed by great enterprises,
which had taxed the regime to its limits under Kammu, or by the
kind of political and dynastic rivalries that had sent Prince Iyo to his
death under Heizei. The Fujiwara, as always, posed a threat to the
authority and independence of the emperors, but Saga and Junna
managed to avoid marital ties with that powerful clan, thus limiting
its direct, personal influence on the throne. They also succeeded in
maintaining a balance of competing families and clans in the higher
ranks and offices of the government. For example, of the thirteen
men who occupied the highest government offices in 841, the year
before Saga's death, only four were Fujiwara, two from the Ceremonials line (one of whom was the ranking court officer, the Minister
of the Left) and two from the Northern. Equally important to note
is that among that select group of thirteen were four members of the
imperial clan or descendants thereof (Minamoto), and four men
from lower-ranking families who were known chiefly for their sinitic
learning or administrative abilities. The presence of learned men and
practical administrators in the highest offices of government not
only is a clear reflection of Saga's well-known interest in, and emphasis on, Chinese learning and Chinese conceptions of government, but is also evidence of a desire to maintain a group of leaders
in the government amenable to imperial direction.
During Saga's time the personal material fortunes of the emperor,
of his family, and of the privileged group of nobles closest to him
began to acquire a generous and independent foundation. This was
the result of large-scale grants to them of agricultural lands developed or reclaimed by provincial governments and worked with the
labor of local farmers. In the case of emperors and retired emperors,
the lands so granted, known for historical reasons as "later close"
(goiri), grew into a substantial patrimony that became a chief support for the reassertion of imperial rule (chiefly in the person of the
retired emperor) toward the end of the eleventh century.
Although the political energies of the court during the time of
Saga and Junna were expended for the most part on efforts to adapt
the institutions of the statutory regime to changing social and economic circumstances, their success was limited. Even that limited
Cambridge Histories Online © Cambridge University Press, 2008
36
THE HEIAN COURT, 794-IO7O
success was placed in jeopardy just one week after Saga's death in
842, when the crown prince was deposed and replaced by a nephew
of Fujiwara no Yoshifusa (804-72), the head of the Northern House.
The events surrounding the crown prince's removal from office,
which are sometimes referred to collectively as the Jowa Incident
(Jowa no hen) after the name of the calendar era in which they occurred, marked the beginning of the end of imperial rule in Japan.
The emperor remained an important and sometimes controlling
force at court into the tenth century, but the main focus of political
interest after 842 is the rise to power of the Northern House of the
Fujiwara.
At the time of Saga's death, there appear to have been two rival
groups at court, one centered around the emperor, Saga's gifted son
Nimmyo, and the other around the crown prince, Tsunesada
(825-84), Junna's son by a daughter of Saga. The core of Nimmy5's
party was the emperor himself; his mother, Saga's consort Tachibana no Kachiko (786-850); his son, the fifteen-year-old future Emperor Montoku (827-58; r. 850-58); and Montoku's uncle Fujiwara
no Yoshifusa. Most important among Tsunesada's supporters were
the officials in his household office, especially Tomo no Kowamine
and Tachibana no Kachiko's cousin Hayanari (d. 842), and his
father-in-law, Yoshifusa's uncle Arachi (787-843). It is noteworthy
that the rival groups were not defined by family or clan lines: it was
Tachibana (Kachiko) against Tachibana (Hayanari), Northern
House (Yoshifusa) against Northern House (Arachi), and imperial
grandmother (Kachiko) and uncle (Nimmyo) against grandchild
and nephew (Tsunesada).
The situation was ripe for conflict. Nimmyo, who doubtless
wished to see his own son succeed him on the throne, was no longer
restrained by the wishes of his father, and Tsunesada, whose position
had been greatly weakened by the loss of his own father (Junna) two
years earlier, now found himself in an even more precarious position
as a result of the death of his last protector, his grandfather Saga.
Furthermore, and perhaps even more to the point, Nimmyo's son
Montoku was the nephew of Fujiwara no Yoshifusa, who not only
could expect to prosper greatly upon Montoku's accession to the
throne but who may have also feared that the leadership of the
Northern House would permanently pass to the line of his uncle
Arachi if Tsunesada became emperor. The crown prince was married to Arachi's daughter, and his accession would almost certainly
have given Arachi control of the court.
Cambridge Histories Online © Cambridge University Press, 2008
EVOLUTION OF STATUTORY GOVERNMENT
37
It seems more likely on the whole that the ambitions and fears of
the Nimmyo party lay behind the events of 842, but according to the
chronicles, it was Tsunesada and his supporters who actually precipitated the upheaval by plotting a coup aimed at deposing Nimmyd
and placing Tsunesada on the throne. The plot was reported to
Nimmyo's mother, Kachiko, on the very next day after Saga's death,
and she immediately informed Yoshifusa, indicating perhaps that he
was already by that time the leading figure at court despite his junior
position (he was a Middle Counselor, outranked by three other men
not associated with the plot). The government moved quickly
against the plotters, arresting and exiling the alleged ringleaders,
Tomo no Kowamine and Tachibana no Hayanari, dismissing from
office or demoting some sixty officials connected with Tsunesada,
including Arachi, and deposing Tsunesada himself. A few days later
Nimmyo's son was appointed crown prince, and the Northern
House of the Fujiwara under Yoshifusa stood on the threshold of the
period of its greatest power.
EVOLUTION OF THE STATUTORY GOVERNMENT
The statutory regime from its outset underwent numerous modifications in the details of its operations and structure as governmental leaders sought to correct past failings and respond to new needs,
opportunities, and pressures, but it was particularly in the ninth century, when the regime began to falter badly because of changing social, economic, and political conditions, that major alterations of the
governmental structure itself took place. The chief problems seem to
have been declining, or at least less adequate, revenues and a concomitant narrowing scope of political action. The main solutions attempted were retrenchment and a simplification of the governmental machinery, the operations of which were less and less concerned
with nationwide projects and policies and increasingly concentrated
on the capital area and the court itself.
It is difficult to pinpoint with certainty the reasons for the central
government's financial difficulties. Although overspending on the
construction of new capitals and on wars against the Emishi during
the reign of Kammu undoubtedly contributed to the problem, that
may not have been the basic cause. The decay of the state system of
allocated rice land, accompanied in some cases by the abandonment
of capitatim rice tillages (kubunden), and the growth of various kinds
of private landed proprietorships or estates {shoen, "rural garths") enCambridge Histories Online © Cambridge University Press, 2008
38
THE HEIAN COURT, 794-IO7O
joying a degree of tax immunity (legal or otherwise) must have also
contributed to the government's straitened circumstances. But perhaps most important was the emergence in the provinces of a powerful and wealthy class that the central authorities were unable fully to
control or to exploit. That class, whose members frequently held appointments in the district and provincial administrations, appears to
have absorbed an ever-growing share of the country's wealth.
Whatever the reasons, the central government during the first century of the Heian period appears clearly to have been faced with a
worsening financial situation, to which it responded in two chief
ways. On the one hand, it sought to adjust the tax structure to the
changed conditions of the countryside, hoping thereby no doubt to
preserve such sources of income as were still available to it; and, at
the same time, it attempted to improve its control over provincial
governments, where the largest losses in income were occurring.
There was thus during the reigns of Kammu, Heizei, Saga, and
Junna - that is, from the beginning of the period to 833 - an almost
constant tinkering with the tax laws and unremitting attention to
provincial government, especially to the means of ensuring provincial compliance with central law. But despite all efforts, revenues appear to have continued to decline in quality and quantity (they were
mostly rendered in kind), and the central government's control over
the provinces steadily weakened.
At the same time that the government was seeking to avoid further
erosion of its fiscal base, it was also taking various measures to reduce expenditures. The decisions to terminate the campaigns against
the Emishi and to halt construction work at the capital in 805 were
evidently motivated by economic concerns, as noted earlier, and
similar considerations probably lay behind efforts to reduce the size
of the central government that continued from the time of Konin to
the beginning of the tenth century. During the time of Konin and
Kammu, government retrenchment policy was concentrated on
eliminating special posts and offices that had proliferated outside
the framework established by law, but in the reigns of Kammu's sons
(Heizei, Saga, and Junna) an attack on unneeded offices and posts
within the statutory structure itself began. By the end of the ninth
century, approximately half of the central offices originally provided
for by statute had been abolished. The resulting savings were offset
to some extent, it is true, by a simultaneous increase in the number
of supernumerary officials, but since supernumerary positions appear to have been restricted to the more senior titles, there was likely
Cambridge Histories Online © Cambridge University Press, 2008
EVOLUTION OF STATUTORY GOVERNMENT
39
overall a substantial reduction in the size of Heian officialdom. The
result presumably was not only economic but also social. Displaced
officials and disappointed aspirants to such status were forced to
look elsewhere for a living, and some, at least, must have sought their
fortunes outside Kyoto, where they would have found and no doubt
frequently joined the provincial gentry.
Even the imperial clan itself was not immune to the budget cutter's blade. By the beginning of the Heian period the clan had greatly
increased in size, including by legal definition descendants of emperors down to the fifth, and in some instances the sixth, generation.
The generous economic treatment of imperial princes and princesses provided for by statute placed a heavy burden on the government's shrinking revenues, and even as early as Kammu's reign steps
to reduce the size of the clan began to be taken. In 798 membership
was redefined to exclude imperial descendants in the fifth generation
and after (thus restoring the original statutory definition of the
clan), and in 805, the same year in which the Emishi campaigns and
construction at Heian were halted, more than one hundred princes
and princesses were reduced to noble status and given clan names,
thus removing them from the imperial family. Individual princes had
been similarly removed from the family in the Nara period, and
Kammu himself had earlier taken the unprecedented step of reducing sons of emperors (one of Konin and two of his own) to noble status, but the action in 805 was the first instance of wholesale exclusion. Emperor Saga continued his father's policy, in 814 going so far
as to reduce all of his numerous children except those by higherranking consorts to noble status with the bestowed clan name of Minamoto, an example that was followed by succeeding emperors
down toYozei (869-949, r. 876-84). Second- or later-generation descendants of emperors who were reduced to noble status in Saga's
time and after seem regularly to have received clan names other than
Minamoto, such as Taira and Ariwara, which conferred less prestige
on their holders than Minamoto. The last grant of noble status to
imperial clan members occurred in the reign of Murakami (926-67,
r. 946-67). Thereafter the absence of remunerative bureaucratic
openings for demoted princes and the smaller number of imperial
children that seems to have resulted from Fujiwara control of the
emperor and his harem probably made the practice mostly unnecessary. When an excess of imperial children did occur after the tenth
century, Buddhist holy orders provided a convenient, dignified
refuge.
Cambridge Histories Online © Cambridge University Press, 2008
40
THE HEIAN COURT, 794-IO7O
Although the growth of the imperial clan may have been checked
by the measures adopted by Kammu and his successors, it was not
necessarily stopped. In the latter half of the ninth century, at any
rate, the clan still numbered more than five hundred members who
had a claim on state revenues, a burden so heavy that the government felt obliged to reduce the level of its support, leaving some
peripheral members of the clan in what appears to have been dire
straits. It was those poorer imperial relatives, presumably, who joined
their Minamoto andTaira kindred in the exodus to the provinces that
was enlarging and strengthening the gentry class there.
At the same time that the government was seeking to retrench and
preserve its fiscal base, it was also attempting to reshape the central
governmental structure in order to meet the changing conditions of
the day. The concern of the emperors and their courts seems to have
been increasingly restricted to the control and support of the internal workings of the court itself. There was consequently an overall
tendency in government toward simplification of operations, with
focus on the person of the emperor, and a shift toward dependence
of emperors and governmental offices alike on revenues from specific pieces of land that were taking on some of the characteristics of
private proprietorships. The process is sometimes thought of as the
partial conversion of a semibureaucratic regime supported by general tax revenues to a more personal type of rule by emperors and
their close associates dependent on private or individualized sources
of income.
At the very beginning of the Heian period, the major business of
government seems still to have been conducted as prescribed in the
statutory code, the emperor and his ministers gathering each morning to attend to business in governmental councils (chbsei) at the
Court of Government (Chodoin), a large complex of buildings just
inside the Gate of the Vermilion Sparrow (Suzakumon), the main
entrance gate of the Greater Imperial Palace on its south-central side
(on the imperial palace and its structures, see Chapter 2). By the beginning of Saga's reign, however, the business of such councils may
have already so dwindled in quantity and scope that they were becoming anachronistic in any case. When the emperor established a
new office at the time of the Kusuko Incident in 810 that gave him
more direct control over the bureaucracy, the councils rapidly withered away, remaining then simply as biannual court rituals.
The new governmental organ created by Saga in or about 810 was
probably designed to help the emperor cope with Heizei's attempt to
Cambridge Histories Online © Cambridge University Press, 2008
EVOLUTION OF STATUTORY GOVERNMENT
41
regain the throne. Called the Chamberlains' Office (kurvdodokoro,
"Repositors' Office"), it was placed in charge of the emperor's personal storehouse, which was within the compound of his residential
palace and included the imperial archives. Initially, the chief duties
of the office apparently were to handle legal disputes, to ensure the
confidentiality and integrity of imperial documents, and possibly
also to secure the personal safety of the emperor. But its duties and
officers rapidly expanded in number and importance during the
reigns of Koko (r. 884-87) and Uda (r. 887-97), when it became a
veritable pivot of court government that Uda was able to employ in
his attempt to check Fujiwara power, at least until the clan leaders
gained control of that office, too. Staffed by men who were personally close to the emperor (they had often served in his household office when he was crown prince), the office took over many of the
functions of the Council of State relating to communication with the
emperor and the issuance of imperial orders, decrees, and rescripts,
sidestepping thereby the cumbersome mechanisms of the Council
and giving the emperor better access to the operating offices of the
government. At the same time, the office also took charge of the administration of many of the emperor's personal or household affairs,
supervising functions at the residential palace and the daily activity
of the imperial audience chamber, providing for the emperor's personal needs, and otherwise generally functioning in the capacity of a
steward's office.
The appearance and evolution of the Chamberlains' Office into
one of the chief instruments by which the emperor exercised control
over the court and the government reflected the atrophy of governmental activity in the ninth century and a shift of political power
away from the public, semibureaucratic institutions of the statutory
regime toward a more personal kind of rule that depended chiefly on
the emperor and a narrow circle of his intimates. The joining of state
and household functions in the office, clearly signaled by the location of the office itself at the emperor's residence, further blurred the
distinction between the imperial position and the imperial person,
thus perhaps rendering the court more easily controllable by those
with personal ties to the emperor. Although the office was an important expression of imperial determination to rule as well as reign,
in the end it may have contributed significantly to the success of the
Northern Fujiwara House in establishing its supremacy at court.
After the near disappearance of the daily governmental councils
(chosei) and the establishment of the Chamberlains' Office, the emCambridge Histories Online © Cambridge University Press, 2008
42
THE HEIAN COURT, 794-IO7O
peror seldom met with the ministers of the Council of State to transact court business, but the Council itself continued to fulfill most of
its original functions as the highest organ of government, advising
the emperor, recommending policies, issuing directives to subordinate offices in at least nominal compliance with imperial orders, and
in general directing and supervising the operations of the regime.
Major policies and issues now were discussed mostly not with the
emperor (or his regent), but in meetings of the ministers themselves.
These evolved through various arrangements and places of meeting,
but from the last half of the ninth century increasingly took the form
of what were usually called "Guard-Post Judgments" (jinnosadame).
Following the general shift of government toward the emperor's residential palace, Guard-Post Judgments were held at one of the posts
(Left or Right) of the Imperial Guards located in the residential
compound. The choice of meeting site may have reflected a concern
for security, the palace itself by that time having become a prime target for robbers and arsonists. The meetings were usually convened
by the emperor or regent two or three times a month through the
Chamberlains' Office to consider specific problems. The decisions of
the meetings, made by the senior minister present after discussion
with the others, were reported back through the Chamberlains' Office as recommendations to the emperor or regent, who then caused
such orders as were necessary to be issued. The business of the
meetings covered the range of court concerns: imperial accessions,
foreign relations, military uprisings, appointments to court rank and
office, Buddhist and Shinto rites, tax matters, provincial government, and so on. Despite the eventual domination of court government by the Fujiwara regents, Guard-Post Judgments remained a
vital instrument of court decision-making into the latter half of the
eleventh century.
A similar shrinkage or rationalization of structure also occurred in
the military, police, and judicial organs of government during the
first century or so of the Heian period. The military conscript system provided for under the statutory code had already been recognized as ineffective by the end of the Nara period, and it was probably nearly moribund when in 792 Kammu formally abolished the
system outside Kyushu, the northern part of Honshu, and the island
of Sado. The dismantlement of the conscript system came in the
midst of the campaigns against the Emishi, suggesting that these
campaigns may have relied less on regular peasant conscripts than
on the specially mobilized forces of vagrants and elitefightingmen
Cambridge Histories Online © Cambridge University Press, 2008
EVOLUTION OF STATUTORY GOVERNMENT
43
mentioned in the official chronicles. In the provinces where the conscript system was abolished, Kammu established small bands of
armed men (called "stalwart youths," kondei), selected from among
the families of district officials to serve on a rotating basis as guards
at provincial administrative seats and other governmental installations in the area. They were few in number, however, and in themselves could not have constituted a significant military force. When
military action was undertaken, the government relied instead on
the private forces that were developing in the provinces and that
eventually usurped the authority of the imperial court (on that subject, see Chapter 10).
In the aftermath of the Emishi campaigns and the Kusuko Incident, the safety of the capital itself seems to have become more precarious than ever: unrest, violence, and rumors of rebellion kept the
city in a more or less constant state of alarm. It may have been for
this and other reasons that, probably during Saga's reign, a new imperial police force, whose members were called kebiishi ("Offenses
Investigation Agents"), was established at the emperor's residential
palace. The new force appears initially to have been restricted mostly
to police work and the maintenance of public order within the city
of Kyoto and its immediate environs. After the force became an independent entity in 834, its powers gradually expanded to include
almost all police and judicial authority throughout Kyoto and surrounding areas and also the administration of prisons in the city. The
office became involved additionally in the resolution of disputes concerning agricultural lands, in the collection of unpaid land rents, in
the establishment of shden (estate) boundaries, and in other similar
police, inspection, and judicial functions relating to land and land
revenues. Eventually it seems to have even taken on some street repair and maintenance duties in the capital.
The success of the kebiishi in controlling crime is difficult to judge,
especially since the depredations of the agents themselves may have
occasionally equalled or surpassed those of the criminals they were
supposed to control, but the bureaucratic success of the office, as
measured in terms of expanding powers and growing numbers of officers, was undoubted. It was soon imitated in the provinces, where
kebiishi agents are found as early as the 850s; in one province at least
(Musashi), such officers are known to have been established even as
far down the administrative scale as the district (gun).
Kebiishi agents were a powerful and ubiquitous force in Heian life
through the eleventh century, providing such security as the capital
Cambridge Histories Online © Cambridge University Press, 2008
44
THE HEIAN COURT, 7 9 4 - 1 0 7 0
city had and playing an important role in the functioning of the land
system that supported it. Their success was the result of a rationalization of the statutory structure that brought together under a single authority police, judicial, and prison functions that had formerly
been divided among several offices: the Board of Censors, the Capital Offices, the Punishments Ministry, and the Imperial Guards. But
the success achieved by that rationalization was also another instance of the concentration of governmental authority in and around
the person of the emperor and the narrowing of governmental concern to the immediate interests of the court, developments antithetical to basic principles of the statutory regime.
At the same time that the organizational structure of the central
government was changing in important ways, the financial support
of the structure also underwent alterations in order to accommodate
governmental needs to a persistent insufficiency of income. From
the point of view of the central government, the most striking development in that respect was the beginning of a shift away from reliance on general revenues for the support of central offices and their
staffs toward the establishment of designated pieces of land as
sources of revenue for particular offices. In 879 the government set
aside about 12,000 acres of rice tillage in provinces near the capital
for direct operation by the government or for rental, a portion of the
income in either case becoming revenue for the central government
itself. Within twenty years, over half of those lands had been permanently allocated to particular central offices for their exclusive use
and exploitation. It is noteworthy that it was generally the less important, lower-ranking offices that received such grants of land, but
even so the creation of offices with independent sources of support
led inevitably to a weakening of the regime's control over its own
subordinate organs. The offices endowed with land became in essential respects institutional equivalents of the shoen proprietors,
whose growth otherwise was already affecting the structure of the
statutory fiscal system.
Parallel to the growth of shoen and the appearance of landendowed offices was the emergence during this same period of large
personal land- and other economically valuable holdings of the emperor himself, as noted earlier. The accumulation of personal landholdings by emperors seems to have begun at least as early as the
reign of Saga, and the process continued in Junna's reign, accelerating after the death of the Northern House leader Fujiwara no Fuyutsugu (775-826), who may have acted as a restraint on the imperial
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
45
appetite for private wealth. The holdings grew rapidly thereafter, and
in Nimmyo's reign considerable acreages of waste and vacant land
were also granted to other members of the imperial clan for their
use. As personal imperial wealth grew, the Chamberlains' Office and
the household offices established for retired emperors acquired important functions in the management and administration of the imperial property.
THE ESTABLISHMENT OF FUJIWARA ASCENDANCY,
850-969
The accession of Emperor Montoku in 850 was a critical turning
point in the fortunes of the Northern House of the Fujiwara, placing on the throne the thirteen-year-old maternal nephew of the
house's leader, Yoshifusa, at forty-six the most powerful figure at
court. By the time of his death in 872, Yoshifusa had become the first
regent to the emperor in Japanese history to be appointed from outside the imperial clan, establishing a claim to the post for the Fujiwara Northern-House line that was never thereafter successfully
challenged by any other clan or family. He was also by that time well
along in the process of excluding other clans from the ranks of the
government's senior ministers (kugyo). In 850 four of the nineteen
senior ministers were Fujiwara; at Yoshifusa's death in 872, seven of
eighteen were; a century later, in 972, eleven of nineteen were; and
at the time of Michinaga's death in 1028, the ratio stood at twentytwo of twenty-five, while the remaining three, all Minamoto, were
closely related by marriage to Michinaga.The genealogies of the emperors (Figure 1.1) and the Northern House of the Fujiwara (Figure
1.3) reveal the domination of the Fujiwara.
But the ascendancy of the Fujiwara was in another sense nothing new. From the very beginning of the statutory regime in the
seventh century the clan had been much more than just another
ministerial lineage at court. It had usually been the leading noble
clan in the central government, its men often occupying the highest or most powerful positions there and its women sometimes exercising great influence in the imperial harem. The clan's founder,
Kamatari (614-69), had played a dominant role in the events surrounding the Taika Reforms of 645; Kamatari's son Fuhito (658 or
659-720) had been a key figure in the institution and codification
of the statutory system; Fuhito's daughter Komyoshi (701-60) had
been empress of Shomu and one of the most influential figures in
Cambridge Histories Online © Cambridge University Press, 2008
4
6
THE
HEIAN C O U R T ,
794IO7O
THE
FUJIWARA A S C E N D A N C Y
47
Fusasaki
681-737
I
2 generations
I
Fuyutsugu
775-826
1
Nagara
802-856
1
1
Mototsune Takatsune
I
TStsune
Yoshifusa
804-872
I
I
Takaiko
842-890
I
Tadataka
Yoshinori
Meishi
829-900
Mototsune
836-891
1
~l
1
Sumitomo
d941
Saneyori
900-970
'
~~1
Baishi
Jutsushi
I
1
1
I
1 —
Sanesuke Yoshitaka Yoshichika Kaishi
Akimitsu
957-1046 I
957-1008 945-975 d 1021
Orishi
Onshi
872-907 885-954
I
1
1
1
Kaneie
929-990
Takafuji
(Kanshuji family)
838-900
1
Michitaka Michitsuna Michikanc
961-995
953-995
955-1020
r
Yukitsune Korechika Takaie
Teishi Kanetaka Sonshi
974-1010 979-1044 976-1001
1
I
1
Morozane
1042-1101
"1
~1
Ietada
I
1
Tsunezane
Tadanori
,
Kenshi
Michifusa
d 1044
I
Koshi
d979
Kanshi
1036-112;
Murasaki Shikibu
ca. 1000
Morotada
920-969
1
I
Toshi
d975
Naritoki
Michinaga
966-1028
I
Choshi
d982
1
Yorimichi Yorimune
992-1074 d 1066
1
1
Senshi
961-1002
Suishi
I
1
r ^—
1
Kenshi
Ishi
Kishi
Yoshinobu Norimichi Nagaie Shoshi
996-1075
988-1074 994-1027 999-1036 1007-25
I
Genshi
1016-39
Nobunaga
d 1097
1
1
Kanshi Shinshi
Kanshi
1057-84
Tadazane
1078-1162
Tadamichi
1097-1164
I
I
1
Motozane
Motofusa
(Konoe family) 1144-1230
1143-66
I
1
r
I
Motomichi
1160-1233
Iefusa
1
Takatada
I
I
Kanezane
Kanefusa
(Kujo family)
1149-1207
I
I
Z-
£
Jien
Seishi Ikushi Teishi
1155-1225 1121-81
Yorinaga
1120-56
_i
Kanenaga
1138-58
Moronaga
d 1175
Tashi
1140-1201
Regents in boldface
Adoption
1
Moroie
1172-1240
Figure 1.3. Genealogy of the Northern House of the Fujiwara.
Adapted from Kodansha Encyclopedia of Japan.
Cambridge Histories Online © Cambridge University Press, 2008
I
Hoshi
Seishi
d 1025
Tadanobu Kiminobu Teishi
1
Michifusa
Toshimoto
4 generations
Kinsue
Anshi
(Kan'in family) 927-964
957-1021
Tamemitsu
942-992
—I
Tokimitsu Asateru
d 1015
d 995
Koshi
Morouji
917-970
Kanemichi
925-977
Koretada
924-972
Junshi
809-871
Yoshiyo
I
Keishi
Yukinari
972-1027
Moromichi
1062-99
Tamishi
Morosuke
908-960
Naritoshi
I
1
Yoshikado
1
Tagakishi
Raishi
Tokihira Kanehira Nakahira Tadahira
871-909 875-935 875-945 880-949
I
Yoritada
Sanesuke
924-989
957-1046
H
1
1
Kinto
Teishi Junshi
966-1041
Yoshisuke
813-867
48
THE HEIAN COURT, 794-IO7O
court both as rewards for their services to the throne and also, it has
been suggested, in order to strengthen their position against wealthy
clans resisting implementation of the Taika reforms. That largesse
gave the two men an enormous personal fortune over and above the
huge stipends and emoluments they received from their official ranks
and posts. Although the sustenance grants were all returned to the
court in 820, during the time of Yoshifusa's father, Fuyutsugu, the decline of other court clans in the meantime and the prominence of
Fujiwara men in the higher ministerial ranks suggest that the clan's
relative economic standing may have remained largely unchanged.
The special social status occupied by the Fujiwara was given formal
recognition in 729 when Komydshi became the first woman not of imperial descent to receive the title of empress (kogo), and again in 792
when Kammu issued a directive permitting men of the clan to marry
daughters of emperors (that is, female descendants in what was calculated as the second generation of descent), while limiting other noblemen to granddaughters (third generation) or after.
Yoshifusa's position in 850 was based, however, not only on a long
history of Fujiwara leadership, preeminence, and preference at
court, but also more specifically on the favorable situation in which
the Northern House of the clan found itself in the post-Kammu
years. Members of the house were the chief governmental ministers
during the reigns of Heizei, Saga, and Junna, and during the latter
two reigns Yoshifusa's father, the astute Fuyutsugu, laid the groundwork of a nearly impregnable position for his son. Fuyutsugu had
been close to Saga ever since the emperor's days as crown prince,
and after Saga's accession he rose rapidly, becoming with Kose no
Notari (794-817) the first head of the Chamberlains' Office in 810
and then, as Major Counselor, the chief minister in the government
on the death of his senior clansman Sonohito (756-819). He buttressed his own position and prepared the way for lasting NorthernHouse domination of the court by establishing close marital links
with the imperial clan, giving a daughter, Junshi (809-71), to Saga's
son, the future emperor Nimmyo, and taking one of Saga's daughters as a wife for his own son Yoshifusa. Fuyutsugu's premature
death in 826 left the young and still relatively low-ranking Yoshifusa
in a somewhat precarious position, but after the Jowa Incident in 842
the future clearly belonged to him and his house, and he did not hesitate to claim it.
Yoshifusa demonstrated his power at court and his willingness to
use it in his own and his family's interests during the very first year
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
49
of Montoku's reign. The emperor had taken Yoshifusa's daughter
Meishi (829-900) as a consort while he was still crown prince, and
in the month of Nimmyo's death and Montoku's accession, Meishi
had borne to her husband a son, the future emperor Seiwa (850-81,
r. 858-76). Montoku already had a son by a woman of the Ki clan,
the six-year-old Prince Koretaka (844-97), but eight months after
the birth of Seiwa, Yoshifusa appears to have forced his nephew
Montoku to pass over Koretaka's claims and declare Meishi's infant
son crown prince, thus assuring Yoshifusa even greater power in the
next reign as maternal grandfather of the emperor.
During most of Montoku's reign the chief ministerial post (Minister of the Left) was held by a powerless son of Saga named Minamoto noTokiwa (812-54), but after Tokiwa's death Yoshifusa became in title as well as in fact the leading minister at court. In 857
he was named Chancellor (daijo daijin or dajjb daijiri), a momentous
appointment that is regarded by some as the beginning of the Fujiwara regency under which Japan was ruled during most of the next
two centuries or more.
According to the provisions of the statutory code, the office of
Chancellor was the highest post in the government, but it was supposed to remain vacant unless a man could be found whose outstanding character and moral probity qualified him to serve as a
model of conduct for the emperor and for the officialdom of the
regime. Although the functions of a Chancellor were formally limited to the role of exemplar, the prestige of the post was unmatched
by any other office at the government's disposal, and it had been
filled in the past either by influential imperial princes, or, exceptionally, by imperial favorites with extraordinary power at court (that is,
Fujiwara no Nakamaro [706-64] and the Buddhist monk Dokyo
[d. 772]).The significance of Yoshifusa's appointment as Chancellor
was emphasized by the fact that the office had remained vacant ever
since the time of Dokyo, ninety years earlier.
With his appointment as Chancellor, Yoshifusa became probably
in everything but name regent to the emperor, and on the accession
of his grandson Seiwa in the following year (858), his control of the
court was virtually complete. Not until eight years later, however,
was he formally invested with the title of regent (sessho). In that year,
866, the Gate of Obedience to Heaven (Otemmon), the main gate to
the imperial palace's Court of Government, burned under mysterious circumstances. One of the two Major Counselors at the time,
Tomo no Yoshio (809-68), charged that his adversary and rival, the
Cambridge Histories Online © Cambridge University Press, 2008
5O
THE HEIAN COURT, 7 9 4 - I O 7 O
Minister of the Left Minamoto no Makoto (810-69), an affinal relative of Yoshifusa, was responsible for the burning of the gate, and
he persuaded Yoshifusa's brother, the Minister of the Right Yoshimi
(orYoshisuke, 813-67), to issue an order for Makoto's arrest. At the
time, Yoshifusa was living in seclusion at Shirakawa just outside the
capital, but he intervened immediately on Makoto's behalf and
halted the proceedings against him. Shortly thereafter a witness with
a grudge against Yoshio opportunely turned up to swear that the fire
was actually the work ofYoshio and his confederates, who included
a man from the Ki clan. After investigation, Yoshio, his son, and a
number of other Tomo and Ki men were found guilty and sent into
exile, bringing to a decisive end the long struggle of those ancient
clans to maintain the important positions they had once occupied at
court. (Yoshio was a member of the Otomo clan, but since the time
of Emperor Junna clan members had used a shortened form of the
name - Tomo, or "Ban" in the Sino-Japanese reading of the character - in order to avoid Junna's personal name, Otomo.)
The events and motives involved in the burning of the Gate of
Obedience to Heaven are by no means clear, and even insofar as we
have a record of them, their authenticity cannot be assumed, since
they are contained in a chronicle compiled under the editorship of
Yoshifusa's grandson. The results of the series of events that flowed
from the fire are, however, quite clear. Two of the oldest clans at
court were removed from the path of Fujiwara ambition, and in the
midst of it all, following Yoshio's arrest and interrogation but before
his exile, Yoshifusa was formally appointed regent by the sixteenyear-old emperor, who may have been anxious to make amends for
his too ready acceptance ofYoshio's charge against Makoto.
Individual clans (the Kazuragi, Wani, Heguri, Otomo, Mononobe,
and Soga) had come to exercise preponderant power at court in the
sixth century as the emperor's strength seems steadily to have waned
in relation to the constituent clans of theYamato state. In that sense,
Fujiwara domination of the court had ample historical precedent.
But the institution of the statutory regime in the seventh century,
while by no means giving despotic authority to the imperial line,
brought some restitution of the line's authority, an authority that
was markedly expanded in the late eighth and the ninth century
under emperors Kammu and Saga on the basis of governmental reform and retrenchment, and it was the newly expanded power that
Yoshifusa exercised as regent to the emperor.
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
51
By the time of his death in 872,Yoshifusa had achieved heights of
power and prestige unprecedented for a court noble, his position
differing but little in form or substance from that of the emperor
himself. His chief failing, his inability to produce a son, had been
remedied by the adoption of Mototsune (836-91), a son of his
brother Nagara (or Nagayoshi, 802-56), who succeededYoshifusa as
the principal power at court, with appointment to the title of regent
coming at the beginning of 873. Emperor Seiwa, now in his early
twenties, sought to rule in his own right for a while, but bothered
by bad health and a series of ominous portents, he abdicated in 876
in favor of the seven-year-old Yozei, his son by Nagara's daughter.
Mototsune, who was also the emperor's maternal uncle, was reappointed regent.
Yozei's accession at the age of seven and Seiwa's at the age of eight
marked a significant turning point in the history of the imperial clan.
They were the first child-emperors to come to the throne in Japan,
clearly indicating that by their time the imperial position was distinct
from the power to rule. In the Nara period and earlier, when an emperor died before a male heir, or the preferred male heir, had come
of age, it seems to have been the custom for the mother or a sister to
take the throne until the heir had reached a suitable age. Now, however, with regency power at court in the hands of Northern-House
leaders, such precautions were apparently no longer thought necessary; they were indeed, from the point of view of the Fujiwara, undesirable, since mature emperors were generally more difficult to
control. We find, consequently, that of the thirteen emperors who
reigned from the time of Seiwa to Go-Ichijo (1008-36, r. 1016-36),
nine were seventeen or younger at the time of their accessions, and
seven, or more than half, were no older than twelve. It may also be
observed that among that group of thirteen, with the exception of
Koko (830-87, r. 884-87), whose special circumstances are noted
later in the chapter, it was the emperors who came to the throne at
the age of twenty or more who were most independent of, or clashed
most seriously with, the Fujiwara: Uda (867-931, r. 887-97), Murakami, and Sanjo (976-1017, r. 1011-16).
Despite his youth, Emperor Yozei seems to have been more than
Mototsune could easily handle. Although the official chronicle of his
reign is discreet to the point of obscurity, it appears that the young
emperor was wild in the pursuit of unorthodox interests, perhaps
not quite a Japanese Caligula but possibly sharing the attitude re-
Cambridge Histories Online © Cambridge University Press, 2008
THE HEIAN COURT, 7 9 4 - I O 7 O
TABLE 1.1
Heian emperors
Numerical
order
50
51
52
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
Emperor
Kammu
Heizei
Saga
Nimmyo
Montoku
Seiwa
Yozei
K6ko
Uda
Daigo
Suzaku
Murakami
Reizei
En'yu
Kazan
Ichijo
Sanjo
Go-Ichijo
Go Suzaku
Go-Reizei
Go-Sanjo
Shirakawa
Horikawa
Toba
Sutoku
Konoe
Go-Shirakawa
Nijo
Rokujo
Takakura
Antoku
Go-Toba
Birth and
death dates
737-8o6
774-824
786-842
810-850
827-858
850-881
869-949
830-887
867-931
885-930
923-952
926-967
9501011
959-991
968-1008
9801011
976-1017
1008-1036
1009-1045
1025-1068
1034-1073
1053-1129
1079-1107
1103-1156
1119-1164
1139-1155
11271192
1143-1165
1164—1176
11611181
1178-1185
1180-1239
Reign dates
781-806
806-809
809-823
833-850
850-858
858-876
876-884
884-887
887-897
897-930
930-946
946-967
967-969
969-984
984-986
986-1011
10111016
1016-1036
1036-1045
1045-1068
1068-1073
1073-1087
1087—1107
11071123
11231142
11421155
1155-1158
1158-1165
1165-1168
1168-1180
1180-1185
1183-1198
Source: Kodansha Encyclopedia of Japan (1983), vol. 2, p. 202.
vealed in that vicious despot's warning: "Remember that I can do
anything to anybody."11 Although not so bloodthirsty as his Roman
predecessor, Yozei was very likely the murderer of one of his own
courtiers. It was in protest of such behavior, no doubt, that Mototsune sought to resign his regency powers on several occasions and
in 883 refused for a number of months to appear at the palace, bringing government to a standstill. The murder appears to have been the
final straw, however, and shortly thereafter Mototsune expelled from
11 John Boardman, Jasper Griffin, and Oswyn Murray, eds., The Oxford History of the ClassicalWorld (Oxford and New York: Oxford University Press, 1986), pp. 9-10.
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
53
court the emperor's more flamboyant companions, forced the emperor himself to abdicate, and installed in his place on the throne a
fifty-four-year-old son of Nimmyo, Emperor Koko.
If it is assumed that Mototsune was pursuing chiefly his own interests and those of his family, his choice of Koko was at first sight
extraordinary and in some respects inexplicable. A more natural
choice from that point of view, one might have thought, would have
been one of the two younger brothers of Yozei who were related
through their mothers to Mototsune, one a ten-year-old nephew (a
son of Koshi) and the other a six-year-old grandchild. But perhaps in
order to avoid another regency and to help bring the court back to a
certain normalcy after the bizarre events ofYozei's reign, Mototsune
selected a prince with a reputation for intelligence and probity and
who was related to him only rather distantly (they were cousins
through their mothers). The choice therefore may have been at least
partially an act of statesmanship, as subsequent historians have
sometimes treated it, but it was perhaps at the same time also firmly
rooted, after all, in a clear calculation of family self-interest. Before
the new emperor's accession to the throne, his seventh son (later
Emperor Uda) had been adopted by Mototsune's childless adoptive
sister, Shukushi. The decisive influence that sisters seem often to
have exercised on their brothers in Heian society provides grounds
for thinking that her Machiavellian hand may have been involved in
the choice of the aged emperor, whose early demise, she might have
correctly foreseen, would lead under Mototsune's influence to the elevation of her adopted son to the imperial title and to exceptional
honors for herself. (She became, in fact, the first woman to hold the
Junior First Rank, a rank so high that it had been rarely awarded to
anyone.) Koko's selection may also be interpreted to mean, on the
one hand, that the power of the Northern House was so firmly entrenched by this time that it did not need to depend on a close blood
relationship with the emperor, and, on the other, that Koko was not
personally inclined toward the assertion of his authority. The latter is
illustrated by his decision (if the decision was indeed his) to reduce
all his sons to noble status, including the son adopted by Shukushi - rendering them at least formally ineligible as heirs to the
throne and thus avoiding possible conflict with the Fujiwara - and
also by his investment of Mototsune with what amounted to regency
powers.
Koko's reign was brief, and an heir to the throne had still not been
chosen when he lay on his deathbed in 887. According to a nearly
Cambridge Histories Online © Cambridge University Press, 2008
54
THE HEIAN COURT, 794-IO7O
contemporaneous document, the emperor had frequently sought to
make Uda his crown prince, but Mototsune did not act, and it was
thanks largely to the efforts of Shukushi that her adopted son was finally restored to his princely title and raised to the throne on the
same day that he was named imperial heir. The selection of Uda, as
we shall see, eventually led to some difficulties for the Northern
House in its domination of the court, but as long as Mototsune
lived, the absence of a close blood connection with the emperor
seems to have made no more difference than in the days of Koko.
One of Uda's first acts was, in fact, to invest Mototsune with regency
powers, the edict of investment employing for the first time the title
kampaku ("internuncio"), the name by which a regent came to be
called when the emperor was an adult, contrasting with sessho ("viceregent"), the title of a regent to a minor emperor.
That there was, nevertheless, at least some friction between Uda
and Mototsune is suggested by a curious and seemingly trivial controversy about nomenclature that erupted following Motosune's appointment as kampaku. In offering the post to Mototsune a second
time after he had formally declined the first edict of appointment as
required by custom, Uda - or, rather, Tachibana no Hiromi (837 or
838-90), the erudite of Chinese letters who drafted Uda's edict equated kampaku with the Chinese title a-heng (Japanese ako), which
in early China had designated the chief governmental minister. Although the same word had been used of Mototsune without incident
in an imperial rescript three years earlier, this time Mototsune's
steward, a Fujiwara named Sukeyo (d. 897 or 898), also learned in
Chinese, persuaded the Northern-House leader that ako actually referred to a purely nominal post without substantive duties and that
the equation of the two titles implied the dismissal of Mototsune
from government. Mototsune thereupon refused to take further part
in court affairs, bringing major governmental business to a halt, just
as he had done during Yozei's reign. The controversy raged on for six
or seven months, with one group of literati supporting Hiromi and
another supporting Sukeyo, but finally the chastened Uda was forced
to apologize and delete the offending word from his edict.
Although the Ak6 Controversy was perhaps fueled chiefly by the
ambitions of rival learned families, Mototsune may have also welcomed it as an opportunity to rid Uda's court of the influential
literati with whom he had surrounded himself. It may have also had
some other important political results. In the first place, although on
the one hand it clearly demonstrated to Uda his inability to oppose
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
55
Mototsune's wishes, at the same time it seems to have left a legacy
of imperial bitterness that encouraged a brief challenge to the
Northern House's control of the court during the generation of Mototsune's sons. Furthermore, the controversy ended in disgrace for
Hiromi, who as the father of Uda's consort and the grandfather of
her two sons posed at least a potential threat to the Fujiwara position. It may have been to allay Mototsune's fears in that respect that
just four months after Uda's forced apology, Mototsune's daughter
Onshi (872-907) was allowed to become a consort of the emperor.
Finally, among the learned men defending Hiromi was Sugawara no
Michizane (845-903), who after Mototsune's death figured prominently in the efforts of Uda to free himself and the imperial clan
from Fujiwara control. Michizane's defense of Hiromi took the form
of a written statement of opinion addressed to Mototsune and dated
after the incident had already been resolved. He argued that persecution of Hiromi for what he contended was an unintended affront
would have a chilling effect on the learned world, and also that it was
not in the interests of the Fujiwara to attack a man who was grandfather to two of the emperor's children.12
The death of Mototsune in 891 ushered in about a decade of imperial ascendancy at court. Immediately after Mototsune's death, an
aged, ineffective brother of Yoshifusa became the senior Fujiwara
minister; and Mototsune's eldest son, Tokihira (871-909), was still
too young and too junior in status to control the senior ministers,
whose ranks were dominated by Minamoto men. Effective leadership in the Council of State was in the hands of a Minamoto Major
Counselor who was a son of Emperor Montoku, which left Uda with
a freedom of rule that had not been enjoyed by an emperor since the
days of Saga and Junna. In his determination to control his own
court and to maintain his independence of Fujiwara power, Uda
strengthened the Chamberlains' Office, expanding its staff and codifying its procedures, greatly enlarged the kebiishi office, and also
created landholdings for the personal use of retired emperors. To assist him, he employed men from the middling ranks of the nobility
unconnected, or only distantly connected, with the Fujiwara leaders,
men like the literati-official Sugawara no Michizane, who was appointed head of the Chamberlains' Office in the month following
Mototsune's death, and the celebrated provincial administrator Fu12 For Michizane's role in the Ako Controversy, see especially Robert Borgen, Sugawara no
Michizane and the Early Heian Court, Harvard East Asian Monographs, 120 (Cambridge,
Mass.: Council on East Asian Studies, Harvard University, 1986), pp. 173-81.
Cambridge Histories Online © Cambridge University Press, 2008
56
THE HEIAN COURT, 794-IO7O
jiwara no Yasunori (825-95), whose series of provincial appointments testify to his remoteness from the center of Fujiwara power.
(He belonged to the Southern House of the clan.)
Michizane, in particular, became a key instrument in Uda's quest
for control of the Council of State, perhaps having earned the emperor's confidence and gratitude by his letter to Mototsune in the
Afco Controversy (although, in fact, it is not known whether either
Mototsune or Uda ever saw the letter). With Uda's backing, Michizane found himself suddenly elevated from the status of an unemployed ex-provincial official at the time of Mototsune's death to
the ranks of the senior ministers just two years later. By 896, when
one of his daughters entered Uda's harem, Michizane had risen to
the post of Middle Counselor, and before his imperial patron abdicated in the following year he had been appointed Provisional Major
Counselor, just a step behind his junior by twenty-six years, Fujiwara no Tokihira. When Uda abdicated in favor of his twelve-yearold son Daigo (885-930, r. 897-930), whom he had earlier named
crown prince in consultation with Michizane and who had no direct
family relationship with Tokihira, Michizane and Tokihira were directed to share the supervision of the government with powers similar to those of kampaku (that is, nairan, or "private inspection,"
powers). In 899, on the same day that Tokihira became Minister of
the Left, Michizane was appointed Minister of the Right, the second
highest regularly filled office in the government and a totally unprecedented honor for a member of the Sugawara clan. In an age
when clan and family membership had come generally to determine
the offices for which a man was eligible, Michizane was understandably alarmed by his newfound prominence, which even he considered beyond his station in life.
Michizane's rise to eminence not only implied a weakening of the
Fujiwara position at court but also represented the emergence in
high court councils of a man with practical experience in provincial
administration, source of the most critical governmental problems of
the day. Much against his will, he had served a term as governor of
Sanuki (on Shikoku) from 886 to 891, acquiring there, and also in
his association with Fujiwara no Yasunori and other experienced
provincial governors, a practical knowledge of the problems confronting provincial administrators that made him unique among the
court's chief ministers. Those problems centered on the question of
how to collect revenues from and administer a province under social
and economic conditions strikingly different from the ones enviCambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
57
sioned in the statutory code. Although the virtual abandonment of
the land distribution system, the emergence of a provincial gentry
possessed of large landholdings, a conspicuous increase in the vagrant population, and the corruption of population registers had
made much of the statutory provincial administrative structure inoperable or useless, the court under Fujiwara leadership had continued to insist for the most part on the literal observance of statutory law and procedures. Caught between those demands and the
actualities of local conditions, provincial administrators seem regularly to have resorted, even at the risk of punishment, to subterfuge
and illegal or extralegal measures in order to fulfill their governmental responsibilities. Needless to say, private gain was no doubt
also sometimes the object of such measures. If a single instance relating to the dispatch of provincial tax inspectors can be accepted as
indicative of the general lines of Michizane's thought on the subject,
he argued at court that the provincial administrators were entirely
justified in the freedoms they were taking with the law. He said in
essence that the provinces were ungovernable under the existing
provisions of the statutory code and that it was necessary to adjust
provincial administrative aims and procedures to existing local circumstances. That pragmatic view seems to have found little favor in
court councils during Michizane's own day, but not many years
later, shortly before Tokihira's death in 909, it became the policy of
the court, contributing vitally to the survival of that institution as a
functioning organ of government.
Despite Uda's abdication, which may have been a political tactic
designed to strengthen his own position and that of his heir, the direction of court affairs apparently remained largely in his hands. But
by that time (897) Tokihira was approaching thirty and beginning to
demonstrate the gift for leadership and decisive action that had
made his father and grandfather before him such dominant figures.
It had probably been clear to him for some time that if Fujiwara control of the court was to be firmly reestablished and maintained,
Michizane, as an instrument of imperial power and possibly also as
a contender for power in his own right, would have to be removed
from his influential position. But it was not until 901 that Tokihira
found, or perhaps created, an opportunity to move against the upstart Minister of the Right. His action may have been precipitated by
a proposal that, according to a twelfth-century source, Uda and Emperor Daigo had made to Michizane the previous year suggesting
that the dual leadership shared by him and Tokihira be ended and
Cambridge Histories Online © Cambridge University Press, 2008
58
THE HEIAN COURT, 7 9 4 - 1 0 7 0
that supervision of governmental matters be entrusted exclusively to
Michizane. Although Michizane prudently refused to consider the
proposal, Tokihira, if he learned of it, could not have been pleased by
so marked a sign of imperial favor. At any rate, early in 901 he persuaded the sixteen-year-old emperor that Michizane was conspiring
with a deluded Uda to force Daigo to abdicate, presumably in favor
of Uda's son and Michizane's son-in-law Prince Tokiyo (886-927),
and an imperial decree was hastily issued demoting Michizane and
exiling him to Kyushu. There is some evidence to suggest that the
chief architect of the conspiracy, assuming that a conspiracy did in
fact exist, may have been Uda himself. If so, the ex-emperor's defeat
and humiliation were complete, for when he raced to the imperial
palace on learning too late what had happened to Michizane, he was
refused access to, or communication with, the emperor, despite the
all-day vigil he mounted in the palace courtyard.
In the years following Michizane's death in exile, calamities at
court and in Tokihira's family - including especially a spectacular
lightning storm that struck the imperial palace in 930, killing or
injuring several high-ranking courtiers - were attributed to Michizane's angry spirit. In an effort to placate the spirit, Daigo rescinded
the rescript ordering Michizane's exile and restored him to his original office and rank. It was probably for the same reason that perhaps as early as the 940s Michizane had begun to be worshiped at
Kitano in Kyoto, an area earlier associated with a god of thunder and
lightning. By the 980s he had acquired the name Heavenly Deity
Temman ("Heaven Filling") and was being called the "ancestor of
the Way of Letters," becoming thus the divinity of a cult that has survived vigorously to the present day.13
After the ouster of Michizane and the isolation of Uda from governmental affairs, Tokihira enjoyed a period of unchallenged power
at court as the sole possessor of "private inspection" authority until
his death in 909, although he never received appointment as regent.
It is worthwhile noting that Tokihira's line of the Northern House
seems to have been so powerful by this time that he was able to continue the family's ascendancy at court despite the presence on the
Council of State of a member of a collateral line of the house who
was both senior in age to him and also the maternal uncle of Daigo.
Northern-House lineage, a blood relationship with the emperor, se13 On Michizane and his cult, see Borgen, Sugawara no Michizane and the Early Heian Court,
PP- 307-36Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
59
niority, and ministerial rank were not necessarily sufficient to give a
man control of the court.
Tokihira was succeeded in family and governmental affairs not by
his eldest son, who was only about seventeen at the time of Tokihira's death, but by his younger brother Tadahira (880-949), who
dominated the court during the remainder of Daigo's reign but
seems to have been a somewhat less powerful or autocratic figure
than his brother had been, never wholly succeeding in usurping the
emperor's authority. He had to contend not only with the emperor
himself, who was in his mid-twenties by that time, but also with the
ex-emperor Uda, who had become a Buddhist monk in 899 and had
remained in seclusion at NinnajiTemple following Michizane's exile.
Uda now reasserted himself and through Tadahira, who was married
to the ex-emperor's daughter (or possibly his half sister - there is a
question about the exact relationship), exercised a powerful influence at court and in the government. Tadahira did succeed, however,
in imposing his nephew, Daigo's son by Mototsune's daughter
Onshi, on the emperor as his crown prince, and when that prince became emperor at the age of seven in 930, Tadahira was appointed regent (sessho). When the emperor, Suzaku (923-52, r. 930-46), was
eighteen in 941, Tadahira's regency title was changed to kampaku,
beginning the practice observed thereafter of distinguishing a regent
for a minor emperor {sessho) from that of an adult emperor (kampaku) . From 944 until Tadahira's death in 949, the three highest offices in the government (regent and Chancellor, Minister of the Left
and Minister of the Right) were held by Tadahira, his brother
Nakahira (875-945), o r his s o n s Saneyori (900-70) and Morosuke
(908-60). The Fujiwara in the line of Mototsune were firmly in the
saddle of power, seemingly impregnable to almost any challenge
from within the court.
During most of Tokihira's time at the head of the central government, the court, ignoring Michizane's prescient advice, continued a
basic policy of attempted institutional revival or restoration first adumbrated in Uda's reign. It sought in a flurry of Council of State
orders to check the growth of a wealthy landed gentry class in the
provinces, and to restore and maintain the system of taxes and publicly allotted rice tillage provided for by the statutory code. Toward
the end of Tokihira's life, however, he and his court seem finally to
have recognized the futility of that policy and to have sought instead
to adapt governmental practice to provincial reality. In pursuit of
that aim, three basic changes in past policy were instituted:
Cambridge Histories Online © Cambridge University Press, 2008
60
THE HEIAN COURT, 7 9 4 - 1 0 7 0
1. A de facto abandonment of the statutory land allotment system
and a substitution of land taxes for capitatim taxes.
2. A new approach to private landed proprietorships that accepted
their existence and growth but also attempted to limit the expansion of tax-exempt land within the proprietorships in return for
increased exemptions on legally exempt land.
3. The recognition of provincial governments as quasi-autonomous
entities with contractual tax obligations to the central government, a notable weakening of a cardinal principle of the statutory
system, which had established a centrally defined uniform tax
system and a thoroughgoing centralized control over the provinces based on the appointment of central officials to the chief
administrative posts.
Pushed forward also under Tadahira, the court's new, more pragmatic approach to land, taxes, and provincial administration enabled the regime to survive the outbreak of large-scale revolt and
piracy in the eastern and western parts of the country in the late
930s and early 940s, and to evolve and endure for another two centuries or more as the nation's central instrument of government.
The change in governmental policy seems in large part to have been
simply a confirmation of earlier tentative steps in the same direction
and a recognition of practices that had already been adopted by the
provincial governors themselves acting on their own authority. If the
change was not therefore particularly sudden or radical, it represented nevertheless a major political turning point. The new policies
brought the last period of the old statutory regime to an end, recognizing, and in part creating, a new organization of the country in
which the political system was increasingly characterized by personalized rule at court, descent-determined office, and provincial autonomy. The land allotment system was gone and the main tax burden had shifted from individuals to land; large private landed
proprietorships had become an ever more prominent feature of the
economy; and the society was marked by a growing military class in
the provinces. Modern historians often use the term "Royal-Court
State" (pcho kokkd) to distinguish the new system from the early
statutory regime.
A quarrel among the Kammu Taira descended from the Emperor
Kammu, who had become great landowners in the eastern provinces, escalated from small armed conflicts in 935 to a full-blown rebellion when Taira no Masakado in 939 seized the provincial headCambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
6l
quarters of Hitachi. He grandiosely set forth to control all of the
eight provinces of the Kanto region and began to appoint new provincial officials. Even more shocking, he declared himself the "New
Emperor" (jshinno) and claimed imperial powers for himself throughout the east. The court, greatly alarmed, appointed a military commander to lead an expedition force against Masakado, but before it
reached the east, Masakado was attacked and killed by Taira no
Sadamori, his cousin, and Fujiwara no Hidesato.
Just at the height of the Masakado crisis late in 939, the court was
beset by a resurgence of piracy in the Inland Sea (Seto Naikai) to the
west. Fujiwara no Sumitomo, once an official in Iyo Province, carried out audacious raids on ports and ships in the Inland Sea, disrupting the transportation of tax revenues and products from the
western regions to the capital. It was not until 941 that an expeditionary force organized by the court, with the assistance of defectors
from the pirate bands, was able to corner and defeat Sumitomo.
The Masakado revolt and the Sumitomo piracy are referred to by
modern Japanese historians as the "Discord of the Johei [931-38]
and Tengyo [938-47] Eras." Masakado's rebellion was the more significant as a harbinger of future developments. It was the first attempt by a local leader to challenge the authority of the central government by trying to establish an autonomous government. The
suppression of the revolt did not establish the central government's
authority in the eastern provinces but, rather, confirmed the
strength of the local provincial leaders who defeated Masakado. In
the conflicts among the eastern warriors can be seen an early stage
in the development of the bushidan, the private "warrior bands,"
based on family alliances and mutual regional interests, that shaped
warrior society in the eleventh and twelfth centuries. (For a fuller account of the Johei-Tengyo disturbances, see Chapter 10.)
In 944 when Emperor Suzaku was twenty-one years old and still
had not produced an heir, his mother, Onshi, persuaded him to
appoint his younger brother crown prince, and two years later he
abdicated.
The new emperor, Murakami (926-67, r. 946-67), was twenty at
the time of his accession and inclined to take a stronger hand in government than his brother had, trying, mostly unsuccessfully, to improve the government's financial situation and to restore security in
the provinces and the capital. The period centering on his reign is
notable as the graveyard of several important activities of the statutory regime. The last official chronicle of the state was commisCambridge Histories Online © Cambridge University Press, 2008
62
THE HEIAN COURT, 794-IO7O
sioned then but never completed; the last attempt at compiling governmental laws and procedures was undertaken but also abandoned
before completion; and the last coins of the regime were minted. By
Murakami's time, much of the statutory system survived only as an
empty shell.
Following Tadahira's death in 949, the office of regent fell vacant,
but the court continued to be dominated by the Northern House, as
represented then by Tadahira's sons Saneyori, Morosuke, and Morotada (920-69), all of whom placed daughters in Murakami's crowded
harem. Morosuke's daughter Anshi (927-64) was the most successful of the Northern-House imperial consorts, bearing three sons Prince Tamehira (952-1010) and the two future emperors Reizei
(950-10 n , r. 967-69) and En'yu (959-91, r. 96g-84)-and also, according to a twelfth-century source, exercising an unusually powerful influence over the emperor himself. Reizei was named crown
prince just two months after his birth, passing over a luckless elder
half brother, Motokata (888-953), whose mother was a daughter of
the Southern-House Fujiwara. Reizei, however, was a sickly and perhaps mentally disturbed child, and after the birth of Anshi's second
and favorite son, Tamehira, in 952, she and Murakami came to feel,
with most of the court, thatTamehira should replace Reizei as crown
prince. But nothing had been done by the time of Anshi's death in
964, and after 966, when Tamehira made the mistake of marrying a
non-Fujiwara (a daughter of his mother's learned brother-in-law and
confidant, Daigo's son Minamoto noTakaakira [914-83]), the Fujiwara were opposed to Tamehira's selection as crown prince at any
time. When Reizei came to the throne in 967, the new crown prince
therefore was not Tamehira but his younger brother En'yu.
Morosuke had been the ablest and most influential of Tadahira's
sons, but his early death in 960, when his eldest son Koretada (or
Koremasa, 924-72) was still relatively junior in rank and office, left
the senior Council of State positions in the hands of his older brother
Saneyori (Minister of the Left), Tokihira's son the Minister of the
Right Akitada (898-965), and Minamoto noTakaakira (Major Counselor). Thus, when Reizei came to the dirone following Murakami's
death in 967 and it became necessary to name a regent for the ailing
emperor, the choice fell on Saneyori, an aged and somewhat weak
figure who was not closely related to Reizei. So much of the actual
power at court during Saneyori's time seems to have been in the
hands of others, especially those of his younger brother Morotada,
that even Saneyori referred to himself in his diary as a "nominal" reCambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA ASCENDANCY
63
gent. It is nevertheless from 967, the year of his appointment to that
post, that historians often date the beginning of the "Regency Period" (sekkan jidai), since the office of regent, which had been established only sporadically until that time, was thereafter almost
continuously in existence, occupied invariably by a Northern-House
leader holding the title ofsessho or kampaku, or exercising "private inspection" (nairari) authority. The Regency Period is usually thought
of as lasting to the accession of Emperor Go-Sanjo in 1068.
Although it seems unlikely that Minamoto no Takaakira posed any
real threat to the interests of the Northern House, his abilities, his
sway at court, his marital connection with a possible imperial heir
(Tamehira), and his high pedigree and office (he became Minister of
the Left at the beginning of 968) could scarcely have endeared him
to Saneyori and the other Northern-House men. His eventual downfall in the Anna Incident of 969 (so named after the current era
name), however, may have occurred at least in part because of his
position. This made him a convenient target of an ambitious military
leader in the capital at that time who seems to have been seeking
closer ties with the Northern House as well as a means of reducing
or eliminating the influence of his military rivals of Kyoto.
The military leader in question, Minamoto no Mitsunaka (91297), was the son of a warrior who had fought against Masakado and
Sumitomo. He had latterly been employed by the court in police and
military functions in the capital area together with Fujiwara no Chiharu, a son of the Hidesato who had helped defeat Masakado. In the
early spring of 969, Mitsunaka and another man reported to the government that Minamoto no Tsuranu and others were plotting treason. In the absence of sources, it is impossible to say with any certainty what the plot was about, if it existed at all, but since Tsuranu
was closely related to Takaakira, who was both his cousin and his
brother-in-law, it is possible that an attempt (fictional or otherwise)
to place Tamehira on the throne was involved. In any case, Takaakira
and a number of others, including Mitsunaka's chief rival for military
leadership in the capital, Chiharu, were implicated in the plot and
sent into exile. Chiharu went to the island of Oki, and Takaakira, following in the footsteps of Sugawara no Michizane, another Fujiwara
victim, to Kyushu, where he remained for three years before being allowed to return to a quiet life of study at the capital. The immediate
result of the Anna Incident was thus the removal of the last real or
potential rivals of the Northern House and Mitsunaka, but the longterm importance of the affair is probably that it marked the first sigCambridge Histories Online © Cambridge University Press, 2008
64
THE HEIAN COURT, 794-1070
nificant appearance of warrior figures in the politics of the capital. It
was also the beginning of an alliance between the Northern House
and Mitsunaka's warrior family, the Seiwa Genji line of the Minamoto, from which eventually came the founder of the Kamakura
Bakufu.
A few months after Takaakira's exile, the sickly emperor Reizei abdicated the throne to his younger brother En'yu, and on the same
day Reizei's infant son by Fujiwara no Koretada's daughter Kaishi
(945-75), the future emperor Kazan (968-1008, r. 984-86), was
named crown prince. Reizei's decision to abdicate was encouraged
by Koretada, who was anxious to see his grandson appointed crown
prince before a competitor for the position could be produced by another of the emperor's consorts, Choshi (d. 982), a daughter of Koretada's younger brother Kaneie (929-90).
THE FUJIWARA REGENCY, 97O-IO7O
In the year following En'yii's accession in 969, the aged regent
Saneyori died, and the regency passed to Morosuke's eldest son, Koretada, who was the maternal uncle of the emperor and the maternal grandfather of the crown prince. But three years later Koretada
himself was dead, and another shift in the regency was required, the
choice lying between Koretada's brothers, Kanemichi (925-77) and
Kaneie. (For a genealogy of the Fujiwara Northern House and its regents, see Figure 1.3.)
Although Kanemichi was four years Kaneie's senior, he had not
been well thought of by Murakami, Reizei, or Saneyori, and his career had languished behind that of his vigorous and able younger
brother, who had been head of the Chamberlains' Office under
Reizei and had progressed rapidly up the Council of State ladder to
the post of Major Counselor by the year of Koretada's death. The
choice of Kaneie as Koretada's successor may therefore have been
regarded as a foregone conclusion by the courtiers, but the wily
Kanemichi, a late source asserts, had obtained from his influential
sister Anshi before her death in 964 a statement written in her own
hand declaring that the regental succession was to follow the order
of seniority among her brothers. Kanemichi is supposed to have
shown the document to Anshi's son, Emperor En'yu, while Koretada lay dying at the end of 972, and the emperor, it is implied, felt
constrained to honor his mother's wishes. When Koretada resigned
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA REGENCY
65
his offices a few days before his death, Kanemichi, who held the
comparatively lowly post of Provisional Middle Counselor, was ordered to take charge of the government. At the beginning of 973, a
few weeks after Koretada's death, he was promoted over the heads
of six other nobles, including his brother Kaneie, to the office of
Palace Minister and named regent (kampaku; the regental appointment may not have come until the next year).
Following his precipitous rise to the regency, Kanemichi ruled the
court in cooperation with his cousin, Saneyori's son Yoritada
(924-89), and prepared the way for an eventual blood link with the
imperial line by marrying a daughter to the young emperor. But his
plans for the future and the enjoyment of his triumph over Kaneie
were cut short by his premature death at the age of fifty-two in 977.
Just before his death, however, Kanemichi exacted a final revenge on
Kaneie, arranging matters so as to humiliate him and, as he no
doubt hoped, to permanently debar him from the regency. In debarring Kaneie, Kanemichi violated that same testament of Anshi by
which he is said to have won the post for himself. Angered, a late
source says, by the unfeeling haste with which Kaneie presented
himself at the imperial palace to claim the regency when he mistakenly concluded that his brother was already dead, Kanemichi roused
himself from his deathbed, rushed to the imperial palace, and obtained the appointment of Yoritada as regent and the demotion of
Kaneie from Major Captain of the Right (a distinguished post that
he held concurrently with his office of Major Counselor) to the inferior title of Minister of Popular Affairs. Whatever the precise circumstances, it seems likely that by Yoritada's appointment and
Kaneie's demotion Kanemichi was paying off old scores accumulated during the years of humiliation he had suffered before becoming regent.
Yoritada, who may have been an innocent pawn in the angry rivalry between Kanemichi and Kaneie, had no blood ties with the
emperor and little stomach for confronting Kaneie in a struggle for
power at court. He soon moved to make amends to his powerful
kinsman, appointing him Minister of the Right in 978, the year following Kanemichi's death, and the two seem to have got along together amicably enough thereafter. But it was undoubtedly a frustrating time for the ambitious Kaneie, whose marital links with the
imperial line placed him in an excellent position eventually to claim
the regency and leadership at court. His daughter Choshi, a consort
Cambridge Histories Online © Cambridge University Press, 2008
66
THE HEIAN COURT, 794-IO7O
of the ex-emperor Reizei, had already borne three potential heirs to
the throne. Another daughter, Senshi (or Akiko; later titled Higashisanjo-in; 962-1002), produced Emperor En'yvi's first son in 980.
A break in the impasse, as Kaneie must have thought of it, occurred in 984, when Emperor En'yu, acting under pressure from
Kaneie, abdicated the throne to Reizei's son Kazan, and Senshi's son
by En'yu, the future Emperor Ichijo (980-1011, r. 986-1011), was
appointed crown prince. Kaneie was then but a single abdication or
death away from the status of maternal grandfather to an emperor.
The mother of the new emperor had been Koretada's daughter
Kaishi (she had died in 975). In the absence of strong leadership
from Yoritada, who now retreated almost entirely from active involvement in government, it was, according to a thirteenth-century
source, Koretada's son Yoshichika (957—1008) who, as a maternal
uncle of Kazan, wielded the principal power at court, though never
himself holding high ministerial office or the post of regent. A young
and active man, Yoshichika is reputed to have plunged with enthusiasm into the business of government, aided by an equally young and
vigorous distant kinsman, Fujiwara no Koreshige (953-89), whose
exceptional authority during Kazan's reign earned him the sobriquet
of Fifth-Rank Regent. During the reign of Kazan the court attacked
some of the major governmental problems of the day, attempting to
increase the supply of currency, stabilize prices, and regulate the
growth of shoen, and it is possible that those measures were the
handiwork of Yoshichika and Koreshige. But however much influence they may have actually wielded at court (and there is reason to
doubt that it was as extensive as late sources would suggest) their
power was of brief duration. After less than two years on the throne,
Kazan abdicated; two days later, on the same day that Yoshichika
and Koreshige joined the newly tonsured ex-emperor in holy orders,
Kaneie was appointed regent, the first maternal grandfather of a
reigning emperor to hold that post since Yoshifusa's time.
The circumstances surrounding Kazan's abdication are not
recorded in contemporary sources, but a twelfth-century work attributed it to the machinations of Kaneie. He is said to have employed his son Michikane (961-95) to play upon Kazan's grief following the death of a well-loved consort and persuade him to
abdicate the throne for Buddhist monkhood. Whatever the truth of
that famous story may be, the identification of Kaneie as the chief
instigator of the abdication seems plausible, given his ambition, his
ruthlessness, and the rewards he reaped from the change of emperCambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA REGENCY
67
ors.With the new reign, he became maternal grandfather of the emperor; a few weeks after Ichijo's accession, another of Kaneie's
grandsons was appointed crown prince (the future Emperor Sanjo,
Reizei's son by Kaneie's daughter Choshi); and still another of Kaneie's daughters immediately became a consort of the new crown
prince (who was also her nephew). The accession of Ichijo clearly
marked the beginning of the high tide of Fujiwara fortunes.
Following his elevation to the post of regent, Kaneie operated with
a high hand at court, treating the emperor himself as the cipher he
was, rapidly promoting his three sons over the heads of other nobles
to high posts in the Council of State, and further blurring the distinction between public weal and private fortune. When he fell fatally
ill and resigned his offices in 990, after four years of extravagant and
often arrogant rule, his powers did not pass in bureaucratic fashion
to the senior officials in the Council of State but devolved in royal
style on the family heir, his eldest son, Michitaka (953-95).
Michitaka continued the marital and family politics of Kaneie, favoring his relatives, especially his second son, Korechika (974-1010),
in official appointments and promoting his daughter Teishi (9761001) from the position of imperial consort to empress. When the
hard-drinking regent fell ill in 995 and was no longer able to attend
to court business, he obtained private inspection powers for Korechika but failed in his attempt to have his son named regent, possibly because of the opposition of his sister Senshi, the emperor's influential mother. Michitaka was succeeded as regent, therefore, by
his disgruntled younger brother Michikane, who, a late source says,
had earlier expected to follow Kaneie in the post because of the service he had rendered in persuading Kazan to abdicate. Michikane,
however, was already gravely ill at the time of his appointment, and
he enjoyed the office for only nine days before his own death raised
the question of regental succession again. (Counting from the date
on which he formally accepted appointment as regent in audience
with the emperor, Michikane's tenure lasted only seven calendar
days; he is thus sometimes called the "Seven-Day Regent.")
The contenders for the regency following Michikane's death were
his nephew Korechika and his younger brother Michinaga (9661028). Korechika was eight years Michinaga's junior but senior to
him in office. Thanks to an epidemic that had decimated the ranks of
the senior court officers, he was also, as Palace Minister, the highestranking member of the Council of State. His claims to the regency,
later sources say, were supported by his sister Teishi, Ichijo's empress,
Cambridge Histories Online © Cambridge University Press, 2008
68
THE HEIAN COURT, 794-IO7O
while those of Michinaga were forcefully pressed by the emperor's
mother, Michinaga's sister Senshi. Whatever doubt there may have
actually been about the choice was resolved three days after
Michikane's death when Michinaga was granted private inspection
powers, beginning a thirty-year rule that marked the apex of Fujiwara
power and influence. Although Michinaga is usually regarded as the
Fujiwara regent par excellence, he actually held that post for only a
little more than a year in 1016 and 1017 and otherwise exercised his
control of the government mainly through the private inspection authority he held and through his ministerial posts, or, following his resignation from office and entrance into holy orders, through his son.14
By the time of Michinaga's investiture with private inspection
powers in 995, rivalry between him and Korechika over offices and
family succession seems to have led to such deep and open hostility
between the two kinsmen that both could not remain comfortably
together at court. It was not long before Michinaga found, or was
given, a means of consigning his nephew to the same fate that had
overtaken Sugawara no Michizane and Minamoto no Takaakira: demotion and exile to Kyushu.
Korechika's fall began with an armed attack instigated by him
against the ex-emperor Kazan, the result apparently of Korechika's
mistaken belief that Kazan, despite his holy orders, was a rival for
the affections of a lady in whom Korechika was interested. The exemperor was not injured and is said to have had no interest in pressing charges against his assailants, especially since the attack occurred
as he was returning from a tryst with the sister of Korechika's lady.
Even so, the court, led by Michinaga, launched an investigation. Evidence of wrongdoing by Korechika - including the harboring of
warriors in the capital, the use of black magic against Senshi, and the
performance of Buddhist esoteric rites forbidden to all but the imperial house - was soon uncovered, and Korechika, his brother
Takaie (979-1044), and others were condemned and exiled for treason. The banishment of Korechika and Takaie left their sister Teishi,
Ichijo's consort, isolated at court, and she soon took Buddhist holy
orders. The brothers were pardoned in a general amnesty in 997 and
allowed to return to the capital, but Korechika never succeeded in
regaining his position at court.
14 Michinaga's diary, Mido kampaku ki, survives in large part and has been translated by
Francine Herail, Nous journalieres de Fujiwara no Michinaga, ministre a la cour de Heian
Cgg5~ioi8): Traduction du Mido kanpakuki, Hautes etudes orientales II, 23, Institut des
hautes etudes japonaises (Geneva: Librairie Droz, 1987).
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA REGENCY
69
With Korechika safely out of the way, Michinaga turned his attention to marital politics. In 999 Teishi, nun though she had become, bore Emperor Ichijo's first son, but just a few days before the
birth Michinaga had placed his eldest daughter Shoshi (or Akiko;
later titled J6t5mon-in; 988-1074) in the emperor's harem. In the following year he shocked even his admirers by arranging for the unprecedented appointment of Teishi (or Sadako) and Shoshi as concurrent empresses of the same emperor, Teishi holding the usual title
of "Lustrous Heir-bearer" (kdgo) and Shoshi that of "Inner Palatine"
(chugu), a toponymically derived equivalent coined for the occasion.
By the time of Ichijo's abdication in i o n , Shoshi had borne him two
sons, the eldest of whom, the future Emperor Go-Ichijo, then became crown prince under the new emperor, Sanjo, Reizei's son by
Michinaga's sister Choshi. Although Teishi's son was Ichijo's eldest
and also a favorite of Shoshi, who had raised him after his mother's
death in 1001, he had little or no chance of ever being chosen heir
apparent, since he lacked powerful maternal relatives at court.
The new emperor, Sanjo, at thirty-five was four years older than
his predecessor and less closely connected with Michinaga because
his mother, Michinaga's sister, had been dead for nearly thirty years.
It was perhaps for that reason and also because Sanjo stood between
the Northern-House leader and the status of imperial maternal
grandfather - just as his half brother, the unfortunate Kazan, had in
Kaneie's day - that relations between him and Michinaga seem to
have been strained. Before his accession, Sanjo had taken as consort
a daughter of Michinaga's cousin Fujiwara no Naritoki (941-95),
who had borne him four sons, including his eldest, Prince Atsuakira
(994-1051). In the years just before Sanjo's accession, however,
Michinaga had given the emperor as a consort one of his daughters,
Kenshi (994-1027), and soon after the accession Kenshi, despite her
youth and lack of children, was appointed empress (chugu), clearly
at the behest of Michinaga. Two months later, Naritoki's daughter
Seishi (972-1015) was also appointed empress {kdgo), but under humiliating circumstances arranged by Michinaga, who made certain
that virtually no one attended the ceremony of appointment.
That kind of pressure on Sanjo seems to have been nearly constant.
Michinaga and his intermediaries took every opportunity to impress
on the emperor the wisdom of abdication, especially after his eyesight
failed. Sanjo was resentful of the pressure, but at last agreed in 1016 to
do as Michinaga wished in exchange for the designation of Atsuakira,
his son by Naritoki's daughter Seishi, as the next crown prince. MichiCambridge Histories Online © Cambridge University Press, 2008
JO
THE HEIAN COURT, 7 9 4 - I O 7 O
naga's maternal grandson Go-Ichijo then came to the throne at the
age of eight, fourteen years younger than his heir apparent. Atsuakira's
only hope of ascending the throne probably lay in the survival of his
father until the death or abdication of Go-Ichijo, but Sanjo died in the
year following his abdication (1017), and just three months later
Atsuakira, probably succumbing to pressure from Michinaga, renounced his title in return for appointment to the status of a retired
emperor (he was known then as Koichijo-in) and marriage to still another of Michinaga's numerous daughters. The new crown prince was
Go-Ichijo's brother, Go-Suzaku (1009-45, r- 1036-45)> who, like him,
was a son of Shoshi and a grandson of Michinaga.
Michinaga was now at the peak of his power. He had been appointed regent (sessho) on the day of Go-Ichijo's accession, and he
ruled over a Council of State composed almost entirely of relatives
by blood or marriage. His ties with the imperial line were as complex as they were close, involving at times relations that were just
short of incestuous. Two of his daughters, the highly influential
Shoshi and Sanjo's willfully extravagant consort, Kenshi, were exempresses; the emperor (Go-Ichij5) and the crown prince (GoSuzaku) were his grandsons; in 1018 his nineteen-year-old daughter,
Ishi (999-1036), became the empress (chugu) of her ten-year-old
nephew Go-Ichijo, who thus became son-in-law as well as grandson
to Michinaga. Three years later the same complicated set of relationships resulted when another daughter, Kishi (1007-25), became
a consort of the crown prince. At the same time, Michinaga's eldest
son, Yorimichi (992-1074), was being promoted in office and rank
with unprecedented speed, becoming a senior noble (kugyo) at the
age of fifteen (his peers were mostly in their thirties, forties, fifties,
and sixties) and receiving appointment as Provisional Middle Counselor at eighteen.15 It was during this period that the great regent
composed his most famous poem, a crow of pride on the occasion
of the appointment of his daughter Ishi as Go-Ichijo's empress: "No
waning in the glory of the full moon — this world is indeed my
world!" His own diary also shows him a happy and triumphant man
a few days later when he received at his Tsuchimikado mansion simultaneous visits from his grandsons, the emperor and the crown
prince, and his daughters, the three empresses.16
15 Tsuchida Naoshige, Ocho no kizoku, vol. 5 of Nihon no rekishi (Tokyo: Chuo koronsha,
1981), pp. 11314) 279301.
16 Shoyuki, part 10, vols. 1-11 of Dai Nihon kokiroku (Tokyo: Tokyo Daigaku Shiryo Hensanjo, 1959-86), vol. 5, p. 55, Kannin 2 (ioi8)/io/i6.
Cambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA REGENCY
71
Michinaga carried all before him at Kyoto, crushing his enemies,
disposing of people and official posts mostly as he saw fit, and on occasion employing state resources for his own private purposes: for example, in the rebuilding of his mansion following a fire in 1016 and
in the construction of his great temple, the Hojoji, after he had become a Buddhist monk in 1019. He and his principal wife, Minamoto no Rinshi (964-1053), whose family wealth and connections
may have paved the way for Michinaga's initial successes, were
granted almost every imaginable court honor, stopping barely short,
it sometimes seemed, of the imperial position itself. So complete was
the family's domination of the court and the capital, so luxurious its
life, and so rich the civilization its patronage spawned, that subsequent generations of the Fujiwara nobility tended to look back on
Michinaga's time as the golden age of court society and to draw from
it the standards and precedents by which court and noble life were
regulated. It was in that important sense the classical age of Japan.
Michinaga resigned the regency to his eldest son, Yorimichi, in
1017 and took Buddhist holy orders because of a grave illness in
1019, but his domination of the court continued nearly unchanged
until his death early in 1028, as his sobriquet "the Sacred Hall Regent" suggests. Much of his success may be attributed to the fertility of his wives, who produced seven sons and eight daughters for
him, and also to the remarkable ability of his daughters to bear male
offspring to their imperial husbands. His heir, Yorimichi, was not so
fortunate. Yorimichi's adopted daughter was married to Go-Suzaku
but managed to produce only girls, and the single daughter that
he himself fathered during his long life proved barren when she was
placed in the harem of Go-Suzaku's successor Go-Reizei (1025-68,
r. 1045-68). Yorimichi's brother Norimichi (996-1075) also attempted to preserve the family's maternal link with the imperial
house by giving a daughter to Go-Reizei, but she had no better luck
than Yorimichi's daughter. Consequently, although Yorimichi's position remained secure during the reigns of Go-Ichijo, Go-Suzaku,
and Go-Reizei, who were all sons of his sisters, when Go-Suzaku abdicated the throne to Go-Reizei in 1045, the former emperor, in the
absence of a Fujiwara grandchild, was able to install as crown prince
his son by an imperial princess. The next accession, that of Go-Sanj6
(1034-73, r. 1068-73), brought to the throne therefore the first emperor born of a non-Fujiwara mother since the accession of Uda in
887. Because neither Yorimichi nor Norimichi, who succeeded his
brother as regent in 1068, seems to have been an able enough politiCambridge Histories Online © Cambridge University Press, 2008
72
THE HEIAN COURT, 7 9 4 - 1 0 7 0
cian to maintain the position of the Northern House without the
support of blood ties to the emperor, control of the court slipped
from regental hands as the imperial line, chiefly under the leadership
of retired emperors, reasserted its authority for the final century of
court rule and brought to an end the period of the Fujiwara regency.
(On rule by retired emperors, see Chapter 9.)
Despite Yorimichi's failure in marital politics, he was not, apparently, the inconsequential figure that historians have sometimes
painted him. It was during his time and probably under his leadership that an important reform of the land and tax system took place.
This reform recognized further contraction of the government's
fiscal base, but also seems to have succeeded in containing the growing strength of the provincial gentry and in securing the revenues
with which the court supported itself during the final century of its
autonomous rule.
The reform was initiated in the 1040s, probably in response to
strong pressure from cultivators and landholders in the provinces
closer to Kyoto. At least some of them had been in open conflict with
the provincial governors since the last part of the tenth century and,
by about the beginning of the eleventh century, were abandoning at
an alarming rate tillage in the public realm for other occupations or
for newly opened lands that had been commended to Kyoto patrons
as privately established (i.e., taxable, not officially sanctioned) shoen.
The chief causes of complaint, as registered in one famous appeal
addressed to the central government in 988, were high taxes and illegal exactions. Provincial governors since the beginning of the tenth
century had had the legal authority to set tax rates at their own discretion. Impelled by the need to meet the demands of the central
government and of their patrons at the capital, and no doubt also by
a desire to fill their own pockets, they seem frequently to have resorted to higher and higher taxes, or to exploitive measures that had
the same effect as higher taxes. The court government, probably generally ignorant of the conditions in the provinces that bred the conflict, initially responded vigorously to complaints brought against especially greedy governors, disciplining them by dismissal from office.
Under Michinaga, however, there seems to have been a realization
that the court's revenues depended on just such rapacious officials,
and a more lenient attitude appeared, encouraging some governors
to suppress complaints against them by force. But continuing pressure from the cultivators and landholders, sometimes accompanied
by violent assaults on the governors or arson at their Kyoto resiCambridge Histories Online © Cambridge University Press, 2008
THE FUJIWARA REGENCY
73
dences, convinced the court eventually that relief had to be given.
The result, it is thought, was the reform of the 1040s, which can be
seen as an attempt to appease the cultivators and landholders in the
provinces while preserving a more assured, if reduced, income for
the government and its officials. The chief known features of the reform were the initiation of fixed taxation rates not subject to change
by provincial administrators and recognition by the central government of all but the most recent shoen that had been formally established with tax immunities by the authority of provincial administrators. The reform seems also to have required the return to the public
realm of the privately established, taxable shoen that had been set up
in great numbers in the name of the regent and other powerful patrons at the capital after the beginning of the tenth century.
The official recognition of all but the most recent shoen represented an important retreat from the court's long-standing effort to
restrict shoen to those officially established before 902. That retreat,
together with the fixed taxation rates, must have been especially well
received in the provinces. In return for those concessions, the government obtained a certain measure of peace, probably better cooperation in the collection of taxes, and an increase in the taxes leviable on shoen lands returned to the public realm.
Whether the reform actually changed the total amount or distribution of revenue flowing into Kyoto is unknown. Insofar as the exploitive habits of the provincial governors were checked (and success
in that respect was by no means complete), the regent's house, in
particular, may have suffered a loss of income, but if the efficiency of
tax collection was in fact increased by the more willing cooperation
of the provincial gentry, who were the primary tax agents, and if the
privately established shoen were actually incorporated into the public
realm, revenues may have held steady or perhaps even increased. On
the other hand, further court efforts in 1055 and 1069 to halt the
growth of provincially authorized, tax-immune shoen suggest that despite the government's policies, its tax base continued to erode.
The reforms of the 1040s also included the beginning of the transformation of the statutory administrative-area system that led eventually in most areas outside the east to the elimination of subprovincial districts {gun) as functioning administrative units and the direct
subordination of lesser units with various designations (known generally as betsumyo, "special nominals") to the provincial governor's
office. It was during the same period that the several specific names
of statutory taxes also disappeared from use.
Cambridge Histories Online © Cambridge University Press, 2008
74
THE HEIAN COURT, 7 9 4 - 1 0 7 0
Following the suppression of Taira no Masakado's revolt in 940,
his conqueror, Taira no Sadamori, and his descendants were the
leading warriors of the eastern provinces. The next serious incident
in the east began with the seizure of tax revenues in Kazusa and Shimosa by the warrior Taira noTadatsune (d. 1031) in 1027, which escalated into a revolt not unlike that of Masakado's. The court sent a
punitive force led byTaira no Naokata, but when, after several years
of campaigning, Naokata had failed to capture the rebel, he was relieved of his command and replaced by Minamoto no Yorinobu
(968-1048), a son of Mitsunaka of the Anna Affair, who immediately
effected Tadatsune's surrender. From this feat the reputation of the
Seiwa Genji rose as a major force in the east.
Later in the eleventh century, more extensive campaigns were
fought in the far northeast against the Emishi (or Ezo) who had
halted the payment of taxes. Two long campaigns were mounted,
known as the Earlier Nine Years'War (1051-62) and the Later Three
Years'War (1083-87), before a stable order was reestablished in the
region. The first campaign was ordered by the court, which appointed Minamoto no Yorinobu's son Yoriyoshi (988-1075) as commander, but the second war was an enterprise of the latter's son,
Yoshiie (1039-1106), which the court refused to authorize or support. (A detailed accounrof-this military history can be found in
Chapter 10.)
REGENCY GOVERNMENT
The Fujiwara regency was in many aspects simply a prolonged and
institutionalized phase in the cyclically shifting balance of power between the imperial line and the noble clans that had for shorter periods and in less formal ways been characteristic of Japanese government since at least the seventh century. It was distinguished from
earlier phases of noble domination of the court by its long history,
by its inheritance in a single family line, by the greater absoluteness
of its domination, and by its coinciding with a period when changing economic and social conditions had much reduced the scope of
the central government's authority and made personal relationships
and private purposes salient features of governmental operations. Its
essential legitimacy, however, continued to be derived from the person of the emperor and from the by then hallowed, if much truncated and changed, framework of the statutory regime. Its authority
and power were, moreover, never absolute: not even the most powCambridge Histories Online © Cambridge University Press, 2008
REGENCY GOVERNMENT
75
erful regent was able to govern autocratically in complete disregard
of the emperor and the Council of State. The view of perhaps most
modern historians is, in fact, that the terms "regency government"
and "regency period" are misnomers in the sense that the regency
was not the significant distinguishing characteristic of the period but
simply an adjustment in, or relatively limited transformation of, the
central government mechanism centering on the Council of State. It
was, in that view, the change in the government's policies toward
provincial administration and taxes at the beginning of the tenth
century that marked a genuinely new phase in Heian history.
The period of the regency was a conservative time when the chief
object of the court was to preserve such authority and resources as
it had, rather than to strike out on radically new paths or to seek expanded wealth and power. Within those limits, however, the government played an active and often effective role in directing the affairs
of the nation. It is undeniably true, as has often been observed, that
much of the central government's attention was focused on the internal affairs of the court itself, especially on ceremonial and ritual,
which were conceived to be essential, and indeed perhaps the most
essential, elements of governmental operations. That was inevitable
perhaps in a society of hereditary nobility where status defined all
roles and where the chief means of coping with any major crisis,
whether political, economic, social, or personal, was recourse to divine intervention, which all agreed was superior in efficacy to anything that man might otherwise do. But the court, its ceremonies,
and its rituals were by no means the whole of regency government,
as has sometimes been alleged. The regent and the Council of State
ministers were actively engaged in the formulation and implementation of practical, mundane policy for a broad range of problems with
nationwide import. It was they in their Guard-Post Judgments and
other councils who directed Japan's diplomatic, cultural, and commercial relations with foreign states and supervised defenses against
foreign piratical invaders; who oversaw efforts to cope with domestic revolt and warfare in the provinces; who instituted a series of attempts to regulate and halt the growth of shoen; who continuously
adjusted the statutory institutions to the evolving social and economic conditions of the day in order to preserve a flow of income
from the lands remaining under public control; and who dealt with
countless other major governmental, judicial, and police matters.
The quest for the regency and control of the court was not an empty
game played simply for honors and personal advantage. Although
Cambridge Histories Online © Cambridge University Press, 2008
76
THE HEIAN COURT, 794-IO7O
the statutory regime was by this time but a pale shadow of its former
self, the regency still represented the greatest concentration of governmental power and wealth in Japan.
The establishment of the first Fujiwara regency seems to have
been the result of a search for a means of giving institutional recognition to the overwhelmingly dominant position occupied by Yoshifusa at court. It did not represent, in other words, the seizure of new
power through an existing office but, rather, the confirmation of existing power through the creation of a new office. The uniquely sovereign powers Yoshifusa exercised were first formally recognized by
his appointment to the post of Chancellor in 857. But the vague, exemplary duties given to that post by the code may have come to
seem inadequate after Yoshifusa's grandson, Emperor Seiwa, came
to the throne in 858, and especially after the emperor reached adulthood. It was therefore probably to give more substantive content to
the Chancellor's position that an imperial rescript was issued in 866
specifically directing Yoshifusa to take charge of the government as
regent, the title itself (sessho), however, being only implied in the rescript. That was an ad hoc arrangement without legal basis in the
statutory code or in any of the amendments thereto, and it is important to note that even after the regency was firmly established, its
ad hoc nature persisted. The holder of the title had to be reappointed
to his post by each succeeding emperor, instead of continuing from
reign to reign, as with regular governmental appointees.
The problem of defining the Fujiwara leader's position at court
seems to have continued in the time of Mototsune, and was perhaps
in some way a chief cause of the otherwise rather mysterious Ako
Controversy in 887 (see earlier in this chapter) and the similar impasse that developed when Mototsune was appointed Chancellor in
880. Mototsune evidently wanted to create a position that stood
apart from and superior to the regular statutory organs of government but was not rendered ineffective by isolation from them. His
solution was basically the same as Yoshifusa's: a combination of the
ad hoc office of regent with a statutory ministerial post. The chief
difference was that after Mototsune's appointment as regent in 876
while holding the title of Minister of the Right, the chancellorship
was clearly differentiated from the regental position. Eventually it
was even possible for the regency and the chancellorship to be held
by different men, as when Yoritada retained his title of Chancellor
after Kaneie succeeded him in the regency in 986. That the regency
continued for some time, nevertheless, to be viewed chiefly as the
Cambridge Histories Online © Cambridge University Press, 2008
REGENCY GOVERNMENT
JJ
definition of the duties of a regular statutory minister is suggested by
the fact that until the time of Kanemichi a regent who held, for instance, the title of Minister of the Right was considered inferior for
protocol and other purposes in the Council of State to a nonregental Minister of the Left (the senior title). The same view is reflected
also in the fact that before Kaneie's day the chieftainship of the Fujiwara clan was always held by the clansman with the highest statutory office and rank, regardless of who was regent. But the position
of regent had become fully independent of Council of State offices
by the time of Kaneie, who resigned his office of Minister of the
Right after he was appointed regent and during most of his tenure
held no other Council title.
It was inTadahira's time that a clear distinction was made between
the titles of sessho, a regent to a minor emperor, and kampaku, regent
to an adult emperor. Initially, the distinction was a significant one,
since the sessho was able to act as the emperor in approving official
documents, in the conduct of ceremonies, and in all other official capacities, whereas the kampaku only acted for the emperor in his dealings with the Council of State and was not otherwise empowered to
represent him. The sessho in effect became the emperor, as a twelfthcentury courtier said, while the kampaku remained a minister to the
emperor. The practical significance of the distinction seems to have
been lost, however, as the post of regent came to be regularly filled
and the alternation between the titles of sessho and kampaku became
simply a mechanical function of the emperor's age. The terminological distinction itself, nevertheless, was always faithfully maintained.
A third regental title, nairan ("private inspection"), first employed
in the time of Tokihira and Sugawara no Michizane, gave its holders
powers mostly identical with those of a kampaku. It appears to have
been of lower official status, though, and to have differed significantly in that a nairan regent regularly took an active part in the
business and meetings of the Council of State, unlike regents holding the titles of sessho and kampaku, who were generally removed
from direct involvement in the Council's affairs. There was probably
always a danger, recognized perhaps as early as Mototsune's day,
that as a result of its isolation from the operating organs of the government, the regency might, like the imperial throne, become largely
a figurehead or honorary position devoid of actual power. It was perhaps because of such considerations that an active politician and
leader like Michinaga preferred the nairan title during most of his regental career. He took advantage of the freedom it gave him to exerCambridge Histories Online © Cambridge University Press, 2008
78
THE HEIAN COURT, 794-IO7O
cise direct administrative leadership over the Council of State, appearing at court almost daily and presiding over most important
Council meetings. There was consequently little of importance that
went on in the government of which he was not directly aware,
which no doubt helps explain the extraordinary power he enjoyed as
regent.
Government finances under the Fujiwara regency continued to
depend on tax revenues derived through the provincial administrations from public lands, but the period saw substantial growth in the
areas of rice tillage under private control, and shoen came consequently to play an important role in the economy of the capital.
Most of the new privately controlled tillage seems to have been land
recently opened to cultivation by provincial gentry, who, to prevent
incorporation of their lands into the tax units of public land called
"nominals" (myo), commended them as privately established shoen
to influential nobles and religious institutions in the capital area.
Such shoen were not officially recognized by the provincial governments and were therefore still subject to the taxes levied on newly
opened tillage, but by their exclusion from the "nominal" system,
they escaped the heavy taxes based on the public land in the units of
that system. The most influential potential patron of shoen was, of
course, the regent, who became therefore the beneficiary of what
was undoubtedly the largest share of such commendations. Chiefly
by that process, apparently, the regent's house accumulated such a
large number of shoen in Michinaga's time that one of his contemporaries complained hyperbolically that there was not a needle's
breadth of land in the country that did not belong to him. Nevertheless, despite the size of the regent's shoen holdings in the time of
Michinaga and Yorimichi, most of the commendations appear to
have been largely fictitious arrangements, and only a few shoen actually yielded significant income to the regental purse. Until past the
middle of the eleventh century, the family probably still obtained the
major part of its income from the official posts and ranks of its leaders. The middle and lower ranks of the nobility, on the other hand,
who had been the first to suffer when the decline in governmental
revenues in the ninth and tenth centuries forced a curtailment of expenditures, relied heavily on shoen income during this time to replace or supplement their governmental stipends. A similar fate
awaited the regent's house itself in the following century, when control of the central government passed into other hands.
The regent's income was supplemented by gifts of labor and
Cambridge Histories Online © Cambridge University Press, 2008
REGENCY GOVERNMENT
79
goods that he frequently received from provincial officials. The tax
system adopted at the beginning of the tenth century and the quasiautonomy of provincial administration seem to have made provincial
posts so highly lucrative that the provincial administrator became a
common metaphor in the contemporary literature for ostentatious
new wealth. Since the regent usually exercised a controlling influence in the selection and appointment of such officials, it may be assumed that their munificent gifts were made in recognition of past
favors and in hopes of future appointment. Michinaga, for example,
received a nearly continuous stream of horses and oxen from provincial officials, and, as mentioned earlier, much of the financial responsibility for the reconstruction of his Tsuchimikado mansion and
for the construction of his H6j5ji Temple was assigned to provincial
governors. The governor of Iyo, Minamoto no Mitsunaka's sonYorimitsu (948-1021), supplied all the furnishings for the reconstructed
Tsuchimikado mansion, which were so fabulously extravagant that
they left even the wealthiest nobles agog, and there was competition
in the nobility to obtain copies of the catalogue of gifts that accompanied the furnishings. A wall of onlookers, straining to get a
glimpse of the gifts, is said to have lined the streets when they were
transported from Yorimitsu's house to Tsuchimikado. Such income,
however, was only as reliable as the provincial officials' control of
their provinces and the regent's control over appointments, and both
types of control suffered drastic change in the last half of the
eleventh century.
Probably as a direct consequence of the growth of their shoen
holdings, the Fujiwara regents found it necessary to expand greatly
the household administrative office that they were allowed to staff
and maintain under the provisions of the statutory code. As provided
for under the code, such offices were staffed by at most half a dozen
titled officials holding various court ranks, some as high as the ranks
held by the heads of regular government bureaus or governors of
provinces. By Yorimichi's day, however, the staff of the regent's
household office had grown to include at least twenty court-ranked
officials, who held a wide variety of specialized titles and managed
what was clearly a large private empire of mansions, palaces, and
land. Under Michinaga, most such officials were at the same time
provincial administrators, who must have been especially useful in
filling the coffers of the house and in enforcing its orders at the shoen
level. But from Yorimichi's time on, the number of provincial administrators on the household staff decreased and the posts tended
Cambridge Histories Online © Cambridge University Press, 2008
80
THE HEIAN COURT, 794-IO7O
to become hereditary. This reflected perhaps both the expanding
role of shoen in the economy of the regent's house and the consequent increased need for experienced administrators, as well as the
loss of regental control over the appointment of provincial officials
during the reign of Go-San jo and afterward.
Perhaps because of the size and far-ranging activities of the regent's household office, which may have occasionally impinged on
governmental prerogatives, and because of the central position
occupied by the regent in Kyoto political life, one eleventh-century
diarist asserted that in his day the regent's house had become the
imperial court. There was probably a good deal of truth in that
observation, since political power flowed from the regent and much
governmental business was transacted at his house. But it is also true
that the court, with its regularly established offices, its fixed procedures and its documents, remained both the formal and, for the
most part, the actual center of government.
FOREIGN RELATIONS, 794-IO7O
Japan's relations with the other countries of East Asia during the
Heian period were driven by the familiar twin engines of fear of external power, on the one hand, and desire for material and cultural
gain on the other, and they were typically structured by trilateral interrelationships among Japan, China, and the Korean peninsula. The
proximate source of the fear tended to be Korea, in the affairs of
which Japan had at times been deeply involved since early historical
times. The source of desire, although more evenly distributed, was
mostly concentrated on the riches of China, either directly or as filtered through the states of Northeast Asia. During the period of present concern, the desire never developed into territorial ambition,
and therefore Japan's neighbors did not become the victims of aggression by forces of the Japanese government. Nevertheless, the
poorly informed Kyoto authorities were sometimes alarmed by rumors of impending foreign assaults on Japan, and raids on Iki,
Tsushima, and Kyushu by Korean and other pirates occasionally
brought those fears to a white heat.
Japan's earliest substantial foreign relations were with the Korean
peninsula. When those relations first dimly appear in historical
sources for the fourth century, the peninsula was occupied by the
three principal kingdoms of Koguryo in the north, Paekche in the
southwest, and Silla in the southeast. In the latter half of the fourth
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
8l
century, the early Japanese kingdom of Yamato appears to have held
or occupied in some fashion a territorial base (called Mimana in
Japanese sources) at the south-central tip of the peninsula between
Paekche and Silla (the area called Kaya). What the nature of the
Japanese interest in Korea was at that time is unknown (ethnic and
cultural affinities may have been the main elements), but it led first
to military alliances with both Paekche and Silla against the expansionist pressure of Koguryo, and then, when Silla leagued itself with
Koguryo, to support of Paekche against the other two. The Japanese
launched repeated attacks on Silla, apparently both from their Korean base and also directly from Tsushima. But at the beginning of
the fifth century their activities in Korea were checked by a major
defeat at the hands of Koguryo armies, and they were expelled from
the peninsula in the mid-sixth century.
Beginning in the late sixth century, Koguryo came under attack by
Chinese armies (first those of the Sui dynasty [581-618], then of
T'ang [618-907]), giving Paekche an opportunity to launch an assault on its longtime enemy Silla, now unsupported by the otherwise
occupied Koguryo forces. In response, Silla allied itself with China,
which in 660 sent a naval force against Paekche. The Paekche king
sought and received military support from Japan, but in 663 the
Japanese forces were crushed by the Chinese in a naval engagement
off the southwest Korean coast at the mouth of the Kum River (the
Battle of Hakusuki, or Hakusan, Estuary, as it is known to Japanese
historians), and Japanese military intervention in the peninsula was
soon at an end, not to be renewed for some nine centuries.
With the destruction of Paekche and Koguryo by Chinese and Sillan attacks in the 660s, Silla was finally able to unite the peninsula
south of P'yongyang under its rule, the former territory of Koguryo
to the north being then occupied by the Tungusic state of Po-hai, as
it is generally known (Parhae in modern Korean). It was that resulting distribution of power in the peninsula, together with the long
history of Japanese-Sillan hostility, that provided some of the chief
determinants of Japanese foreign relations until as late as the last
half of the tenth century.17
Japan's relations with China began very early but did not rival
17 There is disagreement among historians, especially between Korean and Japanese historians, about the nature and extent of early Japanese involvement in Korean affairs. The present account owes more to the commonly accepted Japanese interpretation, as summarized, for example, in Inoue, Nihon rekishi taikei, pp. 16-18 and 273-93, which also presents
the chief alternative views.
Cambridge Histories Online © Cambridge University Press, 2008
82
THE HEIAN COURT, 794-IO7O
those with Korea in practical importance until the reunification of
China under the Sui and T'ang dynasties made Chinese cultural
and military influence throughout East Asia so overwhelming that
even a seagirt nation like Japan was compelled to place China at the
forefront of its attention. As Sui and T'ang armies became involved
in the complicated affairs of the Korean peninsula, threatening
Japanese allies and interests there, and as Japan became ever more
deeply involved in the assimilation of sinitic culture and institutions, official relations with the Chinese court became indispensable. In 600, when a Japanese army was confronting Sillan forces in
territory at the southern tip of the Korean peninsula claimed by
the Japanese court, a Chinese chronicle records the arrival of an
embassy from Japan at the Sui court, which was allied with Silla
against their common enemy Koguryo. From that time on for more
than two centuries the Japanese government, spurred by military
and cultural concerns, maintained official relations with the Chinese court, becoming in form a tributary state in the Chinese system of international relations.
But by the end of the eighth century much had changed, internally
as well as externally. The compilation of theTaiho andYoro codes at
the beginning of the century had put a capstone on the sinitically inspired structure of the statutory regime and rendered less pressing
the need for study and observation of the operations of the Chinese
government. Several generations of officials had provided a base of
experience and learning, and it was no longer as necessary for student-officials to undertake the long journey to the Chinese capital at
Ch'ang-an in search of the knowledge, books, and techniques required by the court's governmental machinery. National amour propre and pragmatic diplomatic aims had been served by Chinese
recognition of Japan's high status in the Chinese tributary system.18
If Japan could still benefit greatly from intercourse with China, that
was less in the realm of government, where official relations might be
most useful, than in economic, cultural, and intellectual matters,
which were perhaps more amenable to private routes of exchange.
Externally, whatever justification there may have been for Japanese fears of the T'ang armies in Korea after the Japanese defeat in
663, such concerns were presumably much ameliorated in the following decade when the Chinese forces in Korea withdrew north of
18 Charlotte von Verschuer, Les Relations officielles du Japon avec la Chine aux viif et iyf siecles,
Hautes etudes orientates, 21 (Geneva: Librairie Droz, 1985).
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
83
the Taedong River (at P'yongyang), leaving a unified peninsula
under the rule of Silla. True, the outbreak of the An Lu-shan rebellion in China in 755 worried the Nara court sufficiently to cause it
to look to its defenses in Kyushu, and Japanese suspicions about Sillan intentions reached such a feverish pitch at the same time that the
government even laid plans for a punitive military expedition against
the peninsula. Fortunately, however, nothing came of them, and, in
fact, internal troubles in both Silla and China were making it less
and less likely that either state would ever again be able to menace
even its nearest neighbors, much less a nation like Japan, whose
main islands were protected from the continent by over a hundred
miles of intervening sea.
Japanese geography placed the country in the enviable position of
being able to regulate its relations with the adjacent continent largely
according to its own internally generated needs and desires. Although the court sometimes convinced itself that invasion by continental forces was imminent, Japanese defenses before the thirteenth
century were never tested by anything more formidable than marauding pirates, and, even more important, the country never faced
an enemy invasion. The court thus was generally able to determine
the pace and depth of its official relations with the outside world unrestricted by the fierce pressures that characterized relations between states on the continent. For the same reason it could also continue a certain semblance of equality with the Chinese court, which
had been expressed by Empress Suiko in 607 when she began her
message toYang-ti, the mighty emperor of the Sui dynasty, with the
famous salutation, "The Child of Heaven of the land where the sun
rises sends a message to the Child of Heaven of the land where the
sun sets."
The relative freedom of action that the Japanese government enjoyed in that respect, lessened the need for, and attraction of, official
relations with China. Possibly other more humdrum factors, such as
the great cost of equipping and dispatching an embassy of several
hundred people to the continent and the perils of sea travel for a people who apparently did not yet fully understand the prevailing winds
of the East China Sea, combined in the eighth century to reduce the
frequency of embassies to the T'ang court from one every fifteen or
sixteen years in the early part of the century (itself a quadrupling of
the interval in the immediate post-Taika years) to the leisurely rate of
one in every twenty-five or more years toward the end of the century.
And the trend continued after the establishment of the capital of
Cambridge Histories Online © Cambridge University Press, 2008
84
THE HEIAN COURT, 794-IO7O
Heian. Emperor Kammu sent his first and only embassy to the T'ang
court in 804, twenty-seven years after the dispatch of the previous
embassy in 777; and another thirty-four years passed before the next
and final Japanese embassy to China for many centuries departed in
838.19 Nearly sixty years later, in 894, Sugawara no Michizane was
chosen to head another embassy to China, perhaps in response to a
request relayed from Chinese officials (or so it was made to seem),
but before the embassy could be dispatched Emperor Uda and the
Council of State accepted Michizane's recommendation that official
relations with China be terminated. Japanese intercourse with China
thereafter was abandoned entirely to private hands, except for a few
exchanges of messages with the king of the southern Chinese coastal
state of Wu-yiieh around the middle of the tenth century and another
exchange with the Sung court in the 1070s.
By the beginning of the Heian period, the chief remaining reason
for maintaining state relations with the T'ang court seems to have
been trade. The earlier quest for knowledge of Chinese culture and
institutions and the desire to keep abreast of the developments in the
East Asian international world, and to participate as a leading member in the Chinese diplomatic order, had been largely fulfilled. With
the growing disorder in China after the middle of the eighth century,
the attraction of, and need for, relations waned. The material wealth
of the continent was as eagerly sought as ever, but private trade was
available to supply that need at no cost to the court and without the
personal and diplomatic risks of official missions. Under those circumstances, it was not in intent a particularly momentous decision
when the court canceled Michizane's mission of 894. There is no indication that that decision represented the adoption of a new policy
of permanent diplomatic withdrawal. Official relations with China
had been petering out for more than a century, and under different
subsequent historical conditions in both Japan and China they
might, in the natural course of things, have eventually been resumed.
But such was not the case. The cancellation of the embassy of 894
turned out to be, in fact if not in intent, the end of the exchange of
official envoys between the two nations that had begun in 630 and
numbered by the end more than thirty missions to China from Japan
(including the "sending-off" missions that accompanied Chinese
embassies back to the T'ang court). The century of turmoil in China
19 On the mission of 838, see Edwin O. Reischauer's two volumes: Ennin's Travels in T'ang
China and Ennin's Diary: The Record of a Pilgrimage to China in Search of the Law (New
York: Ronald Press, 1955).
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
85
that followed the collapse of theT'ang empire at the beginning of the
tenth century and the shift of an impoverished statutory regime's attention to internal affairs left little room for the practice of traditional diplomacy. In his proposal of 894, Michizane cited as justification for jettisoning a foreign policy that had endured by his time
for nearly three centuries the chaotic disorder accompanying the decline of the T'ang dynasty. He also mentioned the hazards of travel
to China:
A Request That the Members of the Council of State Decide on the Dispatch of a Mission to the T'ang
Last year in the third month, the merchant Wang No brought a letter
from the monk Chiikan, who is in China. It described in detail how the
Great T'ang is in a state of decline, and reported that the emperor is not at
court [because of the rebellion] and foreign missions have ceased to come.
Although Chiikan is merely a wandering monk, he has shown great loyalty
to our court....
Investigating records from the past, we have observed that some of the
men sent to China have lost their lives at sea and others have been killed by
pirates. Still, those who arrived safely in China have never yet had to suffer
there from hunger and cold. According to Chukan's letter, however, that
which has never yet happened now seems likely to occur. We humbly request that his letter be distributed to all members of the Council of State
and the professors at the university so that they may carefully read it and
consider the merits of this proposal. This is a matter of national importance
and not merely of personal concern
The fourteenth day of the ninth month, in the sixth year of the Kampyo
era [894].2°
Although it may be suspected that Michizane did no more than
state a generally accepted view of the current diplomatic situation,
and although he submitted his request simply to provide a basis for
the formal adoption of a policy that had already been decided on,
the implied reasoning was fundamentally sound and especially convincing, very likely, to a somewhat impecunious court that may have
been less than eager to undertake the huge expenses of outfitting
and dispatching an embassy.
By 894 the T'ang dynasty was tottering toward its final collapse in
907, dragging with it the remnants of the relatively stable and orderly
empire of which it had been the founder and center. The hardpressed Chinese court at the time was in no shape to receive foreign
embassies, and its once great empire was no longer the military
threat it had been in the seventh and eighth centuries, when it and
20 Robert Borgen, Sugawara no Michizane and the Early Heian Court, pp. 242-43, slightly
modified.
Cambridge Histories Online © Cambridge University Press, 2008
86
THE HEIAN COURT, 7 9 4 - 1 0 7 0
its Sillan allies seemed at times on the verge of attacking Japan. War,
revolt, banditry, and piracy were endemic in the continental countries of East Asia, preparing the way for the vast upheavals of the
tenth century, when all the old regimes were swept away by new and
sometimes very different powers. It was a dangerous, confusing
world, and a country that could elect to stay clear of it was doubtless well advised to do so.
The perils of the voyage to China mentioned by Michizane were,
as he himself recognized, nothing new. Only one of the earlier Japanese embassies to the T'ang court seems to have made the crossing
and return completely unscathed, and some suffered catastrophic
losses of life and property. But the dangers of the trip may well have
been even more intimidating in Michizane's day than they had been
in the seventh century, when most embassies seem to have followed
the longer but safer northern route across the Korea Strait, along the
west coast of Korea, and then over to the vicinity ofTeng-chou at the
base of the north coast of the Shantung peninsula. The seventh century was also a period of relative stability in East Asia, when strong
governments in China and Korea were presumably able to exercise
some control over the piracy that Michizane cited two centuries later
as one of the hazards of sea travel.
In the 660s, however, the west coast of the Korean peninsula fell
under the control of Japan's longtime adversary, the increasingly
hostile state of Silla. Thereafter, embassies apparently found it prudent usually to follow a southerly route, making for ports on the
coast of central China either indirectly via the islands south of Kyushu or, later, directly across the East China Sea. The latter direct
route could be quicker if all went well but was also more dangerous,
involving two hundred miles or more of navigation across a body of
water notorious for its great storms. The route became even more
hazardous after the middle of the eighth century as disorder grew in
China and Silla, relaxing whatever restraints had been imposed on
piratical activity, and travel within China itself was dangerous and
hard. By the last half of the ninth century, pirates were making even
the passage along the Seto Naikai from Naniwa (in present-day
Osaka) to Hakata unsafe for official travelers.21
21 On Japan's official relations with China in the eighth and ninth centuries, see the work by
Charlotte von Verschuer cited in note 18. Chapter 5 (pp. 161-86) examines in some detail
the circumstances of the Japanese decision to cancel Michizane's embassy and constructs
a narrative that reconciles apparent inconsistencies in the sources and speculates about additional reasons for the cancellation. On the same subject and to much the same purpose,
see Robert Borgen's study of Michizane, pp. 240-53.
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
87
The cessation of official relations with China did not bring a halt
to intercourse between the two countries. Private Chinese traders
had been a familiar sight in the Dazaifu's port on the Bay of Hakata
since the first half of the ninth century, and they continued to come
now as before carrying the material and intellectual products of
the continent and also providing incidentally transport for Japanese
Buddhist monks traveling to China for study. During the first half of
the turbulent tenth century, it is true, the Japanese government, apparently for reasons of economy and in response to fears of piratical
incursions and foreign attack, adopted a semi-isolationist policy severely restricting the frequency with which Chinese traders were allowed to visit Japan; by the second half of the century unauthorized
travel overseas by Japanese had also been banned. But by the end of
the century enforcement of trade and travel restrictions, which had
been sporadic in any case, was being further undermined by a weakening of the central government's control of the provinces and by the
emergence of alternative ports free to some extent from the supervision and exactions of government officials. Although the trade seems
to have been picayune compared, for instance, to that conducted by
contemporary Arab traders of the Umayyad and Abassid empires, it
meant that Japan remained open to the stimulation and influence of
its surrounding world.
Under the statutory system of private foreign trade that the government sought to enforce in the tenth century, Chinese merchants
were restricted entirely to the Dazaifu port in Kyushu, which they
were allowed to visit only once in three years. The conditions under
which trade was conducted at the port worked further to the disadvantage of merchants, forcing them to sell their choicest goods on
interest-free credit to the government at prices determined by it.
Thus they were exposed to the often realized threat of confiscation
and placed at the mercy and whim of corrupt officials. The appearance of unofficial ports within shoen partially immune from government taxes and law offered Chinese merchants a more attractive and
profitable alternative to the Dazaifu trade. By the eleventh century
they had begun to take full advantage of that opportunity, providing
through the sMen-port proprietors in and around the court a supply
of imported goods for Kyoto noble society that was quite possibly
steadier and more abundant than anything the purposely restrictive
official system of trade had ever permitted. The principal private
ports engaged in the China trade during the eleventh century were
at Hakata, Hakozaki, and Kashii, all just across an intervening river
Cambridge Histories Online © Cambridge University Press, 2008
88
THE HEIAN COURT, 794-IO7O
from the Dazaifu trading and diplomatic office (the Korokan) on the
Bay of Hakata, but there were similar ports elsewhere in Kyushu and
also opposite the capital on the Japan Sea coast of Honshu.
The China trade seems to have brought to Japan mainly aromatics, medicines, fancy silk fabrics and other luxury items, and manuscript and printed texts in book format (the latter imported as early
as 986, when a Japanese monk returned from China with a printed
edition of the Buddhist Tripitaka), but it also may have included
some items like those exchanged during the eighth and ninth centuries between the Japanese and Chinese courts in their official relations or purchased by individual members of the embassies in
China: court costumes, arms and armor, musical instruments, and
such utilitarian objects as an iron measuring rule. In return, the
Japanese are known to have sent to China by the same official and
semiofficial routes pearls, yellow amber, and agate; and Japanese
regulations specified silver, silk thread, "prisms," camellia oil, liana
juice (a sweetener), and gilt lacquerware as part of an embassy's
"tribute" to the Chinese court.22
In the first half of the tenth century, following the collapse of the
T'ang empire, Japan's chief commercial and cultural ties with China
appear to have been concentrated in the successor state of Wu-yiieh
(907-70), one of the "Ten Kingdoms" occupying an economically
rich area in southern China - the Chekiang area, which included the
premier overseas trading ports of Ming-chou (the modern Ningpo,
south of Shanghai) and Kuei-chou. It also included the famous
Buddhist complex on Mount T'ien-t'ai, which, together with the
Wu-t'ai mountains in the north, was a principal pilgrimage objective
of Japanese Buddhist monks during the Heian period. Perhaps encouraged by the founder of the state, Ch'ien Liu (852-932), himself
a former salt merchant, aggressive traders from the area early on
established commercial relations with the Khitan, Po-hai, Silla,
and Koryo. Soon they were also in Japan, where their vessels are
recorded as having arrived on nine occasions between 935 and 959.
More than once the Wu-yiieh king sought to establish official relations with the Japanese court, dispatching personal letters and gifts
to the emperor and his ministers; but the gifts for the emperor were
returned and the king's overtures rebuffed.
The kings of Wu-yiieh were devout Buddhists, and the fifth in the
line, Ch'ien Shu (r. 948-78), seeking to reassemble the texts of the
22 Von Verschuer, Les Relations officielles, pp. 134 ff.
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
89
T'ien-t'ai school lost during the proscription of Buddhism instituted
in 845 by theT'ang emperor Wu-tsung (r. 840-46), sent a request for
replacement texts to the Enryakuji Temple, headquarters of the
Tendai (T'ien-t'ai) school in Japan. In 953 the temple responded by
dispatching the monk Nichien with an unspecified number of texts
for the king. Nichien returned to Japan four years later, bringing
with him Buddhist and other texts and also one of the 84,000 small
stupas made by Ch'ien Shu in an act of devotion imitating the great
Indian king Asoka, who was celebrated for his piety.
China during the early Heian period posed little or no real military
threat, but in contrast, the court was apprehensive about its immediate continental neighbor Silla, which lay on the other side of the
Korea Strait just thirty-five or so miles distant from the island of
Tsushima. In the last half of the eighth century and throughout the
ninth century the government repeatedly ordered the strengthening
of coastal defenses in anticipation of Sillan attacks. Fears were fueled
in one instance in 870 by a Japanese fowler who escaped a Sillan jail
(he had been caught in Sillan waters) and brought back stories of
large-scale Sillan military preparations for an attack on Tsushima.
Ancient animosities, exacerbated by Japan's pretensions to suzerainty
over Silla, by the latter's preference for its strong tributary ties to
China, and by the increasingly bold attacks of Sillan corsairs on the
Japanese coast in the ninth century were at the root of the prickly and
sometimes hostile relations between the two countries.
Official relations between Japan and Silla had been close, closer
than between Japan and China, in the early part of the eighth century. But the last Sillan envoy to reach the Japanese court arrived in
779, and thereafter the relationship became fairly remote and strictly
unilateral, continued only by the inclusion of envoys to Silla in the
Japanese embassies to China of 804 and 838. (The last full-scale
Japanese embassy to Silla was dispatched in 799.) Intercourse between the two countries during the ninth century was maintained
chiefly by Sillan traders, by large numbers of refugees from the revolts and banditry that were bringing the kingdom to its end in 935,
and by the ever-present Sillan pirates. The historic hostility between
the two states seems to have been replicated even among the Sillan
refugees in Japan, who were first settled in the eastern provinces of
Honshu, where harsh conditions and treatment led to uprisings and
revolts (a revolt of 820 is reported to have involved 700 Koreans),23
23 Inoue, Nihon rekishi taikei, p. 741.
Cambridge Histories Online © Cambridge University Press, 2008
9O
THE HEIAN COURT, 7 9 4 - I O 7 O
and then in 824 resettled in the even more remote former Emishi
territories of Mutsu.
The number of Sillan traders visiting Japan grew from the 820s
on, their principal partners in trade being the Kyoto nobility and
local magnates in the Kyushu area. Alarmed especially by the
prospect of alliances between Sillans with their advanced weapons
technology and Kyushu and Tsushima magnates, suspicious as always of Sillan territorial ambitions, and shaken by Sillan piratical
raids and rumors of a Sillan invasion of Tsushima, the court sought
to limit the trade and retreat within its semi-isolationist walls.
The coolness in Sillan-Japanese relations stood in contrast to the
warm relations between Japan and Po-hai. Bordering China on the
west and Silla south of P'yongyang, Po-hai was a large state that at
its greatest extent occupied the area of present-day northeastern
China (in Chi-lin and Hei-lung-chiang), North Korea, and the
Russian Maritime Province. Po-hai claimed to be the legitimate heir
to the old state of Koguryo, from which a powerful Tungusic leader
in the present area of Chi-lin Province had declared his independence in 698 and in 713 had been enfeoffed king of the Po-hai Commandery by the T'ang court. As the state expanded in the reign of
the second Po-hai king, it came into conflict with its tributary lord
T'ang China and also with the T'u-chiieh, or "Turks," to the north
and Silla to the south. Pressed on all sides by hostile forces, the Pohai government dispatched an embassy to Japan in 727, apparently
intending to ally itself with what it considered a tributary equal in
the East Asian international system. But the Japanese court seems to
have misunderstood Po-hai intentions, mistakenly concluding that
the Koguryo successor was submitting tribute in recognition of the
imperial court's suzerainty. That misunderstanding led to various
diplomatic contretemps, but by the last half of the eighth century
Japan was accepted as the tributary lord in the relationship. In any
case, by that time Po-hai's relations with the T'ang court had improved, the emphasis of the relationship with Japan shifting to trade.
The official relations between the two states that had begun in 727
continued at a brisk pace until the destruction of Po-hai by the Khitan in 926. During that time, more than thirty official envoys arrived
in Heian from Po-hai, reciprocated by some fifteen Japanese missions to Po-hai.24 The Po-hai embassies, sailing directly across the
Sea of Japan from ports south of present-day Vladivostok, succeeded
24 Ueda Takeshi, Bokkaikoku no nazo, Kddansha gendai shinsho, 1104 (Tokyo: Kodansha,
1992), pp. 64-66.
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
91
in making land mostly in the provinces closest to the capital (somewhere between Izumo and Noto), but a number landed far north in
Dewa, and one came ashore as far south as Tsushima. Although the
Japanese government stopped dispatching its own officials to Po-hai
after 811, Po-hai envoys continued to arrive in Japan until 920, when
the last of the state's diplomatic missions landed in Wakasa.
The later embassies were, as the Japanese themselves recognized,
primarily trading missions. Japanese enthusiasm for them seems to
have waned, however, as the men from Po-hai reaped the profits of
the trade while the Japanese court bore the heavy expenses of transporting, feeding, housing, and receiving in suitable style the hundred
or so persons who made up an average embassy. At any rate, beginning in 824 the court tried with limited success to impose a rule restricting Po-hai embassies to one in every twelve years, but by 871 it
had found it expedient to permit Po-hai trading in the capital of
Heian itself. To judge by scanty evidence, the embassies brought for
trade chiefly furs (tiger, leopard, bear), honey, ginseng, and other domestic goods and products; but they also may have regularly supplied items from China, like the copy of the Chinese Hsuan-ming
calendar that was brought by an embassy in 859 and remained the
official calendar of Japan, with growing inaccuracy, from 862 to 1684.
(It was replaced in China at the end of the ninth century.) The Pohai embassies took home with them a variety of luxury products and
goods acquired in Japan: silk fabrics, silk wadding, silk thread, gold,
mercury, lacquer, camellia oil, crystal prayer beads, and other goods.
Po-hai played what was clearly a vital intermediary role in Japan's
relations with China, including the importation of Chinese culture,
although most details of that role are missing. Po-hai itself, both as
a successor state to Koguryo and in its own right, was under heavy
sinitic influence. The Po-hai governmental structure and its chief
capital, the walled Upper Capital at Lung-ch'uan-fu (in present Heilung-chiang Province), were both modeled onT'ang prototypes, and
its officials were versed in Chinese and Chinese poetry. Goods from
even more exotic sources arrived in Japan through Po-hai. A record
is preserved, for instance, of a sake cup made of tortoiseshell that
had originated in the vaguely defined "South Seas." (The shell was
that of the hawksbill tortoise, a widely distributed denizen of tropical and subtropical seas.) Musk is also known to have reached Japan
by the same route.25 Japanese dancing girls, goshawks, falcons, and
25 Tajima Isao, "Bokkai to no kosho," in Hashimoto Yoshihiko, Komonjo no kataru Nihon shi,
vol 2: Heian, p. 255.
Cambridge Histories Online © Cambridge University Press, 2008
92
THE HEIAN COURT, 794-IO7O
at least one fine chalcedony chest figure in the tributary gifts proferred by Po-hai to the T'ang court, and these or similar goods may
have also been items in the Po-hai-Japanese trade.26 Communication
with Japanese living or studying in China and the transmission of
goods to them was sometimes accomplished through the use of Pohai intermediaries, and Japanese in China were able to use Po-hai
visitors there to communicate with, or send goods to, Japan. On at
least one occasion a Japanese monk traveled to China aboard a Pohai ship. Japanese are also mentioned as having been resident in Pohai for study purposes.
During the Heian period, embassies from foreign countries were
mostly similar in their personnel and received much the same kind
of treatment in Japan. Their chief formal purpose was usually purely
diplomatic and ceremonial: to convey expressions of goodwill and to
observe, or avoid, as circumstances dictated, the linguistic niceties of
an established suzerain-subject relationship. That purpose was fulfilled in the conveyance of a message or messages between the foreign and Japanese courts, always couched in the ornate language of
Chinese diplomatic intercourse. We may assume that more substantive communication sometimes took place at the banquets and receptions regularly held for embassies in Heian. Despite linguistic
barriers, discussions between members of the embassies and more
senior Japanese officials must have taken place on such occasions,
aided by interpreters (both foreign and native are known to have existed) and especially by written Chinese being a language common
to all embassies and familiar to Japanese courtiers. All that remains
to whatever informal discussions that did take place, however, is
a few poems exchanged between embassy members and lowerranking Japanese court officials, poetry that succeeds in avoiding
mention of, or allusion to, any diplomatic or governmental matter.27
The embassies might consist of one hundred or more members,
including, in addition to the envoy who headed it and his assistant,
miscellaneous officials and clerks, interpreters, traders, a goodly
number of seamen, and, in the absence of the mariner's compass,
astronomers to navigate the embassy's ships to Japanese shores. The
many traders could account for more than half the embassy mem26 Edward H . Schafer, The Golden Peaches of Samarkand: A Study of T'ang Exotics (Berkeley
and Los Angeles: University of California Press, 1963), pp. 56,94, 228.
27 For examples of the poems, see Bunka shurei shu in Kojima Noriyuki, ed., Kaifusb. Bunka
shurei shu. Honcho monzui, vol. 69 of Nihon koten bungaku taikei (Tokyo: Iwanami shoten,
1964), pp. 228-29, z 2 5-
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
93
bers, at least in the missions from Po-hai. They landed at various
ports in Japan, the Chinese mostly in northern Kyushu or at the
western tip of Honshu - but sometimes on the Sea of Japan coast of
Honshu at about the latitude of Heian or a little north of there - the
Po-hai ships almost entirely along the Sea of Japan coast of Honshu. When their arrival was reported to the court at Heian, minor
officials were appointed to look after and supervise them during
their stay in Japan. In dealing with the Po-hai embassies, one of
those officials was dispatched to the port of arrival to obtain copies
of the embassy's official messages in order to send the copies to the
court in Heian so that their language might be checked for acceptability. (An earlier verbal miscue of 772 was the origin of that cautious practice.)
After permission to enter the capital was granted by the court, the
embassy began its journey to Heian. A few miles outside the city
they were met by an official deputed for their care and supervision,
who performed for them a ceremony of welcome and expressed the
court's concern for their welfare. Under his guidance, the embassy
then entered Heian and was lodged in the two Korokan buildings on
Suzaku Avenue. From the time of the embassy's appointment until
its arrival at the Korokan, as much as half a year or more might have
elapsed.
During the days immediately after the embassy's arrival at the
Korokan, the court sent frequent messengers to inquire after its
members and to transmit provisions of food and clothing, which had
also been supplied at the time of the welcoming ceremonies on the
outskirts of the city. But soon the embassy was escorted to the imperial palace, where an audience with the emperor was held (usually
in the Chodoin), and the envoy presented the chest containing the
message from his own sovereign. In the case of the Po-hai embassies,
in the ninth century it became customary for the king's message to
be accompanied, or to be replaced, by a message from a responsible
office in the Po-hai government, which might be turned over to
Japanese officials at the Korokan before the palace audience. Gifts
from the foreign ruler were also presented on the occasion of the audience, and the envoy might later make his own private gifts of "local
products" to the court. Two formal banquets for an embassy were
provided: one by the emperor himself, at which it was customary to
award court rank to the envoy and other embassy members; the
other by the Council of State. There often may have been private
banqueting as well. Subsequently, the envoy was entrusted with gifts
Cambridge Histories Online © Cambridge University Press, 2008
94
THE HEIAN COURT, 794-IO7O
for the foreign sovereign and a message of reply, the conveyance of
the latter to the envoy marking the final ceremony of the visit.28
The fall of Po-hai in 926 was followed just nine years later by the
final collapse of Silla, and in 936 by the reunification of the Korean
peninsula under the new state of Koryo (918-1392). The Koryo king
immediately sought to establish official relations with Japan, but his
overtures were twice rejected by the wary Japanese court, and intercourse was left as before in the hands of Korean refugees, pirates, and
merchants, who were joined occasionally by Japanese traders in defiance of the court's ban on unauthorized overseas travel by its subjects.
After decades of disunity, China was finally reunified under the
Sung dynasty during the years between 960 and 979 and entered a
period of rapid agricultural and handicraft-industrial development
that stimulated vigorous trade with all the nations of East Asia. This
trade was actively fostered by the Sung court, where the imperial
coffers depended heavily on customs duties collected from overseas
traders and on the monopoly the court reserved for itself in the sale
of aromatics and other luxury items. Based chiefly around the port
of Ming-chou, the Sung merchants early made their way to Japan,
crossing the East China Sea to Hakata in Kyushu. There the Kyushu
authorities at the Dazaifu determined the status of the merchant, the
object of his visit, and what cargo he carried, reporting the information to the court in Kyoto, which determined the allowed length of
the merchant's stay in Japan and whether or not he would be permitted to trade. If trade was permitted, the Kyoto government exercised its right of first purchase either directly through a specially dispatched official, the Foreign Goods Commissioner (karamono no
tsukai), or indirectly and increasingly through the Dazaifu office. It
was the growing authority of the Dazaifu in the trade that encouraged the Sung merchants to seek out private ports in Kyushu.
After the cessation of its official relations with the continent, which
can be dated to the year 920, when a Po-hai embassy is last known to
have reached Kyoto, the Japanese court's chief foreign problem apart
from trade was piratical brigandage. Large-scale attacks by Koryo
and Amami Island pirates on Tsushima, Iki, Kyushu, and other
nearby islands between 997 and 999 resulted in heavy losses of life
and property.
28 For official messages presented by the Po-hei mission of 841-42, see Shoku Nihon koki in
Nihon koki. Shoku Nihon koki. Montoku jitsuroku (Shintei zohoj Kokushi taikei, ed. Kuroita
Katsumi, vol. 3 (Tokyo: Yoshikawa kobunkan, 1934), kan 11, pp. 129-30, Jdwa 9 (842)73/6.
See also Tajima, "Bokkai to no kosho," pp. 243-58.
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
95
Twenty years later, the ferocious attacks of a people previously
unknown to the Japanese spread even greater havoc. The Japanese
called the mysterious new marauders "Toi," a Korean term, it is
said, meaning "barbarian" borrowed from Koryo prisoners who had
been impressed into Toi service. It was only subsequently that the
Japanese authorities learned from the Koryo government that their
attackers were actually a Tungusic Jurchen people from the maritime region northeast of the Korean peninsula. The attacks came in
the spring of 1019, when fifty large ships loaded with several thousand Toi pirates ravaged Tsushima, Iki, and the northern coastal
areas of Kyushu for seventeen days, killing more then 350 people,
including the governor of Iki, taking nearly 1,300 prisoners, and
looting and burning countless buildings. Dazaifu forces at the Bay
of Hakata and local warriors in Hizen put up a stiff resistance and
finally succeeded in expelling the invaders. Koryo, which had also
earlier suffered from Toi depredations, deployed armed ships at several places along the Korean coast and inflicted heavy damage on
the piratical fleet as it sailed homeward. The Koryo forces captured
eight of the Toi ships and sent back home 270 or so Japanese prisoners on board (mostly women), a friendly gesture that the Japanese authorities at the Dazaifu acknowledged with a gift of gold.
Trade between Koryo and Japan grew during the tenth and
eleventh centuries despite the refusal of the Japanese court to enter
into formal relations with the Korean government, the trade forming part of a significant, if unquantifiable, volume of trilateral commerce among China, Koryo, and Japan. It was presumably at least
in part the importance of the trade and the more favorable Japanese
attitudes toward the Koryo government following the Toi attacks
that finally forced the court at Kyoto to emerge somewhat from its
isolationist shell in the last half of the eleventh century. At that time
the Sung court in China, its treasury strained by the southward pressure of the Khitan state of Liao (916-1125), repeatedly sent envoys
to Japan seeking the opening of state relations and trade (the latter,
as usual, under the fiction of tribute rendered to the Sung emperors,
with "gifts" sent in exchange). Although the Japanese were still unwilling to enter into a formal relationship, they now at least responded to the Chinese imperial messages and sent gifts in return.
Insofar as the content of the China-Korea-Japan trade is known
(and that is not very far at all), the Japanese exported such natural
products as gold and gold dust, mercury, pearls, sulfur, pine, cryptomeria, and hinoki cypress, and also various handicraft items, inCambridge Histories Online © Cambridge University Press, 2008
96
THE HEIAN COURT, 794-IO7O
eluding different types of fancy lacquerware, hinoki cypress folding
fans, folding screens, and swords. They imported from China brocades, damasks and other rich silks, ceramics, writing implements,
books, paintings, and copper coins; from Koryo came chiefly ginseng and saffron; from Southeast Asia, dyes, medicines, aloeswood,
and other aromatics. The items of trade, in other words, seem to
have been chiefly low in bulk and high in cost, as would be expected.
The general nature of Japanese foreign commerce remained much
the same in the twelfth century, except that domestic and external
problems in Koryo lessened the level of trade with that state, creating an almost entirely bilateral trading relationship between Japan
and Sung China, which by that time had lost its northern territories
to the Chin and was centered on the valley of the Yangtze River. The
importance of the trade to Japanese leaders at Kyoto grew markedly
when the imperial court came increasingly under the domination of
Taira no Kiyomori (1118-81) and his family in the last half of the
century. Much of the Taira military strength was in the Inland Sea
and Kyushu areas, where local warrior leaders were often heavily involved either directly or indirectly in overseas trade, and it clearly
served Taira interests to protect and develop that trade. Kiyomori
himself was notably active in that regard, undertaking a large-scale
redevelopment of Owada-no-tomari, the port for his estate at Fukuhara on the Inland Sea coast near modern Kobe, where he succeeded in developing a brisk commerce with Sung merchants and,
according to literary sources, reaping rich rewards for his efforts. In
1171, Kiyomori and the retired emperor Go-Shirakawa received one
of the Sung merchants in an audience at Fukuhara, much to the dismay of some conservative courtiers at Kyoto, and in the following
year he and the retired emperor were also recipients of messages and
gifts from the Sung emperor. (For a more detailed discussion of foreign relations at the end of the Heian period, see Chapter 9.)
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 2
THE CAPITAL AND ITS SOCIETY
SITE OF THE NEW CAPITAL
When Emperor Kammu dispatched Fujiwara no Oguromaro (73394) and Ki no Kosami (733-97) to determine the auspices of a site
for a new capital in the spring of 793, he took the first official step in
creating one of the longer continuous urban traditions in world history, stretching nearly twelve hundred years down to the present
day.1 The site was at Uta, the mausolea area for Kammu's imperial
lineage (that of Tenji) in the upper end of what is now called the
Kyoto basin, 115 square miles of land and water. The area had attracted human habitation ever since Jomon man had settled down
on the edges of its marshes and swamps to harvest aquatic life there
while continuing to hunt and gather in the thickly forested hills and
steep valleys surrounding the basin on the east, north, and west. As
the watery areas retreated and dried up, the basin became ideal ricegrowing country, relatively flat, blessed with rich alluvial soil, and
well watered by streams flowing out of the mountains to the north,
which caught moist winds from the Sea of Japan only 35 miles away.
Rice agriculture appeared in the basin inYayoi times, followed by
Tomb culture with its more complex social and political institutions,
its greater wealth, and its expanding intellectual horizons. The area
appears to have been incorporated into the Yamato state in the
fourth or fifth century, and with the establishment of the statutory
regime in the seventh century it became the heart of what was called
Yamashiro Province. The rice lands there were eventually brought
under the public land system with its checkerboard pattern of fields,
which can still be traced in the agricultural areas of the basin today.
By the time the basin emerged on the historical scene in the sixth
1 The chief source for the physical description and history of Heian presented here is Kyotoshi, comp., Kyoto no rekishi, vol. 1 (Tokyo: Gakugei shorin, 1970). A much older but still
mostly reliable English-language study is R. A. B. Ponsonby-Fane, Kyoto: The Old Capital of
Japan, yg4~i86g (Kyoto: The Ponsonby Memorial Society, 1956; first published in article
form 1925-28).
97
Cambridge Histories Online © Cambridge University Press, 2008
98
THE CAPITAL AND ITS SOCIETY
and seventh centuries, its most prominent inhabitants were the Hata,
members of a rich and powerful clan that claimed Chinese descent
but seems to have come most immediately from Korea. No doubt
making use of the technology of the continent, the Hata appear to
have brought under control the streams and rivers in the western and
southeastern parts of the basin and developed the area into prime
agricultural land. They were especially connected with silk making
and weaving, but they were also sake brewers and probably accomplished hydraulic and construction engineers. An early Heian source
associates them with the building of a large weir (pi) on the Katsura
River, whence perhaps the present name of the river west of central
Kyoto (Oi). Hata men built the palace wall and the wall around the
Council of State compound at Nagaoka.
The neighbors of the Hata in the southern part of the basin were
members of another powerful clan, the Haji, who had been known
as clayworkers in earlier times and had long enjoyed close ties with
the imperial line. The presence of the Hata and Haji in the basin may
have been one of its chief attractions for Kammu and his advisers.
Kammu himself was the maternal grandson of a Haji woman and
seems to have lived with her in his youth; and two of his closest
associates, Fujiwara no Tanetsugu and Oguromaro, were intimately
connected with the Hata, Tanetsugu's maternal grandmother and
Oguromaro's wife having been women of that clan. Tanetsugu had
been a leading advocate of the move to Nagaoka, which was on the
southern edge of the basin in Haji territory. Oguromaro, who had
been appointed with Tanetsugu in 784 to determine the auspices of
the Nagaoka site, seems to have played a similar, if somewhat less
central, role in the move to Heian, where the imperial residential
palace, a tenth-century source says, was eventually located on the
site of a Hata leader's house.
When Oguromaro and Kosami reported the results of their survey
to Kammu at Nagaoka, they informed him that the site at Uta was
a natural fortress formed by surrounding mountains and streams,
and that it matched the geomantically auspicious features of "corresponding-to-the-four-gods" topography: a great river on the east, a
great highway on the west, a mountain in the north, and marshy lowlands to the south. Apart from its geomantic virtues, Uta was indeed
in many ways well situated for a capital city. The steep, thickly timbered hills and mountains on the east, west, and north formed a skyline generally between 1,500 and 2,500 feet above the basin floor and
in combination with the lake and marsh region known as OguraCambridge Histories Online © Cambridge University Press, 2008
SITE OF THE NEW CAPITAL
99
noike to the south (now reclaimed and dry) and the river systems
that converged on that area (chiefly the Kamo from the northeast,
the Katsura from the northwest, and the Uji from the east) provided
defensible positions against hostile attack. At the same time, there is
reason to think that there were already well-established roads leading out of the basin in all directions, making communication with
the rest of the country reasonably convenient, and theYodo River in
the south gave easy water access to the Inland Sea (Seto Naikai).
The site was, moreover, mostly level, easing problems of layout and
construction, and it was well irrigated by numerous streams and easily tapped underground water, which served not only mundane
needs but also made possible the creation of elaborate gardens.
The site's chief disadvantages were its climate and inadequate
drainage. Modern residents of Kyoto, it is said, endure the miseries
of summer and winter in return for the glories of the spring and
fall, and one can easily imagine that both miseries and glories were
greatly intensified by the more natural environment of the eighth
century. Although the average temperature in Kyoto now through
the year is an equitable fifty-seven degrees Fahrenheit, hot, breathless summers, with high temperatures in the upper eighties and
often climbing into the high nineties, and relatively cold winters that
bring the thermometer down into the low twenties or below, combine with high humidity (averaging over 70 percent throughout the
year) to make the Kyoto basin famous for its muggy, unbearable
heat and its piercing cold. It was probably more the pains of the climate than its pleasures that moved the eleventh-century poet Izumi
Shikibu to write:
If only the world
Into spring and fall
We could forever make
And summer and2winter
Were never more.
Heavy precipitation (over 60 inches a year in modern times) and
heavy runoff from the surrounding hills place severe demands on the
drainage system in the Kyoto basin. The general slope of the land in
a southwesterly direction and underlying strata of gravelly sand gave
most of the eastern half of die Heian site fairly good drainage. But
the western half, which was low-lying and generally underlain by
2 Izumi Shikibu shu, in Zoku kokka taikan, ed. Matsushita Daizaburo (Tokyo: Kadokawa
shoten, 1963), no. 40575.
Cambridge Histories Online © Cambridge University Press, 2008
1O0
THE CAPITAL AND ITS SOCIETY
clay or clayey sand, must have often been so wet and marshy as to
be scarcely fit for human habitation under the engineering and architectural conditions of the day. A story from a twelfth-century
source tells about a rich and clever man who succeeded in building
a house on a fill of rushes and earth at a watery site in the western
part of the capital. Few ordinary people, though, could have managed so ambitious an engineering feat, and one wonders even so how
long the foundations of the clever man's house lasted.
Following the advice of Wake no Kiyomaro (733-99), Kammu
himself two or three months before the dispatch of Oguromaro and
Kosami had twice used the pretext of a hunting expedition to visit
and inspect the Uta area. His decision, just a few days after the submission of their report, to transfer the capital to the new site suggests
he was simply waiting for geomantic confirmation of a choice already made. Oguromaro, who was immediately appointed to supervise the construction of the new imperial palace, was faced by a multitude of difficult and complex tasks. One of the first and greatest
must have been the diversion and control of the numerous streams
that flowed through the site, a project in which his Hata in-laws, with
their wealth, experience, and engineering skills, may have proved indispensable.
The Kyoto basin at one time had been an inlet of the Inland Sea
and it was still in the late eighth century a very watery place: streams
and rivers ran across it at the site of the new capital in a generally
southwesterly direction, and ponds and marshes dotted the landscape throughout. One of the more prominent bodies of water
would have been the famous pond that was later incorporated in the
great imperial preserve south of the palace called the Park of the Divine Spring (Shinsen'en), a remnant of which precariously survives
in modern Kyoto. A small river called the Kamo ran along the eastern edge of the site, its course shifting within a broad bed that directly adjoined what was to become the eastern limit of the city.3
Perhaps to simplify bridge building and to reclaim usable land, the
riverbed was narrowed and straightened with the aid of dikes. That
task and the work of eliminating or realigning other smaller streams
at the site presumably began early and may have continued for many
3 Ishida Shiro, "Kyoto bonchi hokubu no senjochi: Heian-kyo sentoji no Kyoto no chisei,"
Kodai bunka 34, 12 (December 1982): 1-14, has demonstrated that the Kamo andTakano
rivers have occupied what are substantially their present courses ever since Jomon times,
thus laying to rest a previously widely held belief that the rivers originally flowed directly
across the site of the new capital.
Cambridge Histories Online © Cambridge University Press, 2008
SITE OF THE NEW CAPITAL
101
decades. At any rate, parts of the Kamo were still being diked and
moved eastward as late as the eleventh century, resulting in an elevated streambed and frequent flooding of the city when the dikes
were breached or flood water clogged the bridges spanning the river.
Although his palace was still in the early stages of construction,
Emperor Kammu, anxious perhaps to be gone from the gloomy and
threatening atmosphere of his already half-dismantled palace at Nagaoka, moved to the new imperial seat in the late autumn of 794.
Two or three weeks later he issued an edict conferring on the capital city its official name and renaming (or, rather, selecting different
Chinese characters for the name of) the province in which the site
was located:
Enclosed collar-and-sash by mountains and streams, the province here
makes a natural citadel. Because of that configuration, we devise a new designation for it: let this Postmontane [Yamashiro] Province be renamed the
Province of the Mountain Citadel [Yamashiro]. Moreover, the joyfully
flocking people and the singers of praise raise their different voices in iden-4
tical words, naming this the Capital of Peace and Tranquillity [Heian-ky6] .
Despite the supposedly popular origin of the new capital's name,
it seems never to have enjoyed much vogue except among latter-day
historians, most people in the following centuries preferring to call it
simply "the Capital" {miyako; more literally, "imperial seat"). Kyoto,
a sinitic synonym of miyako, had been applied as a common noun to
earlier Japanese capitals and was similarly used of Heian during the
first centuries of its existence. But the term also began to function
sporadically as a proper noun in the late tenth and eleventh centuries
and had become a fairly common name for the city in everyday types
of writing by the thirteenth century. There were several other Chinese-derived names used of the city, mostly in fancy or learned writing. Among the most common were "Rakuyo" and "Raku," the latter
an abbreviation of the first.
"Rakuyo" is the Sino-Japanese reading of the characters for Loyang, the name of the Eastern Capital of theT'ang dynasty as paired
with the Western Capital at Ch'ang-an. Since Heian was modeled on
neither Ch'ang-an nor Lo-yang, as often erroneously supposed, but
on the earlier Japanese capital at Fujiwara — which seems to have
taken its inspiration from the capital of the southern Chinese dynasties at Chien-k'ang (Nanking) - the application of "Rakuyo" to
Heian is unexpected. But Japan had also known a period of dual
4 Quoted in translation from Nihon kiryaku in Kyoto no rekishi, vol. 1 (1970), p. 238.
Cambridge Histories Online © Cambridge University Press, 2008
1O2
THE CAPITAL AND ITS SOCIETY
capitals at Heizei and Naniwa, and although that period had passed
with the establishment of the capital at Nagaoka, the tradition was
revived in terminology about the time of the sinophile emperor
Saga, when the literati started referring to the western half of the city
(Ukyo) as Chdan (i.e., Ch'ang-an) and the eastern half as Rakuyo
(Lo-yang).5 As the western half withered and failed and the eastern
half became the heart of the capital, Rakuyo began to function as a
name for the whole. (The abbreviation "Raku" much later became
especially familiar in Western art history circles because of the brilliant sixteenth-century paintings of scenes in and around Kyoto
called "Rakuchu rakugai zu.")
PLAN OF THE CITY
Knowledge of early Heian and its palace rests chiefly on a collection
of government regulations and procedures dating from the early
tenth century {Engi shiki), on twelfth- or thirteenth-century plans
and maps thought to reflect the city mainly as it existed in the ninth
and tenth centuries, and on some other literary and documentary
evidence. Since every surface vestige of Heian disappeared long ago,
and since the site is now overlaid by the densely populated city of
Kyoto, severely limiting archaeological investigation, physically verifiable knowledge of the city is far more restricted than in the case of
the earlier capital, Heizei-kyo (Nara).This capital for the most part
quickly reverted to agricultural land after the removal of the imperial seat to Nagaoka and remained therefore not only relatively undisturbed for the modern archaeologist but also more easily accessible.
Consequently, although it is possible to describe Heian toward the
date of its founding with considerable confidence in the general reliability of the detail, one must bear in mind in reading the following
account that the city plan as a whole and some particular features,
such as the walls that are said to have lined the streets and avenues,
may have been only partially realized in practice.
The site of the new capital, after an extension of its northern boundary in the last half of the ninth century, was a rectangle measuring
approximately 3.3 miles north and south and 2.8 miles across, or
9.24 square miles, which gave it roughly the same area as that of the
5 KishiToshio, 7070 no seitai, vol. 9 of Nihon no kodai (Tokyo: Chuo koronsha, 1987), pp. 5458, and 23-37.
Cambridge Histories Online © Cambridge University Press, 2008
PLAN OF THE CITY
103
old capital at Nara or of small modern university towns like Berkeley (10.4 sq. miles) and Cambridge, Massachusetts (6.2 sq. miles),
but just one-third the size of Ch'ang-an, at the time perhaps the
largest metropolis in the world. A rectangular space of about 1,300
by 1,500 yards (six-tenths of a sq. mile) at the north-central edge of
the site was reserved for the Greater Imperial Palace (Daidairi), the
location and general contours of the area being identical to those of
the imperial palaces at Ch'ang-an and at previous statutory capitals
in Japan. The modern city of Kyoto includes the entire area of what
was Heian, but it is, of course, much larger. The geographical center
of Heian was located at about what is now the intersection of Shijo
and Sembondori Streets in Kyoto, its northeast corner corresponding almost exactly to the southern parts of the site presently occupied by the Imperial Palace (Gosho) and the Kyoto Imperial Gardens, its southern border lying on a line extending along the
southern edge of Toji Temple, and its northwest corner falling a
short distance northwest of the Myoshinji. The southeast corner of
the Heian Greater Imperial Palace was at almost the exact center of
the present Nijo Castle.
Outside the Greater Imperial Palace, the layout of the city was
modeled fairly closely on the grid system used in the allocation of
agricultural land. Thirty-three north-south and thirty-nine east-west
streets (koji) and avenues (pji) traversed the site at regular intervals,
intersecting at right angles and dividing the site checkerboardfashion into blocks (machi) of equal size. Low earthen walls about 6
feet wide at their bases lined both sides of each street or avenue. Except in two instances where streams shared the roadway (Horikawa
and Nishihorikawa), the streets were of uniform width, measuring
nearly 35 feet across, or about the width of a modern street accommodating two lanes of automobile traffic and parallel parking on
both sides. Most of the twenty-four avenues (thirteen running east
and west, eleven north and south) were a little more than twice the
width of the streets, but several in more prominent positions (at the
city limits and leading to or past palace gates) were between 90 and
110 feet wide. The avenue paralleling the south face of the palace
(Nijo) measured a little over 160 feet across; and the great axial avenue (Suzaku) that ran north to south at the exact center of the city
and led to the main palace gate was a mall-like thoroughfare 270 feet
wide. One street lay between every pair of parallel avenues in the
sections of the city due west, east, and south of the imperial palace;
elsewhere the street interval between avenues was three.
Cambridge Histories Online © Cambridge University Press, 2008
THE CAPITAL AND ITS SOCIETY
104
1. Chuwain
2. Emperor's Residential Compound
3. Court of Abundant Pleasures (Burakuin)
4. Court of Government (Chodoin)
Hfl
HOK
UKYO
SAKYO
CHIJO
Greater Imperial Pala ce
(Daidairi)
1
3
IO
z
2
4
IJO
SANJO
Su. zakumon
X
GOJO
Vi
OKUJ
IO
a:
IO
SHIC
X
IO
KUJO
HAC
X
mo
Ichijo
Ogimachinokoji
Tsuchimikado
Takatsukasanokoji
Konoe
Kadenokoji
Nakamikado
Kasuganokoji
Oimikado
Reizeinokoji
Nijo
Oshigakoji
Sanjo bomonnokoji
Anegakoji
Sanjo
Rokkakunokoji
Shijo bomonnokoji
Nishikinokoji
Shijo
Ayanokoji
Gojo bomonnokoji
Takatsujinokoji
Gojo
Hinokuchinokoji
Rokujo bomonnokoji
Yamamomonokoji
Rokujo
Sameushinokoji
Shichijo bomonnokoji
Kitanokoji
Shichijo
Shionokoji
Hachijo bomonnokoji
Umenokoji
Hachijo
Harinokoji
Kujo bomonnokoji
Shinanokoji
Kujo
Figure 2.1. The Heian capital (Heian-ky5). Names in capital letters
are districts; names in italics are gates. Adapted from McCullough
and McCullough, A Tale of Flowering Fortunes, p. 834.
It must be added, however, that Heian streets and avenues were
not entirely devoted to traffic. All were lined on both sides by ditches
and narrow paths or strips of vacant land (called "dog runs," inubashiri) that occupied from 15 to 35 percent of the total width. In the
Cambridge Histories Online © Cambridge University Press, 2008
PLAN OF THE CITY
105
case of a street, for example, the area next to the walls on either side
was occupied by a "dog run" 3 feet wide and beyond that by a ditch
of the same width, leaving a usable roadway of about 23 feet between
the ditches. The median roadway on Suzaku Avenue was similarly
constricted to a width of about 230 feet. The avenues on the four
sides of the Greater Imperial Palace were even further restricted
where they passed the palace, which was separated from them by a
vacant space 26 feet wide and an 8-foot fosse.
The capital outside the Greater Imperial Palace was divided physically and administratively into western and eastern halves, the line
of demarcation being Suzaku Avenue and, north of its terminus at
Nijo Avenue, the Greater Imperial Palace. (Suzaku Avenue corresponded mostly with the modern Sembondori Street in Kyoto, except that it ended at the imperial palace 200 to 300 yards due north
of what is now Nijo Station.) The avenue was on a line that ran
due south from Funaoka, a low (368-foot) hill less than a mile outside the northern city limit that seems to have served as a chief reference point in the orientation and planning of the city. Since the
geographical perspective of the city was, in keeping with Chinese
practice, southward-facing from the palace, its eastern half was frequently called the Left Capital (Sakyo) and the western half the
Right Capital (Ukyo).The two halves, Left and Right, were further
divided by avenues into nine parallel east-west belts of equal area
called "zones" (jo) and a single narrower belt at the city's northern
edge called the North Edge (Kitanobe). Numbered in order from
north to south, the nine zones straddled the city from border to border except where interrupted by the Greater Imperial Palace, each
being about 560 yards wide on its north-south axis (the North Edge,
however, was just half that).
Each zone south of the palace was divided by north-south avenues
into eight equal "quarters" (bo), each about 17 acres in area. The
four in each half of the city were numbered separately in order beginning with the quarter next to Suzaku Avenue. Zones One and
Two, which were interrupted by the palace, contained only six quarters each, three on either side of the palace. Each quarter was usually divided by intersecting north-south and east-west streets or avenues into sixteen blocks (machi) numbered boustrophedonically
(as in the route of a plow ox) from north to south away from Suzaku
Avenue, and each block was further divided into thirty-two rectangular house lots (henushi). There were eight house lots north to
south in a block and four east to west. Each lot measured in theory
Cambridge Histories Online © Cambridge University Press, 2008
1O6
THE CAPITAL AND ITS SOCIETY
about 49 by 98 feet, but some were apparently reduced in size to accommodate the one to three 10- or 15-feet-wide alleyways that penetrated the blocks north to south and gave access to interior lots.
Lots were identified within each block by a grid system of four numbered columns (ko) and eight numbered "gates" {mori), the columns
running east to west in the Right Capital and west to east in the Left
Capital, and the gates, which corresponded to individual house lots,
running north to south. It was thus possible to identify precisely any
piece of property in the capital after the fashion of the following description, which comes from a deed of sale dated in 912: "A single
area of four house lots in all (located in Gates 4, 5, 6, and 7 North,
Column 1 West, Block 15, Quarter 1, [Zone Seven, Left Capital])."
Thanks to the precision and clarity of the system, we can determine that the four lots here described were in a row facing Kushige
Street just west of the present site of Nishihonganji and that they
measured altogether approximately 98 by 196 feet.
The broad, straight streets and avenues of the new capital must
have lent an agreeable air of openness and spaciousness to its vistas,
affording unobstructed views of the surrounding mountains and creating small and large squares at road intersections throughout the
city. The largest such square was directly in front of the Vermilion
Sparrow Gate (Suzakumon) at the palace, where the intersection of
Suzaku and Nijo Avenues created an open space 160 feet north to
south and 270 feet across (just about an acre in area). Another, measuring 270 by 110 feet, was at the intersection of Suzaku and Kujo
on the southern edge of the city just inside Rampart Gate (Rajomon,
also Rashomon),6 formally the main entranceway to Heian.
The openness of the city was enhanced by the low profile of its
prevailing architecture, which was with but few exceptions uniformly single-storied, and especially by the absence of encompassing
city walls. In the tradition of Heizei and earlier statutory capitals in
Japan, but unlike the heavily fortified major cities of medieval Europe and China, Heian itself was almost certainly unwalled, except
for a small garden-like structure about 6 feet high on the city's
southern border that served as a setting for Rampart Gate. That extremely modest "rampart," only about a third as high as the great
walls that surrounded Ch'ang-an, was paralleled by two ditches or
6 The original pronunciation of the gate's name seems to have been "Raseimon." A common
pronunciation since thefifteenthor sixteenth century has been "Rashomon," as in the cinema title.
Cambridge Histories Online © Cambridge University Press, 2008
PLAN OF THE CITY
107
moats a little less than 10 feet wide, one inside the wall and the other
outside. The remainder of the city's boundaries is thought to have
been delineated by nothing more formidable than extensions of
those moats and perhaps some kind of simple earthwork. Permanent, fixed bridges apparently spanned the moats at various points,
leaving the city open to the countryside.
The builders of Heian may have also deviated from their Chinese
model in the disposition of walls within the city. The walls on either
side of streets and avenues formed enclosures for each block
(machi), but inconclusive evidence suggests that, unlike at Ch'angan, the quarter (bo) may have been walled and gated only where it
abutted Suzaku Avenue south of the imperial palace and at the city's
eastern and western limits (Higashikyogoku and Nishikyogoku).
Suzaku Avenue, which exactly bisected the Heian site north to
south and led directly from Rampart Gate to the main entrance of
the palace at Vermilion Sparrow Gate, was the chief ceremonial
thoroughfare in the city. If quarter walls were erected only along the
sides of that avenue, their chief function may have been to enhance
the dignity of the main approach to the palace. A foreign envoy entering Rampart Gate was undoubtedly meant to be impressed by
the resemblance of the city to its Chinese prototypes, especially
Ch'ang-an, the acknowledged queen of East Asian capitals. The
great, two-storied Rampart Gate, n o feet wide, 25 feet deep, and
perhaps 70 feet high; the Chinese-style bridge spanning the moat
outside the gate; the vastness of willow-lined Suzaku Avenue,
flanked near at hand by two imposing temples (Saiji and T5ji) and
in the middle distance by the paired lodgings for foreign embassies
(the Korokan), and bounded on either side by continuous rows of
quarter walls pierced at regular intervals by gates; and, finally, far in
the distance at the northern end of the avenue, the soaring, twostoried Vermilion Sparrow Gate - all would have been reassuringly
familiar to an official visitor from the continent, confirming Japan's
vaunted reputation as a country of sinitically learned "superior
men" (chun-tzu) who could be counted on to understand Confucian
rites and principles.7 He might not have ever fully realized that the
city's southern "rampart," the quarter walls, and the tiled roofs of
the buildings along the southern edge of the palace precincts were
mostly ambitious facade.
7 Charlotte von Verschuer, Les Relations officiettes du Japon avec la Chine aux viit et ix1 sticks
(Geneva: Librairie Droz, 1985).
Cambridge Histories Online © Cambridge University Press, 2008
1O8
THE CAPITAL AND ITS SOCIETY
GREATER IMPERIAL PALACE
Not much is known about the actual process of construction at the
new capital, the sources for the years in question being particularly
meager, but there can be little doubt that the scene of greatest activity was at the Greater Imperial Palace itself, which was both the residence of the emperor and his household and the seat of the government's central administrative organs. Work there was directed at first
by specially appointed commissioners, and then from 796 by a semipermanent Office of Palace Construction staffed by some 150 officials and technicians, who employed and oversaw a larger labor force
levied for one year of recompensed service from the provinces. In the
early stages of construction before public labor levies were organized,
noble families were also obliged to contribute laborers for work on
palace and city projects, and prisoners were used as well. The chief
responsibilities of the Office of Palace Construction were, apart from
the construction of palace buildings and government offices, the expropriation of land for the city's site, the construction of streets and
the layout of house lots in the city, the diking and channelization
of streams and rivers, and the organization of labor for those various
tasks. The office remained in existence until early 805, when economic exhaustion led to its abolition. Even then, it appears, not all of
the projected palace buildings had been completed. The size of its
task and the burden the office placed on the economy are suggested
by an early-tenth-century estimate that during Kammu's reign threefifths of the central government's expenditures were devoted to the
construction of the Heian palace and princely residences.
The spacious site of the palace was surrounded at its periphery by
an outer fosse 8 feet wide, a median strip of vacant land 26 feet
across, and an inner earthen wall a little over 6 feet high. Fourteen
gates provided access to the grounds, the largest and most impressive by far being the tile-roofed Vermilion Sparrow Gate, whose fanciful name was borrowed from a corresponding structure at the
Ch'ang-an palace. The "vermilion sparrow," a mythical creature associated in Chinese thought with the south, was said to appear to
holy men and rulers of exceptional merit and power as a harbinger
of good, and it served therefore as an auspiciously apt symbol for the
main southern entrance of an emperor's palace.8 It was the Yomei
8 Edward H. Schafer, The Vermilion Bird: T'ang Images of the South (Berkeley and Los Angeles: University of California Press, 1987), p. 267. Although the Chinese-derived names by
which the main palace buildings and gates at Heian are commonly known were mostly
Cambridge Histories Online © Cambridge University Press, 2008
GREATER IMPERIAL PALACE
10g
("Brilliance," "Sun") Gate, however, standing on the east side of the
site at the entrance most convenient to the Emperor's Residential
Compound (Dairi, "penetralia"), that the emperor and his courtiers
most commonly used in their comings and goings. The senior nobles
in particular tended to live in the quarter of the city that lay directly
east of the palace, and by using the Yomei Gate they were within a
ten- to twenty-minute walk of the Emperor's Audience Hall, less
than half the distance of the route through Vermilion Sparrow Gate.
As construction progressed, the palace grounds were filled with
perhaps as many as two hundred or more buildings, gateways, towers, and connecting corridors situated within numerous walled enclosures in a setting of courtyard gardens, winding streams, ancient
trees, and occasional broad, open spaces. Space within the grounds
was differentiated functionally both in the general layout and in the
details of particular areas. Reflecting the dual character of the
palace as an imperial residence and administrative seat, the buildings were arranged in two principal groups: (1) government offices
and facilities, which occupied most of the southern two-fifths of the
site; and (2) the imperial residence and associated household offices, which were located in an area of equal size north of the first
group. A third group of buildings, consisting mainly of governmental storehouses, was concentrated on the remaining land at the
northern extremity of the site.
The first group of buildings provided space for both the workaday
and ceremonial business of government, but it was the ceremonial
space that, characteristically of the age, received the greatest emphasis in the palace plan. The two chief compounds of ceremonial
buildings, the Court of Government (Chodoin) and the Court of
Abundant Pleasures (Burakuin), were the largest of all the walled
enclosures in the palace; their buildings were the grandest and most
ornate of all palace structures; and the more important of the two,
the Court of Government, was centered exactly in the north-south
median line of the palace grounds directly opposite Vermilion Sparrow Gate to the south.
The Court of Government (or Court of the Eight Ministries Hasshoin - as it was also called) was on a generally rectangular site
surrounded by its own wall, the main entrance through which was
the celebrated Obedient-to-Heaven Gate (Otemmon) on the southadopted about twenty-five years after the founding of the city (during the reign of Emperor
Saga), Vermilion Sparrow Gate was probably so called from the beginning.
Cambridge Histories Online © Cambridge University Press, 2008
THE CAPITAL AND ITS SOCIETY
no
Bureau of
Military
Storehouses
(Hyogoryo)
Storage
(Nuriya)
Palace
Women's Family
Office
(linemen
Tsukasa)
I
I
OJJke
(Okimim *
Tsukata) ,1
""
1 Storage
j (Okura)
Treasury I
Ministry J
(Okurasho)T
11
I
Storage
(Okura)
1
j
Storage
(Okura)
Storage
(Okura)
1
r
Sjorage
(Okura)
Storage
(Nagadono)
1
|
(
Bodyguards
of the
Right
(Ukon'efu)
Bureau of
Books and
Drawings
(Zushoryo)
I
|
Tea plot
(Cha'en)
Guards
Office
ir'""-^
(Otonoi)
Storage
•• 1(Ritsubunzo)
KltSUDUnZO) •
I
|
Josaimon
Bureau of
Grounds
(Toriomoryo)
Sjorage [
(Okura) j
(Otonoi)
|
I
Female
Dancers' and
Musicians'
mOffice
mitwiu
(Naikyobo)
Office
Jowmon
(Naikyobo)
Folk Music
Office
(Outadokoro)
Itnputncfi
Popular Affairs Ministry
I
(Minbusho)
Q
rH
Kbgamon
Figure 2.2. Greater Imperial Palace (Daidairi). Names in italics are
gates. Adapted from McCuUough and McCullough, A Tale of Flowering Fortunes, p. 835.
Cambridge Histories Online © Cambridge University Press, 2008
GREATER IMPERIAL PALACE
111
central side of the enclosure. Originally designed as the center of
statutory government, the place where the emperor was to meet in
daily audience with his chief ministers to conduct the business of the
state, where foreign envoys were to be received, and where the great
state ceremonies and rituals were to be conducted, the Court consisted of three subprecincts. The smallest, at its southern end, was
occupied by two buildings used as waiting rooms by senior nobles.
The middle and largest area contained a broad courtyard surrounded on its eastern, western, and southern sides by twelve symmetrically arranged buildings where members of the bureaucracy assembled when business was being conducted at the Court; and
finally a somewhat smaller section at the northern end was the site
of the imperial-throne building, the Great Hall of State (Daigokuden).The latter, a soaring Chinese-style edifice of vermilion pillars,
green roof tiles, and dolphin roof finials, was the most magnificent
building in the entire palace complex, measuring nearly 175 feet east
to west and 65 feet north to south. (The present Heian Shrine in
Kyoto is a five-eighths scale replica of a 1072 reconstruction of the
hall. It was built in 1895 to commemorate the eleven-hundredth anniversary of the founding of the city.) The peculiar name of the hall
(literally, "grand culmen"), borrowed from the T'ang imperial
palace, was derived from one of the appendices to the I-ching (Book
of Changes), where "grand culmen" was used to signify the source
of the universe, or absolute existence. The name implied a view of
the emperor as the source of all things, a veritable pivot of the world.
The Court of Abundant Pleasures also stood within its own
walled enclosure about 90 feet west of the Court of Government, occupying an area slightly larger than the latter's. Built as the principal
imperial banquet facility, the site of the chief festivals (sechie), and a
center of court cultural life, the Court contained ten buildings disposed in a pattern similar to that of the Court of Government. The
main structure was the Celestial Presidence Pavilion (Kenrinkaku),
which corresponded in position and function to the Great Hall of
State. (The Pavilion was later renamed the Hall of Abundant Pleasures, Burakuden; its site is one of the few in the palace that have
been excavated by modern archaeologists.) From there the emperor
presided over banquets and festivals, watched archery contests and
wrestling matches, and participated in other events on the regular
court calendar.
The remainder of the southern area of the Greater Imperial
Palace was given over mostly to buildings housing governmental
Cambridge Histories Online © Cambridge University Press, 2008
112
THE CAPITAL AND ITS SOCIETY
REGENT
COUNCIL OF STATE
(Daijokan)
SPECIFIC YORO CODE AUTHORITY
CENTRAL AFFAIRS MINISTRY (Nakatsukasasho)
Empress's Household Office (Chugushiki)
Palace Storehouse Bureau (Kuraryo)
Bureau of Divination (Onmyoryo)
Bureau of Skilled Artisans (Takumiryo)*
Bureau of Imperial Attendants (Otoneriryo)
MINISTRY OF CEREMONIAL (Shikibusho)
CIVIL AFFAIRS MINISTRY (JibushS)
Bureau of Music (Utaryo)
POPULAR AFFAIRS MINISTRY ( M i n b u s h o )
Bureau of Computation (Kazueryo)
Tax Bureau (Chikararyo)
WAR MINISTRY (Hydbusho)
GENERAL AUTHORITY
BOARD OF CENSORS (Danjodai)
SIX GUARDS HEADQUARTERS
(Rokuefu)
IMPERIAL POLICE (Kebiishi)*
CHAMBERLAINS' OFFICE
(Kurododokoro)*
PROVINCIAL ADMINISTRATIONS
CAPITAL OFFICES (Kyoshiki)
KYUSHU GOVERNMENT OFFICE
(Dazaifu)
ALL OTHER OFFICES
PUNISHMENTS MINISTRY ( G y o b u s h o )
TREASURY MINISTRY ( O k u r a s h o )
IMPERIAL HOUSEHOLD MINISTRY ( K u n a i s h S )
Palace Table Office (Daizenshiki)
Carpentry Bureau (Mokuryo)
Bureau of Grounds (Tonomoryo)
Bureau of Housekeeping (Kamonryo)*
Imperial Table Office (Naizenshi)
* Extra-Code office
Figure 2.3. Major Heian governmental organs, ca. c.E. 1000.
Adapted from McCullough and McCullough, A Tale of Flowering
Fortunes, p. 801.
ministries, bureaus, and offices arranged left and right of the Court
of Government according to the statutory table of organization. A
walled compound lying directly east of the Court of Government
was the administrative heart of the government, containing the three
connected buildings of the Council of State. The positions of ceremonial prominence on either side of Vermilion Sparrow Gate were
occupied by the Ministries of Ceremonial and War, while the bureaucratically superior Department of Shrines was relegated to the
southeastern edge of the site.
Cambridge Histories Online © Cambridge University Press, 2008
EMPEROR S RESIDENTIAL COMPOUND
113
EMPEROR'S RESIDENTIAL COMPOUND
The most radical departure of the Heian imperial palace from its predecessor at Nara was in the physical separation of the Emperor's Residential Compound from the Court of Government. At Nara the residential compound had stood on the palace's north-south median line
directly north of the Court, and its outer encircling wall had included
the Great Hall of State itself, thus linking in an obvious way the person of the emperor and the imperial position. At Heian, however, the
residential compound, following the example of the Nagaoka palace,
was not only completely distinct from the Court of Government but
also removed to a directionally intermediate position in the eastern
half of the grounds off the median north-south line.
It is tempting to see in that physical arrangement of the palace a
symbolic expression of a newly conceived, or more clearly recognized, distinction between the emperor as a man and as an institution, a distinction that became important, no doubt, as regents imposed broadly on imperial authority after the middle of the ninth
century. The removal of the residential compound from a position
astride a cardinal axis of the palace site might even be interpreted as
a diagrammatic subordination of the imperial person to the imperial position. But it may be closer to the truth to view the repositioning of the residential compound at Heian as having less to do with
the symbolization of abstract political notions than with a practical
desire to avoid the frequent moves and rebuildings of the imperial
palace that had characterized Japanese history since early times.
The reasons for those moves are unknown, but most speculation
centers around considerations of ritual purity or the short life-span
of the lightly constructed wooden palace buildings. In the case of ritual purity, the motive would have been to remove the ritual pollution caused by the death of a previous emperor or to maintain purity through periodic rebuilding, after the manner of the similar
custom still practiced at some Shinto shrines. Although the early
"palaces" were presumably quite simple buildings that could have
been rebuilt at manageable cost, the growing size and complexity of
the palace from the sixth century on must have caused a vast increase in the financial burden of reconstruction. It would have been
natural for a politically ambitious emperor like Kammu to seek a
way of containing palace-building costs, just as he was also seeking
retrenchment in expenditures elsewhere in his government.
Cambridge Histories Online © Cambridge University Press, 2008
THE CAPITAL AND ITS SOCIETY
114
Inmeimon .
Bodyguards I
I I
(Konoc no Jin)
K°roden
Figure 2.4. The Emperor's Residential Compound (Dairi). Names
in italics are gates. Adapted from McCullough and McCullough,
A Tale of Flowering Fortunes, p. 840.
Cambridge Histories Online © Cambridge University Press, 2008
EMPEROR S RESIDENTIAL COMPOUND
115
West of the median north-south axis of the Heian palace, in a position corresponding exactly to the residential compound on the east,
was a large open space called the Banquet Pine Grove (En no Matsubara).The arrangement was strongly reminiscent of those Shinto
shrines where a vacant space is reserved as a site for the next periodic
rebuilding of the shrine's sanctuary. At such shrines, when a new sanctuary is built the old sanctuary is disassembled and its site left vacant
until the next rebuilding, the location of the sanctuary thus shifting
back and forth regularly between the two sites. Given that model,
which originated in pre-Heian times, it is possible to conclude that the
designers of the Heian palace may have shifted the position of the residential palace compound in order to create a balanced pair of sites,
one in the known position of the compound on the east and the other
at Banquet Pine Grove on the west, where rebuildings of the compound could be alternated after the fashion of the Shinto shrines. If
that inference is correct, the positioning of the residential compound
and the Banquet Pine Grove can be seen as an attempt to integrate
into the structure of the palace the conflicting native tradition of a
shifting imperial seat and the sinitic concept of a fixed capital, an issue
that was by no means fully resolved at the beginning of the period.
The expanded scale of the statutory government and its capital
made frequent transfers of the capital intolerably costly, and the
planners of Heian may have seen in the Banquet Pine Grove site a
convenient means of bringing them to a permanent end while still
providing for whatever was achieved, or thought to be achieved,
through new construction. If so, the failure in practice to use the alternative site for a new residential compound is puzzling. It might be
attributed to the straitened finances of the regime at the time of
Kammu's death and possibly also to changing attitudes toward native custom during the reigns of his sons, when Chinese cultural influence was exceptionally strong, but the absence of contemporary
comment on the problem is difficult to explain.
The residential compound was enclosed within a double set of
walls and gates, the outer set also encompassing household offices,
storage areas, and the Court of Central Harmony (Chiiwain), a
small, walled area of Shinto ritual buildings that included the geographical center of the Greater Imperial Palace. The residential compound proper, measuring about 710 feet north and south and 560
feet east and west, contained more than thirty named buildings.
Those in the northern two-fifths of the site housed the various imperial consorts and female officials; the remainder in the southern
Cambridge Histories Online © Cambridge University Press, 2008
Il6
THE CAPITAL AND ITS SOCIETY
portion were devoted to the imperial residence itself, a ceremonial
building, and various household offices and storage structures.
The core of the residential compound consisted of two buildings:
the Purple Sanctum Hall (Shishinden), said to have been on the site
of the residence of a seventh-century Hata leader named Kawakatsu;
and the Benevolent Longevity Hall (Jijuden).
The name of the Purple Sanctum Hall was borrowed from a similar structure at the T'ang imperial palace, "purple sanctum" signifying the presence chamber or building of an emperor. The largest
structure in the compound, measuring about 100 by 85 feet, the hall
served as the site for lesser rites and ceremonies involving the emperor, occupying a position in the south-central portion of the compound physically and functionally analogous to that of the Great
Hall of State within the Greater Imperial Palace. Facing south across
a courtyard toward the main entrance gate to the compound, the
large, simple building contrasted sharply, however, with the grandiloquently sinitic lines of the Court of Government building, its unpainted surfaces, cypress-bark roof, and high plank floors echoing
the architecture of Shinto shrines. The imperial-throne chair occupied the central chamber of the hall, facing south in front of the Panels of the Sages, a series of removable partitions decorated with representations of meritorious Chinese ministers that may not have
been installed, however, until the reign of Emperor Saga. Complementing the Great Hall of State, where the great ceremonies of state
were held, the Purple Sanctum Hall and its southern courtyard were
used to accommodate such events as Buddhist services, coming-ofage ceremonies for the emperor and crown prince, and the ordinary
ceremonies and rituals of the court's annual calendar.
The Benevolent Longevity Hall, its name alluding to a Chinese
classic that associated the Confucian virtue of benevolence or humanity with longevity, was a cypress-bark-shingled building located
directly north of the Purple Sanctum Hall at the exact center of the
compound. With perhaps 15 percent less space than the Purple
Sanctum Hall, it was the usual residence of the emperor, its somewhat cramped central chamber being further divided into two rooms
separated by a corridor. On the north, it communicated through another building with the quarters of the imperial harem.
OTHER PUBLIC BUILDINGS AND SPACES
Outside the palace, the city was occupied chiefly by individual residences, but there were also a number of public buildings, facilities,
Cambridge Histories Online © Cambridge University Press, 2008
PUBLIC BUILDINGS AND SPACES
117
and spaces that in their sum accounted for perhaps as much as 5 to
10 percent of the total area. The public facilities included the city's
two prisons, disposed slightly asymmetrically east and west of the
palace; the building of the government's Academy of Chinese learning (Daigaku-ryo, "Bureau of the Greater Learning") and an official
granary directly south of the palace; the Carpentry Bureau near the
southeastern corner of the palace; the lodgings and work space for
the labor levies allocated to the various government offices, concentrated east and west of the palace; the offices of the city's administrative organs on either side of Suzaku Avenue just south of the
palace; and probably some charitable institutions for the sick, destitute, and the orphaned that are known to have existed from at least
the 820s on. But the most prominent public spaces were the paired
foreign embassy lodgings, the markets, and the Buddhist temples
that flanked Suzaku Avenue in the southern half of the city, and the
spacious Park of the Divine Spring (Shinsen'en) directly south of the
palace on the eastern side of Suzaku.
Two temples, the only Buddhist institutions permitted in Heian,
were at the southern edge of the city on four blocks (over 14 acres)
of land each, about 300 yards east and west of Rampart Gate. Commonly known because of that arrangement as the West and East
Temples (Saiji and Toji), the two establishments were built at government expense and under government supervision to obtain divine protection for the state and its capital and to serve other official
needs, such as the performance of mourning rites for emperors.
Athough construction of the multitude of buildings required for the
temples may have begun as early as 796, progress was slow, hindered
partly by the large scale and luxuriousness of some of the structures,
but chiefly no doubt by the competing demands of construction at
the palace and by the financial strains of the EmishiWars. Both temples appear to have been in operation by about 816, but construction
continued long after that, the pagoda of the East Temple not being
completed perhaps until the 870s or 880s and that of the West Temple not until 906. The heights of the pagodas are unknown, but once
completed they must have been among the most conspicuous features on the Heian skyline. The East Temple survives on its original
site in modern Kyoto, but all of its buildings are much later reconstructions, its famous 180-foot pagoda, for many the symbol of historical Kyoto, dating from 1644. The West Temple disappeared in the
thirteenth century, but the site has been extensively excavated, and
the general features of the temple are known.
About a half-mile north of the temples lay two official markets,
Cambridge Histories Online © Cambridge University Press, 2008
Il8
THE CAPITAL AND ITS SOCIETY
similarly called the East and West Markets, about 600 yards east and
west of Suzaku Avenue. The markets were, interestingly, among the
earliest features of the Heian landscape, having been transferred
there from Nagaoka three or four months before the arrival of Emperor Kammu in 794. Walled and gated, it appears, and distinguished architecturally by a tower or loft structure, each market was
four blocks in area, the same size as the temple sites, and contained,
in addition to the stalls, warehouses, and residences of the merchants, the offices of the market administrators.
At the same distance north of the temples but facing each other
directly across Suzaku Avenue in Shichijo ("Zone Seven") were the
two Korokan Lodgings, the government's residences for visiting foreign embassies. The name of the lodgings, which was borrowed
from the title of a Chinese ministry in charge of foreign relations, is
said to have meant "transmission of the voice," implying direction
and assistance of (presumably) foreign peoples. Falling administratively under the Bureau of Buddhism and Aliens, the residences occupied walled sites that were probably equal in size to those of the
temples and markets. Since the lodgings were used only for the very
occasional embassy from Po-hai, they usually stood empty, and special supervision by the city's administrative offices was required in
order to prevent vandalism and occupation by squatters. The East
and West Markets lay due east and west of the lodgings, separated
from them by little more than 300 yards. Given the highly commercial nature of the Po-hai embassies, it may be assumed that the
propinquity of lodgings and markets was not coincidental. A medieval statement that the Koorokan originally occupied the sites of
the East and West Temples and were moved to their Shichijo locations to make way for the temples appears to be mistaken, but if
such a move did in fact take place, it may have been as much to
bring the foreign embassies closer to the markets as to accommodate the temples.
Finally, east of the northern terminus of Suzaku Avenue directly
south of the palace lay the Park of the Divine Spring, a 30-odd-acre
stretch of water and woods that probably preserved something of
the original natural landscape of the Kyoto basin. On the shores of
the spring-fed pond stood the park's chief architectural feature, the
Celestial Presidence Pavilion (Kenrinkaku), a central building with
two connected wing structures on either side where the emperor and
his courtiers gathered for banqueting, archery exhibitions, flower
and autumn-leaf viewing, music, dance, and poetry composition, or
Cambridge Histories Online © Cambridge University Press, 2008
RESIDENTIAL DISTRICTS
119
simply to escape the heat of a summer day. Hunting and fishing also
occasionally took place in the park, the latter perhaps from an angling hall that stood near the main pavilion, and visitors could enjoy
the view of a waterfall from another nearby structure. There was an
island in the pond, probably connected to the shore by a bridge. The
water deity later known as the Good Dragon Queen (Zennyoryuo)
may have already been enshrined there, perhaps providing the impetus that during the course of the ninth century transformed the
park from an imperial pleasure ground into a sacred place where
holy men prayed for rain or for the cessation of rain, or sought to
soothe the angry spirits that caused epidemic disease.
RESIDENTIAL DISTRICTS AND POPULATION
An allocation of house lots was made to prospective residents of the
city in the fall of 793, and since most of the court must have accompanied Emperor Kammu when he transferred his seat to Heian a
year later, it may be assumed that the construction of private
dwellings on the lots proceeded rapidly. Little is known about individual land occupancy and use in the early days of the city, but if the
practices associated with the agricultural land system were applied,
house lots may initially have been allotted to families for limited periods of time. If so, the system soon broke down, for by the beginning of the tenth century lots were being treated as private property
that could be inherited and sold. Although it is not clear that commoners (those who did not hold court rank) were provided for at all
in the original city plan, or, if they were, how and where they acquired dwelling sites, the administrative nomenclature itself-house
lot (henushi) - seems to suggest that each family was entitled to one
lot. If commoners did in fact receive house lots, they were almost
certainly those located farthest from the imperial palace and off the
major avenues.
In accordance with a Nara-period precedent, individuals who held
court rank may have been entitled to varying amounts of land, depending on their ranks. According to the earlier usage, the very highest ranks (First to Third), which in 793 were held by only five men,
were supposed to receive a maximum of thirty-two house lots,
the area of an entire city block (3.5 acres). The next highest ranks
(Fourth and Fifth), numbering perhaps seventy-five or so men in
793J were entitled to a maximum of half that amount, and those with
lower ranks (Sixth or below) to half that again, that is, to a maximum
Cambridge Histories Online © Cambridge University Press, 2008
12O
THE CAPITAL AND ITS SOCIETY
of eight lots.9 If the Nara-period precedents were in fact followed, it
is clear that not everyone received his maximum share of lots, since
the five thousand to seven thousand lower-ranking officials estimated to have been in the capital would alone have been entitled to
over forty thousand lots, and there were only about thirty-two thousand lots available in the entire city plan. We may suspect that most
of the senior ranks received approximately the maximum areas to
which they were entitled, as a certain amount of later evidence suggests, whereas the lowest ranks rarely, if ever, did so. However, even
if the lower ranks received just one house lot each, their residences
accounted for about a sixth of all the lots in the city plan. Their
prominence in the capital would, in fact, have been even greater than
that fraction suggests because, as we shall see, only about half of the
city area seems actually to have been fully developed and inhabited.
Such in outline were the physical circumstances of Heian during
its first decades of existence. The city clearly was devoted to but one
purpose: the provision of living and work space for people associated
with the workings of the central government, from the emperor
down to the lowliest laborer. It was to be a capital city and nothing
more, and its plan revealed that exclusive goal with exquisite clarity.
The largest single area, amounting to nearly 7 percent of the total
space, was occupied by the Greater Imperial Palace, which sat at the
end of Suzaku Avenue, the great thoroughfare that led directly from
the city's main entranceway at Rampart Gate. The palace was the
symbolic center of the city, and the symbolic center of the palace was
the ceremonial center of government, the Court of Government and
its Great Hall of State, which lay on a line extending due north from
Suzaku Avenue just beyond the main entrance to the palace precincts. One could scarcely imagine a more graphic representation of
the city's function.
At the same time, there was little else except government buildings and private residences in the rest of the city. Religion, as represented in the East and West Temples at the city's southern edge, was
confined to less than 1 percent of the area; commerce existed in limited form on similarly narrow sites at the East and West Markets, and
industry outside the government's own workshops was not represented at all in the city plan.
Despite the narrowness of its conception and flaws in its planning,
9 Yoshioka Saneyuki, "Kizoku shakai no seijuku," in Hashimoto Yoshihiko, ed., Komonjo no
kaiaru Nihon shi, vol. 2: Heian (Tokyo: Chikuma shobo, 1991), pp. 93-138.
Cambridge Histories Online © Cambridge University Press, 2008
RESIDENTIAL DISTRICTS
121
Heian may rightly be called the first successful city in Japan, not only
because it survived and prospered, but also because it is there that
we can see clearly for the first time in Japanese history the distinctive characteristics of what may have been the most important product of the statutory system of government: urban life and civilization. The capital at Heizei was undoubtedly urban in essential ways,
but the brevity of its life and the paucity of sources leave only scant
knowledge about actual conditions in the city. Thanks to the work of
archaeologists, the physical layout of Heizei and its palace is generally well known; the Shosoin and the older temples of the Nara area
reveal in their structures and treasures much about the physical,
artistic, and intellectual environment of the imperial court and the
Buddhist clergy; and written sources tell about government and its
problems. But there is little that can be learned with certainty about
the society of Heizei itself, about what it meant in concrete terms to
be an inhabitant of the city. With Heian, however, and especially
from the tenth century on, an increasingly abundant and varied supply of written sources begins to reveal at least the outlines of life at
the capital, a picture skewed certainly toward the imperial court and
the nobility but full enough nevertheless for us to recognize in it an
ever more urban society.
The distinctively urban character of the society appears in several
ways, but the most fundamental were, on the one hand, the diversity
of the population, and, on the other, the population's removal, for
the most part, from primary modes of production.
How large that population was at the city's founding or at any
other time during its first few centuries of existence can be estimated
only in the crudest fashion, but one can at least say that the frequently cited figure of 200,000 is almost certainly too high for the
early decades of the city's life. (Inflation of early-city population figures seems to be a common failing of historians, who often mistakenly extrapolate from the size of households determined either on
the analogy of much later social history or by guesswork.)10 There is
reliable evidence to show that in 829 there were only a little over 580
blocks (machi) in the city, or, in other words, not much more than
half of the approximately 1,100 blocks originally planned in the area
outside the imperial palace. Since a fairly large number of blocks was
reserved, as we have seen, for public institutions and spaces, residential space in the city probably amounted to only about 500
io Several urban historians have made the point, including Gideon Sjoberg, The Preindustrial
City, Past and Present (Glencoe, 111.: Free Press, i960), pp. 80-85.
Cambridge Histories Online © Cambridge University Press, 2008
122
THE CAPITAL AND ITS SOCIETY
blocks, which, at thirty-two lots {henushi) per block, would have contained sixteen thousand lots. If we assume for the moment that one
household occupied one house lot, as the city's planners may have
envisaged, and if we accept somewhat uncertain evidence for 665
households in the year 871 indicating that the average size of a Heian
household was 6.2 persons, we can calculate that the maximum population of the city outside the imperial palace during the ninth century was around 100,000. It may be true that many house lots were
occupied by more than one household, but since there were other
households that occupied several lots each, and since it also seems
probable that there were unoccupied lots even in the developed
areas of the city, the assumption that there was an average of one
household per lot is perhaps as reasonable as any other. It is also true
that the average household size may have been larger than the figures for 871 indicate. There is evidence to suggest that the average
household size in the western section of Heizei during the Nara period was 9.4, but even if we use that figure in our calculations, the
population of Heian would still be no more than 150,000.
In sum, the maximum population of Heian proper outside the imperial palace in the ninth century seems most likely to have been
somewhere in the neighborhood of 100,000 to 150,000. If we accept
a more conservative estimate based chiefly on the number of government officials and employees in the city, the figure may have actually been as low as 70,000. Such numbers are unimpressive by the
standards of present-day megalopolises, but if they are even approximately correct in indicating the size of the city's population, they
made Heian one of the larger cities in the world of its time and gave
it an urban role that was even more central to the country than huge
cities like Paris and Tokyo are to their modern-day societies. Just as
a Wyoming city of 50,000 is, because of its physical and demographic
context, a more important urban place in almost every sense than a
bedroom town of similar size on the outskirts of London, so also undoubtedly was a ninth-century capital of 100,000 people, located in
a country otherwise sparsely inhabited and almost wholly rural in
population, a more significant center of urban civilization than a metropolis of millions is in a modern industrial state.
It has been estimated on the basis of fairly good evidence that at
the beginning of the tenth century there were between 5,000 and
10,000 people who held titles in the organs of government at Heian
or were employed at the imperial palace. If that figure is correct,
most of the population in the city during its early days must have
Cambridge Histories Online © Cambridge University Press, 2008
IMPERIAL CLAN AND COURT NOBILITY
123
been either directly in the service of the court and its governmental
apparatus or resident as family members or servants in the households of those who were so employed, numbering altogether perhaps
as many as 50,000 to 60,000 people. Although that population was
uniform in the sense that it was employed by, or was indirectly dependent on, the court and the government, the uniformity encompassed a diversity of occupations and social distinctions. These
clearly distinguished the society from that of the countryside, where
in any particular location one occupation (usually agriculture)
tended to monopolize the economy, and specialization of labor and
social differentiation were limited.
IMPERIAL CLAN AND COURT NOBILITY
The most conspicuous and best-known part of that large official or
courtly population was, of course, the political and social elite: the
imperial clan and the court nobility.
It is possible that the statutory regime is more accurately characterized as a rein on imperial power than as an expression of it. The
emperor was far from being a despot under the regime's code, which
in rule and practice gave great power to the nobility, who controlled
entry to their own and the lower ranks of officialdom and were the
conduit for most official documents issued under the imperial seal.
If the regime was not quite the creature of the nobles, it seems to
have served them nearly as well as it did their sovereign. But the emperor stood unquestionably at the apex of the social and political
structure. Although his position and authority were defined chiefly
by implication in the statutory code, all governmental action was
taken ultimately in his name, and the more important actions required his direct, personal approval, or that of his regent. The emperor presided over the chief court rituals and ceremonies, which
were widely considered the most significant contribution a ruler
could make to his own and the general welfare, lying as they did at
the very foundation of a healthy state. And the emperor was the universally acknowledged arbiter of social status, a recognized incarnation of divinity, greater or lesser proximity to whom usually defined
the standing of his subjects. His authority and prestige were nearly
as absolute as his power and influence were inconsequential later
under the Fujiwara regents, which may be a principal reason that the
line survived.
Succession to the throne was not regulated by law, but custom
Cambridge Histories Online © Cambridge University Press, 2008
124
T H E
CAPITAL AND ITS SOCIETY
provided unmistakably clear guidance. Although there had been a
number of reigning empresses in the eighth century and before bearing the imperial title of tenrio, the throne throughout the Heian period was occupied exclusively by males, all of whom were sons of emperors and succeeded one another in the masculine line: from father
to son in a bare majority of cases, but also frequently from brother
to brother and exceptionally from uncle to nephew or nephew to
uncle, cousin to cousin, or to a great-uncle (a great-grandfather's
son). The chief qualification was that of imperial son, and the preferred succession was from father to son or brother to brother, but
any line was possible in case of need or for the sake of political convenience. (Kammu's father, Konin, was, it will be remembered, an
exception in the Nara period to the rule of succession by imperial
sons: he was the paternal grandson of an emperor but not the son;
similar exceptions had occurred earlier.) Imperial succession was
unusually flexible. For example, a brother might succeed to the
throne even when a son was available, and yet he did not necessarily do so. Such flexibility made almost any imperial son a potential
heir to the throne and thus sometimes led to, or was used as a device for the creation of, conflict, but it worked on the whole with
surprising smoothness during the Heian period, which was spared
the war and bloody intrigue among imperial sons that might have
been expected from such circumstances. One or two attempts on
the throne may have been made during the period (the details are
not altogether clear), but succession to the position was never sullied by bloodshed or physical violence.
The imperial clan shared materially, socially, and sometimes politically in the emperor's exalted position. Until 798, the clan included
by law male and female descendants of emperors in the male line
down to the fifth generation, but in that year, probably for reasons of
economy, the original provision of the statutory code excluding the
fifth generation and beyond was restored. The clan remained very
large, including all patrilineal descendants of an emperor down to the
generation of his great-grandchildren's children, so that a relative as
remote from a reigning emperor as the child of the great-grandchild
of the father of the emperor's great-grandfather (his fourth cousin?)
was a member. Those numerous imperial relatives were distinguished
by their titles into two groups according to the degree of lineal
proximity to the emperor: (1) the "near-princes" (shinno) or "nearprincesses" (naishinno), and (2) the "princes" (0) or "princesses"
(nyoo). The near-princes and -princesses were imperial children and
Cambridge Histories Online © Cambridge University Press, 2008
IMPERIAL CLAN AND COURT NOBILITY
125
siblings, and originally all relatives of that degree of relationship received one of the titles. But, again for reasons of economy, during the
course of the Heian period it came to be the practice that only those
children of an emperor who had been specifically granted the titles
by imperial decree were allowed to use them and enjoy their emoluments and perquisites. All other members of the clan bore the lesser
title of prince or princess, titles that might also be used by descendants in the fifth generation but without the right to court support.
Male descendants in the sixth generation and beyond sometimes
used the tide of prince, but the text of the statutory code is not clear
on the legality of that practice.
By law, the clan had originally tended toward endogamy, as had
apparently many of the nonimperial clans, but the emperor himself
early took the lead in ignoring legal restrictions, and in the Heian period emperors and their male descendants married freely outside the
clan. Female imperial descendants, on the other hand, seem to have
been bound by endogamy rules until well into the period. The first
known instance of the marriage of a princess to a man outside the
clan did not occur until the generation of Emperor Nimmyo's granddaughter, and the first known marriage of a near-princess to an outsider occurred in the reign of Emperor Daigo (897-930). Thereafter,
however, such marriages were common.
The size of the imperial clan at any particular time in the early
part of the Heian period is unknown, but in the latter half of the
ninth century, when efforts had already been made to reduce its
number, there were over five hundred people with princely titles that
qualified them for government support, and toward the beginning of
the tenth century there were upward of seven hundred. The level of
support was generous, sometimes amounting, through appointment
to high court rank and remunerative offices, to an annual income of
as much as twelve hundred times that of an ordinary agriculturalist.
That support, which often included as well valuable perquisites
other than those derived from rank and office, became a very heavy
burden on the court, and steps were taken, as we have seen, to reduce the level of support and further check the clan's growth. A critical turning point came in the reign of Emperor Saga (809-23), when
for the first time large numbers of imperial sons and daughters were
removed from the clan and reduced to the level of the court nobility. That step had the effect not only of immediately paring the size
of the clan by seventeen near-princes and fifteen near-princesses,
but also of eliminating the possibly two hundred or more princes (0)
Cambridge Histories Online © Cambridge University Press, 2008
126
THE CAPITAL AND ITS SOCIETY
and princesses (nyoo) who might have sprung from the lines of the
imperial sons during the next three generations. Saga's policy, continued by succeeding emperors down to Murakami (r. 946-67), the
smaller size of the imperial harem under the Fujiwara regents, and
the immaturity of many emperors after the tenth century so greatly
reduced the size of the clan that only a single prince is known to have
been alive in 1143. It is symptomatic of the clan's changed circumstances that although the imperial genealogy lists a total of eightytwo descendants with princely titles in the first two generations
under Emperor Kammu, it shows only sixteen for the same span of
generations under Go-Sanj5 (r. 1068-73).
An imperial consort, like all Heian wives, remained a member of
her natal clan and was not formally assimilated to that of her husband. But her status even when she was not of imperial origin was
in most ways equivalent to that of imperial clan members both during the lifetime of her husband and, if she survived him, after his
death as well. The statutory code provided for ten imperial consorts
hierarchically arranged under four titles of descending prestige conferred by decree of the emperor. The highest-ranking title, that of
empress (kdgo), was held by only one consort at a time, a woman of
the imperial clan chosen from among the two women holding the
next lower title in the hierarchy, who were also members of the imperial clan. The rule restricting the two upper titles to women of the
imperial clan had already been breached in the eighth century, however, and the empress was in fact normally a Fujiwara woman
throughout the Heian period. As indicated by the literal meaning of
the empress's title, "Lustrous Heir-bearer," it was originally conferred on a consort who had already borne the emperor a son, but
in the reign of Reizei (967-69) an exception was made for the cherished only daughter of Emperor Suzaku, and thereafter it was common for childless women to receive the title. It remained the case,
however, that no woman was ever directly appointed empress, always first holding one of the lower consort titles. Until the time of
Ichijo (r. 986-1011) it was also the case that not every emperor had
an empress.
The lesser consort titles of the code, which were held by women
from the clans of the court nobility, disappeared and were replaced
by other, office-derived designations (notably nyogo, "female attendant") during the ninth and tenth centuries. As a result of Fujiwara
political need a second empress's position and title (chiigii, "Inner
Palatine") became available at the beginning of the eleventh century.
Cambridge Histories Online © Cambridge University Press, 2008
IMPERIAL CLAN AND COURT NOBILITY
127
But despite such changes in substance and nomenclature, the code's
system of a hierarchically arranged group of imperial consorts drawn
from a restricted group of clans remained basically intact, usually
making at least half a dozen princely and noble consorts simultaneously available to a mature emperor.
Those women lived with their ladies-in-waiting, their servants,
and the numerous female court officials and servants in the twelve
connected buildings of the "rear palace" (kokyu) north of the emperor's residential palace, where, especially from the tenth century
on, they presided over one of the chief centers of noble social life.
The total female population of the rear palace in the Heian period
has been estimated at one thousand, but that seems high for the
available space. Those living and employed there may have numbered that many, but it seems unlikely that all would have been present at one time, since the rather small buildings would otherwise
have been virtually wall-to-wall with people.
Although the rear palace was by no means freely accessible to any
noble, neither was it a sultan's seraglio jealously guarded and disciplined by a corps of eunuchs. The consorts were not prisoners of the
palace, and even while there they were not isolated from society.
They frequently returned for visits or childbirth to the homes of
their parents or other close relatives, and while at court they were
freely visited in their quarters by a variety of male and female relatives. There were probably few higher-ranking nobles who did not
have some kind of access to the rear palace. Since each consort usually sought to make the physical trappings of her quarters as attractive as possible and to surround herself with particularly accomplished and beautiful ladies-in-waiting, the result was at least
sometimes a highly stylish salon where men and women were able to
meet and entertain themselves with considerable freedom. The
parentage of many nobles and the literature of the period suggest
that the freedom frequently enough included sexual trysts, and it is
clear that the consorts themselves were not always blameless in that
respect.
Principal consorts during the Heian period came from a handful
of noble clans in the capital that ranked in social status below the imperial clan but included in their number the great Fujiwara lineage
and its Northern House, whose power and wealth for much of the
time rivaled or eclipsed that of the emperor. The term "Northern
House" (Hokke), which does not appear in sources until about the
twelfth century, refers to the Fujiwara line descending from Fuhito's
Cambridge Histories Online © Cambridge University Press, 2008
128
THE CAPITAL AND ITS SOCIETY
son Fusasaki (681-737), the line of the Fujiwara regents. There were
three other lines descending from Fusasaki's brothers: (1) the Southern House (Nanke) from Muchimaro (680-737); (2) th e Ceremonials House (Shikike) from Umakai (694-737); and (3) the Capital
House (Kyoke) from Maro (695-737). Of the three, only the Southern and Ceremonial houses figured prominently in Heian history,
and even they quickly faded from view as the Northern House established its ascendancy at court. (For Fujiwara genealogy, see Figures 1.2 and 1.3.)
The "clan" in the early part of the period and during the Fujiwara
regency was a loosely knit, patrilineal kin group of nobles whose
members shared an ancestral or guardian deity, bore a common
patronymic (except for the imperial clan, which had no name) and
hereditary title of status, acknowledged a common clan chieftain (uji
no choja), and were usually buried together in a clan cemetery. It was
a survival of what very scanty and problematical evidence suggests
to have been in pre-Taika days a more substantive and powerful kin
and fictive-kin organization with hereditary political and economic
functions that seems to have been created in the consolidation of
central rule in Japan as an instrument of imperial power. "Clan" is
used here conventionally to refer to what was called in Japanese the
uji, a term of possibly Korean origin. Since it is clear that the preTaika uji was not characterized by some of the major features of the
clan, as that term is traditionally used (it was not, for instance, exogamous and may not have been unilineal), some have been reluctant to call it a clan at all. But the concept of P. Kirschoff's "conical
clan" seems to fit the early Japanese uji fairly well, and the use of the
term may be less misleading than sometimes supposed.11 It should
be noted, however, that the early uji differed significantly from the
conical clan in that succession to the chieftainship was not primogenitary but shifted continuously among collateral lines.12
The inauguration of the statutory code deprived the uji clan of
most of its direct political role, but it remained the broadest kin
group to which a noble belonged and continued to play an important role in the lives of individual clan members. Membership in a
clan was itself a definition of nobility in the broadest sense of the
11 PaulWheatley and Thomas See, From Court to Capital: A Tentative Analysis of the Japanese
Urban Tradition (Chicago: University of Chicago Press, 1978), pp. 94-95, appear to have
been the first to call attention to the conical clan in connection with the uji, but scholars
in Japan have since reached the same conclusion.
12 Pointed out by Yoshida Takashi, "Uji to ie," Sasaki Junnosuke and Ishii Susumu, eds.,
Shimpen Nikon shi kenkyii nyumon (Tokyo: Tokyo Daigaku shuppankai, 1982), pp. 31-58.
Cambridge Histories Online © Cambridge University Press, 2008
IMPERIAL CLAN AND COURT NOBILITY
129
term; it determined the court positions to which a member might
reasonably aspire; in part it made possible the achievement of those
positions; and it was a controlling influence in the religious life of
members. The clan chieftain - whose position was not hereditary in
a direct line but passed by imperial decree usually to the clan member holding the highest court rank and office - was responsible for
the worship of the clan deity or deities; for the administration of clan
shrines and temples, including fairly extensive police and judicial
functions; for the execution of imperial orders addressed to the clan
(orders requiring, for example, the presentation of clan women for
service at court); for the discipline of members by expulsion from
the clan; and, to some ill-defined extent, for the physical and educational welfare of clan members. (In the case of the Fujiwara, the
Kangakuin, the "Learning Promotion Court," a clan dormitory and
school for members enrolled in the government's Academy of Chinese learning, was under the chieftain's jurisdiction.) He also enjoyed certain privileges symbolic of distinction, such as burial in the
manner usually reserved for holders of the Third Rank or above. But
from at least the tenth century his most important prerogative was
the nomination each year of a Sixth-Rank clan member for promotion to the Fifth Rank. That prerogative in effect gave him the power
to create a primary member of the nobility, for it was the elite stratum of holders of the Fifth Rank and above who supported their
clans with their incomes, controlled the government, and led society.
It is important to note that the clan seems neither in its earlier history nor in the Heian period to have been an organization found
throughout Japanese society. It is thought to have been restricted
from the outset mainly to nobles at, or closely tied to, the court of
the emperor, and in the Heian period the clans, insofar as they are
known, were all centered on the capital. By the middle of that period
only a few survived: chiefly the Fujiwara, the Taira, the Tachibana,
the Sugawara, the Otomo (Tomo), the Takashina, the Onakatomi,
the Imbe, the Urabe, the Wake, and the Ochi. There was also the recently founded Minamoto clan - a single clan despite the diverse
imperial origins of its various lineages; its clan chieftainship was held
in the Saga line up through the mid-Heian period, then by the Murakami line, and finally by what is usually identified as the Seiwa
line, that is, the line of the Ashikaga andTokugawa shoguns.The imperial, or "princely" (o), clan, consisting of the princes and princesses (near and otherwise), also survived, of course, but it was distinguished from the noble clans by the special conditions attaching
Cambridge Histories Online © Cambridge University Press, 2008
130
THE CAPITAL AND ITS SOCIETY
to membership in it (descendants in the fifth generation and beyond
outside the direct line of imperial succession were excluded); by its
lack of a patronymic; and by its peculiar relationship to the focus of
kinship in the clan, the emperor, who seems himself not to have been
a member. There were, in addition, many other presumably patronymic lineages found among the lower reaches of officialdom and in
private employment and occupations (over eighty such names can
be found, for example, in the diary of Fujiwara no Michinaga), but
although little can be said with certainty about them, it is nearly certain that they were not clans of the sort just described.
All members of the clans were "noble" in the sense that they bore
imperially conferred hereditary titles of honor called kabane that
seem in practice to have been the essential qualification for appointment to one of the five highest court ranks. Although those ranks
were not hereditary in law and often not in fact, the statutory regulations governing appointment to rank ensured that once a member
of a clan had been appointed to the Fifth Rank or above, his lineage
would continue to receive appointment at that level. The hereditary
nature of rank was the result of a system of preferential treatment for
men whose fathers or grandfathers held, or had held, noble rank
(Fifth or above). Such men on reaching the age of twenty-one calendar years were entitled by the fact of their birth to appointment to
relatively high rank (Fifth down to Eighth), the precise rank depending on the rank of the father or grandfather. That privilege gave
even those appointed to die Sixth Rank or below (the overwhelming
majority) practical assurance of promotion to noble rank, their rise
being accelerated by other kinds of preferential treatment, and their
eventual appointment to the critical Fifth Rank facilitated by the direct control over such appointments that seems to have been exercised by the Council of State, where sympathetic kinsmen were frequently to be found. Sons of men who never rose above the Sixth
Rank, on the other hand, started at the bottom court rank at the age
of twenty-five calendar years (four years older than noble sons at
the time of their initial appointment to rank) and moved slowly up
the ladder through an elaborate review and evaluation process that
effectively barred all but the most exceptional or fortunate from the
upper ranks. It might take a man starting at the bottom rank as long
as thirty or more years to reach the rank at which the son of a FifthRank noble began his career (the Junior Eighth Rank Lower).
As a result of such privileges, the number and identity of noble
clans (those whose male members regularly achieved the Fifth Rank
Cambridge Histories Online © Cambridge University Press, 2008
IMPERIAL CLAN AND COURT NOBILITY
131
or above) varied little from the beginning of the statutory system up
to the early Heian period. The highest and mightiest nobles, who
held the Third Rank or above and were distinguished under the code
as "the august" (ki), came from only about 20 clans, while holders
of the lower noble ranks (the Fourth and Fifth), who were known
under the code as "the august equivalent" (tsuki), represented another 150 to 200 clans. Few new clans were admitted to the noble
ranks, and few of the established clans lost status before the ninth
century, when the atrophy and decline of the statutory system and
the growing monopolization of offices by the Fujiwara began to reduce the number of clans to the dozen or so already mentioned.
Even then, however, the system of preferential appointment to rank
continued to operate in defining the nobility, although Fujiwara
power and the inflation of ranks soon led to much higher initial
ranks for noble sons at much earlier ages.
But the formal definition of the nobility in terms of court rank was
no longer entirely valid by the tenth and eleventh centuries, for by
then shrinking governmental revenues and a tendency toward inflation of ranks had erased much of the sharp difference that had separated the Fifth from lower ranks. More appointments were being
made to the noble ranks, and the lower ranks, deprived very often of
any economic meaning at all, were falling into disuse. Reflecting the
personalization of government and the shift of the center of governmental activity to the Emperor's Residential Compound, the chief
distinguishing feature of the nobility then was a combination of rank
(Fifth or above) with the privilege of attendance in the imperial audience chamber, a privilege that was enjoyed ex officio by holders of
the top three court ranks and by Consultants (sangi) of the Fourth
Rank (a group known collectively as the kugyo, "lords and ministers") and that was granted by imperial decree to selected holders of
the Fourth and Fifth Ranks and to certain other officers who required access to the audience chamber in the performance of their
duties.13
The nobility at the beginning of the Heian period was still generally
what it had become under the impact of the Taika Reforms and the
13 In the years around 1000 there appears to have been a fixed number of senior nobles
(kugyo): sixteen. The number radically increased thereafter. Senior nobles holding appointment at the level of Consultant (sangi) or above were called "currently active" (genniri), and it was they who took part in the deliberations of the Council of State (daijokan).
Yoshioka, "Kizoku shakai no seijuku," p. 112. See also William H. and Helen Craig McCullough, A Tale of Flowering Fortunes: Annals of Japanese Aristocratic Life in the Heian Pe-
riod, 2 vols. (Stanford, Calif.: Stanford University Press, 1980), pp. 790-91.
Cambridge Histories Online © Cambridge University Press, 2008
132
THE CAPITAL AND ITS SOCIETY
institution of the statutory state: a partially bureaucratized social and
economic elite concentrated in the capital and almost totally divorced from the direct use and exploitation of rural land. It was the
core of the urban population whose culture defined the civilization
of the period, controlling the governmental apparatus, reaping the
major economic rewards of that apparatus, dominating society, setting standards of behavior that were recognized by most as the society's ideals, writing the literature, practicing and patronizing the
arts, and creating the religious and intellectual life of the day. It fluctuated in size through the period but seems usually to have consisted
of somewhere between 100 and 125 men, who together with their
families may have accounted for 1 or 2 percent of the capital's population. Women also held ranks among the nobility, but since they
normally acquired those ranks only as consanguinal or affinal relatives of an emperor or regental noble and did not usually hold substantive office within the statutory structure, they are perhaps better
excluded from a tally of the nobility's primary members.
Although the statutory government was in law a largely meritocratic bureaucracy, insofar as the regime survived in the mid-Heian
period and after, its offices tended to become hereditary in particular sublineages of clans called "houses." That was especially true
initially of those lower offices requiring what might be regarded as
specialized or technical skills. The subjects of Chinese history and
poetry at the Academy of Chinese learning, for example, were taught
by a Professor of Literature (monjo hakase), an office hereditary in
the Sugawara or Oe houses; key posts in the Secretaries' Office
(gekikyoku) of the Council of State requiring a knowledge of Confucian texts went to Nakahara and Kiyohara men; and yin-yang professors expert in calendrical and astronomical matters were drawn
from the Abe and Kamo. Such hereditary "house status" (kakaku)
determined not only the houses of clans from which appointments
to particular titles could be made, but also usually the careers to
which a man might aspire. A Sugawara noble did not, if he was prudent, aim to be a Minister of State (daijiri). A "house status" designated, in fact, an entire career, both a hereditary office and the particular route of promotion thereto. The system eventually extended
also to the major governmental posts, evoking a nomenclatural system that by the late Heian period in the twelfth century and after included in descending order of rank and prestige: (1) the "five regental houses" (gosekke) of the Northern House of the Fujiwara clan;
(2) the "limpid flower" (seika) houses whose members were the
Cambridge Histories Online © Cambridge University Press, 2008
IMPERIAL CLAN AND COURT NOBILITY
133
hereditary holders of Minister of State posts; and (3) the "feather
grove" (uriri) and "name" (met) houses whose members similarly
held the Middle and Major Counselor posts.14
The senior nobles, those "august" ones who held the highest three
ranks and occupied the chief positions of the court government,
numbered a scant five at the beginning of the period, but economics and family ambition had quadrupled that figure by the eleventh
century. At the same time, the several old-line clans that had still
been represented among the senior nobles in the early part of the period were gradually excluded from their number as the Northern
House of the Fujiwara established its ascendancy. By the last half of
the ninth century the group was made up almost entirely of Fujiwara
clan members and men of the imperial Minamoto families, who
were often marital kin of the Fujiwara.
The ordinary nobles of the "august equivalent" Fourth and Fifth
Ranks (or, in the mid-Heian period, those of them who held audience chamber privileges) were not usually directly involved in the
highest councils of government. However, they commonly held key
administrative or ceremonial positions that gave them access to, and
influence with, the emperor or his chief ministers, and they were full
participants in court life, "dwellers above the clouds" (unjo-bito) at
the very apex of society. From the ninth century on, however, the
shrinkage of the statutory regime and its revenues made them increasingly dependent on the favor, protection, and patronage of the
senior nobles, whom they sometimes served in personal capacities.
The differences in status among the nobility and the lower officialdom (those whose highest ranks were the Sixth or below) were
numerous and profound, and there was a similar, if slightly less
sharp, difference within the nobility itself among the senior and ordinary nobles. The differences occurred in the treatment accorded
each group in criminal law, in mortuary matters, in imperial audience privileges, in marital access to the imperial clan, in ceremonial
behavior, in the size of residences, in costumes, and in almost every
other branch of life. The nature of the differences, which, of course,
always favored the superior strata, is most easily seen in the income
allotted to the various ranks under the statutory code. According to
the provisions of the code, the lowest-ranking noble (one of the Fifth
Rank) received approximately ten times the income of the highestranking lower official (one of the Sixth Rank), and there was an
14 Yoshioka, "Kizoku shakai no seijuku," p. 126.
Cambridge Histories Online © Cambridge University Press, 2008
134
THE CAPITAL AND ITS SOCIETY
equally great difference in income between the lowest rank of the senior nobles (the Third Rank) and the highest of the ordinary nobles
(the Fourth Rank).Those incomes changed markedly during the period as the government's economic situation deteriorated, but it is
probably safe to assume that the disparity in wealth between the two
levels of the nobility remained at least as pronounced as in the code's
provisions, and that the gap between the nobles and whatever remained of the lower officialdom by that time was similarly maintained or widened.
THE NOBLE FAMILY, MARRIAGE, AND THE
POSITION OF WOMEN
Despite the continuing importance of the clan in the early and midHeian periods, the chief focus in the day-to-day life of the nobility
was the clan sublineage and the coresidential family or household,
the partial divorce of which from each other distinguished both
from the medieval and later "house" (ie). Sublineages within a clan
seem to have been referred to most commonly as "gates" (kado) or
"houses" (ie), the latter a term excessively familiar to students of
Japanese social history. The word ie was used in Heian vernacular
texts up to the eleventh century mainly in the meaning of "house"
or "home" (the physical place, including the various buildings and
grounds), but sometimes also in reference to the dwellers in a
house, or, in other words, to a family or household. The ie as family had genealogical but not to any very important degree proprietary continuity. One spoke, for example, of a noble or base house,
but not of a rich or poor one; of house lineage but not of house patrimony. Misapplication of the later ie concept to Heian noble society has led to some confusion on the subject.
An examination of the Heian noble family may begin, therefore,
with the observation that the family residence was unconnected with
income or income-producing property. A noble had neither a country seat that was the administrative center of a surrounding landed
estate nor, of course, any kind of urban commercial, trading, or
manufacturing enterprise. His income came almost entirely from
governmental revenues or the revenues from scattered pieces of agricultural land, most of which he never saw, and with the operation
and oversight of which he usually had little or nothing to do. Status
and income were totally divorced from family residence, adhering
instead to individuals, and residential arrangements were therefore
Cambridge Histories Online © Cambridge University Press, 2008
THE NOBLE FAMILY
135
free to follow other principles. A man acquired rank and office
largely through the patriline, but since that rank and office, with the
attendant income, followed him wherever he chose to live, his residence could be, and perhaps even ordinarily was, unrelated to lineage. Residence was determined partly by social convention and
partly by circumstance.
A chief determinant of Heian noble residence was the nature of
the marriage institutions of the time. Before taking up the direct
bearing of marriage on residence, however, three preliminary points
may be made.
One is the very early age at which initial marriages frequently occurred, often at about the age of puberty (girls might be only eleven
or twelve, boys a year or two older). In such marriages, the ages of
the bride and groom were usually fairly close to each other, but the
bride was often two or three years older than her husband, and in
some cases as much as seven to ten years older. In second and subsequent marriages, men often married women much junior to themselves (sometimes mere children), and the marriage was apt to be
the first for the bride. Women were apparently less likely to marry a
second time (their high mortality rate in childbirth and the general
absence of violence between men accounts undoubtedly for part of
the difference - there must have been many fewer widows than widowers), but when they did, the husband seems typically to have been
older and the marriage seldom, if ever, his first.
A second feature of the period's marriage institutions was the
practice of polygyny. Men were permitted to have more than one
wife, although by no means all of them did, even among the wealthiest and most puissant. Women, on the other hand, were not permitted to engage in polyandrous marriage. They could, through divorce
or by the death of a spouse, have more than one husband, but not
two or more at the same time. When a man had more than one wife,
the woman first married seems to have had a strong presumptive
claim to be a principal wife in the sense that (1) her husband's usual
or expected residence was with her, (2) her male children rose higher
in the official hierarchy than the sons of other wives, and (3) the title
used of her by others was that associated with a wife distinguished by
the first two characteristics. The title of a principal wife, "northern
quarter" {kita no kata), was defined and confirmed, however, neither
by law nor by sacred writing, and the status of such wives may have
been as vague in practice as it seems in modern formulation. Although the ranked titles of the imperial harem might lead one to exCambridge Histories Online © Cambridge University Press, 2008
136
THE CAPITAL AND ITS SOCIETY
pect a similar hierarchy of wives in noble polygynous marriages,
there are quasi-historical cases of polygynous marriages in which the
principal-wife title was used of concurrent wives of a single husband,
suggesting equal or nearly equal status for the wives.
A final point about Heian marriages is that they were accomplished solely through domestic rites, and, as far as is known, neither
the government nor any religious authority played a role in sanctioning or confirming them. Furthermore, there appears to have
been no written marriage contract or other similar instrument executed by the families or the principals - the existence of a union
rested entirely on familial and social recognition. There was consequently no legal or moral bastardy, and there is also no evidence that
children of informal unions who were recognized by their fathers
suffered greater social or career disadvantages than was the common
lot of children by secondary wives. Perhaps, in fact, paternal recognition of children was the necessary condition for social regularization of a sexual relationship, in other words, for marriage.
From a social point of view, the most critical aspect of a marriage apart from its basic function of uniting a man and woman for the
production and rearing of children - is apt to be the location of the
married couple's residence. Anthropologists and sociologists often
identify four chief types of marital residence found among the diverse
societies of the world: (1) at or near the husband's parental home
(virilocal, or patrilocal, residence); (2) at or near the wife's parental
home (uxorilocal or matrilocal); (3) at a house separate from either
spouse's parental home (neolocal); and (4) at the respective parental
homes of the wife and husband, the husband visiting his wife at her
house but continuing to live at his own (duolocal). Other types of residence are, of course, possible, and there may be shifting from one
type of residence to another in the course of a single marriage (especially likely in the case of duolocal residence, it seems), but societies
appear usually to exhibit a preference either in practice or in the ideal
for one of the above four chief types. It should be noted, however,
that few if any societies follow exclusively one type of residence pattern. Each of the four residential types and their various combinations is undoubtedly found at least occasionally, for example, in
present-day American society, where neolocal marriage is, nevertheless, usually the goal and practice. Individual circumstances and
needs seem commonly able to overrule social norms in such matters.
During most of the Heian period, and especially during its middle century and a half (950-1100), the prevailing modes of marital
Cambridge Histories Online © Cambridge University Press, 2008
THE NOBLE FAMILY
137
residence in noble society at the capital were uxorilocal, neolocal,
and duolocal. It appears that the society's ideal marriage, the marriage that took place when all things were normal and as they should
be, was uxorilocal, or uxorilocal followed by neolocal residence, and
it is likely that such marriages were also numerically predominant.
They were, in any case, very common among the nobility.15
When a marriage was the first for both bride and groom and the
bride's parents were present and leading nonreligious lives, the typical residence rule seems to have been uxorilocal. Such marriages
began with a period of duolocal residence that lasted for varying periods of time, from a few days up to a year or more (the birth of the
first child may have usually marked the upper limit of the duolocal
phase), but eventually the husband took up residence at the house of
his wife's parents. As the youthful couple matured and began to assume greater responsibility for their own lives, the wife's parents
often moved into another house, usually nearby, or provided a separate house for the couple, and sometimes the husband supplied a
new house for himself and his wife.
The uxorilocal and uxorilocal-neolocal unions were the marriages that might be regarded as the most orthodox. The extreme
youth of the bride and groom, together with the general absence of
opportunities for social intercourse between high-ranking boys and
girls, meant inevitably that such marriages were arranged by the
families of the couples. The wishes of the principals in the matter
were but little if at all consulted (the couple often, perhaps even regularly, met for the first time during their wedding rites, which involved lying together but not necessarily the consummation of the
marriage). The primary considerations were, rather, the usual ones
of noble societies: rank, political or social advantage, wealth, and so
on. The bride's family was especially interested in obtaining a husband whose family connections promised high rank and office for
him at court and thus eventually a large income and broad influence. The groom's family looked for a wife whose family could provide adequate support and care for their son and assist him in his
career at court. Both families also sought to strengthen their political and social standing through marital ties with leading court figures. The natural result of such considerations was fairly strict class
endogamy and, as the Fujiwara came to control most of the choice
15 William H. McCullough, "Marriage Institutions in the Heian Period," Harvard Journal of
Asiatic Studies 27 (1967): 103-67.
Cambridge Histories Online © Cambridge University Press, 2008
138
THE CAPITAL AND ITS SOCIETY
court offices, even a tendency toward clan endogamy, especially at
the upper levels of the nobility. (Incest rules were narrowly drawn,
the only clearly disapproved marriages being those with a direct ascendant or descendant and with a full or half sibling.)
If a family had more than one daughter, the second and subsequent daughters might be married uxorilocally, either in the same
residence with the first daughter or in the house to which the parents later removed, or, especially perhaps when the groom was an
adult, they might be provided by the parents with neolocal residences. As one would assume, when the groom was still a child, the
initial marital residence seems never to have been neolocal.
It was not uncommon for an adult groom or his parents also to
provide the initial residence for a neolocal marriage. That was least
often the case, probably, when special circumstances (a fire, for example, at the wife's parents' house) made the arrangement necessary
or convenient even under what might otherwise be considered normal circumstances (i.e., when the wife's parents were alive and of the
noble group from which the husband might be expected to take a
principal wife). Groom-supplied neolocal residences may have been
the rule, on the other hand, for imperial princes, and also when the
bride was severely disadvantaged: an orphan, or of a station or situation so inferior to that of the groom that her parents could not provide living quarters and support commensurate with their son-inlaw's superior status. If uxorilocal marriages were the most orthodox
and the least sentimental in origin, the latter type of neolocal marriage, in which the husband supplied the residence for an orphaned
girl or a woman of inferior rank or circumstances, was the most romantic, and it is perhaps inevitable that much of the fiction of the period centered on such unions. They originated almost exclusively, the
fiction and inference tell us, in the man's affection for the woman,
since there was otherwise nothing to attract him to a marriage that
yielded neither important economic nor social benefit to him. The
woman, on the other hand, had much to gain from such a marriage:
economic support, a rise in social status, and perhaps even some
promise of long-term security. It was in her interest clearly to encourage and maintain her husband's love.
If a man took a secondary wife, the marriage was almost always
duolocal, since wives did not usually share houses and the husband
normally resided with his principal wife. There is no historical support for the well-known fictional case in The Tale of Genji where the
living quarters of three wives are found disposed on three sides of the
Cambridge Histories Online © Cambridge University Press, 2008
THE NOBLE FAMILY
139
husband's mansion, nor is there any undisputed historical case of
permanent duolocal marriage with a principal wife. A woman in a
duolocal marriage was typically inferior in status to her husband, and
it was no doubt as often the promise of an improved economic and
social situation as of love that induced her to accept the position of
a secondary wife. For the husband, on the other hand, the chief motive for the marriage was again likely to be affection, although family
politics or the desire for children also sometimes played a role.
The absence from literary and historical sources of a single, unequivocal case of virilocal residence among the Heian nobility, except in imperial marriages, is striking and noteworthy.16 We may assume that virilocal marriage did occur exceptionally - just as, for
example, duolocal marriage occurs exceptionally in modern Western
societies - but the circumstances giving rise to it must have been
very special indeed. The near absence of virilocal residence meant
that as a rule a noble wife never lived with her husband's parents.
Since children remained with the mother in case of divorce and
often with a maternal relative in the event of the mother's death, it
meant also that grandchildren seldom lived with their paternal
grandparents. A further consequence was that most men left their
parental homes at an early age, often while still children.
The matrimonial residence patterns of the Heian nobility produced, in combination with other characteristics of the marriage institutions and of the society, distinctive common patterns of personal relationships. The most important pattern emerged from the
uxorilocal marriages, where the chief potential family members were
the primary couple (the maternal grandparents), their married
daughters and the husbands and children thereof, and their unmarried children. Since the unmarried sons of the grandparents could
be expected soon to marry out of the household, and since the second and subsequent daughters were often established in marriages
elsewhere (a separate uxorilocal or neolocal residence), the longterm core of the house tended to be composed of maternal grandparents, a daughter of the grandparents, and the daughter's husband
and children, and it was between them that the strongest family ties
seem to have been found. The relationship between the maternal
grandparents and grandchildren was often particularly close, perhaps because the grandparents played an especially large role in rais16 The existence of such virilocal residences has been alleged by, among others, Sumi Toyo,
Zenkindai Nihon kazoku no kozo (Tokyo: Kobundo, 1983), but the cases cited in support of
that view all seem to be neolocal.
Cambridge Histories Online © Cambridge University Press, 2008
140
THE CAPITAL AND ITS SOCIETY
ing the children when the parents themselves were still very young,
as was often the case. Father and daughter were also often particularly close, the result, perhaps, of living together longer than any
other family members. The mother frequently died early in childbirth, leaving the father as the sole parent (except when he remarried and moved to his new wife's residence or to a neolocal establishment). And while a son married out at an early age, a daughter
married uxorilocally and might continue living with her father until
she was well into her adult years. The relations between sisters and
brothers also seem usually to have been close, as were those between
the husband and his wife's parents (the husband was, however, never
adopted by his in-laws, as happened in later Japanese history and
sometimes even today). On the other hand, the ties with paternal
grandparents, uncles, and aunts, who almost always lived elsewhere,
were usually fairly remote. Even those between father and son and
brother and brother seem to have been somewhat distant, the result
in each case very likely of separate residence or of only a short period of coresidence.
The particular pattern of personal relationships characteristic of
uxorilocal marriages varied somewhat with neolocal and duolocal
marriages, but the closeness of familial ties through, among, and to
females and the relative remoteness of the corresponding male connections remained basically unchanged. Those characteristics of
noble society had a pervasive influence on the period's history and
civilization (as seen most dramatically in the example of the Fujiwara regency), exercised partly through the control of infant and
child emperors by their maternal relatives, and perhaps also in the
flourishing of female letters, which depended in part on a recognition of feminine worth that must have been supported and encouraged by the women-centered relationships of the society.
A summary of Heian noble marriage residence rules draws particular attention to the social and economic position of the wife. The
general tenor of the marriage institutions and the nature of the resulting personal relationships within the family suggest a fairly elevated status for a principal wife vis-a-vis her husband, and that impression is strengthened and confirmed by several additional
considerations.
To begin with, the psychological position of a principal wife in an
orthodox, uxorilocal marriage must have been fairly strong. Instead
of moving to a strange house among the unfamiliar and possibly cold
or hostile relatives and servants of her husband's family, she reCambridge Histories Online © Cambridge University Press, 2008
THE NOBLE FAMILY
141
mained in her natal home. There she still very likely played in the
early years of her marriage as much or more the role of daughter as
of wife, and was surrounded by brothers, sisters, and faithful servants. Marriage meant for her no radical shift of physical or social
scene, and thus she escaped the severe psychological trauma that
was a common fate of Japanese brides in the virilocal marriages of
later periods. It was the husband who became the guest or intruder,
although even for him the transition was eased to some extent by the
period of duolocal residence that preceded his move to his bride's
house.
The wife's position was also buttressed by a degree of economic
autonomy. Under prevailing inheritance custom, a daughter shared
in the estate of her parents, often or regularly receiving the family
residence and also a share of the other property, a share that sometimes, at least, included almost the whole of the estate. That was a
logical, pragmatic arrangement, since it could be expected that a son
would be largely supported by his court income and that he would
live uxorilocally or in a neolocal residence supplied by his wife's family. A woman's control over the property she inherited was absolute
in the sense that she could sell or transfer it at will (it was not a lifetime holding that reverted elsewhere on the woman's death, as in
later periods). Because there are instances of marital residences that
passed down from mother to daughter through as many as four generations, it may be that real and other property was often inherited
in the maternal line. That point, however, is uncertain. Although female ownership and bequeathal of property is clearly established in
surviving documents, no case of an actual inheritance of a residence
and property by a noble daughter from her mother has been found.
It is possible that the seemingly matrilineal succession to houses was
at least in some cases actually accomplished through and at the discretion of the husband, into whose hands at some point the uxorilocal residence seems often to have passed, a development that negotiations leading to the marriage may conceivably have provided for.
The Heian noble wife also enjoyed certain other customary rights
that tended to give additional substance to the degree of her social
and economic autonomy. She retained her own clan name at marriage and was eventually buried with members of her natal clan (she
did not become, in other words, a member of her husband's clan).
She had and used the right of divorce, and she kept her children in
case of divorce or the death of her husband. Furthermore, she had a
limited measure of sexual freedom in the sense that although preCambridge Histories Online © Cambridge University Press, 2008
142
THE CAPITAL AND ITS SOCIETY
marital sex and adultery were not socially condoned in women, they
seem not to have been uncommon, and even if the relationship was
discovered, the woman was not ordinarily subjected to physical or
other punishment. If she was discovered in adultery, her husband
might divorce her, but that was not certain, and perhaps not even
customary.
Finally, the general nature of the surrounding society in which the
Heian noblewoman found herself provided relatively favorable conditions for the growth and exercise of her talents. The society was intensely urban, thoroughly civil, generally pacific and nonviolent, and
inclined heavily toward aesthetic, literary, and learned pursuits. The
conditions of society may not have been "feminine" in any essential
sense, but they were far more promising for women than those of,
for example, a feudal society geared for war, where the high value
placed on muscle and meanness led inevitably, no doubt, to harsher
conditions for women.
It is clear that the Heian noblewoman enjoyed a relatively high
and strong position in her society, in sharp contrast with the women
of feudal Japan. It is important, however, to stress that her position
remained distinctly and absolutely inferior to that of noblemen. She
did not occupy what were clearly the preeminent positions in the society: there were no reigning Japanese empresses during the Heian
period, no female government ministers, no female heads of Buddhist sects, no female chieftains of noble clans. The relative position
of men and women is well illustrated by the names commonly used
of women, names frequently drawn from their roles as daughters,
wives, or mothers (as, for instance, "the Mother of Michitsuna," the
name of a famous writer who was the mother of a totally undistinguished son); those, indeed, were the chief functions in life for most
women. Moreover, it must be admitted that the noblewoman's concept of herself was usually as an inferior being: unintelligent, emotional, incompetent, and dependent.
LIFE IN THE MANSION OF A GREAT NOBLE
The houses in which the Heian nobility married and lived were
highly perishable wooden structures that fire and war completely
obliterated centuries ago, leaving not a single physical trace of their
existence for the historian. But thanks to abundant literary and pictorial evidence and to scanty architectural survivals in religious institutions outside Kyoto, the chief features of the houses are well
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
143
known. Modern scholars have even been able to produce with some
exactitude floor and ground plans of several of the chief noble residences in the mid-Heian period, and scale models are found in more
than one Japanese museum. The plans and models are based, of
course, on a fair amount of guesswork and inference, but their degree of reliability appears to be generally high. It is unlikely that a
resurrected Heian courtier would have much difficulty in identifying
them as the species of house in which he had once lived.
The architectural style of the Heian noble mansion deriving ultimately, it is said, from China, has been known since the nineteenth
century as "dwelling hall construction" {shindenzukuri), taking its
name from the building, the "dwelling hall" (shinderi), where the
main inhabitant or inhabitants slept and lived. A fully developed establishment consisted of several buildings, whose construction was
much the same as that of the common Japanese domestic architecture of more recent times: rectangular, single-story, post-and-beam
structures with probably unpainted surfaces, gable or hip roofs, and
plank floors elevated several feet above the ground on posts. The
roofs were plank or, in the better establishments, shingled with cypress bark; tile seems not to have been used except on roof ridges.
Interior space in the major buildings was for the most part undivided by permanent partitions or walls, but the configurations of
post alignment, supplemented as occasion required by curtains,
blinds, and sliding or freestanding screens, created a large central
chamber (called the moya) surrounded on four sides by oblong "eave
chambers" (hisashi), beyond which, but still beneath the eaves, were
open verandas running the entire length and breadth of the building. The eave chambers were divided from the verandas by boardbacked latticework partitions that were usually opened or entirely removed during the day, leaving the occupants protected from the
elements only by standing paper screens, curtains, or bamboo
blinds. Living took place mostly on the plank floors, which were bare
except for cushions and movable mats, and furniture was minimal:
armrests (used by people seated on the floor), screens, oil-lamp
stands, an occasional cabinet or table, a curtained dais for sleeping,
charcoal braziers (the only artificial source of heat), and apparently
little else.17
Many ordinary nobles may have had to content themselves with
17 The recognized authority on shindenzukuri is Ota Seiroku, Shindenzukuri no kenkyu (Tokyo:
Yoshikawa kobunkan, 1987).
Cambridge Histories Online © Cambridge University Press, 2008
144
T H E
CAPITAL AND ITS SOCIETY
unelaborated dwelling halls on cramped sites in the back streets of
the city, and some may have been unable to manage even that. But
we know little of such houses, since it is only the grander mansions
of the senior nobles that the literary and historical sources describe
in any detail. Those establishments were distinguished by sites of
generous dimensions (usually up to about 3.5 acres) facing on major
avenues near the grounds of the imperial palace in the northeastern
part of the city, and by a multiplicity of linked and independent
structures that, together with their accompanying gardens and
courtyards, accommodated the ceremonial and ritual, as well as the
domestic, needs of government and clan leaders, and also produced
residences that must have been as singularly pleasing to the eye as
they were uncomfortable to the other senses.
Although there was considerable variation in detail among the
great noble mansions, a basic identity of plan and function allows a
description of one to serve as a guide to the others. For the present
purpose, which is to observe the noble family in its physical context, the well-studied Ononomiya residence of Fujiwara no Sanesuke (957-1046) is particularly suitable: Sanesuke was powerful and
wealthy enough to allow his family and residence to attain a maximum development, but he did not marry a daughter to an emperor,
and the family structure and the functions of the residence were not
complicated by the exceptional arrangements attendant on imperial
marriages. His case, it may be reasonably assumed, was fairly typical
of the senior nobility, except that he was wealthier than most, lived
to be nearly ninety, and was not remarkably successful in the fathering of children.18
Ononomiya, at the peak of its development during Sanesuke's
time, consisted of a main site roughly 400 feet square in the northeastern section of the city and additional adjacent pieces of property
of unknown dimensions across the streets that bounded the site on
four sides. It was a five- or ten-minute walk west from there to the
nearest gate of the imperial palace and only two or three minutes
longer in the opposite direction to Michinaga's famous Tsuchimikado mansion on the northeast.
The main site, which had come to Sanesuke from his grandfather
18 On Ononomiya and Sanesuke's family, see Yoshida Sanae, "Fujiwara no Sanesuke to
Ononomiya-tei: shindenzukuri ni kan-suru ichikosatsu," Nihon rekishi 350 (July 1977):
50-69, and "Fujiwara no Sanesuke no kazoku," Nihon rekishi 330 (November 1975): 69-85.
Both studies are based mainly on Sanesuke's diary, Shoyuki, in part 10, vols. 1—11 of Dai
Nihon kokiroku (Tokyo: Tokyo Daigaku Shiryo Hensanjo, 1959-86).
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
145
Oinomikado Avenue
North Gate
North
Wing
West
Wing
Northwest
Roofed
Hallway
North Opposed
House
SI
0
feet
50
East
Roofed
Hallway
13
Dwelling
Hall
West
Roofed
Hallway
East Wing
West Gate
East Gate
Wing
Chapel
Reizei Street
Figure 2.5. Plan of Ononomiya. Adapted from Yoshida Sanae,
"Fujiwara no Sanesuke to Ononomiya-tei," p. 51.
and adoptive father, Saneyori, was developed over a period of about
twenty-five years following a fire in 997 that destroyed most of the
original structures. Sanesuke seems to have lived in the residences of
his first two wives while they were alive, but within a few years of the
death of the second wife in 998 he took up permanent residence at
Ononomiya, never again formally marrying and moving to the residence of a principal wife.
The chief residential quarters, situated in the northern two-thirds
of the square site, consisted eventually of a central "dwelling hall"
and three adjacent "opposed houses" (tai no yd) on the east, west,
Cambridge Histories Online © Cambridge University Press, 2008
146
THE CAPITAL AND ITS SOCIETY
and north, each connected with the dwelling hall across a space of
about 40 or 45 feet by one, or, as in the western house, by two roofed
hallways (watadono). Steps descended to the main courtyard from
the southern of the two hallways leading to the western house, suggesting that the hallway there was open on the sides. The rectangular design of the opposed houses was the same as that of the dwelling
hall, but each also included one (in the case of the northern house)
or two (the eastern and western houses) rectangular wings ("corridors," ro) extending from corners of the structures at right angles.
Since the houses are assumed to have faced the dwelling hall, as their
name suggests, the wings are believed to have been on the sides of
the houses away from the central structure.
The dwelling hall itself faced south toward a courtyard formed by
the southern wings (ro) of the eastern and western houses. One or
more apricot trees (ume, commonly misidentified as plum trees)
grew in the courtyard, and south of there, occupying about a third
of the entire site, were the man-made ponds and hills of an extensive
garden. The largest of the ponds, broad enough to contain an island
on which a banquet for a sizable group of people could be held, was
directly south of the courtyard; the two smaller ponds seem to have
been west and south of there, near the corner of the property. Two
hills stood at the southern edge of the site, rising behind the ponds
and providing for observers in the residential quarters displays of
foliage and blossoms across the intervening water. The ponds were
probably fed both by an irrigation stream brought in beneath the
roofed hallway connecting the eastern opposed house to the dwelling hall and by a spring southeast of the largest pond.
The garden contained two Buddhist chapels, the larger one at the
foot of the hill next to the spring southeast of the main pond, and the
smaller apparently between the two ponds in the southwest corner.
A stable and a building used for storage stood on the southwestern
edge of the site, just inside a wall that encircled the entire property.
There were three gates in the encircling wall, one each on the east,
west, and north. The west gate, a relatively large, roofed passageway
of the kind described as quadripedal (yotsuashi), was the main formal entrance to the site. Inside the west gate, at the southern end of
the western house and its southern corridor wing, was another gate
(the "inner gate," chumon) leading into the courtyard south of the
dwelling hall.
Such were the known major structures and features of the main
Ononomiya site as they existed in about the year 1030. It was a large
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
147
and handsome establishment, laid out and executed with a sophistication of taste and sensibility uncommon, surely, in the world of medieval domestic architecture. A source from the following century
claimed that the mansion was kept in such beautiful repair that there
was never a day when seven or eight carpenters were not busy there:
the two places where one was always certain to hear the sound of
adzes, the narrator in the source asserts, were the Todaiji Temple at
Nara and the Ononomiya mansion. But it must be kept in mind that
the spaciousness of the site did not necessarily result in commensurately spacious living quarters for the inhabitants. The overall dimensions of the largest structure measured perhaps 80 by 50 feet;
the opposed houses on the east and west were probably about 80 by
40 feet; and the northern opposed house approximately 70 by 40
feet. (All of those measurements are rough estimates.) The total
floor space in the main buildings thus amounted to something over
13,000 square feet, and the hallways and corridors wings may have
added another 5,000 to 6,000, a grand total of nearly 19,000 square
feet of usable space.
Lower-ranking nobles lived in much less spacious surroundings
than those of Sanesuke, of course. A will from the end of the
eleventh century, for example, seems to show that the main residence of an otherwise unknown but obviously wealthy man named
Oe no Kiminaka was crowded onto eight house lots of one of his
three large properties in Heian (each listed at one cho, or about 3.5
acres). The property on which the main residence stood also included in its other sections a library, which would have been especially appropriate for a member of the learned Oe family, and a
chapel.1"
The physical traces of another more humble, probably noble, establishment were revealed at the time of the reconstruction of a
Kyoto high school in 1979. Located in the northeastern corner of
Heian, the site showed in excavation the outline of a rectangular
main hall about 21 by 16 meters, with two smaller buildings on either side and another at the rear. The plan seems to represent the
shindenzukuri design at an incipient stage.20
The Ononomiya compound, in addition to its chief half-dozen or
so inhabitants, also housed, temporarily or permanently, ladies-inwaiting, servants, and miscellaneous workers and hangers-on, so
19 Yoshioka, "Kizoku shakai no seijuku," pp. 134-36.
20 Yoshioka, "Kizoku shakai no seijuku," p. 137.
Cambridge Histories Online © Cambridge University Press, 2008
148
THE CAPITAL AND ITS SOCIETY
that the total twenty-four-hour population of the site may have been
counted in three figures. There was, furthermore, a daily stream of
visitors: kinsmen, court officials, and religious, who came at all
hours of the day and night, often accompanied by their own companions and servants. Between the resident population and the visitors, the chambers, corridors, and wings of Ononomiya must have
been constantly alive with what to modern sensibilities might seem
veritable hordes of people.
The relatively high density of site population and the openness of
interior space would have combined to produce a public style of living. Sleeping or waking, the individual was rarely if ever very far removed from companions, and the occasions when one was completely alone were probably exceptional indeed. Darkness, sleep, and
curtains seem often to have been all that protected even a couple's
sexual relations from kin and attendants in the same room. To be
alone, or nearly so, at night in a dark building was enough to make
even the bravest man feel the sharp talons of goblins snatching at
him, or to frighten a timid girl quite literally to death.
Large numbers of people living together in a relatively limited
space with little provision or opportunity for privacy may have created two partially contradictory tendencies in the personal relationships of family members. On the one hand, the frequency and intimacy of association may have helped strengthen certain ties,
especially those with personal attendants, who sometimes were also
family kin or related through fosterage. (The child of a wet nurse
might become a personal attendant of its mother's nursling.) Sanesuke's diary is too impersonal a document to provide examples of
such relationships, but contemporary fiction contains frequent instances suggesting that the degree of intimacy (if not its precise nature) approached what a modern observer might expect to find between congenial siblings. On the other hand, the multifariousness of
relationships within a very large household, the strength of particular ties with household members outside the family nucleus, and the
absence of a special nourishing privacy for relationships among nuclear family members would have presumably diffused and somewhat weakened affective relationships between nuclear members
themselves. It may be noteworthy that in a society where genealogy
and kinship played such decisive roles there was not a single word or
customary phase in the spoken language that designated the nuclear
(or elementary) family (parents and their unmarried children).
Nuclear family relationships may have been further weakened by
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
149
the practice of wet-nursing, which appears to have been nearly universal among at least the more exalted noble households; few if any
ladies of distinction ever suckled their own children even for the
briefest period. The practice differed radically, however, from a form
of it found at one time in Europe, where a newborn infant of the
wealthier classes might be sent away to a poor, usually peasant,
household for one or two years of nursing by a woman whose poverty, ignorance, and inattention seems frequently enough to have led
to the infant's death. The Heian custom was to select one or more
wet nurses for an infant from among women of relatively good birth
(often from Fifth-Rank families) and, if she was not already in service at the house, to bring the nurse (or nurses) there, where she
might then remain more or less indefinitely, even after her charge
had been weaned. The natural mother was not physically separated
from her child, but her relationship with it would have usually been
weakened, one must suppose, insofar as she delegated nursing and
care to another woman. There was no doubt a good deal of individual variation in the degree of maternal responsibility delegated to a
nurse, ranging from simply a year or two of breast-feeding on up to
what may have amounted at times to a nearly complete surrender of
maternal duties, but the overall tendency would have been toward
some loosening, at least, of the link between a child and its natural
mother.
Nuclear family ties were similarly affected by the presence in the
household of other close relatives, regularly a daughter's husband
(the son-in-law) and her children (the grandchildren), but also occasionally, according to individual circumstances, a brother or sister
of one of the parents. In such cases, however, the effect on nuclear
relationships may have been slight, since the nonnuclear kin are
known sometimes to have lived in subhouseholds in detached or
semidetached buildings. For example, when Sanesuke's daughter
Chifuru married in 1029, she and her husband lived together in the
eastern opposed house at Ononomiya, which had its own household
offices and kitchen in the corridor wings and was probably also
equipped to function otherwise as at least a semiindependent establishment. When Sanesuke's sister moved to Ononomiya in 1005, she
was not housed at the main site but in the Western Residence, which
lay across the street bounding the site on the west. That residence
also presumably functioned more or less independently of the residential quarters of Sanesuke's immediate family.
The ties of the nuclear family could be further loosened by the
Cambridge Histories Online © Cambridge University Press, 2008
150
THE CAPITAL AND ITS SOCIETY
practice of polygyny. Sanesuke, like many or most Heian noblemen,
seems to have been formally married to only one wife at a time, but
some, like Michinaga, had two wives, and it was common for men to
have continuing relations with women other than their principal and
secondary wives, either within or outside their own households. All
such relations, but especially with a secondary wife in a separate
household, would have reduced the frequency of association between
a man and his principal wife and between him and their children,
leading ordinarily to some weakening of nuclear family relationships.
Early mortality had the same general effect, limiting the lifespan of
a nuclear family and reducing the measure of common history and
association that lies at the base of personal relationships. In Heian
noble society, where social violence was relatively rare and warfare
almost unknown, the chief danger to parental life apart from disease
was probably childbirth, and it was therefore the link with the wife
and the mother that was most likely to be severed by early death. Although statistics are lacking, the frequency with which death in, or
shortly following, childbirth appears in the literature and sources of
the period suggests that at the very least a large proportion of all
marriages ended in that fashion. This, taken together with the undoubtedly high mortality rate from disease, may mean that it was the
exception for a husband and wife in their first marriage to live beyond their twenties together. For instance, Sanesuke's first marriage
ended with the death of his wife in 986 shortly after the birth of their
first child, in the twelfth or thirteenth year of the marriage, when
Sanesuke was twenty-nine. Had the wife borne a child earlier, there
is an excellent chance that the marriage would have lasted no longer
than Sanesuke's second marriage, which was brought to an end in its
fifth year by the death of his wife. Sanesuke also had two children by
a woman who may have been in the service of his sister; she, too, died
shortly after the birth of her second infant, which had died immediately after birth.
Such personal histories may have been more nearly the rule than
the exception, and a child probably had to count itself fortunate if it
reached adolescence with even one parent alive. Since the death of
young women in childbirth was so common as to be almost normal,
and since childhood mortality was especially high, the Heian noble
child also often found itself with neither a natural mother nor any
full siblings. Many children, and perhaps a majority, lived for longer
or shorter periods of their lives in foster or adoptive families, or became stepchildren in the household of a remarried surviving parent.
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
It is entirely possible that the nuclear family as a rule was even less
stable among the Heian nobility than it is in present-day societies,
where the frequency of divorce may be at least partly a function of
improved physical health and increased longevity. If stepchildren
and adopted children are excluded, it may also be true that the
Heian nuclear family was little larger than its counterpart today in
Japan or the West: few Heian couples seem to have had more than
two or three children who lived past infancy.
Finally, ties within the family were probably also affected by the
tendency of the Heian noble to change residences often. The divorce
of residence from income and status, the relative simplicity and
ephemerality of the noble house, and the prevailing pattern of marital residence led to a high degree of residential mobility within the
narrow confines of the capital city. A great nobleman like Michinaga
or Sanesuke was likely to live in a succession of houses as one marriage followed another, new houses were acquired by purchase or
transfer, and fire destroyed the old. Michinaga, for instance, owned
or controlled seven houses in the capital, three acquired through his
marital connections and four otherwise, and he had villas at Katsura
(near the present Katsura Detached Palace) and Uji (on the site of
the Byodoin), in most of which he is known to have lived at one time
or another. Fire, storm, and family vicissitude probably also made it
rare even for a woman to reach old age in the house in which she was
born and married. Such absences were occasioned by the interdictions of yin-yang lore,21 by the requirements of official duty at court,
by ceremonial or ritual functions at other noble households, by childbirth and its attendant ritual pollution, and sometimes simply by inclination. Although it is true that the range of physical movement was
restricted, usually involving no more than a few hundred yards of
travel and often simply a shift from one structure to another on the
same property, the frequency of moves and absences from home and
of new construction presumably contributed to at least some diffusion of personal relationships outside the household and a corresponding loosening within. That process would not, however, have affected the two sexes equally. Since the woman was far less likely to
move or to absent herself from home than the man, the overall effect
must have been to reinforce the uxorilocal nature of many households by centering the closest, most intense personal ties on women.
21 See Bernard Frank, Kata-imi et kata-tagae: Etude sur les interdits de direction a I'epoque Heian,
Bulletin de la Maison Franco-Japonaise, Nouvelle Serie, Tome 5, 2-4 (Tokyo and Paris,
1958)Cambridge Histories Online © Cambridge University Press, 2008
152
THE CAPITAL AND ITS SOCIETY
The foregoing suggests there were circumstances in Heian noble society tending toward a diffusion of affective ties away from the nuclear family and a weakening of relationships within it. There is also
at least one small piece of objective evidence that may confirm the
existence of that tendency with respect to children. The death of a
Heian noble adult was marked by a considerable amount of ceremony and ritual leading to and following the cremation or interment
of the body. However, when children under eight calendar years of
age died, parents were advised by experts in such matters that the
proper course was simply to abandon the body in the open, which
seems to have been a frequent practice among commoners even in
the case of adults. References in contemporary sources to the abandonment of children's corpses are frequent enough to indicate that
the practice was commonly observed. The high rate of childhood
mortality undoubtedly argued in favor of a simple, inexpensive
method of disposing of the corpses, and in an age accustomed to the
sights and smells of death and mostly unacquainted with the origins
of disease, abandonment in the open may have seemed the simplest
and most practical means available. But the custom seems to imply,
nevertheless, a certain callousness toward, or distance from, children, a view of them as somehow less important or less human than
the adults who were sent off with elaborate obsequies and mourning.
It seems legitimate to conclude that, generally speaking, the affective ties of the Heian noble with members of his immediate family
were neither so strong nor so enduring as those among members of
a reasonably successful modern family in Japan or the West, and that
those ties were therefore more easily severed or shared than they usually are nowadays. The implications were not usually quite the same
as they would be for modern family members when, for example, a
Heian spouse ended a marriage by taking holy orders or by divorce,
or when a husband took a secondary wife, or when a son left his family to live at his bride's house, or when a child was given up for adoption. The break may often have been more easily accomplished at
that time than now, the sharing more willingly undertaken.
One should not infer from this, of course, that nuclear family relations were unimportant in the practical or emotional life of the
Heian noble. Although a noble husband and wife seem normally to
have been less closely bound by association and affection to each
other and to their children than we like to think the modern married
couple should be, there was usually no other set of relationships that
was of comparable strength. For most people of that age and class
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
153
the deepest and most durable emotional ties were still undoubtedly
within the confines of the immediate family. In individual cases,
moreover, the bond between husband and wife or between parent
and child was as strong as any such relationship commonly is in the
modern family. Or so, at least, the contemporary literature tells us
and some of the historical sources suggest. We may perhaps infer
from Sanesuke's diary, for instance, that he was strongly attached to
his first wife, whom he married when he was sixteen and lived with
for twelve or thirteen years. Although the diary, as usual, says nothing about the author's personal feelings at the time of his wife's
death in 986, it does reveal that he continued to hold annual mourning services on the date of her death for many years thereafter, evidence perhaps of an uncommon degree of devotion (or fear of the
spirit of the deceased?). Sanesuke was even more deeply attached to
a daughter he had by a woman who had been in the service of his
second wife. He was in his mid-fifties when the daughter, Chifuru,
was born, childless except for adopted children and a clerical son in
whom he had never evinced much interest, and he seems to have
poured all of his thwarted parental affection into his daughter's care
and upbringing. He kept her with him constantly, took her on outings, catered seemingly to her every whim, reconstructed and refurbished the eastern opposed house at Ononomiya for her marriage,
and willed both Ononomiya and most of the rest of his estate to her.
There is no indication in any of this that he ever considered using
Chifuru to further his own political ambitions at court. Her mother's
relatively low birth may have made hopes for an imperial marriage
impractical in any case, and for a man in Sanesuke's high position
(he was Minister of the Right during most of her life), there was no
other marital alliance that could have been of much interest to him.
Unhappily, he outlived that daughter, too: she died at about the age
of twenty-seven, when he was eighty-one, leaving behind a two-yearold daughter who eventually inherited Ononomiya.
The literary and quasi-historical sources for the period frequently
describe domestic scenes that would entirely agree with what can be
seen in, and inferred from, Sanesuke's diary: a husband and wife in
free and affectionate converse about their everyday lives; a fond father holding his infant daughter and being amused when his clothes
are wet by her; a father playing horse with his little son; a mother
moved to tears by the beauty of her children; sisters assisting each
other in the conduct of love affairs; or the searing grief caused by a
death in the family, including that of an infant or small child. The afCambridge Histories Online © Cambridge University Press, 2008
154
T H E
CAPITAL AND ITS SOCIETY
fective ties between parent and daughter seem to have been especially strong, and Sanesuke's devotion to Chifuru may also in that
respect be regarded as typical of noble family relationships.
The ways in which space was employed at Ononomiya were also inevitably related to human relations there. Each area of the establishment tended to be associated with particular functions, although the
functional definitions of space were neither so detailed nor so strictly
applied as one might find in the great houses of the wealthy in recent
times, where many or most rooms may be reserved individually for
some one particular activity or purpose: dining, breakfasting, gaming, dancing, sleeping, reading, and so on. At Ononomiya, the main
functional divisions had to do with whether space was chiefly employed for everyday private living or for the more public and ceremonial or ritual aspects of life. At the same time, in limited areas of
the establishment, space was also defined by fairly specific functional criteria. The division of space into public and private parts
mostly concerned the principal family members; the more specific
functional divisions mainly involved the household staff.
The greatest part of the main site at Ononomiya was reserved for
the use of Sanesuke and the members of his family. The space allotted for that purpose consisted of the dwelling hall, all of the opposed
houses, the corridor wing extending south from the western opposed
house, and one or both of the hallways connecting that house with
the dwelling hall.
The central chamber and the northern and eastern eave chambers
of the dwelling hall, together with the opposed houses on the east
and north, were used by family members for everyday living, but beyond that minimal definition the space was not, for the most part,
further differentiated by function: the inhabitants slept, ate, and pursued most of the activities of their waking hours in the same undifferentiated areas. Latrine arrangements of the period are not well
understood, but since there were servants responsible for the removal and cleaning of pots of human waste, perhaps those needs,
too, were often or usually met in the regular living quarters. The central chamber of Sanesuke's dwelling hall included a walled room that
was probably used for storage or possibly as sleeping quarters, but
commonly in noble residences, the chief inhabitant or inhabitants
slept on a slightly raised movable platform enclosed simply by curtains and blinds suspended on a lightly constructed frame, and the
attendant ladies-in-waiting slept nearby on the floor. In the normal
Cambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
155
course of affairs, the only visitors to the living quarters seem to have
been close relatives, one's official subordinates, and household retainers, attendants, and servants. However, in the case of a highly
placed man like Sanesuke, whose household was large and to whom
almost everyone else in the government was subordinate, that group
was very numerous.
Although Sanesuke was at one with his modestly reticent age in
not mentioning the details of his sexual life, it appears that none of
his mistresses or secondary wives ever shared the dwelling hall with
him as a permanent resident after he took up residence at Ononomiya. The mother of his daughter Chifuru may have lived in the
northern opposed house, which is thought to have been a common
location for the quarters of noble wives, who were therefore generally removed from the scene of social intercourse centering on their
husbands in the dwelling hall. Such women, moreover, had far less
occasion than their husbands did to venture forth from their houses.
The chief centers of social intercourse among Heian nobles were the
imperial court and the mansions of the leading court figures, and
since noble wives did not normally participate in the life of the court
or visit other residences, they remained perforce for the most part in
their own living quarters among their children, ladies-in-waiting, and
servants. An occasional pilgrimage to a Buddhist temple or a Shinto
shrine, an outing to witness a festival procession through the streets
of the capital, transfer to another residence for childbirth, a visit to a
parent or sister, and in rare cases a trip to the imperial court at the
time of a daughter's marriage to an emperor - such limited occasions
constituted the bulk of the usual noblewoman's experience of the
outer world. The eleventh-century tale that describes a noble girl
alone in the streets of Kyoto and unable to find her way to her own
nearby house may not be a wholly fanciful representation of contemporary female knowledge of the world outside the home gates.
As a small child, Sanesuke's daughter Chifuru seems to have occupied the eastern eave chamber of the dwelling hall (she lived, in
other words, in the same building with her father), but when she
grew up, she moved to the eastern opposed house and after her marriage continued living there and in one of the house's corridor wings
with her husband and daughter. How living space was shared in the
house and wing is unknown, but the general lack of functional differentiation remained no doubt very much the same as in the
dwelling hall. A common development would have been the removal
of Chifuru's family to the dwelling hall upon the death, retirement,
Cambridge Histories Online © Cambridge University Press, 2008
156
THE CAPITAL AND ITS SOCIETY
or entry into holy orders of Sanesuke, but that was forestalled by
Chifuru's early demise and Sanesuke's long life. Chifuru's husband
returned to his father's house for a while after her death but came
back to live again in the eastern house with his and Chifuru's small
daughter, who had probably remained there during his absence. He
had been only thirteen or so when he married Chifuru in about
1029, five years younger than his wife, and he had lived at Ononomiya with her for nearly ten years. How much longer he remained at
the mansion is unknown, but circumstantial evidence suggests that
his residence may have been more or less indefinite.
The northwest corner of the northern opposed house is known to
have been the residence of one of Sanesuke's five adopted sons, and
others of those five may have also lived there until they married. As
already noted, the northern house is thought to have commonly
been occupied by a man's principal wife and by her small children,
but circumstances frequently altered that arrangement. Sanesuke
himself, for instance, lived for a while in the northern house at
Ononomiya while his dwelling hall was under construction.
The remainder of the family space at the main site (i.e., the southern and western eave chambers of the dwelling hall, and the western
opposed house and its southern wing and hallway) appears to have
been used chiefly for formal occasions, such as the reception of an
imperial envoy or one of the many rites de passage that marked the
life of a great noble family. The chief ceremonial entrance to the estate was the western gate, the side nearest the imperial palace, and it
may have been because of that geographical circumstance that the
western side of noble houses generally was often the place where receptions were held and ceremonies conducted, most such houses
having been, like Ononomiya, located east of the palace. On very
great occasions, such as the visit of an emperor, the marriage of a
daughter, or the celebration of a major Buddhist ritual, the central
chamber of the dwelling hall might also be used. At Ononomiya, the
rites and ceremonies held in the public space were not only for
members of the immediate family but for more distant kin as well.
It was in the corridor wings (ro) of the opposed houses that the
function of space seems to have been given its narrowest definition.
There were located the offices and work areas of the household
staff: the attendants' office (saburaidokoro), the escorts' office (zuijindokoro), the administrative office (mandokoro), the repairs office
(shuridokoro), the servants' office (zoshikidokord), the pages' office
(kodoneridokoro), the kitchens (daibandokoro), and the pantry (zenCambridge Histories Online © Cambridge University Press, 2008
LIFE IN THE MANSION
157
sho). The eastern opposed house had its own separate kitchen and
attendants' offices, and probably some of the other offices as well. If
Sanesuke had had a principal wife living in the northern opposed
house, she, too, might have had some of her own separate offices.
Sanesuke tells us little about the roofed hallways (zvatadono) connecting his dwelling hall to the opposed houses, but if their use was
similar to that of hallways in other noble mansions, they functioned
not only as passageways between the buildings but also, suitably outfitted with screens and curtains, as the living quarters of the mansion's ladies-in-waiting.
The ponds and gardens occupying the southern part of the property seem to have been designed largely for the pleasures of contemplation and strolling. Sanesuke, like others of his noble contemporaries, also personally participated in the layout, construction, and
upkeep of his garden, and he frequently took his ease there, especially, on hot days, near the spring at the southeastern corner of the
largest pond. The garden also occasionally became an extension of
the formal space at Ononomiya, serving as the site for a banquet or
some other type of entertainment. Sanesuke, to judge by his diary,
seems to have been little interested in the more vigorous kinds of
amusements, but at other mansions of the aristocracy the courtyard
immediately south of the dwelling hall was sometimes the site of
cockfighting or kickball (ketnari). At Michinaga's establishment, for
instance, part of the garden was given over to a horseracing course.
The Buddhist chapels in the garden were devoted, of course, to
religious purposes, but not exclusively so. The larger chapel near the
spring had quarters for the chapel monks in a corridor wing and had
its own kitchen and bathing facilities, functioning thus independently of the main buildings. Perhaps because of the nearby spring
and the pleasant surroundings, Sanesuke spent considerable time at
the chapel, using it as an extension of his regular living space and
sometimes taking his meals there. It is possible that the building also
served as a library. That would explain the absence of any mention
of a library in Sanesuke's diary, although he was a learned man, and
men of far lesser means than he managed to maintain a separate library building in their gardens.
A twelfth-century source gives particular attention to the larger
Buddhist chapel in a description of Ononomiya narrated by a fictional contemporary of Sanesuke's:
The mansion [Sanesuke] has erected is a splendid sight. Wings, a main hall,
and galleries are common enough, but he also has a Buddha Hall to the
Cambridge Histories Online © Cambridge University Press, 2008
158
THE CAPITAL AND ITS SOCIETY
southeast, three bays long on each of its four sides, with monks' quarters in
all the corridors. The monks' bathhouse is furnished with two huge threelegged cauldrons, plastered with earth, under which fires are lighted every
day. In the Buddha Hall itself, there are innumerable golden images, and
the receptacles in front of them always contain thirty koku [about sixty
bushels] of rice. People walk to the Buddha Hall from the front lake, following a path through a great park laid out to resemble a wild meadow,
which is filled with trees and plants chosen for their seasonal flowers and
autumn color - or else they can row across the lake. Those are the only ways
of approaching it. Every one of the monks is either a distinguished scholar,
a special sutra chanter, or an expert in the mystic Shingon rites. Sanesuke
gives them summer and winter vestments and sustenance allowances, and
tells them to pray for the extinction of his sins, the growth of his virtues,
and the safety of Her Ladyship his daughter.22
Although relatively little is known about the subsidiary Ononomiya properties across the streets bounding the main site on four
sides, they seem to have had two distinct uses. The properties on the
west and north - referred to sometimes by Sanesuke as the Western
Hall (nishidono) and the Northern Residence (hokutaku) - seem to
have been reserved for use by members of the family as occasion required. Sanesuke himself and Chifuru and her husband made some
use of the Western Hall, but it seems to have been occupied chiefly
by, first, Sanesuke's sister, then by a nephew (an elder brother's son),
after whose marriage it may have been used by one or more of Sanesuke's five adopted sons, all of whom were sons or grandsons of his
two elder brothers. All that is known about the Northern Residence
is that Sanesuke ceded it to his adopted son Sukehira. The properties on the east and south contained, on the other hand, dwellings of
household staff members and perhaps also the storehouses, kitchens,
and workshops that seem commonly to have been grouped together
in great mansions in a separate area of their own called the "storehouse row" {mikuramachi).The storehouses there and elsewhere in a
noble establishment contained both the everyday necessities - food,
cloth, clothing, paper, dyestuffs, household furniture, dishes, utensils, and so forth - and also various luxury items and treasures: aromatics, precious metals, religious statuary and paintings, Buddhist
scriptures, medicines, ritual ware and objects, heirlooms, musical instruments, and the like. Workshops known to have been associated
with noble mansions included those of seamsters, hatters, pictorial
artists, lacquerworkers, carpenters, joiners, metalworkers, and metal
22 Helen Craig McCulloughj trans., Okagami, The Great Mirror: Fujiwara Michinaga and His
Times (Princeton, N.J.: Princeton University Press, 1980), p. 110.
Cambridge Histories Online © Cambridge University Press, 2008
OFFICIALDOM AND ITS FUNCTIONS
159
casters, but not all of those were necessarily staffed with full-time artisans. Metal casters, for example, seem to have been called in only
as the need arose.
OFFICIALDOM AND ITS FUNCTIONS
Although the great nobles in Sanesuke's elevated sphere of society the Ministers of State (daijiri), the Counselors (nagori), and the Consultants (sangi) - were the chief makers of governmental decisions
and often prided themselves on their detailed knowledge of court
procedures, ceremonies, and rituals, it was the nobles whose careers
culminated at the Fourth or Fifth Rank who provided most of the
workaday administrative direction of the governmental offices and
much of the special knowledge and skills that fueled their operations.
They were the working (as well, usually, as the titular) heads of the
eight ministries that directed the activities of all central organs of the
government under the Council of State. They were, as well, the chief
officers in a host of other key administrative or technical offices and
bureaus, including those responsible for the administration of the
capital; the reception of foreign embassies; the computing, budgeting, and disbursement of government revenues; the construction and
repair of public buildings; the supervision of the Academy of Chinese
learning; divination, purification, and other matters relating to yinyang arts; the management of Buddhist and Shinto affairs and of
imperial mortuary matters; the maintenance of the imperial library;
the teaching and performance of court music and dance; the direction of the hundreds of Imperial Attendants who saw to the domestic and personal needs of the court; house- and groundskeeping at
the palace; medical treatment and the preparation of medicines; the
supply of furniture and other crafted items to the court; and the provision of food and drink for official banquets. Their number also included the secretariat or principal staff for the Council of State and
in the Chamberlains' Office; they held other pivotal posts, professional and administrative, in the various ministries, offices, and bureaus; and they sometimes attended in person on the emperor. It is
perhaps not too much to say that most of the day-to-day practical
work of the court and its government fell in the first instance on their
shoulders.
During the first several decades of the Heian period, the roughly 95
percent of the capital's population that belonged to neither the
Cambridge Histories Online © Cambridge University Press, 2008
l6o
THE CAPITAL AND ITS SOCIETY
princely nor the noble ranks was, nevertheless, largely official, and
perhaps overwhelmingly so. Those were the people who occupied at
the peaks of their careers one of the court ranks below the Sixth or
worked in menial government jobs without court rank. The higher
ranks of that large group included the heads of suboffices in charge
of such matters as the imperial clan register, the East and West
Markets, the two government prisons, the palace weavers, and
various domestic or housekeeping tasks concerned with the emperor, the crown prince, or the palace (i.e., meals for the emperor
or crown prince, drinking water, sake making, women servants in
the imperial residential compound, the crown prince's stable and
equestrian equipage, etc.). The middling and lower ranks held secretarial and subordinate administrative positions in the ministries
or the higher offices and bureaus, and they shared with the higher
ranks the many professional, technical, or skilled-labor posts that
supported government operations and palace life. They were teachers of the various subjects at the Academy of Chinese learning
(Chinese letters, Confucian classics, Chinese pronunciation, calligraphy, law, and mathematics), physicians, yin-yang practitioners,
acupuncturists, astrologers, calendrical specialists, herbalists, healers by incantation, masseurs, veterinarians, musicians and dancers,
painters, dyers, weavers of fancy silks, investigators and judges of
crime, disbursing officers, storekeepers, and so forth. Below them
and under the control and supervision of diverse superiors was the
army of servants, guards, armed men, attendants, and laborers who
supplied much of the skilled manual work in the palace and its organs, kept the watches and walked the patrols, and performed all
the menial tasks. Most of the laborers, guards, and armed men
seem to have been levied from the general population, but the servants and attendants were regularly recruited from among the children of officials.
By the mid-Heian period, however, a large part of the official class
had disappeared; its income and duties had either vanished in the
contraction of the statutory regime or been absorbed by superior offices. The Seventh and lower ranks seem rarely, if ever, to have been
used anymore, and many offices either ceased to exist or remained
as little more than paper organizations. Government workshops
closed or curtailed operations, and the great number of menials, laborers, and guards that had once populated the palace seem to have
been sharply reduced, leaving many palace buildings outside the residential compound uncared for and unprotected. The absence of
Cambridge Histories Online © Cambridge University Press, 2008
THE CITY'S ECONOMY
161
guards at most of the palace gates (especially at night) gave free run
of the grounds to arsonists and robbers, and even the personal safety
of the courtiers was no longer assured.
The shrinkage of the official class probably resulted in the departure from the capital of large numbers of people, especially those
who had been brought in on labor levies, most of whom presumably
returned to their rural homes and were not replaced, but also some
of whom were descendants of court-rank officials who failed to obtain substitute employment in the city and found themselves forced
to seek their fortunes elsewhere. Key segments of the class remained,
however, and contributed importantly to the diversification of the
general, non-noble and nonofficial population, which toward the
date of the city's founding had been small and also, in contrast with
the highly diversified noble and official population, fairly uniform in
composition. The largest part of the early general population may
have consisted of the swarms of servants and laborers employed by
the great princely houses, the nobility, and the more elevated members of the official class. A few other elements can also be identified
or surmised to have existed, including a scattering of artisans and
skilled laborers, who for the most part do not appear in the sources
but must have been available to some extent for private construction
and other similar needs in the city (some or most may have been
moonlighters from the palace); a small group of merchants, to whom
we shall return; and a priestly class, which was very small within the
city proper, confined mostly to the Buddhist monks at the East and
West Temples, but considerably larger if the many Shinto shrines and
the growing number of Buddhist temples in the immediate neighborhood of Heian are taken into account.
THE CITY S ECONOMY
The relative uniformity of the general population in early Heian was
a product of the city's peculiar economy which, in its nearly complete dependence on the court and the government, was more like
that of a large military base than of a fully evolved city. As a purely
political center of administration, Heian was almost exclusively a
city of consumers, drawing into itself the produce, manufactures,
and labor of the countryside; sending back little in return except orders, officials, and superfluous members of its own society; and producing solely (or virtually so) for its own uses.
The city's basic source of income was governmental tax revenues
Cambridge Histories Online © Cambridge University Press, 2008
162
THE CAPITAL AND ITS SOCIETY
in kind and labor levies, which supplied in themselves through the
individual incomes of princely personages, nobles, and officials most
of the goods and services required by the population. A description
of the early city economy becomes therefore primarily an account of
the distribution of income.
Under the statutory system, all holders of rank and office received
varying amounts of goods and services according to their particular
ranks and offices. The system was complex in detail and underwent
numerous modifications, but in outline there were three primary
categories of payment. For the noble ranks and offices and for certain specialists (physicians, yin-yang experts, teachers at the Academy of Chinese learning, etc.), the tax revenues from specified
amounts of land or from specified numbers of households were assigned to the holders of the ranks and offices. Those revenues consisted of handicraft items (especially cloth) as well as rice and other
food products. A second type of payment, again for the noble and
princely ranks only, consisted of specified numbers of household officials, servants, and guards, provided in part through the government's labor levies. Finally, all levels of officials received outright
stipends paid from the government's treasuries in cloth, iron implements, salt, rice, other foods, and at times limited amounts of cash.
(There were twelve small mintings of primarily copper coins in
Japan, the last in 958; by the twelfth century, the chief currency was
Chinese copper cash.)
The income received by great nobles and the leading princely personages was very large, supplemented sometimes by special grants of
land from the emperor, by privately acquired landholdings, and, in
the tenth century and after, also by commended shoen, the revenues
from all of which would similarly have been in goods and services.
The ordinary noble was also amply rewarded, receiving enough, it is
thought, to support comfortably many or most of the families in his
clan. The large incomes of the nobility and the princely houses were
used in substantial part to employ the great numbers of servants, laborers, nurses, guards, estate managers, and so on, that high social
position both required and made possible. The pay received by such
dependents would, of course, have also usually been in goods.
Most of the early population of Heian therefore received many or
most of the necessities of life directly in their incomes and not
through commerce. Residents at that time also held capitatum
tillages (kubunderi)fromwhich they derived income in kind. The income of many at the lower levels of the economy may have been furCambridge Histories Online © Cambridge University Press, 2008
THE CITY'S ECONOMY
163
ther supplemented by small vegetable gardens and byfishingand
hunting on the outskirts of the capital. Some residents of the city
also cultivated rice and barley fields immediately adjacent to the city
or on vacant city land.
Although the early economy of Heian functioned primarily on income in kind or in labor, not all needs could be met for all people
through their incomes alone, and a certain amount of commerce was
thus provided for at the East and West Markets. The only places in
the capital where trading was permitted, the markets were closely
supervised and regulated by a special organ of the city administration, the Market Office, which fixed prices, weights, and measurements, determined the types of goods to be sold in the markets, and
controlled merchant access, only those merchants registered by the
Market Office and living in the market area being permitted to trade
there. The markets were periodic, but since they alternated with
each other at half-monthly intervals, one or the other was always
open for business.
Both markets, East and West, were subdivided into shop areas
{ichikura), each of which was devoted to a single kind of merchandise.23 There was one and only one shop area for each kind of good
offered in the market, and every area was required to display a sign
indicating what was sold there. The number of shop areas and the
types of goods sold at the time of the founding of the city are unknown, but a century later sixty-seven different kinds were provided
for in official regulations. Some were common to the two markets
and some were restricted to one or the other, so that, with duplications excluded, the total number of shop areas in both markets was
eighty-four: fifty-one in the East and thirty-three in the West. Most
goods sold were fairly practical items: food (rice, barley, salt, bean
paste, a kind of vermicelli, fruits and nuts, seaweed, pungent bulbs,
sweet gluten, etc.); many types of cloth, clothing, and dyestuffs; tools
and implements of various sorts (combs, needles, writing brushes,
charcoal ink, iron and gold implements, lacquerware, wooden implements, pottery); transport animals (horses and oxen); oils; and
equestrian gear. But there were also shops specializing in less humdrum goods, such as jewels (pearls and jade), weapons and armor
(swords, bows, arrows, etc.), medicines and medicinal ingredients
23 Abe Takashi, "Heiankyo no keizai kozo," in ltd Tasaburo, ed., Kokumin seikatsu shi kenkyu,
vol. 2: Seikatsu to shakai keizai (Tokyo: Yoshikawa kobunkan, 1959), p. 74, shows that
ichikura did not refer to "shop" or "stall," as often thought, but to an area of stalls or shops
dealing in the same kind of goods.
Cambridge Histories Online © Cambridge University Press, 2008
164
THE CAPITAL AND ITS SOCIETY
(herbs, cinnabar, etc.), and aromatics. The goods sold in the markets came mainly, no doubt, from surplus governmental and private
incomes in the city, but also apparently to a certain extent from individual producers and merchant suppliers who began to appear in
the market area and on the outskirts of Heian, and perhaps also from
foreign trade. Some cash circulated in the markets during the ninth
and tenth centuries while the government was still minting coins,
but even then the chief means of trade was undoubtedly barter.
In the quest for recognition as a civilized society by the seemingly
advanced countries of the adjacent continent, Japanese leaders followed the example of Korea and China in minting the government's
own coinage, despite the apparent absence of a vigorous domestic
commerce in need of money currency. The leaders also probably
hoped thereby not only to encourage and facilitate such commerce
as existed, but perhaps as well to reap the profits that currency manipulation made possible. The first minting was the well-known
Wado kaiho coin of 708, which was produced just seven years after
the adoption of theTaiho code of 701. Eleven new coins followed in
the next two and a half centuries (until 958), eight of them during
the Heian years. Minting at various places but mainly in copperproducing regions like Suo and Nagato, they were mostly made of
brass, but some were silver, and there was one gold coin, the Kaiki
shoh5 coin of 760.
The coins were legally valued at more than the worth of their
metallic content. At the time of new mintings they were also customarily exchanged at the rate of one new coin for ten old coins, the
result being that the authorities experienced considerable difficulty
in getting them accepted, resorting frequently to fiats and rewards
directed toward that purpose. Their cause was not helped, one assumes, as the coins were steadily debased in the ninth century, becoming smaller and lighter, with more lead and less tin, and eventually what were in effect lead coins. Gresham's law no doubt led to
the hoarding of more valuable coins, and despite draconian measures against it, private minting was also practiced, further confusing an already chaotic currency situation. The government seems to
have tried to regain control of the currency by its frequent minting
of new coins, and also by reducing the disparity between the legal
value of the coinage and its actual metallic worth. But despite all efforts, the coins fell rapidly out of use after the last minting in 958, replaced in the late Heian period by imports of Chinese coins, especially the copper coins of Northern Sung.
Cambridge Histories Online © Cambridge University Press, 2008
THE CITY'S ECONOMY
165
The government used coins in its own transactions and in payment of its obligations, and the coins also circulated publicly, mainly
in the Heian region, but also to some extent in the outer provinces,
where they were sometimes hoarded as treasure or treated as magical objects. The currency of trade in the Heian markets was a mixed
affair, the trade relying at times on straight barter, at other times on
values expressed in units of rice or fabrics, and at other times partly
on cash, either Japanese or Chinese.
The relatively uniform nature of the economy and of the general
population in the early part of the period was an obvious consequence of the origins of Heian, a city thrust on a rural landscape
practically overnight to serve a single, wholly political purpose. But
even as the Heian palace was being built, the statutory regime it was
intended to serve was changing, and the pace and depth of the
change only intensified in the following century, soon leading to correspondingly major alterations in the capital's economy and population. The main economic changes were a contraction of government
revenues and the growth in importance among the nobility of private
or quasi-private sources of income, chiefly directly owned land; gifts
of goods and labor made by wealthy provincial governors to leading
nobles; and dues from people and land under tax protection by
noble houses, the protected people sometimes very likely including
merchants and artisans. The most notable population changes were
the radical reduction in the size of the official class already mentioned and a diversification of the general population, in part a consequence of the diminished fortunes of the government.
As the government experienced increasing difficulty in collecting
revenues and levying labor in the quantity and kind called for by
statutory law, it tended to concentrate its remaining resources in the
imperial family and the senior nobility, leaving the ordinary nobility
in what were often perilous financial straits and completely depriving the greater part of officialdom of any income at all. To survive in
the capital, it became necessary for the ordinary nobles and officials
to attach themselves in some fashion to the senior nobles, the chief
source of disposable income in the city outside the government itself. At the same time, as the wealth of the senior nobility - and especially that of the Fujiwara regents and their kin - grew, and as
their economic affairs were made more complex by the acquisition
ofshden and possibly other types of private income-producing rights
and property, their mansions increasingly demanded larger and
Cambridge Histories Online © Cambridge University Press, 2008
l66
THE CAPITAL AND ITS SOCIETY
more professional household administrative staffs. We find, consequently, that by the eleventh century some considerable part of the
ordinary nobility and the upper officialdom had been converted into
at least partial dependents of senior nobles, and especially of the regents. Some became formal members of senior-noble households,
where they served with appropriate administrative titles on a regular basis, tending to the wide range of economic, political, religious,
ceremonial, ritual, domestic, and personal business that occupied a
great noble house. Others served noble patrons in such particular
professions as the yin-yang arts, calligraphy, Chinese learning, the
healing arts, or the like, while many ordinary nobles simply attended
on the comings and goings of senior nobles, participated in their ceremonies, rituals, and religious observances, and played the courtier
at their festivities, for all of which rewards were regularly received
(commonly, silk fabrics or wadding, sets of costly female robes, and
horses - clothing and horses also being the items that seem to have
most attracted Heian robbers). If it was the privilege of a leading
noble to monopolize the more lucrative ranks and posts and to be
the recipient of gifts from those seeking his favor, it was also his
obligation to be generous in his treatment of the men and women
who served and attended on him.
Since most of the officials and ordinary nobles patronized by, or
in the service of, senior nobles continued to hold government titles
and ranks entailing some official duties, they are perhaps better
thought of as intermediate in social and economic status, neither
strictly government officeholders nor entirely private persons. The
same may be said of another important element of the population
that was making a similar transition from public to private status in
the mid-Heian period: the artisan or skilled-labor class. We know by
inference - and from the Heian sake brewers whose vats were sealed
in 806 because of a drought-induced inflation of rice prices - that
even in the early years of the city's history a certain amount of private industry and skilled labor existed to meet everyday needs of the
population that were not supplied in their incomes or through their
own labor, but it was at the palace and in its workshops that industry, and especially the more skilled forms of industry, appears to
have been concentrated. There were found the makers of most of the
material culture of the Heian court: the furnishings and draperies for
palace chambers, utensils of all types and materials (lacquer, silver,
bamboo, pottery, wood, etc.), weapons, tools and implements,
palanquins and carriages, balls and clay dolls, cloth and clothing,
Cambridge Histories Online © Cambridge University Press, 2008
THE CITY'S ECONOMY
167
roof tiles, paintings, paper and writing materials, and so on in an almost endless list. And when nobles had need of specially crafted objects, it was (fiction from the ninth and tenth centuries tells us) to
the palace artisans that they sometimes turned, supplying them with
the necessary materials and rewarding them for their work.
By the beginning of the tenth century, the number of palace artisans had been drastically reduced, perhaps by as much as half or
more. The chief causes of the reduction were probably declining governmental revenues (including a diminished supply of the materials
used in the workshops) and the avoidance of, or flight from, palace
duty by artisans who were finding more favorable working conditions
at or under the protection of great noble and princely houses. By the
eleventh century the wealthier households of princely personages
and senior nobles (especially the regents) regularly included a number of artisans and workshops. The artisans, it is thought, came
mostly from government shops or traced their lineages to government artisans, and some at least may still have been formally liable
for duty at the palace. The weavers of fancy silks, for instance, remained in the service of the palace perhaps longer than any other artisans, preserving a monopoly of their trade until the middle of the
eleventh century, but already in 1013 one of their number was working concurrently in the household of a near-prince. Few if any such
workers achieved autonomous status during the period, however,
since an artisan still had to be attached to the palace, to a noble or
princely house, or to a temple or shrine in order to receive the support and protection that made his work possible or profitable. His
first duty, therefore, was to meet the requirements of his patron or
patrons, but beyond that his work could be, and perhaps often was,
for the market.
The emergence of the artisan from the palace was very likely
among both the causes and effects of the more open market and the
expanded scope of commercial operations that had developed by
the eleventh century. The market by then had long ceased to be the
minutely regulated, carefully controlled institution the government
had sought to impose on the city at its founding. The simple dynamics of urban life, with its constantly shifting social and economic
patterns, almost immediately generated powerful forces for change
in the rigid concept of commerce embodied in the East and West
Markets. The trend was toward decreased interference in, and control of, the market by the government and significant overall expansion of commerce, leading eventually in the twelfth century to the
Cambridge Histories Online © Cambridge University Press, 2008
l68
THE CAPITAL AND ITS SOCIETY
emergence of the characteristic form of medieval market organization, the guild.
The official market system was already in trouble by 835, when the
government was forced to take steps to shore up the faltering West
Market. Its location in the sparsely populated western half of the city
seems to have affected its fortunes from the outset, and evidence of
near anarchic conditions in the markets suggests that even at that
early date the government's control of them may have been tenuous
at best. Thirty years later, the Market Office reported to the Council of State that merchants were placing themselves under the protection of princely personages and nobles, forming "gangs" (shiirui,
possibly a forerunner of the guild) and running roughshod over the
officials. Attempts were still being made at the beginning of the tenth
century to control trade and to restrict it to the official markets, but
by then, or certainly not long afterward, such attempts had become
largely futile, and private need or convenience was probably the chief
determinant of the content, the location, and the agents of trade in
the city.
No doubt the government's inability to maintain the official market system was generally the result of its own growing weakness and
the opposing strength of more or less natural market forces; but one
specific factor also must have contributed greatly to the process.
That was the government's loss of control over the bulk of the goods
moving in commerce in the capital. Under the statutory system, as
we have seen, most of the identifiable wealth entering Heian during
its early years came in the form of governmental revenues. It was the
surplus from those revenues and the surplus production of the government's own workshops that are thought to have constituted, in
one way or another, the major part of the goods traded in the official markets. The shrinkage of revenues and the curtailment of operations in government workshops in the ninth and tenth centuries
must have entailed, therefore, an at least proportionate contraction
in the quantity of official goods in the market. At about the same
time, there seems to have been a marked increase of goods in private
trade, coming chiefly from the growing private incomes of the nobility (including goods imported from the continent); from the
riches brought or sent to the capital by provincial administrators;
from the provincial gentry to whom much of the government's former revenues had been transferred; and from the emerging class of
semiautonomous artisans. In short, merchants were no longer so
heavily dependent on the government for the goods of their busiCambridge Histories Online © Cambridge University Press, 2008
THE CITY'S ECONOMY
169
ness, and control of their activities was neither so advantageous to
the court nor so readily accomplished.
The collapse of the official market system seems to have been accompanied by an expansion of commerce in the city, marked in the
tenth century by a tripling of the physical area of the markets themselves and in the following century by the growth of commerce and
industry in areas outside the markets. The latter development had
already led by the middle of the eleventh century to the appearance
of several new street names that probably derived from local concentrations of merchants and artisans in the southeastern part of
the city, names like Salt Street, Oil Street, Needle Street, Brocade
Street, and Damask Street. The official markets were converted to
private ownership and disappeared in all but name during the
course of the twelfth century, the chief commercial and industrial
quarters shifting to a north-south street near the center of the eastern half of the city called Machi (the modern Shimmachi), and especially to areas near the intersections of that street with Sanjo,
Shijo, and Shichijo Avenues.
The deregulation and degovernmentalization of Heian introduced
many new elements into the population or gave added prominence
to previously existing groups. Most conspicuous perhaps was the
greatly increased presence of Buddhist monks in and around the city,
subverting the highly secular atmosphere imposed in the days of the
Confucian-trained emperor Kammu. The founding of chapels on the
estates of great nobles like Sanesuke, the erection of large temples by
Fujiwara leaders on the eastern and northern edges of the capital,
the appearance of other temples nearby, and the growth of the
Tendai establishment on Mount Hiei must have resulted in a vast increase in the population of Buddhist monks in the city or within a
day's walk of it, their numbers almost certainly reaching into the
thousands. They were by then a vital economic and cultural force in
the capital, omnipresent in the life of the court and in the personal
affairs of the nobility. They supported a wide variety of artisans in the
construction, decoration, furnishing, and operation of their temples,
creating an informed demand for art and books, playing the role of
cultural mediators between China and Japan, and serving generally
as intellectual and cultural catalysts. The great bulk of them were
probably of fairly humble origins, but noble sons were found frequently in the higher Buddhist ranks and offices, and most of the
monks who figured directly in the lives of the court and the nobility
seem to have been of that class. The noble presence in Buddhist ternCambridge Histories Online © Cambridge University Press, 2008
170
THE CAPITAL AND ITS SOCIETY
pies and the catering of Buddhism to noble interests and needs were
so pronounced that historians sometimes speak of the "nobilization
of Buddhism" in this period.
Other new or expanded elements that can be identified in the
eleventh-century population of Heian include a small but important
group of warriors up from the provinces to serve chiefly in the kebiishi office or under regimental command; gangs of robbers, among
whom were found Buddhist monks, declasse or renegade nobles,
and probably also others whose livelihoods had vanished in the contraction of the statutory government; a flourishing society of beggars; male and (especially) female peddlers who roamed the city
streets selling a variety of goods; transport workers, who had grown
in numbers and importance as the flow of goods from the provinces
passed from government to private hands; entertainers, wonderworkers, gamblers, and, at least on the outskirts of the city, prostitutes; rich former provincial administrators whose houses were
sometimes assaulted by infuriated (or possibly calculating) citizenry
of their former provinces; and, of course, as always the huge numbers of servants and manual laborers with which the nobility surrounded itself. In its diversity and complexity, it was a truly urban
population, the most urban that Japan had ever known.
CITY ADMINISTRATION
The population in the early days of Heian was governed by two Capital Offices (Kyoshiki), the Left and Right, which were responsible
respectively for the Left and Right halves (the eastern and western
halves) of the city. Classified with the provincial administrations as
regional organs, each office communicated directly with the Council of State and the central ministries under a chief official of the
Fourth Rank, who directed an administrative staff of 7 and an armed
force of 240 men. Like their provincial counterparts, the offices were
responsible for the entire range of government in their jurisdictions,
including the compilation and maintenance of household registers;
the collection of taxes; police and judicial matters; the repair and
maintenance of canals, ditches, bridges, and quarter walls; the cleaning of streets; the dispatch of the abandoned sick and orphaned to
governmental institutions (it seems to have been the practice to eject
ailing menials from the houses where they were employed, perhaps
to avoid the ritual pollution of death); the removal and disposal of
Cambridge Histories Online © Cambridge University Press, 2008
CITY ADMINISTRATION
171
corpses (an onerous duty in times of epidemic disease, when as
many as 5,000 bodies might be collected from the streets and
riverbeds of Heian); and the distribution of aid to the indigent.
The authority of the Capital Offices was exercised at the local level
through an officer in each zone (jo) called the quarter, or zone, magistrate (borei,jorei), who was supposed to be an influential local resident of good character. Beneath him was a chief (bocho) in each quarter of the zone and also a ward chief (Jiocko) for every five households.
Although the quarter magistrate was a key official intimately involved
in the daily lives of Heian residents, the government experienced
continuing difficulty in finding suitable men for the post. This must
have been partly because of the scanty emoluments attached to it, but
perhaps it was mainly because of the frustrations and dangers the job
might involve for a low-ranking magistrate, who had to cope with
high officials and members of the imperial family living under his jurisdiction. It would have required a brave, even foolhardy, magistrate,
for example, to admonish a mighty court minister or a prince of the
blood for failure to clean up the streets in the vicinity of his house, an
obligation imposed by law on all residents of the city. Yet the magistrate was liable to punishment if he were remiss in the enforcement
of the law.
The problem of recruiting effective quarter magistrates seems to
have been one of the chief causes of the atrophy that affected the
Capital Offices during the course of the tenth century and led eventually to their replacement by other organs of government, chiefly
the kebiishi office. The emergence of the latter as the chief administrative organ for the capital probably also reflected the increasing
insecurity of the city from the tenth century on, when arson, robbery, and murder were epidemic and the exercise of police power
became perhaps the first concern in city government. But even that
office was unable to control the growing criminality of Heian, especially since its own agents were sometimes themselves involved in
the crime they were supposed to suppress. By the eleventh century
the city was, if not lawless, at least a freewheeling place compared
to the carefully governed and regulated capital envisioned under
statutory law.
Almost every aspect of life in Heian was minutely legislated at the
founding of the city, from the symmetrical layout of the city itself
down to the number of trees (seven) to be planted in the blocks sur-
Cambridge Histories Online © Cambridge University Press, 2008
172
THE CAPITAL AND ITS SOCIETY
rounding the Park of the Divine Spring, the architecture of private
houses, the cleaning of streets, the colors and materials of costumes,
behavior in social intercourse, the tethering of horses and oxen, and
so on in nearly infinite detail. In its orderliness, its discipline, and its
nearly uniform dependence on a single industry, the city reminds
one again of a military post or company town. And as with a military post or company town, the withdrawal of its single industry (the
court and its government) during the early years of Heian's existence would have surely resulted, as at Nara and Nagaoka, in the almost immediate disappearance of the city. But the court remained,
and by the eleventh century the capital had acquired in some measure a life of its own, forcing great changes in the plans and conceptions of its founders. The city of Kyoto had begun to emerge through
the debris of the collapsing ideal of Heian.
CHANGES IN THE CITY PLAN
The greatest physical departure from the original city plan was an
eastward and northward shift of the population that left the western
and southern parts of the site only thinly populated, tending toward
a geography in which the principal division was no longer east and
west, but north and south (or, in the parlance of later centuries,
Upper and Lower). The shift was in considerable part chimerical,
representing not an actual movement of people but a simple disjunction between the original plan and its realization in practice. As
noted earlier, it appears that only about half of the total area planned
for the city was actually laid out and developed, and it is clear that
most of the land thus left unoccupied or under cultivation was concentrated in the damp, low-lying western and southern parts of the
planned site. But those sections were by no means totally devoid of
urban development, and their loss of population to the eastern and
northern parts of the city seems to have been real, as the opening
lines of a short essay written in 982 by the religiously inclined Yoshishige noYasutane (d. 1002) indicate:
For the past twenty years and more I have observed the situation throughout the eastern and western sections of the capital. In the western part of
the capital the houses have become fewer and fewer till now it's almost a
deserted wasteland. People move out of the area but no one moves in;
houses fall in ruin but no new ones are ever built. Those who don't have
any other place to move to, or who aren't ashamed to be poor and lowly,
live there, or people who enjoy a life of obscurity or are hiding out, who
Cambridge Histories Online © Cambridge University Press, 2008
NEW BUILDINGS
173
ought to return to their native mountains or countryside hut but don't. But
anyone who hopes to pile up a fortune or whose heart is set on rushing
around on business wouldn't be able to stand living there even for a day.24
Spurred also probably by a cluster of communication routes along
the eastern edge of the city and by the attraction of the imperial
palace and the mansions of the great nobles in the north, the demographic shift noted by Yasutane (doubtless with exaggeration) carried the population beyond the capital's official borders into the alluvial plain of the upper Kamo River to the north and across to the
foot of the Higashiyama Hills on the other side of the river where it
flowed past the eastern city limit. The western edge of the urban area
meanwhile retreated eastward, by the end of the latter half of the
twelfth century coming to rest, it is thought, somewhere between
Suzaku Avenue and the avenue (Omiya) that bordered the eastern
side of the Greater Imperial Palace. The geographical center of the
city was shifting toward the eastern reaches of Sanjo and Nijo Avenues, and the symmetry of the city's original plan was gone. With it
went the earlier zone, quarter, block, and house-lot land divisions,
the zone and quarter ceasing to function as administrative areas, and
the house lot made meaningless by repeated divisions and sales. By
the end of the eleventh century, property was being identified both
by the old land divisions and by a new system of street coordinates
(Nishinotoin and Shijo, for example), and it was not long before the
older system disappeared altogether from documents.
NEW IMPERIAL AND FUJIWARA BUILDINGS
The eastward and northward shift of the population left the Greater
Imperial Palace stranded on the western edge of the capital, its
changed position relative to the city an almost too neat symbol of the
changed status of the regent-ridden emperors. Yet despite frequent
fires, the major structures of the palace itself seem to have remained
much the same until past the middle of the eleventh century. The
original Court of Government and its Great Hall of State were destroyed by fire in 876 but were almost immediately rebuilt. Although
they burned once again in 1058 and fourteen years elapsed before reconstruction was completed, they remained a usual part of the
palace landscape until their final destruction in the great Kyoto fire
24 From "Chiteiki" (Record of the Pond Pavilion), trans. Burton Watson, in Japanese Literature in Chinese, vol. 1 (New York: Columbia University Press, 1975), pp. 57—58.
Cambridge Histories Online © Cambridge University Press, 2008
174
THE
CAPITAL AND ITS SOCIETY
of 1177. The original Court of Abundant Pleasures survived unscathed for more than two and a half centuries, but it, too, finally
succumbed to fire in 1063 and was never rebuilt, adding greatly to
what was probably by that time the growing vacant space within the
palace walls. Even though the two great ceremonial courts of the
palace survived until well into the eleventh century or beyond, their
functions had long been almost entirely usurped by the Emperor's
Residential Compound (or the substitutes therefor). Only the imperial accession ceremony seems still to have been regularly conducted
at the Great Hall of State.
Although not quite so fortunate as the Court of Abundant Pleasures, the Emperor's Residential Compound also led a charmed life
for a century and a half after the founding of Heian, escaping the numerous conflagrations that were a nearly inescapable part of life in a
city where all structures were wooden; curtains, bamboo blinds, and
paper or cloth screens were the chief partitions; and open flame or
coals were the only means of illumination, heating, and cooking. But
at last, in 960, fate caught up with it, too, and the entire compound,
along with the many art treasures, heirlooms, and documents and
books stored there, went up in smoke. The buildings were soon rebuilt, but the disaster was repeated so frequently thereafter (fourteen
times, or an average of a little less than once a decade, between 960
and 1082) that politically or larcenously inspired arson is generally
assumed. The residential compound continued to be used by most
emperors until past the middle of the twelfth century, but by the end
of that century it and the Greater Imperial Palace seem to have been
largely abandoned. A fire in 1227 put a finish to what remained of
the buildings in the compound, and the old palace site turned eventually into vegetable gardens famed especially for their turnips.
Although the Emperor's Residential Compound seems to have
been rebuilt according to its original plan following the fire of 960,
the use of the buildings within the compound differed in one important respect. At the founding of the capital, the emperor's personal residence within the compound, it will be remembered, was
the Benevolent Longevity Hall (Jijuden). But for unknown reasons
emperors by the middle of the ninth century were living not only
there but also frequently in other buildings within, and even on occasion outside, the compound, and not until after the fire of 960 did
a single building again come to be identified as the usual imperial
residence. The building so favored was called the Limpid Cool Hall
(Seiryoden), a name derived from that of an imperial summer palace
Cambridge Histories Online © Cambridge University Press, 2008
NEW BUILDINGS
175
in Han dynasty China. Facing east (rather than south) and located
off the north-south median line of the compound directly across an
enclosed courtyard from the Benevolent Longevity Hall to the west,
the Limpid Cool Hall was, according to Chinese notions, appropriate neither in name nor position for the residence of a reigning emperor. It was even more cramped than its predecessor. The dimensions of the largest chamber in the building, the chamber where the
emperor apparently spent most of his time, seem to have been about
50 by 25 feet; the imperial bedchamber was about a third of that; and
the audience chamber at the southern end of the building where the
emperor received the nobility was a hall only 10 to 15 feet wide and
perhaps 60 feet long. Servants, female officials, ladies-in-waiting,
consorts, and courtiers crowded the building day and night, and the
emperor, like his subjects in the noble mansions, was probably never
alone or more than a few feet from other people even in his most intimate moments.
An emperor who found himself without living quarters as a result
of the frequent fires at the residential compound after 960 usually
took up temporary residence outside the Greater Imperial Palace at
a mansion, or part of a mansion, vacated and placed at his disposal
by a Fujiwara maternal or marital relative, where he remained until
the reconstruction of the compound was completed. The period of
residence in such "town palaces" (saw dairi), as they were usually
called, varied anywhere from several months to several years, but
until the end of the eleventh century the emperor always eventually
returned to the residential compound of the Greater Imperial Palace.
By the beginning of the twelfth century, however, emperors were beginning to live permanently in "town palaces" especially constructed
for the purpose, returning to the Greater Imperial Palace only for
ceremonial occasions, and after the final destruction of the Emperor's Residential Compound in 1227, such palaces became the regular imperial residence. It is from a "town palace" of the fourteenth
century just west of the site of Fujiwara no Michinaga's residence in
the northeastern corner of the old capital that the present Kyoto imperial palace (Gosho) descends.
The architectural landscape outside the imperial palace also
changed in conspicuous ways. Rampart Gate, the towering entranceway to the capital at the southern end of Suzaku Avenue, had
been rebuilt after its destruction by wind in 816, less than twenty-five
years after the city's founding, but it collapsed again during a great
storm in 980 and seems never again to have been restored. The "ramCambridge Histories Online © Cambridge University Press, 2008
176
THE CAPITAL AND ITS SOCIETY
part" along the city's southern border was probably already gone by
that time, as were very likely many of the city's interior walls. The
Korokan Lodgings must have been in an advanced state of decay by
the end of the tenth century, their function lost with the cessation of
relations with Po-hai in the 920s, but some evidence suggests that the
buildings themselves may have survived into the twelfth century.
Gone, too, most likely by about the beginning of the eleventh century or before, were the governmental charity institutions, the public
granary, and the imperial pleasure buildings in the Park of the Divine
Spring. The park had been much favored by the early Heian emperors, but imperial visits declined sharply as the space began to be employed for public religious rites in the latter half of the ninth century.
The frequent use of the large pond there to provide drought relief for
the rural population south of the capital may have further reduced
the natural attractions of the place for its courtly patrons. Left a
wasteland by the great fire of 1177, the park seems at times thereafter
to have become a peculiarly noisome site for the dumping of garbage, human waste, and corpses. The northern half of the park was
incorporated into the site of the Nijo Castle at the beginning of the
seventeenth century; its surviving remnant in modern Kyoto is less
than 5 percent of its original size.
The decay and disappearance of the symbols and monuments of
the old statutory capital were accompanied by the emergence of the
architectural landmarks of the city of the regents, most notably the
luxurious mansions and the great temples of the Fujiwara leaders.
The mansions were located exclusively in the northeastern quarter of the capital, northeast of the intersection of Shijo and Omiya
Avenues, an area that because of them became perhaps the most vital
center of political and social life in the capital. The site of constant
ceremony, ritual, and entertainment, of the birth and upbringing of
imperial children, of imperial visits, of great economic strength, and
of much of the substance (as opposed to the formal aspects) of political power, the mansions exercised a centripetal influence on the
life of the city. They did not by any means replace the imperial
palace, but heavily diluted its once unrivaled role in shaping the capital's civilization. Physically, the Fujiwara dwellings were all generally
of the same mold and on the same scale as Sanesuke's Ononomiya
establishment. The best known, of course, were those of the most
powerful regents: Kaneie's Higashi sanjo residence, Michinaga's
Tsuchimikado mansion, and Yorimichi's great Kayanoin, which was
twice as extensive as the others, a place (in Sanesuke's words) of "inCambridge Histories Online © Cambridge University Press, 2008
NEW BUILDINGS
177
comparable magnificence," polished with peach pits until it shone
like a mirror, and so grand that it seemed to another contemporary
observer as if it belonged to a different world. Grand and brilliant as
such mansions may have been, however, their grounds were only a
fraction of the size of the Greater Imperial Palace, and only Yorimichi's Kayanoin rivaled even the much smaller dimensions of the
Emperor's Residential Compound. The regental world was on a
smaller scale than that of the statutory emperors.
The Fujiwara regents founded two Buddhist temples dedicated
chiefly to Amidist faith at the eastern edge of the city on the narrow
strip of land between Higashikyogoku Avenue, the eastern city limit,
and the Kamo River. One, called the Hokoin (or Hokoin), was
founded in the dwelling hall of a mansion owned by Kaneie just
north of Nijo Avenue, shortly before his death in 990; the other, the
Hojoji, was established by Michinaga directly east of his Tsuchimikado mansion and a short distance north of the Hokoin in 1019.
Illness and apprehension of approaching death were the direct motives for the founding of both temples, but whereas Kaneie died almost immediately after the founding of the Hokoin, leaving its further development in the hands of descendants, Michinaga made a
good recovery from his illness and was able to devote most of the last
decade of his life to the construction, decoration, and outfitting of
the Hojoji, which became in consequence the larger and more famous of the two temples.
Lavishing wealth and attention on the temple, the great ex-regent
turned the 14 or so acres of the site (the same size as the sites of the
East and West Temples) into the scenic and architectural wonder of
his age, creating elaborate gardens and constructing over a dozen
major halls and chapels there. A probably contemporary source provides an impressionistic description of the construction in its early
stages:
A great tile-capped wall was thrown around a four-block area. Michinaga
urged the work on with floods of orders, chafing at the slowness of dawn
and bemoaning the gathering shadows of night. He turned ideas over in his
mind all night long. How should the artificial hill be built up? The lake laid
out? The garden designed? He must go on to construct a whole series of
impressive halls. Nor could the images be run-of-the-mill affairs; there
would be great numbers of golden buddhas sixteen feet tall, arranged in a
row with a passageway running from north to soudi in front of them. Padis
and walks would be needed, and corridors and galleries. . . . Daily levies of
laborers, amounting to from 500 or 600 to 1,000 men, came from the sustenance households and private estates of Michinaga's male connections
Cambridge Histories Online © Cambridge University Press, 2008
i78
THE CAPITAL AND ITS SOCIETY
Great North Gate
Great South Gate
1. Northwest Cloister (Saihokuin)
2. Monks' Quarters
3. Michinaga's Residential Quarters
4. Lecture Hall (Kodo)
5. Sakyamuni Hall (Shakado)
6. Golden Hall (Kondo)
7. Amitabha Hall (Muryojuin;
Nakagawa Mido)
8. Hall of the Five Great Mystic
Kings (Godaido)
9. Healing Buddha Hall (Yakushido)
10. Samadhi Hall (Sanmaido)
11. Bell Tower
12. Sutra Treasury
13. Ten Days of Fasting Hall
(Jissaido)
Figure 2.6. Plan of Hojoji. From McCullough and McCullough,
A Tale of Flowering Fortunes, p. 782.
and the Imperial personages, and everyone was immensely heartened by
the availability of so many hands. Provincial governors competed to provide
the most labor, timber, cypress bark, and tiles for the hall, even at the cost
of falling behind with their rental taxes and tribute commodities; and artisans flocked from near and far, making themselves useful in capacities and
places suited to their callings. In one place, master joiners worked on sacred images, assisted by a huge crew of 100 image-carvers. What assignment
could be more splendid for an artisan! Near the top of the building 200 or
300 carpenters were at work, shouting "Esa! MasaP' in unison, as they
raised massive beams attached to thick cables. In the interior, where gorgeous thrones for the images were being built, forty or fifty men were polCambridge Histories Online © Cambridge University Press, 2008
NEW
BUILDINGS
179
ishing the plank floors with scouring rushes, muku leaves, and peach pits.
Countless cypress-bark roofers, plasterers, and tile makers worked away,
and venerable monks and other aged men were cutting and laying threefoot rocks. Some 400 or 500 men had descended to excavate the lake, and
another 500 or 600 had climbed up onto the artificial hill and were adding
to its height, layer by layer. On the avenues, shouting laborers pulled immense tree trunks roped to work carts; on the Kamo River, raftsmen sang
cheerful, lusty songs as they poled their loads of lumber upstream....Crews
tugged at fragile rafts, which somehow managed to keep afloat under the
weight of mighty rocks as big as cliffs.25
The same source also describes the dedication of the temple's
main building, the Golden Hall, which was attended by Michinaga's
grandson, Emperor Go-Ichijo:
Very much at ease, the Emperor gazed at the scene inside the temple compound. The garden sand glittered like crystal; and [artificial] lotus blossoms
of varying hues floated in ranks on the fresh, clear surface of the lake. Each
blossom held a buddha, its image mirrored in the water, which also reflected, as in a buddha domain, all the buildings on the east, west, south,
and north, even the sutra treasury and the bell tower. Jeweled nets hung
from every branch on the trees bordering the lake; fragile blossoms quivered in the still air. Green-pearl leaves shone with the hue of beryl; elegant
glass branches appeared on the bottom of the lake; delicate clusters of flowers hung as though about to fall. There were leaves of many kinds and colors — green pearl, like pine trees at the height of summer; gold, like late autumn foliage; amber, like mid-autumn foliage; white glass, like a winter
garden mantled in snow. Ripples washed the lake's golden jeweled shores
when a breeze stirred the trees. A bridge made of the seven treasures
spanned the golden jeweled lake, jeweled boats glided in the shade under
the trees, and [artificial] peacocks and parrots played on the island.26
(On the dedication ceremony see also the section "Music" in Chapter 6.)
In a bit of ironical symbolism that even Michinaga may have appreciated, some of the old foundation stones of Rampart Gate were
used in the construction of Hojoji.
A final new element in the city landscape, and one that would
grow in significance in the twelfth century, was the small group of
palaces used by retired emperors. In the Nara period, the abdicated
emperors had continued to live at the Greater Imperial Palace, but
when Emperor Saga gave up the throne in 823, he took up residence
at Reizeiin, a spacious four-block establishment at the southeastern
corner of the Greater Imperial Palace. Subsequent abdicated em25 William H. and Helen Craig McCullough, A Tale of Flowering Fortunes, pp. 499-501.
26 Flowering Fortunes, pp. 553-54.
Cambridge Histories Online © Cambridge University Press, 2008
l8o
THE CAPITAL AND ITS SOCIETY
perors and their empresses followed his example, using either the
Reizeiin or one of several other palaces scattered through the city, including the largest of them all, the six-block Suzakuin directly south
of the Greater Imperial Palace. Landed property was attached to
some or all such palaces, and that wealth and the household organization established for each abdicated emperor are generally thought
to have provided part of the impetus that led to the brief revival of
imperial-family power in the twelfth century.
CEREMONY AND RITUAL
Heian, as a political administrative seat for an imperial line and its
hereditary nobility, was perhaps inevitably an intensely ceremonious and ritualistic city. Whether in the early Heian of Kammu and
his immediate successors on the throne or in the Heian of Michinaga's time, ceremony and ritual were primary concerns of government and private citizens alike. During the period of the Fujiwara
regency, in particular, it might even be argued that ceremony and
ritual were conceived to be at the very center of life and government, more time and wealth being expended on them than on perhaps any other single category of public or social activity at court
and among the nobility.
An order of ritual and ceremony at the imperial court had been
established at least as early as the ninth century, and precedential
rules providing detailed instructions for the conduct of individual
participants were evolving by the tenth century, when knowledge of
the rules became the sine qua non of a successful career at court. A
key figure was Fujiwara no Mototsune, the imperial regent whose
observances became the basis, through his sons Saneyori (Ononomiya) and Morosuke (Kujodono), of two ritual-ceremonial schools
(?yu)> the Ononomiya and the Kujo. An initial dominance of the
Ononomiya (founded by Saneyori's son Sanesuke) came to an end
with the revival of the Kujo under Michinaga, from whom descended a ritual-ceremonial line known as the Mido school, which
was a mixture of Ononomiya and Kujo elements with Michinaga's
own particular contributions. A number of learned treatises were
written on the subject by eminent authorities like Minamoto no
Takaakira for the reference of courtiers, and when a man was called
learned and able, it often had to do with his knowledge of ritual and
ceremony. The existence of such schools and the extraordinary intellectual attention to their subject matter by senior government
Cambridge Histories Online © Cambridge University Press, 2008
CEREMONY AND RITUAL
l8l
ministers may baffle or appall the casual modern observer, but they
were the rational product of minds that conceived of ritual and ceremony as the chief means by which nature could be controlled, society regulated, and personal success achieved in a mysterious, imperfectly understood world.
In the life of Heian, no ritual or ceremony was more representative
of the capital and its society than the regular Kamo Festival, which
took place every year in the fourth month. The festival was both ritual and ceremony, functioning on the one hand to guard the city and
its people, and on the other to affirm and display the bonds that held
the community together. It was also, as almost all rituals and ceremonies were in part, an entertainment, providing one of the great
spectacles of the city's year. In short, it was the Heian equivalent of
the famous Gion Festival that later came to symbolize Kyoto under
warrior rule as a city of artisans, merchants, temples, and shrines.
The explicit purpose of the Kamo Festival was to pay homage to
the Kamo deities and to secure their protection of the court and the
capital. Originally worshiped as deities of the ancient Kamo clan,
the gods were enshrined in two locations in the northern part of the
Kyoto basin: at the Lower Shrine near the confluence of the Kamo
and Takano rivers and at the Upper Shrine a couple of miles farther
north on the east bank of the Kamo.
There were two central figures in the festival: the Kamo Vestal and
the Imperial Messenger. The Vestal was a near-princess chosen by
divination at the beginning of each reign, or whenever the post fell
vacant. Her role is not altogether clear, although it is obvious that it
was modeled on that of the Ise Vestal, who was considered the Chief
Priestess of the Ise Shrines and lived there. The Kamo Vestal, who
was first appointed in the reign of Emperor Saga, lived at a special
palace called Murasakinoin on the northwest outskirts of the capital. Her role seems to have been simply to maintain ritual purity and
to represent the emperor at the festival. The Imperial Messenger was
a Fifth-Rank courtier holding the office of Middle or Lesser Captain
of the Imperial Guards, and thus usually a man destined for high office. He read an imperial rescript at the shrines and presented the
emperor's offerings, which was the chief object of the festival. The
Vestal and the Messenger were accompanied by ten dancers and
twelve musicians, who were also courtiers from the Imperial Guards.
Their duty was to perform an ensemble of music, dance, and song
called "Eastern Music" (Azumaasobi) at the shrines.
The great festival procession began from the Vestal's palace and
Cambridge Histories Online © Cambridge University Press, 2008
182
THE CAPITAL AND ITS SOCIETY
proceeded along Ichijo Avenue on the northern edge of the city toward the shrines. It was made up of a huge throng of military and
civil officials, court ladies, and attendants, some walking and others
mounted on elaborately caparisoned horses or riding in ox-drawn
carriages, and all brilliantly costumed in formal robes, with headgear, mounts, and carriages decorated with flowers and leaves. The
chief figures in the procession were the Vestal carried in a large
palanquin by twenty bearers and surrounded by numerous male and
female attendants; the Imperial Messenger dressed in black silk
robes, wearing a sword and riding horseback; the Messenger's retinue carrying the official vermilion umbrella and an umbrella covered with artificial flowers; the mounted musicians and dancers; an
officer from the Storehouse Bureau in charge of white chests containing the offerings for the shrines, and bearing himself the imperial rescript to the shrine enclosed in a brocade bag; an officer from
the Stables Bureau in charge of fourteen horses that were to be paraded before the shrines; the vice-governor of Yamashiro Province,
the province in which the shrines were located, riding at the head of
troops; and kebiishi police to clear the way for the procession and to
provide ceremonial guard at the shrines.
Along the route of the procession, crowds of townsmen and countrymen filled the streets and overflowed onto housetops and trees.
Gorgeously attired ladies, courtiers, and exalted personages sat in
their lacquered carriages or luxurious viewing stands, while their
lackeys jostled against the commoners in an excited, unruly mass
through which kebiishi agents, marching in the vanguard, cleared a
passage. Houses along the way were all decorated with garlands of
real and artificial flowers, with leaves of the katsura tree, and, especially, with the aoi or "heartvine" leaves (Asarum caulescens Maxim)
that gave the festival its popular name.27
The ritual at the shrines themselves was simple. At both sites the
Vestal paid her respects while the Imperial Messenger intoned the
imperial rescript praising the gods and requesting their continued
favor. The offerings (silk, hemp, etc.) and the dances and songs were
presented, and the horses were paraded and raced. The Vestal spent
the night at the Upper Shrine, and on the following day there was a
less elaborate procession of return to the capital, culminating in a lavish banquet at the imperial palace, with rewards for all participants.
27 "Heartvine" is the ingeniously apt neologistic translation by Edward Seidensticker.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 3
LAND AND SOCIETY
Social and economic change in premodern times is seldom rapid.
But even by premodern standards, change in rural Heian Japan
must at first glance seem glacial. Population did increase moderately
and with it the area of land under cultivation. Millet gave way to barley as the second grain most grown, but rice retained its place as the
primary and most prestigious grain. Irrigation networks, both simple and intricate, grew to feed the expansion in arable land, but with
no discernible advance in irrigation technology - indeed, the great
state-sponsored water projects of previous centuries found no echo
in the Heian period. Provincial handicraft industry grew but little,
and there may even have been a decline in the production of silk.
House construction remained little changed through most of the
country, although in the home provinces there may have been some
shift from excavated floors ("pit dwellings") to raised foundations.
All in all, there is little to show for the passage of four centuries.
Or so it must seem to the modern observer. Yet to an eighthcentury peasant, agricultural life in twelfth-century Japan would
have seemed significantly altered. Iron tools were much more abundant, and included tools and uses unfamiliar to him. Households
would have seemed smaller and less independent than in his own
age, and the force of local elite families on economic life much
stronger. Most of all, he would have found a dramatically changed
tax structure, with taxes now managed by an extensive class of local,
regional, and provincial notables and officials whose powers, initiatives, and frequently only quasi-official status would have struck him
as bewildering and, no doubt, intimidating.
Change is not always progress: of the changes in technology, social
structure, taxation, and landholding to be discussed, perhaps only
the diffusion of the plow can be seen as an unequivocal advance. But
most of these changes worked toward a restructuring of rural Japanese society and economy that was at least as significant in premodern terms as straightforward technological advances and quantitative
183
Cambridge Histories Online © Cambridge University Press, 2008
184
LAND AND SOCIETY
growth. The greater complexity of rice farming in the Heian period
drew households together in greater interdependence, and so began
the slow growth that was to culminate in the highly political and cooperative Tokugawa village many centuries later. Changes in farming
also brought more power to wealthy families, just as developments in
taxation rendered the central and provincial governments more dependent on these local notables to keep the ruling class in the capital properly supplied with the goods and services on which it relied.
The greater power and influence that thus devolved on local notables
did not of itself create the provincial warrior class of bushi that dominated medieval Japan. But bushi growth could hardly have proceeded in quite the same manner without this restructuring of rural
Heian society.
If a single factor can be assessed as most responsible for this restructuring, the leading candidate must be the brisk growth in population experienced in the first third of the Heian period - a growth
that, significantly, was most marked in the eastern and outer provinces where bushi development became strongest. By putting pressure on agricultural resources, growth in population stimulated a
search for means of increasing crop yields, stabilizing cultivation,
and, at the very least, lessening the extreme losses in harvests during
bad years. Growth in population and in agricultural production also
increased the potential tax income that could be collected by local
officials - an increased income most available in provinces distant
from the capital, and hence out of sight of a ruling class dependent
on local officials and notables for even a straightforward continuation of existing levels of tax income.
AGRARIAN TECHNOLOGY
In Chapter 26 of Konjaku monogatari-shu. there is a tale, most likely
from the early Heian period, that begins as follows:
At a time now past, there was a lowly person who lived in Hata District in
Tosa Province. This person held rice fields both on the shore where he lived
and on the shore opposite. To cultivate the fields on the opposite shore, he
planted seed rice in the fields near his home, then carried the seedlings by
boat to be transplanted on the opposite shore. With the seedlings, he
brought by boat hired farm laborers, food provisions for these and for his
family, a single-stem plow (maguwa), a full plow (karasuki), a scythe, hoe,
ax, and other cutting tools.1
1 Konjaku monogatari-shu, vols. 22-6 of Nihon koten bungaku taikei (Tokyo: Iwanami shoten,
1962), vol. 25, kan 26, story 10, p. 443.
Cambridge Histories Online © Cambridge University Press, 2008
AGRARIAN TECHNOLOGY
185
In this tale we can see the adoption of a rice regime that joined
seedbeds and summer transplanting, preparation of the soil by animal-drawn cultivating tools, and, by implication, draining of fields
midseason to allow the crop to mature in dry soil.2 Rice cultivation
was introduced to Japan about the second century B.C.E.; the use of
seedbeds and the transplanting of seedlings to flooded fields in early
summer can be found as early as the second century c.E.;3 both the
simple and full plows were brought to Japan from the continent in
the fourth or fifth century.4 The use of seedbeds and the plow produced greater rice yields; equally important, each allowed rice to be
grown in a greater variety of soils. But both also required a considerable cost, in labor as well as in capital expense. Transplanting seedlings was extra work and had to be done quickly while water was
diverted into each field - hence requiring the extra expense of hired
help. Purchasing and maintaining a plow and draft animal was a
greater expense still. These costs hampered the full adoption of an
improved rice regime available to the Japanese in the fifth century
until as late as the tenth century.
Evidence of planting first in seedbeds, then transplanting to
flooded fields in early summer - a practice known as taue - has been
noted in the famous Toro site in Shizuoka Prefecture as early as the
second century. Some scholars further feel that taue must have been
well known even before this period.5 Yet the simpler practice of direct seeding was still familiar in the seventh and eighth centuries. A
poem, "Where ricefields have been planted along the Sumiyoshi
coast / Alas, that I do not see you from planting even to harvest!" in
the eighth-century anthology the Man'yoshu, for example, seems
clearly to imply that the rice had matured in the same field in which
it had been first planted.6
Under conditions of primitive rice agriculture, the higher yields
resulting from the practice of taue did not offset the greater amount
of labor required to transplant seedlings. In fact, it is probable that
taue was first adopted not for its higher yields, but to conserve water
during the spring growing season and to concentrate seedlings into
2 Iinuma Jiro, Nihon nogyo gijutsu ron (Tokyo: Miraisha, 1971), p. 88, describes this regime.
3 This has been shown at the Hiraide site; see Hiraide iseki chosakai, ed., Hiraide (Tokyo:
Asahi shimbunsha, 1955).
4 Kinoshita Tadashi, "Nogu," in Wajima Seiichi, ed., Nihon no kbkogaku, vol. 3 (Tokyo:
Kawade shobo, 1966), p. 246.
5 KondoYoshiro and Okamoto Akio, "Nihon no suite noko gijutsu," in Ishimoda Sho, ed.,
Kodai no koza, vol. 3 (Tokyo: Gakuseisha, 1962), p. 344. Also see Kinoshita, "Nogu," p. 240.
6 Man'yoshu, vols. 4-7 of Nihon koten bungaku taikei, vol. 6, p. 131 (poem 2,244).
Cambridge Histories Online © Cambridge University Press, 2008
l86
LAND AND SOCIETY
a smaller area that could be better protected from the elements.7
With improved control of irrigation - a notable area of agrarian advance in the fourth to sixth centuries - practice of taue also resulted
in superior yields, however, for seedlings grown in seedbeds were
hardier and more evenly distributed than those sown by direct
broadcast.
Furthermore, fields to which seedlings were transplanted could be
richly prepared by either plow or manual cultivation. This is indicated in the Konjaku tale just quoted by the mention of two such
cultivating tools, the single-stem plow (maguwa) - basically a heavy
hoe drawn by a horse or ox - and the more substantial plow known
as suki or karasuki. Conceivably the single-stem plow found some
use after its introduction to Japan in the fourth century, but its employment was confined, of course, to those few who owned draft animals.8 The more effective full plow, however, remained rare for the
first three to four centuries after its introduction. In an era of iron
scarcity, it was far more effective to use the iron that would have
been needed to make one heavy plowshare to make instead several
hoe or spade blades. It is also likely that the advantages of plow use
in rice farming were not well understood at first.
We should not be surprised by this, for the plow in Asia (as elsewhere) was essentially a tool of dry-field farming. Originally used
in north India and north China for cultivation of millet and barley,
the plow was only later transferred to wet rice culture.9 This association of the plow with dry field farming retarded its diffusion in
Japan, where for several reasons the greatest importance was always
attached to the cultivation of rice. One reason was that the introduction of rice to Japan in fact preceded that of other grains by at
least two centuries. Beyond this, however, rice was actively promoted by the developing aristocratic government from the early
centuries c.E. Aristocrats preferred rice because of its high and relatively secure productivity, its effect on settling - indeed, tying
down - a population into narrow regions easier to control, and because of their own preference for rice as the staple of their diet. As
a result, the best fields were reserved for rice culture; fields growing
other crops, though perhaps as extensive as those growing rice,10
7 Kondo and Okamoto, "Nihon no suito," p. 343.
8 Okamoto Akio, "Nogyo seisan," in KondoYoshiid and Fujisawa Choji, eds., Nihon no kokogaku, vol. 5 (Tokyo: Kawade shobo, 1966), p. 34.
g Iinuma, Nihon nogyo, pp. 73 and 92.
10 Dana Morris, "Peasant Economy in Early Japan, 650-950," Ph.D. diss., University of California, Berkeley, 1980, pp. 137-38.
Cambridge Histories Online © Cambridge University Press, 2008
AGRARIAN TECHNOLOGY
187
were largely confined to marginal lands and to land not conveniently located for irrigation.11
With the continued moderate growth in population, arable land
planted to rice continued to expand in the eighth and ninth centuries. Nonetheless, there is reason to believe that expansion of land
planted to dry-field crops such as millet, barley, and soybeans may
have more than just kept pace with expansion in rice fields. Growth
of paddy was limited by the still primitive state of irrigation technology; this was particularly the case in the more heavily cultivated
home provinces (Kinai).12 In addition, the widespread failure of land
projects sponsored by the aristocracy in the eighth century led to the
withdrawal of aristocrats from land reclamation efforts in the Heian
period. Whereas the ruling class was most anxious to create new rice
fields, the local land developers who took their place were more likely
to respond to peasant demand, and hence create dry fields in at least
equal number to paddy. Also significant was the imposition of a
higher land tax in the tenth century. Because it was applied to rice
fields of all kinds, including those developed privately, the higher
land tax increased the advantage of developing tax-free dry fields.
From this impression of the increased importance of dry-field cultivation in Heian Japan, we may deduce the means by which the utility of the plow in rice farming came to be recognized. As dry-field
cultivation received more attention, perhaps even constituting a
larger percentage of all fields, it made increasing sense to invest in a
plow and animal, since the advantages of plow use in dry-field farming probably were well known. The rising role of local notables in
land development was crucial, for such entrepreneurs were best able
to afford plows and draft animals.
Once brought into use on dry fields, it was a logical extension to
apply the plow to rice fields, particularly considering the magnitude
of the plow investment. Leading an animal through carefully prepared rice fields may have seemed risky at first, but the extra time
spent repairing damage to embankments was more than offset by
the time saved with the plow. When it became apparent that deeper
tillage benefited the soil and increased yields as well, the advantages
of the plow in rice cultivation should have been soon evident.
11 Toyoda Takeshi, ed., Sangyo shi, vol. 10 of Taikei Nihon shi sosho (Tokyo: Yamakawa shuppansha, 1964), pp. 163-69.
12 Kinda Akihira, "Heian-ki noYamato bonchi ni okeru jori jiwari naibu no tochiriyo,"Shirin
61,3 (May 1978): 75-112. Also see FurushimaToshio, Nihon ribgyb gijutsu shi, vol. 1 (Tokyo:
Jichosha, 1947), pp. 142-46.
Cambridge Histories Online © Cambridge University Press, 2008
l88
LAND AND SOCIETY
The kind of plow best suited to rice culture, however, was not that
best for dry-field crops. Dry-field farming required a light plow to
minimize dehydration of the soil - most especially on the Asian continent, where the Japanese plow originated. (Because of its wet summers and volcanic soil, Japan was better suited for use of a heavier
plow even for dry-field farming.) By contrast, a heavier plow, which
cultivated deeper, could be used to advantage when growing rice because the flooding of fields during the first half of the growing season prevented soil dehydration. It was also possible to use a plow
with a moldboard, an attachment set at an angle to the top of the
plowshare to turn over the earth as it is loosened. The action of the
moldboard helped aerate the upper layer of the soil, further renewing fertility. But this action also allowed the soil to dry out, rendering it undesirable in dry-field farming.
References to the plow are few in both Nara and Heian literature.
Despite this, historians have been able to delineate a chronology of
its diffusion on which there is general agreement. There is little sign
of use of the plow in the Nara period, at least in rice culture, and literary evidence of its use is virtually nonexistent.13 Two simple
eighth-century plows survive in the Todaiji warehouse, Shosoin, but
their use appears to have been ceremonial.14 Both the Shosoin plows
lack moldboards. As already noted, plows without moldboards were
better suited to dry-field culture, suggesting that plows of the
Shosoin type were not intended for use in rice culture.
Heian sources, by contrast, reveal much greater use of the plow.
The early tenth-century sources Engi shiki and Wamyo ruiju sho both
speak of the plow as in common use, at least as farming was viewed
from the capital.15 Use of the plow in rice farming also is described
in the tenth-century romance Utsuho monogatari and, as noted, in
Konjaku.l6 The Konjaku story does not reveal whether the plow in
use had a moldboard or how many animals were used to draw it.
Both Engi shiki and Wamyo ruiju sho, however, describe the plow as
being equipped with a moldboard. This is significant as evidence
that the plow was used in Heian times not only in dry-field farming
but also in rice farming.
13 Furushima, Nihon nogyo, vol. 1, pp. 94-95.
14 Iinuma Jiro, "Nihon-shi ni okeru suki to kuwa" (Kyoto daigaku jimbun kagaku kenkyujo),
Jimbun gakuho 32 (March 1971): 10.
15 Engi shiki, in (Shintei zoho) Kokushi taikei, 60 vols. in 66 (Tokyo: Yoshikawa kobunkan,
1929-64), vol. 26, kan 39, pp. 878-81; Wamyo ruiju sho, 3 vols. (Tokyo: Kazama shobo,
1954), vol. 2, kan 15, 8b.
16 Utsuho monogatari, vols. 10-12 of Nihon koten bungaku taikei, vol. 10, pp. 339-40.
Cambridge Histories Online © Cambridge University Press, 2008
AGRARIAN TECHNOLOGY
189
The plow described in Engi shiki was heavier than the Shosoin
plows, and consequently required two persons to operate. The moldboard, which tends to pull the plow off balance, no doubt was one
reason for this. Nonetheless, the Japanese plow was pulled by a single animal, unlike the four to six needed to pull the very heavy plows
of northern European agriculture. The relative lightness of the Asian
plow was a legacy of its origins in dry-field farming in regions where
dehydration of the soil was a major problem. The need to restrict the
number of animals led over rice fields during plowing made it undesirable to develop a heavier plow even for rice farming, where
deeper tilling would have been advantageous. Hence the plow of
Japanese agriculture was more akin to the lighter plow of Mediterranean agriculture than the heavy plow of northern Europe.
Why did the plow, known for centuries before, not become widely
used until the Heian period? Increased attention to dry-field farming was certainly a catalyst leading to its use in rice culture, but this
might have occurred before the Heian period. Knowledge of the
moldboard and its utility in rice cultivation might have been delayed,
but we do not know this for sure: there is, for example, a reference
to the moldboard as early as 772.1? The limited number of oxen
raised in Japan before the Heian period no doubt slowed diffusion of
the plow. But the early ninth-century collection of tales Nihon ryoiki
amply demonstrates that Japanese peasants were familiar with oxen
in the Nara period, although in no case does it show oxen being used
to draw plows.18
Almost certainly the main impediment to diffusion of the plow before the Heian period was the severely limited supply of iron. A
plowshare required much more iron than a simple hoe or spade
blade. Furthermore, the plow was a luxury tool that duplicated, albeit more elegantly, the cultivating job of the hoe.
The spread of iron tools among farm households can be seen in
figures gathered by Harashima Reiji showing the distribution of iron
tools in archaeological sites from the fifth to ninth centuries. It
should be noted that most of these sites are in eastern Japan; iron use
was greater in the central and western parts of the country. Most of
the sites owe their preservation to destruction by natural disaster followed by failure to resettle. Harashima notes that many iron tools
left in these sites must have been lost through corrosion. An even
17 Furushima, Nihon nogyd, vol. 1, p. 94. The reference occurs in a poem in Kakyo hydshiki
(771), cited inTakeuchi Rizo, ed., Nara ibun, vol. 2 (Tokyo: Yagi shobo, 1943), pp. 930-37.
18 See Furushima, Nihon nogyd, vol. 1, p. 59.
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND SOCIETY
igo
TABLE 3 . 1
Distribution of iron tools in farm households,
fifth to ninth centuries
Period"
Goryo (early
5 * C.)
Izumi/Yakuradai
(late 5th C.)
Onidaka (6th to
early 7th C.)
Mama (late 7th
to 8th C.)
Kokubun (gth C.)
Adjusted
percentage
(approximate)
Houses
in sample
Houses with
iron tools
225
5
2.2
5
118
10
8-5
20
223
35
15-7
40
135
25
206
78
18.5
37-9
100
Percentage
40
"Periods refer to the pottery eras to which the sites belong; the time spans in
parentheses are widely accepted approximations, but should not be taken as
incontestable.
Source: Harashima Reiji, Nihon kodai shakai no kiso kozd (Tokyo: Miraisha,
1966), pp. 30 and 314.
more important cause of loss, however, was salvage by the displaced
population, since iron implements were just about the only possessions capable of surviving fire or other natural disaster unharmed.
Hence Harashima's adjustment factor of 2.5 to estimate actual iron
tool rates seems a reasonable minimal adjustment. Harashima uses
this adjustment figure only on his ninth-century figures, but in Table
3.1 it has been extended to earlier periods as well; there seems no
reason to believe the discrepancy between original tool use and what
survives today would have varied from one period to another.
Although the extrapolation from iron tool incidence in archaeological sites to actual tool possession must of necessity be rather
rough, it appears that by the ninth century virtually every farm
household had at least one minor iron tool. (This is so particularly
when we bear in mind how few iron tools must have been abandoned even after homes were destroyed.) It is clearly this increase in
the supply of iron that enabled the diffusion of the plow in the Heian
period. Specifically, it appears that the takeoff point for plow diffusion after the nearly plowless eighth century came only after lesser
iron tools were within the grasp of most peasant households. From
this we may speculate that the principal spur to plow diffusion in the
Cambridge Histories Online © Cambridge University Press, 2008
AGRARIAN TECHNOLOGY
1Q1
Heian period was not excess local elite wealth but a wide demand
among peasants for a more advanced level of technology.
The second major agrarian development of the Heian period was
the practice of keeping animals year-round in stables instead of pasturing them on wasteland. Two factors made this change in animal
husbandry necessary: an increase in number of animals raised owing
to the diffusion of the plow, and a decrease in wasteland available for
pasture. Both factors reduced the capacity of available wasteland to
sustain the number of horses and oxen raised. But of the two, it was
the decrease in available wasteland that most strongly forced owners
to keep and feed their animals year-round in stables.
The early Japanese state, like the aristocracy of medieval Europe,
posed as guardian of the right of public use of wasteland. In 706, for
example, the government issued a decree forbidding monopolization
of public land for private use. Aristocrats were not to block peasant
access to wasteland to collect grasses or firewood, nor were peasants
themselves to block others from free use of uncultivated land other
than land immediately adjacent to their homes and graveyards.19
Public use of wasteland included hunting, food gathering, pasturing
of animals, and collection of wood for building and firewood. Aristocrats, however, were also interested in taking over wasteland for
most of these same uses. Consequently, Nara and Heian aristocrats,
like their European counterparts, began to appropriate large tracts
of waste for private use.
One means of doing this derived from the laws of 743 permitting
permanent possession of rice fields that had been brought under cultivation privately. State-supported temples and some upper aristocratic houses were given large grants of wasteland that they were to
develop into landed estates. Although efforts to create rice fields
from waste often proved fruitless, the land granted these institutions
was nonetheless withdrawn from public access.
From the late eighth century, the central government granted
waste for nonagricultural uses as well. The major recipients of these
grants were temples, shrines, and the imperial household. Significantly, all three were considered part of the "public interest" that
had a natural claim to wasteland use. All three, however, often had
interests that came into conflict with the "lesser" public of peasant
farmers, for the products derived from wasteland grants were the
19 Shoku Nihongi, in Kokushi taikei, vol. 2, Jingo-keiun 3/3/14) pp. 362-63.
Cambridge Histories Online © Cambridge University Press, 2008
192
LAND AND SOCIETY
same as those peasants expected to get from communal wasteland,
such as firewood, lumber, salt, fish, vegetable oils, clay, and chestnuts.20 Private wasteland reserves were also used to pasture animals
and even to raise falcons, although occupation of waste for private
hunting reserves, a major cause of aristocratic encroachment on
public wasteland in Europe, was not a factor in Buddhist Japan.
Appropriation of waste for private use, reclamation for rice cultivation by both estate owners and local notables, and a probable expansion in the proportion of farmland devoted to dry-field crops all
worked to decrease wasteland that could be used by peasants for
pasture of animals. This decrease seems to have been most pronounced in the ninth century. Most wasteland grants were in this period;21 so, too, are the first signs of the changes in tax structure that
made dry fields more attractive to local land developers. By the midtenth century, when plow use had increased greatly, the amount of
wasteland where peasants or even local notables could freely pasture
their animals had been much reduced.
Consequently it became necessary to keep and feed animals in enclosures throughout the year. The most logical source of animal feed
was the stalks of rice and other grains after they had dried to hay. Because allowing animals to forage freely on rice fields after harvest
would have destroyed many of the embankments, rice was harvested
at the ground and the stalks after threshing were brought to the animals as feed. This required an extra step at harvest to separate stalks
from the head. For this reason, rice in earlier times had been harvested at about 15 centimeters from the tip (where the kernels were
concentrated), and the stalks left in the field. The practice came
about partly because ground-level harvesting required the use of a
scythe with an extended handle (known as tokama), which was more
difficult to operate than the simple sickle (kama). Like the plow, the
scythe had been known since the fifth and sixth centuries but was
not widely used until the Heian period.22 Another reason to favor the
simpler harvest method was that rice stalks left in the ground rotted
during winter and could be plowed under in the spring, enriching
the soil. Only the need to keep animals year-round in stables, or to
clear the ground after harvest to plant a winter crop, provided a sufficient inducement to harvest rice at ground level.
20 Okuno Nakahiko, "Hachi kyu seiki ni okeru shiteki tochi shoyusei no rekishiteki seikaku,"
Nihon rekishi, 279 (August 1971): 53-69.
21 See Takeuchi Rizo, ed., Tochi seido shi, vol. 6 of Taikei Nihon shi sosho, pp. 124-28.
22 Furushima, Nikon ribgyo, vol. 1, p. 57.
Cambridge Histories Online © Cambridge University Press, 2008
AGRARIAN TECHNOLOGY
193
In the Nara period rice was still harvested near the tip with a sickle.
An eighth-century register of ceremonies at Ise Shrine describes use
of the long-handled scythe, but only to cut grass.23 The varieties of
mortars and pestles used to process rice harvested with stalks are not
to be found until the ninth century. By contrast, Heian sources
clearly show the harvesting of rice at ground level: both Engi shiki and
Makura no soshi indicate straw was a by-product of threshing, as
would only be the case were rice harvested at the ground.24 It is unlikely that rice was harvested at ground level to clear the ground for
a second crop. Since planting both a summer and a winter crop required a great increase in fertilizer if the soil was not to be quickly depleted, it is probable that few fields were double-cropped until well
into the medieval period. Adoption of the more cumbersome method
of harvesting rice at the ground, therefore, may be taken as a sign of
the need to harvest rice stalks to feed animals kept in enclosures.
Even with the increased use of the plow, it is likely that only a
small number of Heian farm households owned draft animals. But
those who raised animals in stables would have needed more hay
than fields farmed by their own families could have provided. Hence
we may assume that farmers who did not raise animals could sell or
barter hay to those who did. The spread of animal husbandry thus
indirectly created a new source of income for even the poorest farm
households.
An equally important consequence of year-round stabling was the
collection of manure previously lost on wasteland pasture. If the new
harvest method robbed the soil of rice stalks that up to then had rotted and been plowed under, this was surely compensated by the beneficial effects of fertilizing rice fields with manure.
Use of the plow, harvesting of rice at ground level, an increase in
animal husbandry, and the stabling of animals year-round worked
together to transform Heian agrarian life. With the plow it was possible to work heavier soils than could be worked with hand tools, as
well as to loosen and turn soil better and faster. With the plow's
greater speed, it was also possible to turn the soil more frequently
before planting, with a direct effect in increased yields. Greater use
of animals, and most especially their stabling at home, provided manure that also significantly increased soil fertility. Stabling of animals
altered peasant society in other ways as well. Pasturing animals on
23 Kotaijingu gishiki cho, in Gunsho ridju, vol. 1 (Tokyo: Keizai zasshisha, 1904), p. 3. Also see
Furushima, Nihon nogyo, vol. 1, p. 58.
24 See Furushima, Nihon nogyo, vol. 1, pp. 57 and 96.
Cambridge Histories Online © Cambridge University Press, 2008
194
LAND AND SOCIETY
wasteland of necessity took some family members away from the
farm for a substantial part of the year. Now both animals and their
tenders were available year-round at home, thus integrating animal
husbandry more thoroughly into Japanese peasant life.
Perhaps the most valuable single effect of increased use of both
the plow and fertilizer was the securing of year-to-year cultivation of
fields that before had been left fallow after one or more years of cultivation. As use of the plow spread, it became possible to revitalize
fields through better aeration of the soil, more frequent working of
soil before planting, and greater use of fertilizer. This was a considerable improvement in Japanese farming, for the high costs of constructing rice fields were best recovered by continuous year-to-year
cultivation.
PEASANT COMMUNITY
The task of transplanting rice seedlings in early summer was arduous, but could not be time-consuming: as each field was flooded and
the soil worked into a mush, the full complement of seedlings had to
be planted while the soil was neither too hard nor too liquid, and before the next field, flooded at no little expense, required similar attention. The intensive work of harvest in autumn could be staggered
by planting both early- and late-maturing rice, a technique well
known in the Heian period.25 But the work of transplanting rice
seedlings required more hands than even a large peasant household
could muster. In more recent times Japanese farmers commonly
pooled labor for transplanting, the entire village turning out to plant
first one field, then another, until all were planted. Such communal
effort did not come just from a desire for cooperation: it came from
a strong village organization. Such an organization is nowhere evident in Heian or pre-Heian Japan, and in fact does not develop until
the fifteenth and sixteenth centuries.26
Instead, transplanting was accomplished with the help of hired
labor. The Tosa farmer in the Konjaku story quoted earlier took hired
laborers and food provisions to the "opposite shore" along with rice
seedlings, a plow and other tools, and his own family. The very term
for the labor organization used in transplanting, yui, was associated
25 Furushima, Nihon nogyo, vol. 1, pp. 161-63; alsoToyoda, Sangyo shi, pp. 189-91.
26 Nagahara Keiji (with Kozo Yamamura), "Village Communities and Daimyo Power," in
John Whitney Hall and Toyoda Takeshi, eds., Japan in the Muromachi Age (Berkeley and
Los Angeles: University of California Press, 1977), pp. 107-23.
Cambridge Histories Online © Cambridge University Press, 2008
PEASANT COMMUNITY
195
with the term yatou, "to hire."27 Government rice loans in the early
summer (discussed in the next section) quite evidently were related
to the need to pay laborers in rice to help with early summer transplanting.
Very likely farmers hired their neighbors to help in rice transplanting, and were in turn hired to help transplant their neighbor's
crop. Such exchange did not develop into an overall communal labor
organization, however, because the political "village" unit by and
through which to organize such exchange did not yet exist. Households were often grouped into identifiable hamlets, with shared economic interests and even shared worship at the same shrine, but no
village structure arose to integrate hamlets into separate, corporate
territorial bodies, nor was any imposed from above by the government. The units of local administration created by central and
provincial governments in the Nara and Heian periods bear no relation to any imagined "natural village" unit.28
As a result, Heian peasant households were larger than those of
more recent times, maintained a larger and steadier supply of labor
within the household, and generally maintained a greater degree of
independence one from the other. This relative independence is evident even in the physical layout of Heian hamlets. Instead of lying
tightly clustered, as in most Tokugawa and modern communities,
houses were distributed in small, loose clumps, or else stood entirely
apart from each other.29
The well-excavated site at Hiraide in Nagano prefecture reveals a
community of sixteen houses dating from the eighth or the ninth
century. As shown in Figure 3.1, the sixteen houses were spread over
an arc of nearly 800 meters. Only in one area, near a shrine, are seven
found loosely grouped together. Yet even this group was spread over
an area about the size of a modern city block, and nearly five city
blocks separate the two extreme houses of the hamlet.
A similar distribution of houses can be found in other archaeological sites from this period.30 The scattered nature of peasant hamlets can be seen also in land surveys from the late Heian period. Two
of the best-studied cases are shown in Figures 3.2 and 3.3. Although
the twenty households of Wakatsuki-no-sho lie within a 300-meter
radius of each other, they show only the loosest of clustering. The
27 Toyoda, ed., Sangyo shi, pp. 192-93. 28 Morris, "Peasant Economy," pp. 108-19.
29 Kinda Akihiro, "Nara Heian-ki no sonraku keitai ni tsuite," Shirin 54, 3 (May 1971):
80-108.
30 Kinda, "Nara Heian-ki," p. 114.
Cambridge Histories Online © Cambridge University Press, 2008
196
LAND AND SOCIETY
A
A
A
A
A
A
M
/
y
A
*
•
/
A
A
A
—
-
~
_
_
m
1
0
1
100
i
200
Figure 3.1. Eighth- and ninth-century house sites at the Hiraide
site. Adapted from Hiraide iseki chosakai, ed., Hiraide, pp. 258-59.
Figure 3.2. Houses inWakatsuki-no-sho,Yamato Province (twelfth
century). Adapted from Kinda Akihiro, "Nara Heian-ki no sonraku
keitai ni tsuite," p. 88.
eighteen households of Kohigashi-no-sho are even more scattered:
only in two areas can three houses be found within 150 meters of
each other.
Without organized communal labor cooperation, peasant households had to be as economically self-sufficient as they were physiCambridge Histories Online © Cambridge University Press, 2008
PEASANT COMMUNITY
197
Figure 3.3. Houses in Kohigashi-no-sho, Yamato Province (1144).
Adapted from Inagaki Yasuhiko, "Shoki myoden no kozo," p. 10.
cally separate. This can be seen in the great size of Nara and Heian
households. The stem family pattern ofTokugawa Japan produces an
average household size of four to, at most, seven persons; under the
modern nuclear family pattern the average is far less. By contrast,
surviving population records from the early eighth century (the only
source for such information) reveal an average household size of
eight to ten persons.31 While not as large as the average of nearly
twenty persons that the appearance of these records has led some
31 Morris, "Peasant Economy," pp. 67-73.
Cambridge Histories Online © Cambridge University Press, 2008
198
LAND AND SOCIETY
historians to surmise, this is still half again larger than the average
size of early modern Japanese households, and larger still than any
found in premodern Europe.
The principal cause of this enlarged household size was the practice of duolocal marriage. In this pattern, familiar among the Heian
aristocracy, husband and wife continued to live in their households
of birth even after marriage, but the husband visited the wife periodically at night. The children of the marriage were raised in the
mother's household.
In some of the communities found in eighth-century population
records the duolocal pattern of marriage is found for several years
after marriage, after which the wife and children were brought into
the husband's household. But in other communities husbands and
wives lived apart for most or all of the marriage. Of the four regions
found in this sample, the incident of duolocal marriage among newlyweds varied from 25 percent (Kyushu) to 95 percent (Shimosa, in
the east), and the average length of the duolocal phase from five or
six years (Kyushu) to more than fifteen years (the home provinces).32
The effects of duolocal marriage on peasant family life must have
been profound. Children were raised in a household where there
were always other adults to care for them, freeing the mother for
farmwork. The mother, in turn, was likely to have greater autonomy
in her own natal household than in a household headed by her husband or by in-laws. Her husband, held in his household of birth
when he might otherwise have founded an independent household
of his own, was allowed correspondingly less autonomy. Overall,
duolocal marriage served to strengthen and prolong control by the
older generation over the fortunes of the young.
By delaying the departure of married children, practice of duolocal marriage served to retain adult labor in the household. At the
same time, splitting of households by noninheriting sons was delayed until the branch households consisted of a more substantial
body of individuals. It was this double effect that produced the enlarged average household size of Japanese peasant families. Instead
of splitting apart as children reached adulthood and married, setting
up branch households too small to manage a farm independently,
households retained an ample supply of labor during exactly that
stage in the family cycle when labor supply would otherwise have
been at its lowest. The practice of duolocal marriage by some family
32 Morris, "Peasant Economy," pp. 76-80.
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
199
members thus led not only to an enlarged household labor supply
overall, but to a steadier supply of labor throughout the family
cycled
In an era when the diffuseness of the peasant community offered
only limited support to the individual family, such labor considerations were essential if households were to survive as successful farming units. By the late Heian period, however, households with access
to use of a plow and draft animal may have been able to function
with correspondingly less human labor. Since a single plow and animal could be spread among several households - indeed, had to be
if it were to be cost-effective - shared plow use must have promoted
a more tightly organized sense of community than is evident at the
beginning of the Heian period. Hence, greater cooperation during
the transplanting season when plows were in use, led by the plow
owner, may have spread to other periods of the agricultural calendar
as well. We might therefore expect to find signs of more tightly organized peasant communities by the end of the Heian period.
There are, in fact, some signs of more tightly clustered hamlets in
the late Heian period, but these cases are exceptional. Over most of
Japan, compact physical communities did not appear until after the
fourteenth century.34 If plow use promoted greater cooperation of
labor and hence a stronger sense of peasant community, we must
wait for signs of this in the Kamakura period, when plow and animal
use spread most rapidly through Japanese agriculture.35 Until then,
Japanese peasant households continued to maintain a striking level
of physical and economic autonomy from each other.
TAX STRUCTURE
Japan's first national tax system, conceived in the late seventh century, reflected what might be called a manorial view of the economy.
The state claimed ownership of all rice fields, and expected that they
would be planted with seed provided by an ancient system of rice
loans known as suiko. State revenue was to be collected in kind, as
raw materials, handicraft items, and labor. By expressing their tax
needs in terms of industrial products and labor, aristocrats in the
capital were spared the necessity of obtaining from an uncertain
market products essential to their way of life. The new tax levies were
33 Morris, "Peasant Economy," pp. 92—95. 34 Kinda, "Nara Heian-ki," pp. 86-103.
35 Furushima, Nihon nogyo, vol. 1, pp. 197-200.
Cambridge Histories Online © Cambridge University Press, 2008
2OO
LAND AND SOCIETY
all related to rights traditional to the state from ancient times: the
right to direct and promote rice culture, the right to demand tribute
from regions not under direct imperial control, and the right to demand corvee labor or payment in lieu of labor from all adult male
commoners.
Far from turning Japan into a kind of giant manor, however, the
new tax system helped stimulate industrial and commercial development to the point that the manorial conception became no longer
necessary to meet the aristocracy's needs. An increasing awareness of
this in the Heian period led the central government to adopt a more
functional view of the nation's economy and tax structure. Aristocrats in the capital started to view the extraction of tax revenue as a
problem between the central and provincial governments rather than
between the capital and the commoner population. They came to
make their demands for increased revenue not directly of peasants,
but of provincial governments. Provincial officials were given incentives to increase the amount of rice lent as suiko - the most flexible
source of revenue - and new local officials were deputized to expand
suiko and ensure steady production of other tax income. Tax quotas
were set for each province, with provincial officials responsible for
seeing that they were met. To a great extent the capital abandoned
the actual means of tax collection to provincial and local government. This distanced the central government from the peasantry and
greatly enlarged the autonomy of provincial officials in tax collection.
Traditionally, these changes in tax structure have been seen as just
so many steps in the breakdown of centralized, bureaucratic government. Yet each change in taxation in the Nara and Heian periods that is, the expansion of suiko loans, the provincial quota system, the
manufacture or purchase of handicraft products at the provincial
rather than the local level, and the conversion of suiko to a direct land
tax - was a change that better suited the capital's revenue needs, better ensured that these needs would be met, and increased the efficiency of tax assessment and collection. Far from indicating a breakdown in the organs of government, these changes should be seen as
elements of a rationalization of the national tax structure, better fitting it to the Heian rural economy.
Suiko
From ancient times Japanese governments at the local and national
levels had made rice loans in the spring or summer that were repaid
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
2O1
after the autumn harvest at a substantial rate of interest - 50 percent,
in the eighth century. These loans came to be known as suiko. Their
chief purpose seems to have been the promotion of rice culture. But
they had long become equally important as a source of government
revenue. Since neither objective - rice planting and state revenue - was considered optional, the loans involved some degree of
coercion, although there was frequently also peasant demand for
suiko. In times of famine, borrowers were often excused from paying
interest, which allowed suiko to function occasionally as a welfare
system as well.
By the early eighth century, income from suiko interest was at least
equal to the revenue from the second source of income in rice, the
land tax (denso). In the 730s the central government decided to increase suiko income still more, as the growing capital and the expanding provincial governments found existing tax revenues inadequate. This it did by directing provincial officials to increase sharply
the amount of rice lent as suiko, allowing them to keep a portion of
the new interest income as a supplement to their stipends.36 In 745
rough quotas of suiko were set for each province, and a variety of
stipend-5t«'&0 (called kugeio) was devised that made permanent the
use of suiko income as a supplement to the stipends of provincial officials. The idea was that officials would make up from stipend income any deficiencies in the suiko quotas that had been set for meeting provincial expenses; anything that remained could then be
divided among them as supplemental income.37
Unhappy with the job provincial officials were doing as tax collectors, the central government at the beginning of the Heian period
tightened the stiptnd-suiko system to prevent officials from taking
their supplemental stipends without first making up deficiencies in
meeting suiko quotas. In 790 it was decreed that a minimum of 10 to
15 percent of stipend-smTso (varying by province) was to be applied
to suiko deficiencies from previous years for as long as such shortfalls were outstanding.38 A measure of 803 extended this approach to
the use of stipend-swz&o income to pay wage and travel expenses of
provincial corvee laborers.39 These measures served to make standard the amounts from stipend-swz&o income that were to be used to
make up suiko deficiencies. The remainder was thus allowed to be36 Sonoda Koyu, "Suiko: Tempyo kara Engi made," in Osaka rekishi gakkai, ed., Ritsuryd
kokka no kiso kozo (Tokyo: Yoshikawa kobunkan, i960), pp. 412—17.
37 Shoku Nihongi, Tempyo 17/10/5, pp. 184-85. 38 Shoku Nihongi, Enryaku 9/11/3, p. 549.
39 Enryaku kotai shiki, in Kokushi taikei, vol. 26, Enryaku 22/2/20, p. 15.
Cambridge Histories Online © Cambridge University Press, 2008
2O2
LAND AND SOCIETY
come what many officials apparently treated the suiko stipend as
from the beginning: a simple, and very lucrative, stipend system.
With the institution of stipend-suiko, the central government
began contracting provincial governments to collect tax revenue by
quota rather than by assessment of the tax base. This could be justified because suiko, although it acted as a harvest tax, was in theory
a loan rather than a tax. Hence, increased suiko, however coercive,
could be justified by the legal fiction that peasants needed the extra
loans. In the course of the ninth century, with this precedent, no further legal fictions were deemed necessary to extend provincial quotas to other taxes as well.
Of equal importance were changes in how suiko loans were made.
In 795 the interest charge was reduced from 50 to 30 percent; at the
same time, suiko debts were no longer forgiven when borrowers died
before repayment.40 This latter policy had, in the 730s, reduced the
effective return on suiko to about 40 percent, largely because of
canny borrowers who took out loans in the names of persons already
dead.41 It is entirely possible that by 795 the effective return had
dropped to close to 30 percent, so that there may have been little
change in actual suiko income under the new rules. The new approach had a major impact nonetheless, for it was no longer necessary that suiko loans be secured only by individual borrowers.
This meant that suiko could now be "levied" as a straight tax. In
807, for example, the circuit inspector for the northern T5sando
provinces recommended that suiko be levied by household size.42
The following year the central government proposed a more complex approach: each adult male taxpayer would be lent 10 to 100 soku
(36 to 360 liters) of rice, half in spring and half in summer, on a sliding scale according to wealth - that is, the wealthier, though less
needy, would be lent more because of their greater ability to repay.43
The growing recognition of suiko as a tax made it possible to assess suiko by unit of rice cultivation. Levy by land area presented several advantages over securing loans by individuals. As long as suiko
was lent to individuals, it was tied to its traditional justification as a
40 Ruiju sandai kyaku, in Kokushi taikei, vol. 25, kan 14, Enryaku 14HM.7I1, p. 396; Ruiju
kokushi, in Kokushi taikei, vol. 5, kan 83, Enryaku 15/1111.7/21, p. 451. There was a brief return to 50 percent interest from 806, a year of crop failure, until 810, thefirstyear of good
harvest after 806: Nihon koki, in Kokushi taikei, vol. 3, Daido 1/1/29, p. 51; and Ruijusandai
kyaku, kan 14, Konin 1/9/23, p. 396.
41 See FunaoYoshimasa, "Suiko no jittai ni kansuru ichikosatsu: Bitchu no kuni taizei ootaru
shibonin cho o chushin to shite," Shirin 56,5 (September 1973): 74-102.
42 Ruiju kokushi, kan 83, Daido 2/9/21, p. 455.
43 Ruiju sandai kyaku, kan 14, Daido 3/9/26, pp. 395—96.
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
203
welfare measure and so might be resisted by unwilling borrowers.
But once suiko was assessed by unit of rice cultivation, the welfare
pretense was abandoned and suiko became a variety of harvest tax.
Administration of suiko was also made simpler, for it was now assessed on the same basis as the land tax.
The first evidence of suiko assessment by land area is from
Kawachi Province in 822.44 By the end of the ninth century requests
from provinces for permission to levy suiko at the rate of 5 soku (18
liters) of rice per tan of land (0.1 hectare) had become common.45
In 901 the central government instructed all provinces to lend suiko
at this rate as necessary to ensure that suiko quotas were met.46 At 30
percent, suiko interest at the official rate of levy came to 1.5 soku per
tan, which was identical to the land tax (denso) rate.
The central government welcomed the levy of suiko loans, as this
better ensured that provincial suiko quotas would be met. After 900,
provinces were held much more strictly responsible for seeing that
these quotas were met. Quotas were now known as shiki quotas
{shikisu), after the administrative guidebooks, most particularly Engi
shiki (927), in which the quotas were set. From about 905 provinces
were required to petition for permission to lend and collect interest
on less than the quota amounts.47 The amount of the reductions allowed in these quotas is known for seven provinces between 945 and
1093: reductions varied from 10 to 70 percent and were greatest in
the home provinces, and least in the east.48
The next step after levy of suiko loans by land area was to delete
the loan entirely and levy the interest as a direct tax.49 In concept
this was a giant departure from the use of suiko as a welfare measure
and aid to agriculture. In practice, however, it was the logical outcome of the attitude, two centuries old, that acceptance of suiko
44 Ruiju kokushi, kan 83, Konin 13/12/28, pp. 456-57.
45 Sandai jitsuroku, in Kokushi taikei, vol. 4, Genkei 5/3/14, p. 495; Ruiju sandai kyaku, kan 14,
Kampyo 6/2/23, P- 402> ar>d ^an 20> Shotai 4/int.6/25, pp. 636-37.
46 Sandai jitsuroku, Genkei 5/3/14, p. 495.
47 Ruiju sandai kyaku, kan 14, Engi 5/12/25, p. 398, and kan 15, Engi 5/12/25, pp. 397-98;
Besshit fusensho, in Kokushi taikei, vol. 27, Engi 7/8/11, suppl. p. 4.
48 See Abe Takeshi, Ritsuryo kokka kaitai katei no kenkyu (Tokyo: Shinseisha, 1966), pp.
123-27.
49 Levy of suiko interest without loan of principal is reported by Owari Province in 988:
Takeuchi Rizo, ed., Heian ibun, 13 vols. (Tokyo: Tokyodo, 1963:0274), vol. 2, no. 339 (Eien
2/11/8), pp. 473-85; by Yamashiro Province in 1001: Chbya gunsai, in Kokushi taikei, vol.
29,1, kan 26, Choho 3, pp. 533-34; by Tosa Province in 1004: Chbya gunsai, kan 26, Kanko
1/11/20, pp. 534-35; by Kozuke Province in 1030 or 1031: Heian ibun, vol. 9, no. 4609
(Chogen 3 or 4), pp. 3511-12; by Kozuke Province again in 1076: Chbya gunsai, kan 26, Joho
3/12/15, p. 541; and by Sagami Province in 1093: Chbya gunsai, kan 26, Kanji 7/6, pp.
539-4".
Cambridge Histories Online © Cambridge University Press, 2008
204
LAND AND SOCIETY
loans was an obligatory benevolence, and that the state could legitimately count on collecting the same level of suiko interest income
each year regardless of peasant demand or need for loans.
But however logical from the point of view of government needs,
such a major departure from the traditional justification of suiko was
not easily made. What led to the transformation of suiko into a direct
land tax was not just central government demand for steady "interest" income, but, rather, this factor coupled with the depletion or
loss of suiko principal. Time and again, mid-Heian provincial government documents explain that suiko principal was no longer lent
because it was either "totally nonexistent" or in the hands of local
tax managers.50 Most commonly, provincial officials transferred
suiko principal to local tax managers in return for promise of a guaranteed level of "interest" income. Whether local tax managers chose
to lend rice to produce this interest or to keep it and extract "interest" as a direct land tax was up to them. In the latter case the suiko
supplies became "totally nonexistent," but the provincial government received the same level of suiko income nonetheless.
It is not clear how early the transfer of suiko rice from provincial
to local officials occurred, but eleventh-century sources speak of
this transfer as already "long ago."51 From the early ninth century
we find references to two related practices by which local suiko
lenders took control of suiko rice stores to become thereafter permanent managers of local tax collection.52 The first, known as
"partial repayment" (hankyo)> is explained by Sugawara no Michizane: a portion of the suiko supply, theoretically half but possibly
more, was kept by the local tax manager, but interest was delivered
in full.53 The practice of the second, "falsified repayment" (kyond),
was similar: lesser grains, or even straw, were substituted for suiko
principal, but interest was delivered properly in rice. In both cases
suiko principal was retained by local tax collectors with provincial cooperation, but undoubtedly the idea originated with local
officials.
At first the central, and some provincial, governments looked
askance at local control of suiko principal. But by the eleventh cen50 For example, Heian ibun vol 9, no. 4609 (Chogen 3 or 4), pp. 3511-12.
51 Heian ibun, vol. 9, no. 4609 (Chogen 3 or 4), pp. 3511-12.
52 The first reference to these practices {hankyo and kyond) is Nihon koki, Konin 6/12/29, pp.
136-3753 Kanke bunso, kan 4, Kampyo 3/7/5; quoted in Murai Yasuhiko, Kodai kokka kaitai katei no
kenkyii (Tokyo: Iwanami shoten, 1965), pp. 30—31.
Cambridge Histories Online © Cambridge University Press, 2008
TAX
STRUCTURE
205
tury high officials no longer seem concerned over the loss of suiko
principal, even that which local officials themselves admitted was
long gone. This is clear, for example, from a document of 1080 or
1081 from the eastern province of Kozuke, which shows that local
tax managers were committed to certain levels of suiko payment to
the provincial government regardless of the presence or absence of
suiko principal.54 This source, in fact, also claims that the capital
authorized transfer of provincial suiko rice to local officials as early
as 966.
It appears it was local tax managers, then, who initiated the collection of suiko interest as a direct land tax without loan of principal.
Already in the ninth century the "true" land tax {denso), was coupled
with provincial suiko (generally known as shozei) to form a single
land tax known as sozei.55 At 30 percent interest, as we saw, the return on suiko levied at 5 soku per tan was 1.5 soku per tan. This was
identical to the rate at which the denso land tax was levied, so that
the combined sozei tax came to 3 soku per tan. This amounted to 8
to 9 percent of a typical crop yield of 35 soku per tan.
It was clearly to the advantage of local tax managers to levy suiko
"interest" without making actual loans, as this freed them to put
suiko rice stores to their own private use. Less obvious, perhaps, is
that this was to the provincial government's advantage as well, as it
freed provinces to transfer tax collection authority to new officials if
earlier tax managers failed to deliver the level of land tax income assessed of their districts. Were the lending of suiko principal still necessary to collect suiko interest, provincial governments would have
been unable to transfer local tax collection authority without somehow inducing former officials to give up suiko stores to new officials.
Without provincial support, local officials might have been unable to
force peasants to pay suiko interest as a direct land tax without first
making suiko loans. But with provincial and, ultimately, central government officials backing local tax managers, there was little the
tenth-century peasant could do to resist the transformation of suiko
into a direct land tax - the obligation to accept suiko loans and pay
suiko interest regardless of peasant need had been a recognized state
right for at least two and a half centuries.
54 Heian ibun, vol. 9, no. 4609 (Chogen 3or 4), pp. 3511—12 i2f.
55 See, for example, Monwku jitsuroku, in Kokushi taikei, vol. 3, Kajo 3/4/17, pp. 7-8, and
3/4/24, p. 9; Ruiju sandai kyaku, kan 8, Kampyo 5/5/17, p. 343; and Heian ibun, vol. 2, no.
339 (Eien 2/11/8), pp. 473-85.
Cambridge Histories Online © Cambridge University Press, 2008
2O6
LAND AND SOCIETY
The handicraft-produce taxes
The eighth-century tax structure included two taxes that were paid
in handicraft or industrial products such as silk cloth, silk floss, linen
or paper mulberry cloth, iron, salt, fish, seaweed, and many other
local products. One part of Japan, however, departed sharply from
this essentially manorial approach to taxation: in the central
provinces, both taxes were commonly paid in cash.
The cho tax was assessed on adult taxable men, with a full share
paid by men aged 21 to 60, a half share by men 61 to 65, and lesser
amounts by men 17 to 20 according to a separate schedule. The yd
tax, paid in a lesser variety of goods, was set generally at half the cho
rate, and was due from men aged 21 to 65 only, assessed as for cho.
Technically, yd was the commutation of a corvee labor requirement
of ten days a year (five days for men 61 to 65), but, as in China, the
corvee requirement seems to have been only theoretical. From the
beginning, therefore, the two taxes were of one nature. Although
the requirement could be satisfied by products in kind as well as by
handcrafted items, for convenience we will refer to the cho and yd
taxes as handicraft taxes, after their most important element.
The original plan was that handicraft items would be produced by
the taxpayers from whom they were due. But already in the eighth
century there is evidence that taxpayers pooled their resources to
hire handicraft specialists to make the items required.56 This soon
became the key principle by which the handicraft-tax system was
transformed in the ninth and tenth centuries.
From about 770 the central government began to complain of a
serious decline in both the quality and quantity of handicraft tax
products. In 785 several provincial officials were punished for sending payments that were both late and of poor quality.57 In 797 the
capital complained that hoes sent as handicraft payments were so
thin and of such poor material that they were utterly unsuited to
agricultural use; provincial and district officials were accused of
keeping all the good hoes for themselves.58
Several steps were taken to prod provincial officials to more diligent delivery of the handicraft taxes. Officials were to pay more attention to registering migrants (Jiirdntn) to be sure that persons farming land in shoen (estate) property did not evade tax payments. More
56 Abe, Ritsuryo kokka, p. 141.
57 Shoku Nihongi, Enryaku 4/5/24, p. 508, and 4/7/28, p. 511.
58 Ruiju sandai kyaku, kan 7, Enryaku 16/4/16, p. 330.
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
207
concretely, the government required that provincial officials after 795
make up from their suiko stipends any deficiencies in payments of silk
(due from handicraft taxes) or rice (due from the land tax).59
Nonetheless, complaints continued that deliveries of the handicraft taxes were late, incomplete, and of poor quality. This led the
central government to institute a major overhaul of the handicraft
tax system in the 840s. At some time before 846, each province was
instructed to make a final measurement of its number of taxable
men, to be used thereafter as a fixed rate by which each province's
handicraft tax liability was to be calculated. As with later suiko quotas, each province had to petition the capital for permission to reduce the number of taxable men used to calculate its handicraft
dues.60 In 846 provinces were also instructed to levy each year 10
percent of the accumulated deficiencies in previous years' payments
of the two handicraft taxes, to be paid on top of what was required
for the current year.61
At either the same time or somewhat later, the central government also set for each province minimum quotas of specific products that were to be included in its annual handicraft tax payments;
these quotas can be found in Engi shiki.62 By these measures the
central government withdrew from concern with the assessment
and collection of the two handicraft taxes, and left the provincial
governments responsible only for meeting specified handicraft tax
quotas.The new system was apparently a success: from this time the
capital rarely complained of lateness or poor quality in handicraft
tax products.63
By setting quotas for the handicraft taxes, the central government
left provincial and local officials free to jettison assessment by population in favor of more efficient methods of collection. Tax liability
could be commuted to payments in rice, which were used to purchase products required by the quotas at the marketplace, or pay for
their manufacture at workshops run by provincial or local officials.
By the early tenth century the central government had set provincial quotas for products outside of the handicraft taxes proper. Engi
59 Ruiju sandai kyaku, kan 8, Enryaku 4/12/9, pp. 339-40, Enryaku 16/8/3, p. 340, and Enryaku
21/8/27, P- 33 1 ; Enryaku kotai shiki, Enryaku 14/7/27, pp. 15-17.
60 Nishibeppu Motoka, "Kyu seiki chuyo ni okeru kokusei kicho no tenkan ni tsuite," Nihonshi kenkyu 169 (September 1976): 30-54.
61 Ruiju sandai kyaku, kan 8, Showa 13/8/17, p. 342.
62 The quotas of items provinces were to include in their annual end and yd payments are
found in Engi shiki, kan 24, pp. 597-622.
63 Nishibeppu, Kyu seiki choyo," pp. 42-43.
Cambridge Histories Online © Cambridge University Press, 2008
2O8
LAND AND SOCIETY
shiki records quotas of paper, brushes, raw linen, leather, and tools
that were to be purchased by provinces with income from suiko.64
These quotas belong to a developing category of taxation, known as
"miscellaneous offerings" (koeki zomotsu), that first attained real importance in the tenth century. Originally such exactions were "tribute" offerings by district officials igunji), who were to purchase them
with income from the suiko stores under their control (gunto). When
district suiko was added to provincial suiko stores in 712, the responsibility for miscellaneous offerings was transferred to the provincial
governments. By the tenth century there seem to have been at least
three ways to meet quotas for miscellaneous offerings: provincial
governments either sent directly to the capital the rice necessary to
purchase the products required, or levied an additional land tax
equivalent to price of miscellaneous offerings and used the proceeds
to purchase these at the marketplace, or levied the items required directly as a tax in kind.65
In short, quotas for miscellaneous offerings were met in just the
same way as quotas for the handicraft taxes proper: direct levies in
kind, indirect levies as an added land tax, purchase at the marketplace, or, in at least some areas, manufacture in government workshops. Despite their different origins, therefore, beginning in the
tenth century the original handicraft taxes and the newer (though
with roots as old) miscellaneous offerings may be considered as one
system. "Tribute" from local officials (koeki zomotsu), "tribute" from
the commoner population (cho), and commutation payments for
labor owed die state by commoners (yd) had all been transformed
into a shopping list of quotas of goods that provinces were to deliver
to the capital.
Corvee
The final component of the eighth-century tax structure to be considered is the provincial corvee labor requirement (zbyo). Corvee
duty was not to exceed sixty days a year for taxable men 21 to 60 (or
thirty days for men 61 to 65, fifteen for men 17 to 20). A taxpayer
could hire others to serve in his place. Generally, it is diought,
64 "Nenryo bekko zomotsu" and "nenryo zakki": see Engi shiki, kan 26, pp. 586-87.
65 Heian ibun, vol. 2, no. 339: Eien 2/11/8, pp. 473-85, and Seiji yoryaku, in Kokushi taikei, vol.
28, kan 57, Tengyo 2/UU.7/5, pp. 437-38. Kbeki zomotsu quotas are given in Engi shiki, kan
26, pp. 591-94-
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
2Og
corvee labor was at first uncompensated, although the evidence either way is rather thin.66 In the course of the eighth century, however, an increasing number of corvee laborers were paid at least subsistence wages.
In 795 the central government complained that provincial and local
officials invariably used corvee labor up to the maximum sixty days
allowed per adult male taxpayer. Since wealthy peasants could pay
others to serve for them, only the poor were actually taken away from
personal business by labor service. This encouraged officials to levy
the maximum so as to collect the commuted payments of those who
could afford not to serve. For this reason, the central government reduced the maximum corvee levy from sixty to thirty days a year.67
From 808 the central government began to require that provincial
officials submit annual reports to all uses of corvee labor. This
brought a temporary halt to the long tradition of local freedom in
the levy of corvee.68 In 822, for the first time since the Taiho code a
century earlier, a national remission of corvee was declared. Ostensibly so that arrangements could be made to hire wage labor this
year, each province was instructed to report for what uses corvee
labor was absolutely essential. But, in fact, the central government
used these reports to prepare a table of legitimate uses of corvee
labor and the number of workers who could be employed for each
use. A few categories of work - repair of government buildings, construction of irrigation canals and embankments, courier service for
government business, and the like-were left to the discretion of
provincial officials. To use more than the prescribed number of
workers for any other purpose, however, provinces had to apply for
permission from the capital.69
With its newly established control over corvee labor, the central
government issued in the 840s and 850s a series of reductions in the
maximum labor requirement. In 864 the requirement was reduced
permanently to twenty days a year.70 As a result of these strictures
on the use of corvee, an increasing amount of government work was
handled by more highly compensated hired labor, financed by income from suiko interest. By the tenth century the levy of corvee
properly speaking had all but disappeared.
66
67
68
69
70
Abe, Ritsuryo kokka, pp. 210-12.
Ruiju sandai kyaku, kan 17, Enryaku \^l\nx..-jl\=,, p. 517.
Nihon koto, Daido 3/8/6, p. 76.
Ruiju sandai kyaku, kan 6, Konin i3/int.g/2o, pp. 278—80.
Sandai jitsuroku, Jogan 6/1/7, PP- 121-3,Jogan 6/1/9, p. 124, and Jogan 6/1/29, p. 130.
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND SOCIETY
210
CENTRAL GOVERNMENT
various services and products
(including miscellaneous
offerings and some rice
from the land tax)
(transmitted
directly)
PROVINCIAL GOVERNMENTS
/
t
Tax:
corvee
suiko interest
land tax
handicraft taxes
Form of
payment:
labor
rice
rice
products handcrafted
and in kind
Unit of
assessment:
adult
male
"demand" for
rice loans
area of
rice paddy
adult male
Figure 3.4. Tax structure in the late seventh and eighth centuries.
The new tax structure of the tenth century
The changes in Japanese tax structure from the seventh to tenth centuries that have just been discussed are summarized in Figures 3.4,
3.5, and 3.6. It will be noted that in the tenth century the former
taxes on land (denso), on handicraft produce (cho and yd), and the
quasi-tax suiko were all reorganized into a single, enlarged land tax
(nengu, or kammotsu). To this unified land tax was added a new set
of duties, levied primarily by provincial and local officials, that were
known variously as "occasional exactions" (rinji zoyaku) or "public
duties" (kuji, or kuniyaku).
The most detailed source for the rate and structure of the enlarged land tax of the tenth century is the famous Owari Province
petition of 988, in which Owari district officials and other local notables requested that their governor, Fujiwara no Motonaga, be removed from office for gross misconduct.71 Because of the nature of
the petition, it is necessary to disentangle those taxes described
which were the norm for all provinces from levies that were the particular excesses of Motonaga. Fortunately, the petitioners made
clear comparisons to neighboring provinces and gave the total land
tax rate for both Owari and neighboring provinces.
The land tax described in the first six articles of the Owari
71 Heian ibun, vol. 2, no. 339: Eien 2/11/8, pp. 473-85. Also see Abe Takeshi, Owari no kuni
gebumi no kenkyu (Tokyo: Shinseisha, 1971).
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
211
CENTRAL GOVERNMENT
handicraft products and
products in kind (handled
. by provincial governments)
various services and
products (as above)
PROVINCIAL GOVERNMENTS
t
Tax:
Form of
payment:
(wage labor,
paid by suiko
interest)
Unit of
assessment:
\
suiko interest
land tax
handicraft taxes
rice
rice
handicraft products
area of rice
paddy or by
quota
area of
rice paddy
quota, based on
one-time-only
census of
adult males
Figure 3.5. Tax structure in the ninth century.
CENTRAL GOVERNMENT
services and products by y
quota (paid from land
tax income)
handicraft products by quota
(purchased from market or
made in provincial
workshops)
PROVINCIAL GOVERNMENTS
(local
tax
\
m a nnaagg e r s )
\ \
/
land tax (nengu, or kammotsu)
(rinji zoyaku, kuji,
or kuniyaku)
(sozei)
additional land
tax (kacho)
"occasional exactions,"
or "public duties"
Form of
payment:
rice
rice
rice, products;
sometimes labor
Unit of
assessment:
area of rice paddy
various, including
area of rice paddy
various, including
by adult male
Tax:
unified land tax
Figure 3.6. Tax structure in the tenth century.
Cambridge Histories Online © Cambridge University Press, 2008
212
LAND AND SOCIETY
province petition was composed of five parts. The first, the old land
tax (denso), was imposed at 1.5 soku per tan. The second, the regular
suiko levy, was based on the suiko quotas set for Owari by the capital. The original quota, as found in Engi shiki, would have produced
a levy of 4 soku per tan of "loan," or 1.2 soku of interest.72 This had
been officially reduced, however, to a quota producing a levy of 0.6
soku per tan of "interest" for which no principal was loaned (Article
One). There was also, however, an "additional suiko levy" of 2.8 soku
per tan to be used for government expenses (Article Four). Other
provinces also levied such additional suiko dues; the complaint of the
Owari petitioners was that Motonaga pocketed the additional levy
for himself. On top of these three legitimate taxes, Motonaga was accused of levying a "secret extra suiko levy" that came to 1.1 soku per
tan (Article One).
The fifth and final element of the tenth-century land tax was an
"additional levy" {kachd) used to purchase the handicraft products
required by central government quotas. In Owari the cho handicraft
tax was assessed by assigning 24 tan of rice fields to produce, at 4
soku per tan, the rice necessary to purchase each bolt of silk required.
In fact, Motonaga manipulated this system so that only 11 tan of
land was to pay for one bolt of silk, forcing local officials to extract
over twice the proper levy (Article Six). In addition, Motonaga
forced producers of handicraft items to accept only half the market
price for silk and other cloth, and to deliver bonus cloth for his own
use (Article Five). Motonaga also made local officials levy kachd
twice, collecting rice a second time to be ground into brown rice
(shomai) that met other state quotas (Article Three). These brown
rice quotas should have been met from the first additional suiko levy
(2.8 soku per tan), but, again, Motonaga was accused of keeping this
income for himself. Hence the total kachd levy of 7.2 soku per tan was
nearly twice what it was supposed to be and was augmented by further illicit payments extracted by Motonaga. For other provinces it is
said the kachd levy was between 3 and 4 soku per tan (ArticleThree).
The above land tax information from the Owari Province petition
is summarized in Table 3.2. The total for Owari, 13.2 soku per tan, is
given in Article Five of the petition itself.
It is a measure of the central government's degree of reliance on
72 The twelfth-century work Shdchiireki, in Zokugunsho ruiju (Tokyo: Zoku gunsho ruijii kanseikai, 1923-30), vol. 32, part 1, p. 101, gives the land area planted to rice in the early tenth
century in Owari Province as 119,400 tan. Engi shiki gives a suiko quota for Owari of
472,000 soku of rice, which would have come to 4 soku per tan.
Cambridge Histories Online © Cambridge University Press, 2008
TAX STRUCTURE
213
TABLE 3 . 2
Land tax rates in the late tenth century
Tax rate per tan of rice field (in soku)
JiHnu tax
components
Owari Province
Other Provinces
land tax (denso)
regular suiko levy
i-5
0.6
0.6
additional suiko levy
"secret" extra suiko levy
by Motonaga in Owari
additional levy (kacho), to
pay miscellaneous offerings,
handicraft taxes, and other
quota items
2.8
2.8
1.1
—
7-2
4.0°
13.2
8.9*
Total land tax
""3 to 4 soku per tan" (ArticleThree).
*"8 to 10 soku per tan" (Article Five).
Note: 1 soku = 3.6 liters; 1 tan = 0.1 hectare. A good annual yield in this
period was 50 soku per tan; an average yield, 35 soku per tan.
Source: "Owari no kuni gunji hyakusei ra ge," in Takeuchi Rizo, ed.,
Heian ibun (Tokyo: Tokyodo, 1974), vol. 2, no. 339, Eien 2/11/8.
provincial officials, local officials, and quasi-official tax managers in
the mid-Heian period that for the first time in at least two centuries
they were permitted to levy new exactions of their own that were not
in any way connected with central government revenue. Somewhat
euphemistically known as "occasional exactions" (rinji zdyaku),
these new taxes were generally levied on individuals rather than by
land unit, and included, significantly, a reappearance of the institution of corvee labor, which in its earlier form {zoyo) had been all but
regulated out of existence in the ninth century.
The earliest reference to occasional exactions is in 889, where
they are described as "new"73 - and, indeed, we do not hear of
them again until 924. For after 924, references become more numerous: "occasional" or not, occasional exactions became a permanent part of the provincial tax structure from about the midtenth century on. In contrast to the unified, expanded land tax
(nengu, or kammotsu), occasional exactions were levied entirely for
the benefit of local and provincial officials. They came under central government control only indirectly, when shden proprietors in
73 Heian ibun, vol. 9, no. 4549: Kampyo 1/12/26, pp. 3464—70.
Cambridge Histories Online © Cambridge University Press, 2008
214
LAND AND SOCIETY
the capital requested that officials on their estates be excused from
all locally imposed duties. In the tenth century occasional exactions
were levied only on products such as silk and other cloths, lamp oil,
sake, and cedar bark.74 Again, unlike the land tax, they were generally assessed by population rather than land area, and were most
likely due only from adult males.
After 1000, however, occasional exactions were expanded to include both a land tax in rice (tammai), separate from the regular land
tax, and corvee labor. It also became increasingly common to assess
occasional exactions by land unit rather than by population.75 The
most common term for occasional exactions in the eleventh century
was "public duties" (kuji), a somewhat misleading term, inasmuch as
these so-called public duties were permitted mainly in order to reward local and provincial officials for providing through land tax income the goods and services required by quota by the central government. One warning, however: "public duties" sometimes denoted
all "occasional exactions," and sometimes all but the corvee element,
which is called "labor services" (buyaku); at other times the term is
used more broadly still to include that portion of the land tax used
to meet quotas for handicraft and industrial products.
By whatever name, the privilege to levy occasional exactions was
part of the price the central government allowed for the successful
functioning of its provincial quota system. For the direct levy of land
tax (denso), suiko loans, and handicraft taxes (cko and yd) in the
eighth century, the central government by the tenth century had
substituted the indirect levy of their equivalents on provincial governments. These indirect levies were translated by the provincial
governments into a broad land tax payable by peasants in rice that,
as shown in Table 3.2, brought together all the exactions of the earlier tax system. Through this system the central government at
length succeeded in stabilizing its income in the course of the ninth
and tenth centuries, but at a cost: quotas were all but frozen at levels thatfitthe tax base as it stood about the year 900. After 900, agricultural output continued to increase through expansion of arable
land, especially in the eastern provinces, and even more through the
greater security of cultivation and higher yields promoted by the diffusion of the plow and by the agrarian regime associated with it. But
as a general rule, it may be argued that none of this growth in pro74 Abe, Ritsuryo kokka, pp. 215-31; Okuno Nakahiko, "Rinji zoyaku no seiritsu to tenkai,"
Nihon rekishi, 255 (August 1969): 32-49.
75 Okuno, "Rinji zoyaku," pp. 32-49.
Cambridge Histories Online © Cambridge University Press, 2008
LANDHOLDING
215
duction was captured as increased tax revenue for the central government; nor should it be thought that estate proprietors (discussed
in the next section) captured more than a small share. Instead, increases garnered from agrarian expansion were captured by local tax
managers as land on which they did not pay public land tax (kammotsu), and by both provincial and local officials through the new
"occasional exactions" (rinji zoyaku; kuji).
LANDHOLDING
In the eighth century, tenure of rice-producing farmland took three
forms: household fields {kubunderi), public fields (koden), and reclaimed fields (konden). Household fields were allotted, or rearranged, each six years under the allotment system known as handen, which apportioned rice fields according to each household's size
and composition. Household fields could not be sold or exchanged,
or used as security in making loans, although their crop potential
could be mortgaged.76 Despite these restrictions, household fields
were legally classified as "private fields," as they were held by households as long as they had labor to cultivate them. This is in accord
with other evidence that private ownership of land in early Japan was
recognized only so long as land was actively in use.77 There seems
every reason to believe, therefore, that the allotment system in early
Japan, though inspired by Chinese example, had traditional roots
that explain the extraordinary differences between the Japanese
practice, which was basically a success, and its Chinese prototype,
which quickly failed.78
Rice fields that remained after the first full national application of
the allotment system in the 690s were classified as "public fields"
(koden). A sizable proportion of these were set aside to support rankholders, officeholders, temples or shrines, or were awarded to individual aristocrats. It is known that in the early eighth century public
fields set aside for these purposes came to 20,000 to 25,000 hectares,
which was somewhat under 5 percent of the total area under rice cultivation, estimated as between 600,000 and 700,000 hectares.79 A very
small proportion of these (some 750 hectares, in the home provinces
and in Kyushu) were at first farmed with corvee labor; after ex76 Takeuchi, Tochi seido shi, p. 56.
77 See, for example, the laws relating to land tenure codified in Ryo no shuge, in Kokushi
taikei, vol. 24, "Den-ryo," pp. 370, 372-7378 Morris, "Peasant Economy," pp. 31-33. 79 Takeuchi, Tochi seido shi, pp. 69-77.
Cambridge Histories Online © Cambridge University Press, 2008
2l6
LAND AND SOCIETY
penses, about half of the crop went to the holder of the fields. A
much larger group, some 9,000 hectares, was assigned to district officials as supplemental "private fields" (on top of their assigned
household fields). Like household and reclaimed fields, these could
be either farmed by the holder's family or rented to others. The remainder of public fields, 50 to 60 percent of the total, was leased by
provincial governments to peasant farmers at a rent set at 20 percent
of the assessed yield of each field. These rents were collected by the
provincial governments and forwarded along with other tax income
to the beneficiaries of the public fields in the capital.
The great bulk of public fields, however, were not assigned to specific beneficiaries but were simply rented at large by provincial governments to local farmers in this same manner - at 20 percent of assessed yield. The rent from these fields was added to the general tax
income forwarded from the provinces to the capital. Nofiguressurvive for the total area of public fields nationwide, but for one district
to Totomi Province in 740 the ratio of public to household fields was
1:5; for the entire island of Kyushu in 823, it was i:6.8°
All rice fields brought under cultivation after the first land distribution in the 690s were classified as reclaimed fields (konden). The
original idea, expressed in the codes, was that reclaimed fields would
be held privately for the remainder of the developer's life, but after
his death would be added to the pool of rice fields available for distribution as household fields. But in 723 the term for which reclaimed fields could be held was first extended to include descendants through the original developer's great-grandchildren. Then in
743, before any fields could have passed so far, reclaimed fields were
made the personal property of the holder and his heirs in perpetuity. From this time on they also could be freely bought and sold.
In view of the central government's tremendous concern with the
allocation and taxation of rice fields, it is a surprising fact that fields
growing other crops were regulated hardly at all until at least the
eleventh century. Dry fields - those planted to crops other than
rice - resembled reclaimed fields in that they could be freely disposed of by sale or inheritance. At the same time, they resembled
household fields in that they generally could be claimed by a family
only so long as they were actively cultivated. But dry fields were not
directly taxed, nor did the government concern itself with their
tenure. Not the least important purpose of government regulation of
80 Nara ibun, vol. 1, pp. 081-88; Rmju sandai kyaku, kan 15, Kunin 14/2/21, pp. 434—37.
Cambridge Histories Online © Cambridge University Press, 2008
LANDHOLDING
217
rice culture was to ensure that adequate rice was produced to satisfy
the aristocratic diet. Other grains grown in the Nara and Heian
periods - in rough order of importance, these were millet, barley,
wheat, and buckwheat - were probably of more importance to the
peasant diet than rice, as were soybean and vegetable crops. But all
of these products concerned the central government only marginally, as "miscellaneous exactions" that could be purchased with
provincial rice income.
Two major changes in land tenure marked the early Heian period:
the cessation of the distribution of household fields, and the reorganization of farmland to meet the changes in tax structure already
discussed. The characteristic product of this reorganization was the
unit of tax assessment and management known as myo.
Under the allotment system instituted in the seventh century,
household fields were adjusted to fit changes in household size and
composition shortly after each population registration (koseki),
every six or seven years. This system worked for about a century.
Then in 794 - the year of the move to the Heian capital - distribution of household fields was held up, and did not take place until
800. The distribution of 800 was the first to include the provinces of
Satsuma and Osumi on the extreme southern frontier. It thus became both the first and last land distribution to be carried out
through the entire country.81
After 800, distributions were carried out only sporadically. Distribution in 828 can be verified for only the home provinces and the
province of Kozuke in the east. After 828, household fields in the
home provinces were not adjusted for fifty years, until 881, when
fields were redistributed principally to enable the establishment of
4,000 hectares of new imperial fields mandated in 87c).82 Some distributions occurred in other provinces in this period, but none at all,
it seems, in the 830s or 840s. The national histories report that land
registers were made in 843 and 844, but that officials were unable to
carry out a redistribution of household fields.83
The process by which the distribution system was terminated in
the ninth century is revealed in two population registers that survive
from the early tenth century.84 These two registers tell a rather re81 ToraoToshiya, Handen shuju-ho no kenkyu (Tokyo: Yoshikawa kobunkan, 1961), pp. 307-16.
82 Nishibeppu, "Kyu seiki chuyo," p. 43. 83 Torao, Handen shuju-ho, p. 326.
84 Heian ibun, vol. i, no. 188: Engi 2, pp. 224-40, and no. 199: Engi 8, pp. 289-305. See Hirata Shuji, "Heian jidai no koseki ni tsuite," in Toyoda Takeshi kanreki kinenkai, ed., Nihon
kodai chusei no chiho-teki tenkai (Tokyo: Yoshikawa kobunkan, 1973), pp. 59-96.
Cambridge Histories Online © Cambridge University Press, 2008
2l8
LAND AND SOCIETY
markable story that, surprisingly, tends to disprove traditional contentions that the land distribution system broke down in the ninth
century because of the resistance of local officials and the misappropriation of public land by aristocrats and temples. The process
was, in fact, rather more orderly than that.
The two registers, from Awa Province in 902 and Suo Province in
908, list an enormous preponderance of women - 88 and 73 percent
of the population, respectively - and of aged persons, including a
number aged over 100. The distribution of ages above 70, however,
is not unlike the curve expected in a normal population starting at
age 1. The distribution from 40 to 70 is nearly stable. Below age 40,
in defiance of demographic logic, the population shrinks with descending age until, below age 10, no younger persons are recorded.85
This remarkable pattern, found in two registers from widely separate provinces, could not have come about through random fabrication. It would appear that local officials, following directives from
above, ceased to record new births from about 830 to 870. During
this period the registered population was regularly advanced in age,
but without note of births or deaths. Then in the 870s births were
again added, but at a declining rate until 894, after which no new
persons were added. It is likely these births were added to take the
place of persons in the register whose recorded age had become improbably advanced.
By making no changes in registered population after the 830s,
other than regular advances in ages, household fields as then distributed were made permanent household possessions. Although redistributions were ostensibly resumed in the 870s and 880s, new persons were added to the registers only to replace names that had to
be removed because of improbably high age. This guaranteed that
there would be no net change in household field eligibility. To further ensure that households would not appear eligible for increased
allotments, officials altered the registration of most men to women,
who were eligible for only two-thirds the allotment due men. This
was clearly done after men had first been registered by their correct
gender: "children" (of whatever age) were listed in two descending
sequences, as in earlier registers, but whereas in the eighth-century
registers the first sequence was of sons and the second of daughters,
in the tenth-century registers both are of daughters.
Manipulation of population registers after the 830s was made pos85 Hirata, "Heian jidai no koseki," pp. 71-77.
Cambridge Histories Online © Cambridge University Press, 2008
LANDHOLDING
21Q
sible by the policy changes in the early 840s, already discussed, that
fixed the number of taxable males in each province at then-current
figures. This fixed the liability of each province for handicraft and
labor taxes. As no further adjustments to the number of men liable
to taxation were to be made, there was no longer any need to keep
population registers other than to determine eligibility for household
fields. From the way population registers were handled, it is clear eligibility for household fields was also to be held constant. The conclusion seems inescapable, therefore, that in the latter part of the
ninth century the central government decided to allow household
fields to remain as permanent possessions of the households then in
possession of them - and not just in a piecemeal or haphazard way,
but as a concerted policy implemented throughout the country.
Hence, the argument that men were disguised as women to avoid
taxes is mistaken. So, too, is the contention that manipulation of
population registers made it impossible to carry out field redistribution. On the contrary, by a curious but consistent bureaucratic compulsion, population registers were manipulated to make the distribution system work on paper in an age when physical redistribution
of household fields was no longer desired.
It is clear, therefore, that household field "redistribution" had lost
any real content after 830. Sporadic "distributions" continued to
take place on paper for some decades thereafter, but ceased even as
a paper institution after 902.86
The underlying cause of the cessation of the land distribution system was an expansion of population without any corresponding expansion in the pool of rice fields available for distribution as household fields. As noted at the beginning of this chapter, the total area
planted to rice in Japan increased by about a third between 740 and
900, from about 650,000 hectares to 875,000 hectares. This suggests
a growth of perhaps 2 million in population (from between, perhaps,
5 and 6 million to between 7 and 8 million). The codes in the early
eighth century had intended that new rice fields would become
available for distribution as household fields after the developer's
death, ensuring a steady increase in the pool of household fields as
population increased. But once this policy was abandoned, and reclaimed fields were allowed to remain permanently as private possessions, population increase could not but lead to a shortage in
available household fields. By the ninth century the government thus
86 Takeuchi, Tochi seido shi, p. 109.
Cambridge Histories Online © Cambridge University Press, 2008
22O
LAND AND SOCIETY
had little choice but to bring the distribution system to an end. The
only alternative would have been to convert most public fields
to household fields, a clearly unpalatable alternative, as the rent
charged on public fields was roughly four times the land tax received
from household fields.
Many other arguments have been offered to explain the cessation
of the eighth-century land distribution (handen) system. As changing land and tax policy was clearly the central social issue of ninthcentury Japan, it is important to consider some of these arguments
individually:87
1. Malfeasance by provincial and local officials, and the fabrication
of population register data, are most often cited as causes of the
decline of the distribution system, but are difficult to verify. As
has been seen, population registers were certainly manipulated,
but in specific and orderly ways designed to implement central
government policy. It is interesting to consider how much of what
has traditionally been seen as brazen illegality by provincial officials in Heian Japan was in fact legitimate behavior by officials
from the capital loyally implementing changing central policy.
2. Another popular theory is that the occupation of land by aristocrats, temples, and shrines led to a shortage of household fields.
For all its importance, however, the percentage of the nation's
land monopolized by aristocratic households and religious institutions was never great. And since these holdings consisted
largely of nonagricultural waste and newly created fields, the
household field system should not, technically, have been affected. Illegal conversion of household to reclaimed fields, while
in theory possible with provincial government help, is nonetheless a supposition that has never been verified.
3. The argument that population expansion led to a shortage of
household fields comes closer to the argument presented here,
but it is rarely noted that the reason for this was that increases in
population were served by increases in reclaimed fields that, after
the new laws of the early eighth century, could no longer be converted into new household fields.
4. Perhaps the weakest argument commonly found is that massive
migration removed peasants from population registers and so
made land redistribution difficult. It first of all is presupposed
that the government viewed migration as illegal, which it had not
87 These traditional arguments are listed inTorao, Handen shuju-ho, pp. 424-6.
Cambridge Histories Online © Cambridge University Press, 2008
LANDHOLDING
221
since the beginning of the eighth century. Migration led to the
loss of eligibility for household fields, but it did not lead to loss
of taxes, as migrants were registered for tax assessment where
they settled. And since migrants did not receive allotments of
household fields, any extensive migration would have made land
distribution easier, not more difficult.
5. Finally, the argument that cultivators themselves resisted redistribution of household fields because they desired a more secure
(that is, private) form of tenure is difficult to accept, for the form
of tenure that replaced household fields provided no more secure rights to the tiller than had been enjoyed under the previous system.
With the cessation of distribution of household fields, it was no
longer as important to distinguish household from reclaimed fields.
Since both were (and always had been) liable to the same tax rate,
late-tenth-century documents such as the 988 Owari Province petition discussed earlier group the two forms of tenure together as
"land tax fields" (sozeiden). The same sources refer to publicfieldsas
"rental fields" (jishiden). The rent (jishi) paid on the latter, ranging
from 6 to 10 soku per tan, had now been matched, however, by the
enlargement of the land tax paid on other fields. Thus, from a tax
standpoint there was little reason to maintain a distinction between
rental (public) and other fields, and after the late tenth century no
distinction was made. From the eleventh century, all rice fields came
to be known simply as "public fields" (koderi), unless they were certified by central government document as tax-exempt "private
fields" (shideri), in which case their tax income was delivered to a private "proprietor" within the central government. (These are the
shden "estates" discussed in the next section.) This later use of the
term "public fields" should not be confused with the earlier use of
the very same term, for the word was now applied to all fields liable
to the provincial land tax, thus grouping together what had earlier
been distinguished as household, public (in the old sense), and reclaimed fields.
From the late ninth century, distribution of rice land ceased to be
dictated by the government. It would be a mistake, though, to assume that farmers could now claim full "private" tenure to the land
they tilled. On the contrary, it was the government view that all rice
land, no matter what its designation, was rented from the provincial
government, unless rent was transferred to designated shden propriCambridge Histories Online © Cambridge University Press, 2008
222
LAND AND SOCIETY
etors, and that rent was paid in the form of the provincial land tax.
This is why some historians, putting the cart before the horse, speak
of land liable to provincial taxation as essentially identical in character to private shoen land. In fact, it was the system of tenure and taxation on public land that provided the model for shoen rental systems, not the reverse.
The unit of taxation, on public and estate land alike, was the
myo - a term that does not yield to translation; a literal rendition
would be "nominal." It used to be common to view myo as units of
land tenure. To some extent they became so in later times, but not
in the Heian period. Myo were simply the units on which taxes were
assessed and collected. As tax liability was calculated by land area,
provincial governments were concerned only with recording the area
of taxable land within each myo, not the number of people tilling this
land or the distribution of cultivation rights among them.
Two examples of the variation in the relationship of myo to land
tenure can be seen in Figures 3.2 and 3.3. The twenty households of
Wakatsuki-no-sho were grouped into fifteen myo. In ten cases myo
were identical with solitary tenures; in the other five, myo included
the holdings of two households, which in only one case were geographical neighbors. By contrast, the holdings of the eighteen households in Kohigashi-no-sho were originally grouped into a single myo
(later reorganized into several wryo).88Thus, myo were at times equivalent to units of land tenure, or else might as easily hold a number
of separate tenures. Furthermore, farm families could, and frequently did, hold tenures in more than one myo.
The taxes owed by the farmers of land in a myo were collected by
a local tax manager (known most simply as fumyo). Some of the tax
revenue was used immediately to purchase special products as required; the balance was then forwarded with the products purchased
to the provincial government. For myo owing dues to a shoen proprietor the procedure was essentially the same, the dues being forwarded by the tax manager to the proprietor instead of to the
provincial government.89 Myo tax managers came in time to be
known as myoshu ("wyo-holders"), a term that first appears in
1047.90 It must be kept in mind that these myoshu were not the own88 InagakiYasuhiko, "Shoki myoden no koz6:Yamato no kuni Ota-no-inumaru-myo ni tsuite,"
in Inagaki Yasuhiko and Nagahara Keiji, eds., Chusei no shakai to keizai (Tokyo: Tokyo
daigaku shuppansha, 1962), pp. 40-41.
89 See examples in Heian ibun, vol. 2, nos. 388-403, pp. 527-31, and vol. 9, nos. 4579-97, pp.
3495-3500, dated from Choho 1 to Choho 3 (ggg-iooi).
90 Heian ibun, vol. 3, no. 646: Eijo 2/10/27, PP- 779~8o.
Cambridge Histories Online © Cambridge University Press, 2008
LANDHOLDING
223
ers or landlords of the land in a myo, but were tax managers who reported myo land area, collected the taxes assessed by the provincial
government, and delivered these in the forms requested. In many
cases a single tax manager collected taxes from several myd.91 In others, such as that of Kohigashi-no-sho, several evidently shared responsibility for tax collection.92 In these latter cases (Kohigashi included) it was common to use a fictitious name as myoshu - this
seems to be the case, in fact, in the 1047 document in which the term
"tnydshu" is first found.93
Right to control cultivation of a given parcel of land was granted
by either provincial governor shden proprietor to persons known as
tato, who may or may not have also held the tax management privilege (fumyo) for the same land. A tato could assign cultivation rights
to others, for whom he would serve as landlord. But in many cases
the amount of land controlled by a tato was no more than enough
for him and his family to farm alone. The one to whom a tato assigned cultivation rights was known as sakushu or sakujin - "cultivator." If a tato retained cultivation rights for himself, the terms tato
and sakushu might be, and frequently were, used interchangeably.
Otherwise, sakushu were tenants of tato, although it must be borne
in mind that tato did not actually "own" land itself, but only cultivation rights to land.
Tato and sakushu are terms that define legal relationships to land
tenure, not social classes. A tato exercising control over the cultivation rights to one parcel of land might at the same time rent other
land as sakushu to another tato. Only when a tato also held the right
of tax collection as well did he stand indisputably in a higher, and
more lucrative, position than other cultivators.
The right of tato to control cultivation of land was revocable at
any time by the provincial government or shden proprietor. Hence
cultivators who obtained tato appointments were not automatically
able to treat land that had been household fields as fully partible
and alienable. On the contrary, even holders offieldsthat had been
developed as reclaimed fields were often less able to assert the full
possessory rights that had attached to reclaimed fields in the eighth
and ninth centuries. This was because provincial governments after
the year 900 no longer issued the certificates of ownership {kugeri)
that had earlier guaranteed these rights to the developers of new
91 One example is Heian ibun, vol. 1, no. 240: Johei 2/9/22, pp. 354—55.
92 Takeuchi, Tochi seido shi, p. 147. 93 Inagaki, "Shoki myoden," p. 70.
Cambridge Histories Online © Cambridge University Press, 2008
224
LAND AND SOCIETY
farmland.94 This change in policy - part of the overall adjustments
in central government land and tax policies in this period - inaugurated a period of at least a century and a half during which land
"ownership" per se was not recorded in the all-important provincial
land surveys (kokuzu). Holders of reclaimed fields who obtained
appointment to deliver taxes from the myb containing their holdings
were generally better able to protect their possessory rights. But
others often stood in danger of being dispossessed of the land they
cultivated.
In the eleventh century some cultivators were able to exert sufficient leverage to win a stronger recognition of their rights to the land
they tilled. These cultivators were known as "permanent cultivators"
{eisakushu) of the land to which they were granted this privilege. In
theory, they were able to sell, exchange, or bequeath their land in
perpetuity.95 Although many more cultivators may have been able to
hold similar privileges in practice, full "permanent cultivator" rights
were not granted routinely by provincial governments or shoen proprietors, but, rather, were an exceptional concession. Full legal
recognition of the right to secure, private land tenure was to develop
only very gradually during the succeeding Kamakura and Muromachi periods.
SHOEN
In 743 the central government decreed that rice fields brought under
cultivation privately - that is, reclaimed fields (konderi) - were to be
held in perpetuity. The maximum acreage of reclaimed fields to be
held by aristocrats of various ranks, and - at the bottom - by commoners, was specified. In 749 similar, but much higher, ceilings were
set for the nation's temples. Starting in 749, the government actively
assisted the major temples, most especially Todaiji, in finding land
that could be developed into estates of reclaimed fields. These estates were the first to be called sho, or shoen (although there is some
use of the term, not entirely relevant, in earlier periods).
These early shoen must be distinguished from those of the Heian
period, when the term applied not just to land actually owned by the
shoen proprietor, but primarily to the transfer of tax incomefromthe
provincial government to a designated aristocratic household or re94 Takeuchi, Tochi seido shi, p. 150.
95 The eisakushu privilege is first seen in a document of 1037: Heian ibun, vol. 2, no. 570:
Choryaku 1/12/8, p. 734.
Cambridge Histories Online © Cambridge University Press, 2008
SHOEN
225
ligious institution. In early usage, shoen may fairly be translated as
"estate." But in the later, far more important use of the term as a
unit of tax revenue, the term "estate" would be utterly misleading.
For this reason, the term sho or shoen (used interchangeably) is used
without translation.
It would be safe to say that the main reason there were a significant number of estates of reclaimed fields developed in the eighth
century was that the central government actively sought their development. This was especially true for the largest of the early shoen proprietors, Todaiji. Between 749 and 756 Todaiji dispatched officials to
several provinces where, with the help of provincial officials, they located suitable land for development. Todaiji was further granted several large tracts of wasteland that they might develop into reclaimed
fields. By 770, when the main impetus to Todaiji land development
was over, the great temple held ninety-two shoen in twenty-three
provinces, most in the home provinces and the Hokuriku region to
the northwest. Convenience of location was all-important; hence, no
sho were developed on Todaiji's behalf in either the eastern (Kanto)
provinces or in Kyushu. About half of the Todaiji sho had their origin in commendation by the central government.96
Land development by aristocrats, apart from the temples, was important too. Several bills of sale survive that show acquisition of reclaimed fields by aristocrats in the home provinces, and in such
nearby provinces as Omi and Kii. Whether aristocrats ever financed
land reclamation directly, as temples did, is far less clear. It seems
likely that aristocrats, even more than temple proprietors, had to depend on hired local managers to acquire fields developed by others,
and to find tenants for them, even when fields were located close to
the capital.
Temples could manage their estates either by sending out their
own representatives or by engaging local notables; the latter were
most often also local officeholders in the government. Generally,
temples preferred to use their own representatives; but even then it
was necessary to win local cooperation. Reclaimed fields in early
shoen were farmed by tenants. In the case of the well-documented
Kuwahara-no-sho in Echizen Province, fields were rented at either 6
or 8 soku per tan, to which was added the land tax of 1.5 soku. When
fields could be certified as "temple fields" (jideri), the land tax por96 Takeuchi, Tochi seido shi, pp. 128-9; s e e a ' s o Kishi Toshio, Nikon kodai seiji shi kenkyu
(Tokyo: Hanawa shobo, 1966), pp. 317-99.
Cambridge Histories Online © Cambridge University Press, 2008
226
LAND AND SOCIETY
tion could be kept by the temple, but the government granted this
exemption very sparingly. On most reclaimed fields owned by temples — and on all fields owned by aristocrats - the regular land tax
(densd) was paid to the provincial government.
With the exception of some estates in the home provinces, these
eighth-century estates did not thrive in the Heian period. Little is
known of most temple estates in the ninth century. But by the tenth
century it is apparent that most consisted of uncultivated waste, or
had disappeared from the landscape entirely. In a 950 Todaiji register, three estates in Asuwa District in Echizen were still listed with
450 hectares of reclaimedfields.Yet a year later the Asuwa chief magistrate reported that the first two of these estates had long been uncultivated, while no one in recent years had so much as heard of the
third.97 Of forty-one Todaiji sho listed in the surviving fragment of a
998 register, only ten were in operation; the others were entirely
waste.98
There were a number of reasons for the decline of temple reclamation estates. One was the withdrawal of central government support. From having assisted in their growth, the government in the
late eighth century turned to limiting any expansion in estates of reclaimed fields. In the ninth century few grants were made to the old
Nara temples. Estates belonging to temples of the now-favored
Tendai and Shingon sects were given preferential tax treatment denied such old temples as Todaiji.99 But even the favored temples did
not receive substantial land grants in the ninth century, as they had
in the eighth. Instead, these grants were made to the imperial house
or to individual members of the imperial family: between 828 and
886, mostly in the 830s, the imperial house was granted nearly 7,000
hectares of land. About 20 percent had already been developed as
farmland; the rest was to be developed into farmland whenever possible. Similarly, from 795 to 878 over 3,700 hectares were granted to
some thirty-three imperial princes and princesses, also to be developed where possible into farmland.100 In the eighth century most
such land would have been given to Todaiji and other temples.
A second reason for the decline of early skoen was lack of cooperation by local officials and estate managers. Proprietors depended on
97 Heian ibun, vol. 1, no. 257:Tenryaku 4/11/20, pp. 372-84, and no. 263:Tenryaku 5/10/23,
P- 386.
98 Heian ibun, vol. 2, no. 377: Chotaku 4, pp. 511-13.
99 Yasuda Motohisa, Nihon shoen shi gaisetsu (Tokyo: Yoshikawa kounkan, 1957), pp. 52-53.
100 Takeuchi, Tochi seido shi, pp. 124—28.
Cambridge Histories Online © Cambridge University Press, 2008
SHOEN
227
the loyalty of their estate managers, who might derive more profit
from developing reclaimed fields for their own use. Managers sent
out from the temple might be more trustworthy, but were less likely
to obtain the local cooperation necessary to enlist tenants for estate
land. Nor were all temple managers reliable: one such manager, sent
by Todaiji to manage Takaba-no-sh5 in Tamba Province in 842, proceeded to embezzle the rents he was to have delivered to the temple,
then disappeared into a neighboring province.101 As a general rule,
the estates that most often failed were those in which the proprietor
played the most active role: those managed by monks or others sent
from the home office, those in which fields were brought under cultivation by the proprietor, and those (found only in Echizen) where
the proprietor attempted to form a geographically compact estate by
trading outlying temple fields for fields at the center of the estate
that had belonged to others. By contrast, estates formed by purchase
of fields developed by others, and managed by local notables, had
a greater chance of lasting into the tenth century and beyond.
But even these survived only if they were fitted into the new tax
structure.
The final, and perhaps crucial, reason for the decline of early shoen
was the transformation of the tax structure in the late ninth century.
As suiko came to be levied as a land tax, it was applied (like the land
tax, demo) to reclaimed and household fields alike. This grew in the
tenth century to a greater land tax that, since it was derived from
suiko, was also applied equally to reclaimed fields and to those that
had been household fields. The application of a large land tax made
reclaimed fields far less profitable to the shoen proprietor unless the
estate was also made tax-exempt. Hence, tax exemption, not land development, became the key to shoen formation in the tenth century.
Most new shoen in the tenth century were formed along a pattern
that has come to be known as the "Kinai sho" pattern. To ensure that
government-supported temples and shrines received their proper
dues from provincial governments, segments of provincial public
land (koden) were designated to provide these dues directly, on either
a permanent or a floating basis. Fields so designated were "excused"
other taxes so that their tax yield could be forwarded directly to the
temple or shrine. Such fields were known as the temple's (or
shrine's) "exempt fields" (menden); when different fields each year
were designated for this purpose, they were known as "floating ex101 See Abe Takeshi, Nihon shoen shi (Tokyo: Ohara shinseisha, 1972), p. 46.
Cambridge Histories Online © Cambridge University Press, 2008
228
LAND AND SOCIETY
empt fields" (ukimeri). Aggregates of exempt fields were called shden
because local tax managers for these fields reported directly to the
proprietor rather than forwarding revenue by way of the provincial
government office. Because most such fields were in the Kinai
(home provinces) region, historians have come to refer to these
shden as "Kinai sM"The term refers, however, not to the location of
a sho, but to its origin in the designation of fields to yield taxes to
support a given religious institution.
Kinai sho proprietors had no direct contact with the land designated as exempt fields, nor with the farmers who paid the taxes they
received as shden dues. Nor did they determine how much they
could collect each year; this was arranged by the provincial government. Furthermore, the taxes forwarded to the shden proprietor constituted only the land tax (kammotsu) component. The balance of
"miscellaneous exactions" {rinji zdyaku) were still paid to the provincial government.
Both Kinai sho and those earlier land-development shden that survived into the tenth century had to be certified by the central government. These certificates (called kanshdfu) specified the limits to
the location (if fixed) and extent of fields that were exempt from
land tax for each of the l-hectare grid squares (tsubo) by which land
was divided for survey and registration purposes. Provincial governments, for their part, recorded the number of such exemptfieldsper
grid square in their provincial land registers (kokuzu). Each year
local officials reported to the provincial government the area in each
grid square under actual rice cultivation that year. Taxes for the area
exempted were then forwarded to the proprietor; taxes from excess
cultivated fields, if any, over the limit of exempt fields allowed by
central government certificate were paid in full to the provincial government.102
There were three ways shden proprietors could expand their influence over shden land and cultivators to embrace more than just receipt of rents equivalent to the land tax that would otherwise have
gone to the provincial government. The first was to levy an additional rent (kajishi) on all exempt fields, parallel to the added exactions provincial governments themselves levied. Frequently these
additional rents could still be collected even when shden privileges
were withdrawn and the basic shden dues proper reverted to the
102 Sakamoto Shozd, Nikon ocho kokka taisei ron (Tokyo: Tokyo daigaku shuppansha, 1972),
pp. 19-28.
Cambridge Histories Online © Cambridge University Press, 2008
SHOEN
229
provincial government.103 Proprietors might also gain limited exemptions from the second component of provincial taxation, the
"occasional" or "miscellaneous" exactions, which they could then
collect themselves. Since these exactions lay entirely outside the
purview of the central government, such exemptions were not covered by shoen certificates and had to be negotiated directly with the
provincial government. In the tenth century, when miscellaneous exactions were levied on individuals, proprietors argued that their
shoen personnel should not have to pay these taxes to the provincial
government. Later, as the exactions came to be assessed by land
area, some argued that shoen land already exempt from land tax
should be exempt from miscellaneous exactions as well. All such requests for exemption were granted solely at the pleasure of the
provincial governor, and exemptions granted by one governor could
be, and frequently were, overturned by his successor.
A final avenue for the expansion of proprietary influence over
shoen land was to assume control of the granting of cultivation rights
to sho fields. In shoen that originated in the development of reclaimed
fields such control was implicit. But for shoen of the Kinai type, the
right to control cultivation rights - the right to tato - was still assigned by provincial officials. Proprietors might gain the right to appoint their own tato, though, through a policy of consolidation of exemptions of both land tax and miscellaneous exactions, obtaining
fixed rather than floating exempt fields, and then using the argument
that the right to receive all major taxes due from a fixed parcel of
land implied with it the right to assign cultivation rights to that land.
For, as should by now be abundantly clear, this concept, which is so
familiar in the West, was not at all a given assumption in Heian
Japan.
It is also important to bear in mind that only a minority of shoen
were developed toward the goals of expansion and consolidation of
tax and cultivation rights. Frequently, the conflict with provincial
governments that such expansion programs inevitably entailed resulted in more losses than gains. Most shoen of the Kinai pattern remained administered through local officials and myoshu, therefore,
with little or no involvement by the proprietor in the actual management of local agricultural affairs. Like provincial governments,
proprietors were generally concerned only with the assessment of
taxable land, as organized into myo, not in the distribution of culti103 Sakamoto, Nihon bcho kokka, p. 44.
Cambridge Histories Online © Cambridge University Press, 2008
23O
LAND AND SOCIETY
vation rights with myd, or even with the identity of tillers. In fact, for
simplicity of administration, it was common to assign myd uniform
shoen dues, although myd themselves were of greatly uneven size, an
approach that was known as "equalized myd" (kintd-myd). While certainly simpler, such proprietor distance left little possibility for full
maximization of additional rents (kajishi), let alone expansion into
new revenue areas. But for many shoen proprietors, just to maintain
a steady income from fields that had been properly certified by central government decrees required a constant struggle with provincial
governments.
In the eighth and ninth centuries the central government had expected shoen to expand through the development of new lands. After
about 900, however, high officials came to feel that no further expansion in the area exempted from provincial taxation was desirable.
In part this was because the new land tax of this period had expanded far beyond the simple 1.5 soku per tan that had been the
limit of shoen tax exemption earlier.104 But the main reason to oppose
further expansion in exempt fields was that under the new provincial quota system the maximum revenue due the capital from each
province had been fixed and was, in fact, subject to reduction upon
provincial petition. Hence, any further expansion in the area of exempt fields directly reduced the ability to meet provincial tax quotas, and thus gave the province a perfect rationale to request reductions in these quotas. This concern was certainly on the minds of
high officials when they inaugurated the first of what was to be a series of campaign to hold back shoen growth and revoke certificates of
exemption that were not in order. There was, moreover, a direct line
between this campaign in 902 and the inauguration after about 915
of the system just outlined for the registration and control of exempt
fields through annual assessment and review by provincial and local
officials.105
Accreditation of exempt fields was always by the central government, not by provincial officials. But under the new system of registration, provincial governments gained the power to recognize or
deny recognition to any new land development claims by shden proprietors or by other land developers, who most commonly were local
officials or tax managers. Provincial governments thus made the crucial decisions in three areas that greatly affected the fate of shden: the
recognition of the certification of exempt fields, the recognition of
104 Sakamoto, Nihon ocho kokka, p. 90. 105 Sakamoto, Nihon ochd kokka,
p.
Cambridge Histories Online © Cambridge University Press, 2008
122.
231
SHOEN
new reclaimed fields, and the exemption of shoen personnel or, later,
shoen land from provincial miscellaneous exactions. In the first matter proprietors could appeal provincial decisions to the central government. The other two areas, however, were entirely under provincial jurisdiction, and a favorable decision by one governor might be,
and frequently was, reversed by a subsequent governor.
The uneasy relationship between shoen proprietors and provincial
governments entered a new stage in the 1040s. In response to a series of squabbles over shoen tax rights - most particularly the case of
the Todaiji estate Kuroda-no-sho in Iga Province - the central government set a standard rate for the provincial land tax at 6 soku per
tan. Shoen proprietors were permitted to collect as rent either all or
part of this rate for those fields certified as exempt fields, but in either case the amount allowed was always to be spelled out as a standard rate. Any provincial land-tax exactions above 6 soku per tan
(and by this time there were many) were to go unequivocally to the
provincial government.106
This did much to clarify the terms of discourse between shoen
holders and provincial governments - generally in favor of the latter.
During this same period the levy of miscellaneous exactions shifted
definitively from persons to land units, and the exactions themselves
came to be known most commonly as "provincial levies" (kuniyaku),
or simply "miscellaneous levies" (zoyaku). This simplified method
for the assessment and collection of taxes favored shoen proprietors
as well as provincial officials: before, shoen proprietors had to argue
case by case the relevance of sho personnel to temple support before
miscellaneous exactions were exempted. Now they could make the
far simpler argument that any field already exempt from (some) land
tax should by extension be exempt from provincial levies as well.
For many proprietary institutions the development of shoen made
up for the decline, or loss altogether, of revenue from provincial government tax receipts. This was perhaps most explicit for the largest
proprietor, Todaiji, whose efforts to establish Kinai-style sho in the
tenth and eleventh centuries followed directly from the failure of
provincial governments to forward income properly due them from
"support households" (Juko), whose regular tax income was to be
set aside for support of the great Nara temple.107 Since all shoen proprietors were entitled to state tax support, shoen established to make
up for previous tax sources now lost were a legal part of the tax
106 Sakamoto, Nihon ocho kokka, pp. 226-29.
10
7 Inagaki, "Shoki my5den," p. 77.
Cambridge Histories Online © Cambridge University Press, 2008
232
LAND AND SOCIETY
structure, and hence were never challenged on grounds of principle
by the central government.
What did lead to challenge by high officials in the capital was the
competition for limited tax sources, a competition in which organized temple bureaucracies held a distinct advantage over the
aristocrat-bureaucrats who staffed the civil government proper.
With the institution of provincial tax quotas, and particularly after
900 when governors were allowed to request each year reductions in
these quotas, aristocrats were forced to divide an ever shrinking pie
of provincial revenue to pay their own stipends and emoluments,
the daily expenses of government operation, the support of the imperial household, and the maintenance of the Heian capital. The
major political households of the capital, none of them ones to
watch their income dry up for want of proper contacts, were thus at
pains to develop ties with career provincial governors and, where
possible, to establish direct links to the source of tax income by setting up shoen, which they could do either in their own names or
through temples and shrines with which they were associated. The
first concerted effort to check shoen growth, in 902, had clearly reflected a land policy aimed at holding all shoen growth in abeyance.
Later efforts at shoen regulation, however, were in fact much more
commonly partisan efforts designed to check the growth of some
temple shoen, the better to enable households currently enjoying
political power, and their clients, to compete for increasingly limited tax resources.
The rise of shoen in its major function as a tax instrument (as distinct from its earlier role as a land-development estate) thus followed
naturally from the evolution of a tax structure centered on shrinking
provincial quotas. From the first important period of shoen development in the middle decades of the tenth century to the end of the
eleventh century, the various elements of the central government,
whether they were government offices, aristocratic households, or
the imperial household itself, received a mixture of support, mostly
indirect at first by way of the provincial governments, and the rest
direct, forwarded by local officials and tax managers from fields designated as shoen.
It would be difficult to calibrate exactly the mixture of these indirect and direct revenue sources, but there seems little doubt that reliance on direct shorn and shoen-like revenue sources increased
steadily in the tenth and eleventh centuries. This was true not only
for the great temples, which became nearly wholly reliant on shoen
Cambridge Histories Online © Cambridge University Press, 2008
SHOEN
233
for support, but also for the great aristocratic households and their
clients, the holders of bureaucratic posts in the capital, most government offices and agencies, and, finally, even the imperial household.
Hence it even seems inevitable, although nonetheless remarkable,
that there would come a point when the much-atrophied provincial
quota system would be jettisoned altogether in favor of wholesale reliance on shoen and s/zoen-like support for virtually all the elements
of the ruling hierarchy. This point came during the last part of the
Heian period, the Insei period (see Chapter 9). Those portions of
"public land" (kdderi) that had not yet been made into shoen - a proportion that has been variously estimated at between 40 and 60 percent of the nation's tax base - were now assigned province by
province to support government offices or household patronage
clusters, including the imperial household, in just the same manner
as shoen. As with shoen proper, the proprietors of a given province's
public land (known as its kokugaryo) received its tax revenue as personal rents, could communicate directly with local tax officials regarding tax delivery, and could appoint officials of their own to supervise local tax collection and accounting.
The key figures in any transfer of tax rights from a provincial government to a shoen proprietor were the local officials and tax managers (fumyo and their equivalents) who controlled local tax and land
management. Provincial governments were totally dependent on
local tax managers to see that taxes were properly assessed, collected, and delivered. Because of that dependence, beginning from
about the 1040s, local officials and myoshu were able to exercise a
rather remarkable right: provided they found a qualified and receptive proprietor, they could "commend" the tax unit (usually a myb")
for which they were responsible to a shoen proprietor, on their own
authority transferring land from provincial administration to inclusion in a shoen.
Owners of privately developed rice land (reclaimed fields) had always been able to commend their land to a religious institution,
which then administered the land as a shoen. In these cases the commendation was an outright gift by a private party for religious merit
or, not uncommonly, by a priest who had developed the land for the
express purpose of commending it to his employer. Commendation
by tax managers in the late Heian period, however, was quite another matter. What was commended in the later instance was not
land owned outright by the commender, but units of public taxation
for which the commender served merely as tax collector, albeit in
Cambridge Histories Online © Cambridge University Press, 2008
234
LAND AND SOCIETY
the expanded role of tax manager. The commender was, as it were,
simply exchanging one master for another. Tax rates and application
remained essentially the same, and, of course, the tax manager retained his lucrative position as the assessor and collector of these
taxes, now deemed shden dues. What made such an act possible in
the late Heian period was the growing central government support
of shden growth and the provincial governments' near total reliance
on local officials as the only people who could guarantee that taxes
would be duly assessed and collected. While provincial officials certainly did not welcome the removal of territory from their tax rolls,
there was frequently little they could do when the land was being
transferred to shden proprietors who outranked them socially and
politically and enjoyed increasing support from the highest quarters
to expand shden.
A number of factors might lead a tax manager to commend the
unit of tax base for which he was responsible to a qualified shden proprietor. One was simply the matter of tax rates. Shden dues duplicated provincial tax rates, but generally lagged behind in terms of
the additional levies provincial officials were able to exact on top of
the land tax proper (kammotsu), which after 1040 was set at a nationally uniform rate. This was simply because provincial officials
were closer to the scene than shden proprietors. Where shden levies
lagged behind provincial taxation, the margin for tax manager profit
increased correspondingly. For the same reasons, most, though certainly not all, shden proprietors were at a disadvantage in supervising
dues collection because of their distance from the shd in question.
This must have been of particular advantage to commenders in the
distant Kanto and Kyushu provinces. The commender's position in
office was also likely to be far more secure under a shden proprietor
who was entirely in his debt for gaining a new shd.
Hence, financial advantage was a considerable factor in motivating a commendation of provincial "public land" to a shden proprietor. But even more considerable in a great many cases was the matter of patronage. By commending a tax source to a central temple,
shrine, or aristocrat household, the commender forged an alliance
with a powerful source of political support. Some profited substantially from the association; others were eventually dispossessed by
their new masters, as indeed they might have been under provincial
administration. And for every case of commendation, there must
have been many other tax managers who used the leverage of potential commendation to win still greater concessions from provinCambridge Histories Online © Cambridge University Press, 2008
SHOEN
235
cial officials. Local tax managers most certainly did not have the
power to keep tax revenues for themselves - any attempt to do so
soon put them in rebellion against the imperial government, as
rebels fromTaira no Masakado on discovered (see Chapter 10). But
through the power of commendation they could determine through
which route the taxes they collected made their way to the ruling
class and thereby jockey for the best possible financial and patronage position.
All of this adds up to a tremendous growth in the power and influence of local magnates. This growth, the central development of
rural society in Heian Japan, was favored by several factors: a significant expansion in population and, consequently, in the utilization of land resources; the spread of use of the plow and of related
agrarian improvements in the use of animals and fertilizer; and the
realization by provincial and central government officials that the
farming-out of tax contracts provided the most secure source of
national tax assessment and collection. And in the outlying
provinces of both east and west (but especially in the east, where
population expansion was greatest) there was a further development: the growth of a warrior-class identity that lent unity and purpose to the aspirations of developing tax managers and local magnates. The power to commend the tax land they controlled to the
political master of their choice was both the symbol and the measure of the growing power of this local tax manager class. More than
this: it was the very substance of power.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 4
PROVINCIAL ADMINISTRATION
AND LAND TENURE IN
EARLY HEIAN
The Heian era, as correctly portrayed in current studies, was characterized by a progressive weakening of centralized control over regional areas. Most of the administrative offices that had been established in the late seventh and eighth centuries for the purpose of
maximizing state power declined rapidly during the early Heian period, with the notable exception of the headquarters offices at the
capitals (Jkokufu) of the sixty-five (or sixty-six) provinces into which
the country was divided. While the central government declined,
these provincial offices (kokuga) retained, and even increased, their
power over local land and people as local elites took over their functions: the collection of taxes, the administration of land, and the
"promotion of agriculture" (kanno). The changes resulted in large
part from movement from below, and, at least in the first two centuries of Heian times (the ninth and tenth centuries), had little to do
with warriors or sheen (estates). The following discussion concentrates largely on those two centuries in an attempt to explain these
changes. A more detailed inquiry, however, requires a preliminary
discussion of local problems during the Heian period as a whole,
and the problems they raise for us.
Managed by officials dispatched for short terms from the capital,
the provincial offices were intended to be a new meeting ground for
capital and local elites, enabling the dynastic state to marshal local
resources and the loyal services of local aristocrats,1 but in practice
the provincial governments turned out to be unique bargaining
i The following abbreviations are used for works cited in this chapter: HIB forTakeuchi Rizo,
comp., Heian ibun, 15 vols. (Tokyo: Tokyo-do shoten, 1968-81); IK for Ienaga Saburo, Ishimoda Sho, Inoue Kiyoshi, Inoue Mitsusada, eds., Iwanami koza: Nihon rekishi, 23 vols.
(Tokyo: Iwanami shoten, 1962-64); KT for Kuroita Katsumi et al., eds., (Shintei zohd)
kokushi taikei, 60 vols. in 66 (Tokyo:Yoshikawa kobunkan, 1929-64); NKBT for Nihon koten
bungaku taikei, 102 vols. (Tokyo: Iwanami shoten, 1957—69).The most authoritative work on
Heian provincial administration is Yoshimura Shigeki, Kokushi seido hokai ni kansuru kenkyu
(Tokyo: Tokyo Daigaku shuppankai, 1979 reprint). See also MoritaTei, "Heian chuki gunji
ni tsuite no ichikosatsu," in Hayashi Rokuro, ed., Ronshu: Nihon rekishi, vol. 3 (Tokyo:
Yuseido, 1976), pp. 208-17.
236
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
237
grounds for the division of resources between capital and countryside, and that function, combined with the functions of the governments as repositories and redistributors of wealth, ensured their survival well beyond the Heian period. Provincial governments did
indeed marshal economic resources, but the share falling to the nobility in the capital steadily diminished. In the meantime, control
over labor and land use fell largely into the hands of local elites,
whose authority was, at the highest level, reinforced by provincial office titles unheard of in the eighth century. The structure of provincial authority changed drastically as local power grew.
The rise of a quasi-autonomous warrior elite in late Heian times
and the proliferation of tax-immune landholdings called shden weakened most provincial administrations and paralyzed one or two, but
60 to 70 percent of the landed wealth remained under provincial jurisdiction until the founding of the Kamakura bakufu in 1185.2 Up
to that time, when the Kamakura regime began to thrust itself between capital and province, the provincial governments remained
unchallenged as centers of local administration, but were profoundly
transformed. This transformation was primarily the result of social
changes in rural society that tended to strengthen purely local
power.3 Changes in local administration have also been seen as reflecting change in the political order as a whole, from centralized
empire to patrimonial state and ultimately to feudalism. Those
changes, including most notably the rise of the warrior class as the
dominant force, occurred at different rates in different places, but
the course of events can, in the light of recent Japanese scholarship,
be divided into three fairly distinct stages.
The first of the stages, which lasted through the ninth century and
most of the tenth, was marked by government edicts for the benefit
of a rural stratum of petty gentry, often called "the rich and powerful" (fugo), who, despite existing rules to the contrary, were able to
insert themselves between the tax-collecting officials and the taxrendering peasantry by assuming the latter's various burdens for the
convenience of the former. By about 900 those de facto arrange2 Murai Yasuhiko, Kodai kokka kaitai katei no kenkyu (Tokyo: Iwanami shoten, 1965), pp.
3 Major works propounding this view include, among others, MiyaharaTakeo, Nihon kodai no
kokka to nbmin (Tokyo: Hosei daigaku shuppankyoku, 1973); Morita Tei, Heian jidai seijishi
kenkyu (Tokyo: Yoshikawa kobunkan, 1978); Murai, Kodai kokka; Sakamoto Shozo, Nihon
ocho kokka taisei ron (Tokyo: Tokyo daigaku shuppankai, 1972); Sato Shin'ichi, Nihon no
chiisei kokka (Tokyo: Iwanami shoten, 1983); andTodaYoshimi, Nihon ryoshusei seiritsushi no
kenkyu (Tokyo: Iwanami shoten, 1967).
Cambridge Histories Online © Cambridge University Press, 2008
238
ADMINISTRATION AND LAND TENURE
ments had been granted almost full legal recognition as "provincial
precedents" by the government in Kyoto.4
"Riches and power" in early Heian provincial life were inextricably
bound up with the official taxation system. Wealth, moveover, was not
yet based primarily on the power of landownership. The nature of
wealth at the very beginning of Heian times is well illustrated by the
following passage from the Nihon ryoiki, an early-ninth-century collection of pious Buddhist stories. The story concerns an impoverished orphaned girl whose problems are solved by the miraculous intervention of Kannon:
When her parents were alive, they enjoyed abundant wealth and goods and
constructed numerous sheds and granaries . . . after die parents died, the
slaves ran away and the horses and cattle died, so diat the goods were lost
and die house impoverished.5
The same story, retold in the twelfth-century Konjaku monogatarishu, presents a revised version of the kind of prosperity that was lost:
In that region lived a district magistrate... .The underlings who served him
all left, and the fields he had held in domain were all seized by odiers so
that there was
no place left in her possession, and her distress grew worse
day by day.6
Early Heian texts do sometimes include land among the constituents of riches and power, along with slaves, animals, and stored
grain, but lands held as a "domain" (ryd) are never an element where
provincial figures are concerned. Power over grain, livestock, or
sources of labor like slaves or, more commonly, dependent clients,
had not yet become incidental to power over land.
Provincial governments during this first phase came to be headed
by a chief executive dispatched from the capital for a short term,
usually four years. That official, holding the formal title of governor
or vice-governor, was the "custodial governor." He alone among the
higher executive staff of the provincial administration assumed responsibility for the rendering of the province's taxes and the conservation of its assets. He was called a zuryo (literally, "custodian") and
could not be discharged from his office on the completion of his
4 Toda, Ryoshusei, pp. 25-45.
5 Kyokai, comp., Nihon ryoiki, in EndoYoshimoto and Kasuga Kazuo, eds., Nihon ryoiki, vol.
70 of NKBT, pp. 276-81; translated by Kyoko Motomichi Nakamura as Miraculous Stories
from the Japanese Buddhist Tradition: The Nihon ryoiki of the Monk Kyokai (Cambridge, Mass.:
Harvard University Press, 1972), pp. 173-74. Cited in Toda, Ryoshusei, pp. 32-33.
6 Minamoto noTakakuni et al., comps., Konjaku monogatari-shu, 16, 8, in Yamada Yoshio, Yamada Tadao, Yamada Hideo, and Yamada Toshio, eds., Konjaku monogatari, vols. 32—36 of
NKBT at 34:438-41. Cited in Toda, Ryoshusei, p. 33.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
239
term without a release from his successor certifying that an audit of
his account showed no irregularities.7 This concentration of responsibility in a single officeholder, rather than in the provincial officials
as a group, relieved many other appointees of any significant obligation whatever, converting their posts to virtual sinecures. The zuryo
system of specially designated custodial officials, a radical departure
from the principles of collective responsibility established by the organic (or statutory) codes (ritsuryo), originated during the Nara period, and by Heian times a complex network of regulations had
grown up around the transfer of custodial authority from one accountable governor to the next in order to circumscribe the resulting opportunities for misuse of public resources.
Fiscal and custodial responsibilities in the provinces were further
complicated by the two-tiered system prescribed by the organic
code. The upper tier was made up of the governors and their official
staffs, all capital aristocrats appointed for a short term, usually four
years. But each province was further subdivided into districts called
gun (or kort), and each gun had its own magistrates who were selected from registered lineages of local nobility to serve for life
terms.8 The most important government stores were dispersed
among the various gun, whose magistrates exercised custody of the
stores jointly with the provincial officers, and during the course of
the ninth and tenth centuries custodial authority for governmental
stores on the local level was further fragmented among other rich
and powerful persons. The income generated by official grain lent
out from such stores for interest was a major source of income for
officials at both gun and provincial levels, making conflicts between
them inevitable, conflicts that were further exacerbated by the demands of the petty gentry for a greater share in these resources.
Although weak, the provincial governments of the ninth and tenth
centuries remained generally under central control. Governors were
sometimes killed in disagreements with local residents, and the more
distant from the capital a province was, the more dangerous it was
likely to be. Nevertheless, governors from the capital continued to
travel to their provinces and extract most, if not all, of the revenue
needed to support the capital and its nobles. The rich and powerful
7 On zuryo governors generally, see Abe Takeshi, Heian zenki seijishi no kenkyu (Tokyo: Shinseisha, 1974), pp. 143-250, 301-41; Izumiya Yasuo, "Zuryo kokushi to nin'yo kokushi," in
Hayashi Rokuro, ed., Ronshu: Nihon rekishi, vol. 3, pp. 173—84.
8 Sakamoto Taro, Nihon kodaishi no kisoteki kenkyo (Tokyo: Tokyo Daigaku shuppankai,
1964), vol. 2, pp. 142-51.
Cambridge Histories Online © Cambridge University Press, 2008
240
ADMINISTRATION AND LAND TENURE
were too well integrated into the existing system of official taxation
to revolt against it. Property, as an institution, was still weak. In the
second stage proposed here, that is, the eleventh century, the rich
and powerful members of provincial society consolidated their control over peasant labor but did not usually legitimate their control in
terms of landholding. Generally known as tato (probably "field
man"), they were primarily managers, rather than owners, of agricultural land, and the term tato designated the performer of a function, not the holder of a claim.9
But, beginning in the late eleventh century, there occurred a general sorting out of agricultural areas into two major categories:
provincial domains (kokugaryo) and shoen. Shoen, the specially designated landed estates of high nobles or religious institutions, had existed even in Nara times, but now that land management had been
largely taken over by local field managers {tato), disengagement from
the authority of provincial officials could be much more complete.
The provincial governors had, during the earlier phase, lost control
of their original function of promoting agriculture (kanrio) to the
local land managers, a loss that in itself meant the end of the ritsuryo
state. The provincial headquarters had nevertheless retained its
power to maintain public order and collect revenue. The removal of
a shoen from the province's fiscal authority meant that its lands and
the obligations of its field managers "belonged" to the shoen lord and
not to the province, making it, in the language of modern scholars, a
mature shoen. Shoen and province enjoyed a de facto, though incomplete, administrative independence that inevitably led to conflict.
The position of the field manager is fairly well illustrated by a decree of the government of Izumi Province issued in 1012 instructing
gun magistrates as follows:
The only basis for reclamation must be the promotion of agriculture; public and private profit similarly depend on the cultivation offields.Now although this province is small in area, the people living here are numerous.
Half concentrate on fishing and have no liking for farm work. Migrants may
sometimes want [to farm], but since they have no claim to the land, they
are unable to contract for its cultivation. The rich and powerful who have
always had fields in their control leave them fallow for years, claiming the
land is infertile. The failing prosperity of the province and the reduced benefits to the people are mostly due to this. Now in considering the situation,
it appears that policies can be adjusted, there being times for laxity and
9 Yoshida Akira, "Tato no seiritsu ni tsuite," Hisutoria, 16 (September 1956): 35-46.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
241
times for strictness. But how can there be private holding of what is public field?
Consequently, in the case of public fields that have been abandoned since
the fifth year of Kanko [1008], petitions of lesser people to cultivate them
are to be allowed even if the fields are claimed as old cultivations of "great nom-
inees. " However, if an "original nominee" who has not abandoned the cultivation of his old fields wishes to bring more land under cultivation, the
gun magistrate must, after an accurate survey of the grid-parcels containing
the new and old parcels, deny the petitions of other nominees.10
The division here of the subject population into residents and
"migrants" (rvniri) originates from a classification of the ritsuryd census system, but had by this time become a kind of legal artifice.
Ronin (literally "wave people") were not necessarily hapless wanderers, and could be both affluent and of long-standing residence. More
significant is the phrase "great and small field managers" in the
heading of the order and the corresponding terms "the rich and
powerful," "great nominees," "lesser people," and "original nominee." For both the terms "field manager" and "nominee" one might,
without much distortion, substitute the word "occupant." There was
nonetheless an important difference of connotation, since the word
"nominee," literally "name" (raa), primarily signified a name on a list
of licensed cultivators, guarantors of official revenue from land they
had undertaken to cultivate. Being a "name" did not mean holding
title to land, but, rather, having responsibility, viewed as a function,
for paying taxes on it.11
"How can there be private holding of what is public field?" That
rhetorical question, echoing the ritsuryd rule that each parcel of agricultural land had to be either "public" or "private" and that only private land could be privately possessed or enclosed, actually betrays
some uncertainty about the position of the field manager and his authority to withdraw a parcel from cultivation, an arrogation of the
kanrio power of the provincial administration. By the terms of this
order, it must be noted, the possessory rights of old cultivators are,
albeit indirectly, given considerable recognition. Despite their legally
precarious position, the field-manager cultivators whenever possible
treated their holdings as if they were private and heritable, and, not
surprisingly, relations between them and their administrative overlords were marked by persistent struggles over both tenure and revenue. By the middle of the eleventh century, a rule called "the law
10 Izumi provincial order of Kanko 9(ioi2)/i/22, in HIB 2:630, doc. 462.
11 Murai, Kodai kokka, pp. 292-97; Toda, Ryoshusei, pp. 247-56.
Cambridge Histories Online © Cambridge University Press, 2008
242
ADMINISTRATION AND LAND TENURE
of apportionment" (rippo) was established in every province, setting
some limits to the dues that could be extracted from holders of
provincial public lands.12
During this period, radical changes occurred within the provincial
administration. In the late eleventh century, for example, in the absence of the governor, vice-governor, or any other member of the officially appointed executive staff, the provincial headquarters began
to issue orders that were everywhere accepted as valid, a sign that
the power of the zuryo was passing its peak. The ritsuryo concept of
responsible officeholding was totally abandoned as governors ceased
to visit their provinces at all, and the interests of the capital were represented locally by gubernatorial delegates, or "deputy supervisors"
(jnokudai).
The provincial headquarters of this period always had a chief executive office, if not a chief executive. Called the "administrative office" (mandokoro), it was sometimes renamed the rusudokoro, or
"custodial office," a term fictively implying the governor's imminent
return. From that office the deputy supervisor could maintain control of the local officialdom.
Starting in the ninth century, the structure of the provincial headquarters changed dramatically. Except for the governor or acting
governor, the vice-governor and other regular officials mandated by
the codes lapsed into insignificance, and instead the officialdom was
divided into a series of functionally differentiated suboffices, such
as the land office (tadokom) and the militia office (kondeidokoro).
This functional compartmentalization of authority and privilege, although to some extent inevitable in any system, was fundamentally
contradictory to the basic premise of the ritsuryo state, that government should be carried out by an upper stratum of omnicompetent,
generalistic officials.
Such divisions had probably existed even during the eighth century, but, now completely staffed by local irregulars, these offices
(tokoro) had eclipsed the regular staff. The custodial-office system
brought the offices and their officials into unprecedented prominence. Called "resident officials" (zaicho kanjin), these functionaries
represented the provincial, as distinguished from the national, elite.
Some came from local lineages, others were descended from capital
nobles who first arrived in the province in an official capacity and
12 On rippo, see Katsuyama Seiji, "Koden kammotsu rippo no seiritsu to sono zentei," Shirin
70 (February 1987): 1-43; Murai, Kodai kokka, pp. 320-36; Sakamoto, Nihon ocho kokka
taisei ron, pp. 223-40.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
243
then stayed on, but all tended to regard their posts as heritable. The
offices of the gun magistrates, whose prerogatives had been in a
steady decline, were in effect amalgamated into a broad provincial
hierarchy through the officials-in-residence posts that many held
concurrently.13
Resident officials belonged to a corporate hierarchy centering on
the kokuga, and by the eleventh century had developed quasihereditary estatist interests in their posts. Their privileges, relatively
safe from cancellation by the Kyoto government, served to reinforce
the new type of landed property that appeared late in the eleventh
century.14
By the time of the second stage, in the eleventh century, provincial governments were, without permission from the capital, beginning to license certain holdings as the specially chartered possessions of their "cultivators," endowed with specific tax preferences.
Such holdings were generally termed betsumyo or beppu myo, literally,
"names by special order," but actually meaning something like "specially named holding." A "named holding," or myo, could cover an
extensive area, including residences as well as both cultivated and
uncultivated fields, and by the late twelfth century the myo had become a major component of larger shoen.
"Names" had been units of tax responsibility for several decades.
Originally, the names referred to were those of provincial gentry
listed as responsible for tax payments from specific areas or groups.
The earliest record of the word "name" (na) in that sense dates to
947.15 By the middle of the eleventh century the myo had become a
unit of specially administered land bearing the name of a fictitious
person. Despite the presumption of the Izumi decree of 1012, such
private holdings in public (i.e., provincial) domain were now in the
making.
The establishment of specially chartered nameholdings was
legally justified in a variety of ways, but usually the principal ground
for granting a provincial special decree was that the founder had
opened the area to cultivation. Reclamation projects on the scale of
nameholdings required, however, the cooperation of the provincial
authorities, as when from 1075 to 1079 one Hata no Tametatsu re13 Morita, "Heian chuki gunji ni tsuite no ichikosatsu," pp. 208-17; Yoneda Yusuke, "Gigunji
ko," in Kodaigaku kyokai, ed., EngiTenryakujidai no kenkyu (Tokyo:Yoshikawa kobunkan,
i960), pp. 215-46.
14 The development of the estatist office is fully discussed in Sato, Nihon no chiisei kokka. See
also Toda, Ryoshusei, pp. 116—65.
15 HIB 3:838-39, doc. 708. See, inter alia, Morita, Heian jidai seijishi kenkyu, pp. 300-316.
Cambridge Histories Online © Cambridge University Press, 2008
244
ADMINISTRATION AND LAND TENURE
opened about 80 cho (1 cho = 11,900 sq. meters) of abandoned rice
paddy in Harima Province using more than five thousand local peasant laborers to restore the irrigation facilities. The holder of minor
gun and provincial titles, Tametatsu became the certified reclaimer,
or "opening lord" (kaihatsu ryoshu) of the Hisatomi Settlement (ho).
Although the term ho, translated here as "settlement," at first meant
a subunit of the ritsuryo village, or go, in late Heian times it always
referred to a special proprietorship created for the benefit of opening-lord reclaimers or their sponsors in Kyoto. The private landholding rights of such people were validated by a customary rule, as
yet only tacitly acknowledged by the capital government, permitting
reclaimers of agricultural land to treat it as alienable property.16
Specially created settlements, of which the Hisatomi is the earliest
known example, were internally complex, as were their equivalents,
the specially organized myo. There were subtenures held by a lord's
"followers," the cultivation and residential rights of which were normally heritable but subject to the obligation to pay taxes to the lord.
That quasi-feudal structure was reinforced by the concept of rights
and powers expressed by the term shiki, usually translated into English as "office," but which had always denoted official function rather
than official title or status (kan). In 1098, when Tametatsu passed on
his rights to his son, his bequest included as a constituent of his proprietorship "the documentation shiki" of Hisatomi Settlement.17 The
appearance of the shiki in the late eleventh century was an essential
step in the development of the medieval domain, and of the medieval Japanese state.
Despite its literal meaning of "office" or, perhaps better, "commission," shiki inTametatsu's case indicated rights to possession and
income as much as it did official responsibilities. By its incorporation of economic benefice into the exercise of administrative power
and management, the term reinforced the still weak concept of domanial property in land, making the enjoyment of lordly powers into
a sort of official delegation of authority. Locally, it integrated the do16 On Hisatomi ho and Tametatsu, see Tametatsu's petitions of Enkyu 3(io7i)/6/25, HIB
3:1077, doc. 1059; Joho 2(io75)/3/i2, HIB 3:1122, doc. 1109; J5ho 2(io75)/4/26, HIB 3:1126,
doc. 1113; Joryaku 3(io7g)/ii/3, HIB 3:1177, doc. 1171; his bequest of Jotoku 2(io98)/a/io,
HIB 4:1351, doc. 1389; the decree of Retired Emperor Toba's chancery of Hoen
2(ii36)/2/n, HIB 5:1982-83, doc. 2339; and the discussion inToda, Ryoshusei, pp. 201-25.
17 HIB 4:1351, doc. 1387; on the shiki, see Nagahara Keiji, "Shoensei ni okeru 'shiki' no
seikaku," in Hogetsu Keigo Sensei kanreki kinenkai, ed., Nihon shakai keizai shi kenkyu,
Kodai chusei hen (Tokyo: Yoshikawa kobunkan, 1967), pp. 249—78; Nakada Kaoru, Shoen no
kenkyu (Tokyo: Shoko shoin, 1949), pp. 185-93; Sato Shin'ichi, Nihon no chusei kokka, pp.
41-45Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
245
manial holding into a larger official network dominated by the
provincial officials in residence. The introduction of a concept of office into the area of landholding, however, also linked it firmly into
the surviving framework of ritsuryo authority in a way that muted the
autonomy of the landholder vis-a-vis the capital nobility. Officeholding still came before property per se.
Long before the shiki appeared, the steady growth of local power
was severely weakening the provincial governors. Another related
challenge to the authority of custodial governors and deputy supervisors, and to the provincial headquarters itself, was the proliferation
of shoen. Although large estates of high nobles or powerful religious
institutions had existed for centuries, now in the eleventh century
they provided certain local lords with a chance to withdraw their
holdings entirely from the fiscal authority of the province. Private
holdings in "public" lands could be transferred to a temple or noble
house at the capital, while the transferor retained hereditary rights of
management and control over the cultivators. That process, called
kishin, or commendation, had a long history, but the transfer of office rights gave it a new significance. When a local lord (ryoshu) conveyed his land title to a high noble, reserving a supervisory office
right for himself, he brought his holding into the sphere of direct
courtier authority, which impaired the bargaining power of the
provincial government when tax and other immunities were asserted
later. Custodial governors and deputy supervisors reaped immediate
profits by their sponsorship of such arrangements, but created problems for their successors in office.
The rapid militarization of the eleventh-century rural elite was intricately related to the development of domanial landholding.18 A
critical step in both processes was the reinforcement of title to land
secured by reclamation with the holding of shiki rights. A secondary
development, the incorporation of shiki rights into the shoen structure and the act of commendation that often accompanied the incorporation,19 facilitated ^Men-holding among court nobles, which,
in turn, stimulated the court in 1069 to issue the first of a series of
decrees requiring nationwide registration and certification of taximmune shoen. In the ensuing century and a half, similar edicts were
periodically issued by the court of the retired emperor (in) as that
18 Taniguchi Kengo, "Mino no kuni Oi no shd ni okeru shokan ichizoku no ryoshusei: Onakatomi-shi no geshishiki soron o megutte," Hosei shigaku 28 (1976): 24-37;Yasuda Motohisa, "Bushidan no keisei," in Kodai, vol. 4 of Ienaga Saburo, IK, pp. 121-60.
19 Nagahara Keiji, "Shoensei ni okeru 'shiki' no seikaku," pp. 249-78.
Cambridge Histories Online © Cambridge University Press, 2008
246
ADMINISTRATION AND LAND TENURE
newly invigorated organ steadily expanded its judicial authority over
rights in land, an authority that up to that time had been almost entirely within the control of each provincial government.
One purpose of the new shden registration system was the elimination of ambiguity concerning the legality of shden through insistence on clear documentation of their establishment,20 and that, together with the judicial decisions of the retired emperor's court, led
to a gradual sorting out of lands into the two categories of shden and
provincial domain (kokugaryo). In the process the fiscal authority of
the provincial governments was largely confined to the provincial
domains, which those governments were obliged to maintain and, if
possible, expand.
What has here been designated the second phase of development
of Heian-period provincial administration (late tenth to early twelfth
centuries) can be seen as the time when domanial property supplanted administrative authority as the dominant factor in the organization of local power. But despite the development of domanial
tenures, the completely tax-immune shden was still a thing of the future, and the distinction between provincial domain and shden was
far from clear in many cases. It was not until the twelfth century that
the medieval type of local administration and land tenure became
pervasive. That necessary preliminary to the foundation of the Kamakura bakufu constitutes the third phase of Heian provincial history as presented here.
During the third stage (the first half of the twelfth century), the
system of land tenure sometimes called "the shden system" reached
maturity.21 Shden almost certainly never occupied the greater portion
of Japanese landholdings, but historically they enjoy the advantage
of better documentation, a record that throws considerable light on
the tenurial system generally. Although many Kamakura warriors
were resident officials (zaichd kanjin) and private lords on technically
provincial domains, the nobility in the capital in the twelfth century
was clearly almost entirely dependent on shden, giving die latter a
political importance far in excess of their relative area.22 Nobles in
Kyoto could, by prearrangement, enjoy a share in die income of the
20 See MuraiYasuhiko, "Shoensei no hatten to kozo," in Kodai, vol. 4, IK, pp. 41-88.
21 Murai "Shoensei no hatten to kozo," pp. 41—88; Murai, Kodai kokka, pp. 237-65; Nagahara
"Shoensei ni okeru 'shiki' no seikaku," pp. 249-78.
22 InagakiYasuhiko, "Ritsuryoseiteki tochi seido no kaitai," inTakeuchi Rizo et al., eds., Taikei
Nihonshi sosho, vol. 5 (Tokyo: Yamakawa shuppansha, 1973), pp. i39-72;Takeuchi Rizo, Ritsuryosei to kizoku seiken, vol. 2 (Tokyo: Ochanomizu shobo, 1958), pp. 436-75;Toda, Ryoshttsei, pp. 241—77; Yoshimura, Kokushi seido, pp. 684—91.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
247
custodial governors, but revenues from the provincial domains were
no longer freely available for the stipends of the court nobility.
The long career of Emperor Shirakawa (1053-1129), who reigned
from 1072 to 1087 and exercised the power of retired emperor from
1087 to 1129, straddles the second and third phases defined here, and
in many ways his regime marked the final transition from bureaucratic to estatist forms of political organization. His s/wen-limitation
edicts of 1075, 1099, 1107, and 1127 were a major factor in the growing body of law concerning shoen and their immunities vis-a-vis the
provincial authorities. Indeed, the announced restrictions and registration requirements operated as invitations to litigate, and shoen proprietors in the capital could sometimes be major beneficiaries of
such contests, which were adjudicated under the guidance of the
many legal experts on the retired emperor's staff.23 The zuryo governorships came to be treated by the court as mere sources of income,
and indeed provinces themselves, in the view of the courtiers, were
becoming mere estates, analogous to shoen. One aging courtier, reviewing Shirakawa's career on his death in 1129, made a short list of
the more regrettable abuses established during his time, including
the simultaneous granting of gubernatorial posts to three or four
children of the same father, the appointment of boys not much over
ten years old, the steep rise in fees, jogo, to be paid to the court for
an appointment, and the refusal of such governors to pay religious
institutions or noble houses the shares of provincial revenues they
were entitled to. The use of governorships to provide sources of official income for court nobles had been fairly common during the
eleventh century and earlier. Concurrent appointments to gubernatorial posts could be granted to capital officials as a kind of compensation. As a means of supplementing their incomes, high nobles
and royalty were also given the power to nominate appointees to certain provinces, that is, power to sell the appointments. In addition, a
governor could assure himself of a renewal of his appointment, or
perhaps a reappointment to an even more rewarding province, by
making a special contribution to the court, typically in the form of financial support for some official project. This meant the increasing
23 See Koizumi Yoshiaki, "Todaijiryo Akanabe-no-sho," in Gifu ken, ed., Gifu kenshi, Tsushihen: chiisei (Gifu: Taishu shobo, 1969), pp. 430-548, for a case where the shoen registration
process resulted in greatly increased immunities and prerogatives for the proprietary temple. Works on the shoen registration efforts of retired emperors and the functioning of
kirokusho include Ishii Susumu, "Insei jidai," in Rekishigaku kenkyukai and Nihonshi
kenkyukai, eds., Koza: Nihon shi (Tokyo: Tokyo daigaku shuppankai, 1970), vol. 2, pp.
193-220; Takeuchi Rizo, "Insei no seiritsu," in Kodai, vol. 4 of IK, pp. 89-120.
Cambridge Histories Online © Cambridge University Press, 2008
248
ADMINISTRATION AND LAND TENURE
politicization of the zuryo office, released by the retired sovereigns
from the bureaucratic oversight to which it had been subjected by
the government of the Fujiwara regency.24
At some time early in the twelfth century a new means of diverting the income of custodial governors to the upkeep of the capital
officialdom appeared in the form of chigyokoku, or "possessory
provinces." Possessory provinces were first awarded to court Counselors (nagori) in lieu of income from "support household" (fuko)
revenues that governors could easily avoid turning over to their designated recipients in the capital. Holders of possessory provinces,
who were never themselves governors, could dispatch "deputy supervisors" to look after their interests. These agents sometimes provoked the hostility of the officials in residence. The large number of
possessory provinces held by members of the Taira clan later in the
twelfth century was an important factor in the uprisings of 1181-85
that ushered in the Kamakura period. In a sense, the zuryo system
had within it the seeds of its own destruction; increasingly, the newly
consolidated (and militarized) local elites could deal with the capital nobility directly, without the governor as intermediary.25
It is a major aim of the following discussion to show how, during
the first two centuries of the Heian era, custodial governors acquired
so much access to free-floating resources in the provinces. That they
did so is beyond argument, and during the eleventh and twelfth centuries their posts were routinely regarded as assets to be tapped for
the redistribution of wealth in the capital. As the appointment of
minor children to these posts indicates, the nominal holder of the office was not necessarily the person responsible for actually collecting
the revenue. The administrative functions of child governors were
carried out by the household superintendents (keishi) of their noble
fathers or grandfathers, and the accountability of the gubernatorial
office was no longer directed primarily toward the offices of the ritsuryo state, but to the persons or institutions that had, in effect,
been given a prior claim on the tax income collected.
In the provinces, claims on the land, including those of the provincial government, were to some degree sorted out by the chancery of
the retired emperors. In 1114, for example, the newly opened Record24 Hashimoto Yoshihiko, Heian kizoku shakai no kenkyu (Tokyo: Yoshikawa kobunkan, 1976),
pp. 172-90; Takeuchi, Ritsuryosei to kizoku seiken, vol. 2, pp. 587-640; Yoshimura, Kokushi
seido, pp. 383-401, 548-649; Fujiwara no Munetada, Chuyuki, entry of Daiji 4(1129)77/25,
cited inTodaYoshimi, Chuyuki (Tokyo: Soshiete, 1979), pp. 245-48.
25 On fuko, see Katsuyama Seiji, "Fukosei no saihen to kaitai: ju-juni seiki no fukosei," Nihonshi kenkyu 194 (October 1978): 1—41.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
249
ing Office, or kirokusho, of Retired Emperor Shirakawa decided a
lawsuit between Toji, a major Kyoto temple, and the governor of the
province of Tamba concerning a property called Oyama-no-sho. The
previous governor had rejected the claims of the shoen, limiting Toji
to three cho of land that had been officially certified exempt from
taxes in the ninth century. Toji's complaint recounted a long history
of fluctuating provincial policies toward its claims to tax-exempt
property, and the nine-year-old governor replied through his representative that most of the estate consisted of "newly established"
shoen lands of the sort prohibited by the retired emperor's edict.
But after further litigation, the Recording Office decided in favor
of the temple in every respect, finding that the earlier documentation
was clear despite the intermittent efforts of governors (and probably
officials in residence) to restrict the tax immunities of the holding.26
The Recording Office, purportedly established to control shoen, had
in the careful exercise of its judicial impartiality actually reinforced,
and probably increased, the tax immunities of the Oyama shoen. The
holdings brought before the tribunal were likely to either win complete fiscal immunity or lose all claim to any immunity whatever.
That development was part of a general trend leading to the emergence of the totally immune shoen from which agents of the provincial headquarters were legally barred. The mature shoen, as it has
been called,27 was usually created through acts of commendation. In
the late eleventh century high-ranking courtiers had been the most
common objects of commendation by local proprietors, but in the
twelfth, the successors of those courtiers were often unable to maintain influence enough to fend off provincial authorities and were
therefore themselves required to commend their lordships to higher
powers, reserving shiki rights for themselves. Those higher powers
usually turned out to be temples or shrines favored by the retired
emperor's court.28
In shoen holdings that had reached the mature stage of development, an original local proprietorship had typically evolved into a
three-tiered hierarchy of tenures. On the lowest level was the successor of the original local proprietor, who was often called the "subofficer" (geshi) or "custodian" (azukaridokord). Above him was the
26 See Miyagawa Mitsuru, Oyama sonshi, Hommon hen (Tokyo: Hanawa shobo, 1964), pp.
77-103; and Takeuchi Rizo, Bushi no tojo, vol. 6 of Inoue Mitsusada et al., eds., Nihon no
rekishi (Tokyo: Chuo koron sha, 1973), pp. 155, 263-67.
27 Yoshimura, Kokushi seido, pp. 684-91.
28 Ishii, "Insei jidai," pp. 193-220; Murai, "Shoensei no hatten to kozo," pp. 41-88; Murai,
Kodai kokka, pp. 373-402.
Cambridge Histories Online © Cambridge University Press, 2008
250
ADMINISTRATION AND LAND TENURE
lord (ryoshu or ryoke) and often, at an even higher level, stood a chief
tenant, almost always a religious institution, called the "principal"
(Jionjo).
That core structure supported many variations of detail, the authority of higher tenants over lower differing widely from case to
case in both degree and kind.29 The principal institution (honjo) typically exercised judicial authority over lower tenancies. It frequently
was entitled to dues of various kinds but in some instances enjoyed
no substantial economic benefits, serving merely as a fiduciary for
the lord's interests so as to preserve the holding's immunities.30
When proprietors commended their land, they reserved to themselves suboffice, custodial, or lordship tenancies, promising faithfully
to render dues to the commendee.The subofficer and custodian categories of tenure, almost always designated as shiki (office) in the
source materials of the period, were generally treated as property interests in land, leading some scholars to regard the shiki as primarily a legal "right" to income from shoen lands. The power and privileges of shoen officers, however, were not entirely insulated from
those of the high proprietors. Rather, relations between the two levels were fluid, and the proprietors could in some cases exercise overwhelming power.31
Shiki still retained, however, some of the characteristics of the local
offices from which they had evolved. Most importantly, they required
certification of appointment (bunin) by a higher authority, usually the
recipient of the original commendation of his successors. In 1164, for
example, a litigant defending his custodian's shiki against a rival for
the same position with prior claim to the title, maintained that "when
a private domain has been commended to another, whatever kinds of
promises may have been made on the occasion, the prevailing practice is to replace a commender who turns against the lord." However
"private" the holding of the original domain, and regardless of the
prerogatives reserved by the commender, typically the right to treat
his position as hereditary, the customary law of the twelfth century
gave the shoen lord the power to cancel a shiki if the holder committed a breach of fides.
The commendation pact was not merely a matter of private property rights. The newly recognized authority of the major proprietary
29 Murai, "Shoensei no hatten to kozo," pp. 41-88.
30 Nakada, Shoen no kenkyu, pp. 60—301.
31 Nakada, Shoen no kenkyu, pp. 98-100, 185-90, and references cited in note 17, above; HIB
8:2948-50, doc. 3836.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
251
lords to award shiki marked a decisive step in the gradual assimilation of state power by particular groups. To be sure, the exercise of
authority over land and people, even on a moderate scale, could not
possibly be regarded as private in the modern sense. Throughout the
eleventh century, for example, dues collected locally on shoen were
fixed by precedent and provincial rules, as if the proprietor were
somehow a surrogate tax collector. But the certification of immune
shoen by the retired emperor's court and the s/zj&z-commendation
procedures of the late eleventh and early twelfth centuries marked
the final step in the appropriation of official power by local and central elites.32
The lands of the mature twelfth-century shoen were made up
chiefly of units called names (myo) or name-fields (mydderi). Typically, each such unit was called by a personal name, so that the designation of entire areas by names like Sanefuji, Motoshige, and so on,
became commonplace by the late twelfth century. Myo holders (myoshu), like their predecessors the field managers (tato), took charge of
cultivation and the rendering of the dues owed on account of their
holdings. These holdings were not ordinarily called domains (ryo),
but the term shiki was, in the century following, often used in connection with mydshu rights.
Areas not preempted by shoen, that is, the provincial domains,
were similarly organized. Shiki proprietorships in fact originated in
the provincial domain as the holdings of the more powerful resident
officials, particularly holdings reinforced by claims of reclamation,
came to be identified with their office functions. In the final analysis, shiki entitlements, on or off shoen territories, were part of a single network of authority centering on the imperial court. In the case
of the provinces, the retired emperor's court, the absentee custodial
governor, the beneficial holder (in the case of possessory provinces),
and the provincial officials in residence were interrelated in a way
not unlike the linkages between principals, lords, and subofficers.
The powers of landed proprietorship on the local level were undoubtedly enhanced by the general delegation of tax collecting authority to local gentry, both on shoen and provincial domain. In both,
there was a great variety of taxes and dues, but legal pleas for exemptions in the eleventh century gradually divided them all into two
broad categories. First was the "official goods" (kammotsu) tax, con32 The document of 1164 quoted above is HIB 7, doc. 3318, pp. 2634-35; on the appropriation of state authority, seeTakeuchi, Bushi no wjd, pp. 269-70.
Cambridge Histories Online © Cambridge University Press, 2008
252
ADMINISTRATION AND LAND TENURE
sisting of a specified quantity of rice and other commodities, and
second were the "irregular," and partially discretionary, taxes
(zokuji), including labor, that the provincial headquarters or shoen
lord demanded annually. The appearance of this new mode of taxation signaled the emergence of truly domanial property. By the late
twelfth century, the ultimate responsibility for both types often fell
on the landholding myoshu.
Not all lands were myoshu holdings, but these holdings, and their
proprietors, were the ultimate surety that taxes or dues would be
forthcoming from local communities, and the myo was the basic unit
of taxation. The provincial domain, which usually contained most of
the arable land, centered on a headquarters dominated by officials
in residence who presided over a number of myo or other shoen-Wht
entities, many bearing older official designations like gun or "village," but in fact managed by shoen-type proprietors.
On both provincial domains and shoen, a new form of levy appeared based merely on residence within a given proprietorship.
Collected by the authorities of gun, shoen, myo, or village from cultivating households on the next lower level, the "resident-householder
levy" (zaikeyaku), including both labor and commodities, could be
required of resident households (zaike) on several occasions during
the year. Zaikeyaku had a distinct resemblance to European feudal
dues; on certain holdings, for example, cultivators were expected to
provide the proprietor with the traditional eggplants and cucumbers
needed on the Festival of Souls (bori). The resident-household levy
has been seen as a natural outgrowth of a established domanial system. Nevertheless, it more likely developed from older practices.
These dues became incidents of residence only after earlier obligations of personal subjection were assimilated into the system of domanial property that matured only in the late eleventh and early
twelfth centuries. The underlying question, then, is: at what time was
domanial lordship established? The discussion that follows is mainly
concerned with the ninth and tenth centuries, when the domain itself was not a dominant feature. It cannot, however, completely
sidestep this problem, as it aims to point out developments that ultimately led to the medieval Japanese domain. The distinction between burdens of tenure and personal obligation is far from apparent in many of the sources, particularly those of the eleventh
century, and it is most likely that there was often little consciousness
of it, but, as in the Izumi document of 1012 already cited, what might
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
253
be called latent domains appeared early in the second phase defined
here.33
It must also be noted that the provincial headquarters, despite the
restrictions on its fiscal powers imposed by shoen, remained the political focal point of the province, especially in police and constabulary affairs. By the middle of the eleventh century, each provincial
headquarters had a resident constabulary, reinforced and countered
by the forces accompanying the acting governor (zuryo) or his
deputies (mokudai). This meant that the more eminent military
chiefs had their own regional bases in addition to commissions to
represent court authority in other provinces.34 The gradual division
of local territory and people between provincial domain and shoen
was an important and final step in the development of estate patrimonialism in early medieval Japan. When this is viewed as a culmination of the changes that took place in the preceding centuries, the
importance of provincial office and the kokuga becomes clear. The
provincial governments themselves proved to be among the most
durable of ritsuryd institutions, probably because from the beginning
they served to integrate the interests of local elites, capital officials,
and court nobility.
One crucial issue is how the relationship of provincial government
to local elites changed over time. In the early Heian years, when
provinces were still repositories of national wealth in the form of
stored grain, the governments could stand apart from local chiefs as
centers of commodity redistribution, public works, and agricultural
development. Then during the ninth and tenth centuries (the first
phase proposed here) those functions of the provincial governments
were appropriated by locally based authorities, a development that
did not, however, undermine the governments. Provincial authority
was gradually reconstituted as the resident officials assumed more
and more authority. From the late tenth to early twelfth centuries
(the second phase), a domanial landholding system evolved and, by
about the year 1100, the subject populace had also been reorganized,
bringing whole communities of cultivators under the patrimonial
control of domanial lords or proprietary officeholders. The stage was
thus set for the third phase, the division of both territory and people
between the provincial domains and shoen.
33 On the tnyoshu, zaike, and village of late Heian, seeToda, Ryoshusei, pp. 190-277,379-402.
34 SeeTodaYoshimi, "Kokuga gunsei no keisei katei," in Nihonshi kenkyukai, eds., Chiisei no
kenryoku to minshit (Osaka: Sogensha, 1970), pp. 3-45.
Cambridge Histories Online © Cambridge University Press, 2008
254
ADMINISTRATION AND LAND TENURE
The transition from the despotic ritsuryo polity of the eighth century to the Kamakura Bakufu at the end of the twelfth was not the
result of uniform, gradual, and continuous change. The course of
change was always sporadic, and not all changes took place everywhere in the country at the same time. The suggested three-stage periodization is consequently highly approximate, based chiefly on
documentation from the home provinces (Kinai) and vicinity, the
area that most Japanese scholars regard as more advanced in development than the "peripheral" provinces of the south and northeast.
In those areas more distant from the capital, local authority, including that created by military power, grew up more rapidly than in the
central regions, whereas nearer the capital, in the provinces of Kii
andYamashiro, the overwhelming presence of temple shoen holdings
had a distorting effect on historical development. Within that erratic,
regionally differing, discontinuous course of change, perhaps the
single most important development was the takeover of provincial
governments by the local officials in residence.
REGIONAL ADMINISTRATION
By the time Emperor Kammu began construction of the Heian capital, the territorial division of the country into provinces and districts
had nearly reached the form that was to endure for many centuries.
The last major change occurred in 823 when the province of Kaga,
on the Sea of Japan north of the capital, was created from the two
northernmost gun of Echizen Province, which were then further split
into four. In the following year the provincial-level island of Tane
(now called Tanegashima) was merged into the nearby province of
Osumi in southern Kyushu.35 No further alterations of provincial
boundaries were made after that, and the total of sixty-six provinces
and two islands persisted into the nineteenth century, albeit in an attenuated sense. The gun units into which provinces were subdivided
were slightly less stable. The total slowly rose from 555 in the early
eighth century to 592 in the early tenth.36
The ritsuryo system had established a rigid classification for both
provinces and gun, based on magnitude of land area, population, and
strategic importance. The sixty-eight Heian provinces (including the
two islands, Tsushima and Iki) were designated as either great, upper,
medium, or lower. This determined the number and ranks of officials
35 Yoshimura, Kokushi seido, pp. 157—58.
36 Yoshimura, Kokushi seido, pp. 219-35.
Cambridge Histories Online © Cambridge University Press, 2008
ADMINISTRATION AND LAND TENURE
255
dispatched from the capital to take command of each provincial government. Great and upper provinces had a provincial administration
(kokushi) containing all four levels of officials provided for in ntsuryo
law: a chief, an assistant chief, and major and minor administrators.
Thirty-five, or slightly more than half, of the provinces belonged to
the upper category, which had one officer at each level of officialdom:
a governor (kami), a vice-governor (suke), an executive officer (Jo),
and an inspector (sakari) ,37 The prescribed rank of the governor was
Lower Junior Fifth; that of the vice-governor, Upper Junior Sixth;
that of the executive officer, Upper Junior Seventh; and that of the inspector, Lower Junior Eighth. Like all other classes of provinces, the
upper province was also given three clerks (shisho) for whom formal
rank was not required, although it might be held. Lower provinces,
by contrast, had only a governor (Lower Junior Sixth Rank), an inspector (Upper Lesser Starting Rank), and the three clerks. This
meant that in the nine smallest provincial units a mere clerk might
function as acting governor if both his superiors were absent or incapacitated. The complete scheme of provincial administration is represented in Table 4.I.38
In addition to administrative officials, the organic law provided that
a physician and a professor be appointed to each provincial capital,
selected either from the lower officialdom of the capital or from the
local population. (During the Nara period, at least, such officers were
ordinarily dispatched from the capital.) Like the three clerks, they
were part of the higher official staff of the provincial government, even
though they did notfitinto the formal four-tier hierarchy of ntsuryo
office structure and were therefore "irregular appointees" (zoniri).39
Divided among eight large administrative regions called circuits
(do), the provinces included all the territory of the country outside
the capital. A total population of somewhat less than 5 million persons was distributed over capital and provinces. Among the provinces, Hitachi and Mutsu in the far northeast had the largest populations, estimated respectively at about 217,000 and 186,000 in the
eighth century.40 They also had the largest recorded areas of culti37 Yoshimura, Kokushi seido, pp. 115-46.
38 Taken from Abe Takeshi, Owari no kuni gebumi no kenkyu (Tokyo: Shinseisha, 1971)) p. 60.
39 Nyuya Tetsuichi, "Zaichi tone no keisei to rekishiteki ichi," in Osaka rekishi gakkai, ed.,
Chusei shakai no seiritsu to tenkai (Tokyo: Yoshikawa kobunkan, 1976), pp. 141-224;Yamada
Hideo, "Sammi no kenkyu," in Sakamoto Tare hakushi kanreki kinenkai, ed., Nihon kodaishi ronshii, vol. 2 (Tokyo: Yoshikawa kobunkan, 1962), pp. 89-128.
40 Population figures for Hitachi and Mutsu are based on Sawada Goichi, Fukkoku: Naracho
jidai minsei keizai no suteki kenkyu (Tokyo: Kashiwa shobo, 1972)) pp. 47-310.
Cambridge Histories Online © Cambridge University Press, 2008
256
ADMINISTRATION AND LAND TENURE
TABLE 4 . 1
Provincial officials mandated by the ritsuryo
Governor
Kami
Vice-governor
Suke
Executive
Officer Jo
Inspector
Sakan
Clerk Shisho
Great Province
Daikoku
Upper Province
Jokoku
Medium Province
Chukoku
Lower Province
Kakoku
One: Junior 5th
Rank Upper
Grade
One: Senior 6th
Rank Lower
Grade
Two: One Chief
Executive
Officer, Daijo,
Senior 7th
Rank Lower
Grade; one
Assistant
Executive
Officer,
Shojo, Junior
7th Rank
Upper Grade
Two: One
Chief Inspector,
Daimoku
Junior 8th Rank
Upper Grade;
one Assistant
Inspector,
Shomoku, Junior
8th Rank Lower
Grade
Three: No
Specified Rank
One: Junior 5th
Rank Lower
Grade
One: Junior 6th
Rank Upper
Grade
One: Junior 7th
Rank Upper
Grade
One: Senior 6th
Rank Lower
Grade
None
One: Junior 6th
Rank Lower
Grade
None
One: Junior 8th
Rank Lower
Grade
One: Senior
Initial Rank
Lower Grade
One: Junior
Initial Rank
Upper Grade
Three: No
Specified Rank
Three: No
Specified Rank
Three: No
Specified Rank
One: Senior 8th
Rank Upper
Grade
None
vated land: according to a dictionary compiled in about 935, there
were approximately 40,000 cho in Hitachi and 51,000 in Mutsu. Although those figures are not trustworthy enough to warrant any conclusions about agricultural productivity,41 they surely indicate a large
concentration of manpower in an area quite remote from the capital
and help to explain why the region quickly became the most difficult
for Heian officialdom to govern.
Kyushu presented special difficulties, not only because of its distance from the capital and its ample potential for hostility toward the
41 Minamoto no Shitago, comp., Wamyo ruijusho (Tokyo: Kazama shobo, 1974 facsimile ed.)
part 5, pp. i5b-i6a, i8a-i8b.
Cambridge Histories Online © Cambridge University Press, 2008
REGIONAL ADMINISTRATION
257
central government, but also because of its proximity to Silla, an
often unfriendly state, and the potential for foreign trade afforded by
its accessibility to the East Asian mainland. To cope with those problems, a Kyushu Government Headquarters (Dazaifu) was established in Chikuzen Province near the present Bay of Hakata. That
large, militarized bureau, headed by a governor general (sotsu) whose
prescribed rank was Junior Third, had a complex structure that in
many ways replicated that of the central government on a reduced
scale. Authorized to exercise broad political control over the nine
provinces of Kyushu and the two island provinces of Iki and Tsushima, the Kyushu Headquarters could affect fiscal policy in the entire area.42
The gun, of which there were at least two in every province, could
not, according to the organic code, contain more than one thousand
"households." Census households were large, somewhat artificial
groupings of related persons, on rare occasions numbering over one
hundred persons. One thousand households thus represented a substantial population, but the code's restriction on gun population size
probably did, nevertheless, set limits on concentration of power in
the hands of the local aristocratic families from which gun magistrates were recruited. Those lifetime appointees, whose office usually
presupposed a documented lineage (fudai) from earlier chieftains,
were something of an anomaly within the ritsuryo bureaucracy, since
in fact, if not in theory, meritocratic norms had little to do with their
selection. Officially certified pedigrees did not always guarantee obedience. As the gentry stratum grew, holders of lineage were sometimes challenged by other prosperous peasants who lacked the
proper genealogy.43 The organic code established five classes of gun
according to size (great, upper, middle, lower, and small), and for
the three larger sorts, at least one officer on all four administrative
levels: chief magistrate (dairyd), assistant magistrate (shoryd), administrative officer (shusei), and secretary (shucho). The official complements of each type are shown in Table 4.2.44
The actual work of governing a province in accordance with ritsuryo standards required hundreds of irregular appointees and auxiliary personnel at the provincial capitals, the gun (district) headquarters, and throughout the provincial territories. In 822 the central
42 On the Dazaifu, see Sasaki Keisuke, "Dazaifu no kannai shihai henshitsu ni kansuru shiron: omo ni zaiseiteki sokumen kara," inTsuchida Naoshige Sensei kanreki kinenkai, eds.,
Nara Heian jidaishi ronshu (Tokyo: Yoshikawa kobunkan, 1984), vol. 2, pp. 245-90.
43 See note 1, above. 44 Taken from Abe, Oviari no kunigebumi no kenkyu, p. 60.
Cambridge Histories Online © Cambridge University Press, 2008
258
ADMINISTRATION AND LAND TENURE
TABLE 4 . 2
Gun officials mandated by the ritsuryo
Great Gun
Daigun
Upper Gun
Jogun
Middle Gun
Chiigun
Lower Gun
Kagun
Chief
Magistrate
Dairyo
Assistant
Magistrate
Shoryo
Administrative
Officer
Shusei
Secretary
Shucho
Small Gun
Shogun
One
Magistrate
Ryb
government in Heian-kyo issued a directive aimed at standardizing
the number of corvee helpers, including craftsmen and a variety of
quasi-official functionaries, who could be fed at public expense while
working for the political authorities. The document reads in part:
Impressed workers to whom food may be granted: General attendants for
the four annual messengers (four for the court report messenger and two
each for the other three).
Scribes for the major-report and tax-grain-report offices (eighteen for
great provinces, sixteen for upper provinces, fourteen for middle provinces,
and twelve for lower provinces).
Paper makers for provincial supplies (sixty for great provinces, fifty for
upper provinces, forty for middle provinces, and thirty for lower provinces).
Brush makers (two per province), ink makers (one per province) and
paper craftsmen (six for great provinces, five for upper provinces, four for
middle provinces, and three for lower provinces).
Chief maker of annual arms supplies (1 per province) and workmen (120
for great provinces, 90 for upper provinces, 60 for middle provinces, and 30
for lower provinces).
Provincial corvee directors (320 for great provinces, 260 for upper
provinces, 200 for middle provinces, and 150 for lower provinces).
Guardians of local branch storehouses receiving tax-grain and the like
(twelve per branch).
Gatherers of black kudzu (two per province; not applicable to provinces
that do not contribute to the imperial table).
Laborers for each member [of the provincial-officer staff] (four serving
men).
Two keyholders, the tax-grain chief, and the officers of official storehouses (three men for each branch granary).
Cambridge Histories Online © Cambridge University Press, 2008
REGIONAL ADMINISTRATION
259
Tax-grain collectors (two per village), two tribute-tax chiefs, and assistant chiefs (one per village).
Commutation-tax chiefs (one per village) and corvee workers (fifteen for
a great gun, twelve for an upper gun, ten for a middle gun, and eight for a
lower gun).
One kitchen chief, fifty corvee workers, two vessel makers, and two
paper-making workers.
Three hay workers, workers providing equipment for two post-station
riders (four for each gun and post station), and grooms for spare horses
(one per gun).
Petitions from the various provinces concerning the aforesaid have not
been uniform, and accordingly the standards have been determined as
stated. Not included in this ruling are master weavers and apprentices making tribute-tax figured silk, shuttle makers (this does not apply to provinces
that do not render tribute of figured silk), transport directors and bearers
delivering miscellaneous tax items to the Council of State, workers accompanying incoming or outgoing provincial officers, or provincial and gun officers delivering tribute taxes, and workers bringing the associated documents, kitchen workers and station workers for post riders together with
grooms, ferrymen and the like, and gatherers of sweet kudzu, honey, and
boar fat, or those delivering such items to the Council of State.45
Although this particular directive is far from complete in its coverage of local government activities, its emphasis being only on
those functions for which the legal corvee-overhead was still unclear, it tells a good deal about the scope and organization of provincial government in early Heian times. The first two items, dealing
with the "four annual messengers" (yodo no tsukai),46 are noteworthy for the light they shed on how the provincial government prepared and delivered its reports to the capital. The messengers in
question were always, in early Heian times, regular officials of province or gun. The four reports they delivered to the capital at different times of the year were:
1. The court report (choshucho), which detailed the conduct of the
provincial administration over the course of the previous year, including the state of public buildings, irrigation facilities, etc.
2. The major accounting report (daikeicho), also called the major
report (daicho), which gave the population of taxable households and able-bodied workers, indicating thereby the theoreti45 Ordinance (kyaku) of Konin i3(822)/int. 9/20, Ruiju sandaikyaku, KT, vol. 25, pp. 279-81.
See also Nyuya, "Zaichi tone no keisei to rekishiteki ichi," pp. 141—224.
46 On yodo no tsukai, see Herail, Francine, Yodo no Tsukai ou le Systeme des Quatre Envoyes
(Paris: Presses Universitaires de France, 1966). See also Sakamoto, Nihon kodaishi no
kisoteki kenkyu, vol. 1, pp. 163-204; Sato Sojun, Heian zenki seijishi josetsu (Tokyo: Tokyo
Daigaku shuppankai, 1977), pp. 57-71.
Cambridge Histories Online © Cambridge University Press, 2008
260
ADMINISTRATION AND LAND TENURE
cal capacity of the province to render tribute {cho) and laborcommutation {yd) taxes.
3. The tax-grain fund report {shbzeicho) which gave a complete inventory of tax grain on hand and amounts collected or disbursed.
4. The tribute-tax report {kochocho) which itemized the tribute
commodities actually delivered to the capital at the time of
delivery.
The second and third reports are of special concern because they
dealt almost exclusively with taxable people and stored grain, regarded by all officials of the time as major capital assets belonging to
the country as a whole but subject to the custodial authority of province and gun. In this area fraught with potential conflict of interest,
liaison with the capital demanded special exertions, and preparation
of the major accounting and tax-grain reports, neither of which was
mandated by the organic code, and the many supplementary documents that were also demanded, had come to require special secretariats or offices, called tokoro (literally, "places"), at the provincial
headquarters.47
Those scribes and many of the other workers listed by the directives were clearly not ordinary corvee laborers, despite their designation as "impressed workers" (yotei) in the heading of the document.
They were among the hundreds of "irregular officials" {zoshiki)
drawn from the upper stratum of the local populace to complete the
myriad tasks that government regulations imposed on the officials of
province and gun.48 These petty gentry thus had from the very beginning substantial representation in the headquarters of both province and gun. Ritsuryo rules prohibiting the provincial-officer staff
from bringing private assistants with them from the capital, generally
disregarded by the early tenth century, originally made local collaborators all the more necessary. As responsibility for the tax revenues
from each province came to be wholly concentrated in a single governor, fiscal, and often military, assistants were an absolute necessity
47 See Kikuchi Kyoko, "Tokoro no seiritsu to tenkai," in Hayashi Rokuro, ed., Ronshii: Nihon
rekishi, vol. 3, pp. 100-54; Morita, Heianjidai seijishi kenkyu, pp. 51-78; Sakamoto, Nihon
kodaishi no kisoteki kenkyu, vol. 2, pp. 163-204; Sato, Heian zenki seijishi josetsu, pp. 67-98.
48 Eleventh-century customary rules regarding zoshiki may be found in Miyoshi no Tameyasu, comp., Choya gunsai, KT29^:517-25; see also Hojo Hideki, "Heian zenki chozei kiko
no ichikosatsu," in Inoue Mitsusada hakushi kanreki kinenkai, ed., Kodaishi ronsb (Tokyo:
Yoshikawa kobunkan, 1978), vol. 3, pp. 121-64; IzumiyaYasuo, "Choyosei no henshitsu ni
tsuite," in Kodaigaku kyokai, ed., Engi Tenryaku jidai no kenkyu, pp. 175-308;Yoneda, "Gigunji ko," pp. 215-46.
Cambridge Histories Online © Cambridge University Press, 2008
REGIONAL ADMINISTRATION
261
in dealing with the local elites permanently ensconced in the provincial headquarters.49
As indicated by the directive of 822, among the irregular officials
were various kinds of tax gatherers, and such officials were, ultimately, the providers of the labor power needed to maintain such
public facilities as irrigation works and post stations, as well as to
provide housekeeping services for provincial officers. Custody of official granaries and of branch granaries (in) was, as noted earlier, of
special importance because the grain in the granaries, let out for interest annually, was a major capital asset, providing the wherewithal
for ordinary provincial expenses, including compensation for the
provincial officers.50 Gun magistrates, acting jointly, presumably,
with the officers in the provincial capital, had been primarily responsible for the granaries, but very early it was decided to disperse
the holdings within the gun, thus diffusing responsibility over a far
greater number of minor functionaries.
Although the reasons given for the change were geographic convenience in the collection and disbursal of tax-rice, as well as the reduction of damage in case offire,the measure was in fact intended
to appease disgruntled local elites jealous of the gun magistrates'
monopoly over a major source of financial power. By conferring
public legitimacy on privately held grain, it systematized the petty
gentry's power over peasant labor by making their residences official
centers for the distribution of rice to a client population. This delegation of official power to "promote agriculture" (kanno) was an important step in the development of the gentry, whose residences and
granaries would later become nuclei for rural domains like that of
Hata no Tametatsu discussed earlier.51
The numerous irregular officials listed in connection with branch
granaries, village tax collection, and the like were officials of the gun,
rather than of the provincial headquarters, and such persons sometimes appear in documents of the period as provisional gun magistrates.52 Although irregular officials could be regarded as menials by
49 Abe, Heian zenki seijishi no kenkyu, pp. 311—18; Choya gunsai, KT 2^:517—25, "Kokumu
jo jo no koto," esp. articles. 5, 6, 34,39, and 40.
50 Nakano Hideo, Ritsuryosei shakai kaitai katei no kenkyu (Tokyo: Hanawa shobo, 1979), pp.
258-59; Sonoda Koyu, "Suiko:Tempy6 kara Engi made," in Osaka rekishi gakkai, ed., Ritsuryo kokka no kiso kozo (Tokyo: Yoshikawa kdbunkan, i960), pp. 397-466.
51 Abe, Heian zenki seijishi no kenkyu, pp. 30-35; Murai, Kodai kokka, pp. 37-59; Toda, Ryoshusei, pp. 94-99.
52 Yoneda, "Gigunji ko," pp. 15-246.
Cambridge Histories Online © Cambridge University Press, 2008
262
ADMINISTRATION AND LAND TENURE
the capital officialdom, as in the document just quoted, their authority within their districts was undoubtedly great: they had the
power to allocate tax burdens, distribute grain revenues, and supervise the manufacture of fine fabrics and other high-quality tribute
tax items required by the court from each province. Other profitmaking opportunities for them included participation in the official
barter called "exchange" (koeki),*3 in which tax-grain was used to
purchase silks and other goods for shipment to Kyoto.
Most irregular officials enjoyed immunity from personal taxation,
adding to the material rewards they were able to gain for themselves
as providers of goods and services. They also sometimes held a grade
of court-appointed rank, obscuring somewhat the distinction in rankhierarchy between them and the regular officers of province and gun.
There were several reasons for discrepancies in rank-hierarchy between local residents and provincial and gun officers. Some local men
earned rank status through official employment as guards or service
workers in the capital, which they were allowed to retain on their return home. Their ranks set them above the lower gun officials, who
were usually rankless, and at times they held ranks higher even than
those of some regular provincial officials (although protocol required
that local magistrates dismount in deference to regular provincial officers, regardless of whose rank was higher).54
The distribution of privilege among this numerous local elite,
often masked by superficially demeaning functional designations,
was thus far from congruent with the hierarchical order presumed
by the organic code. To the wide variety of local people enjoying
some degree of tax exemption must be added the unregistered, or
separately registered, migrants (rdniri) whose services were specially
reserved for specific purposes, irregular officials,55 Buddhist monks
and nuns, Shinto priests and priestesses, militiamen (kondei), and
the twenty to fifty students admitted to each provincial academy.56
Provincial capitals and gun headquarters were built on a scale appropriate to the number and status of functionaries quartered there.
A provincial capital differed from the national capital of Heian in
that it was never, in the census system, taken to be the official residence of a subject household. In theory, all present were merely on
temporary duty. All provincial capitals shared the symmetrical grid
53 Murai, Kodai kokka, pp. 80-116. 54 Yamada, "Sammi no kenkyu," pp. 89-128.
55 For rules on legal exemptions, see Buyaku ryb, especially Toneri shisho no jo, Koremune no
Naomoto, comp., Ryo no shiige, KT 23:416-18.
56 Abe, Heian zenki seijishi no kenkyu, pp. 143—248.
Cambridge Histories Online © Cambridge University Press, 2008
REGIONAL ADMINISTRATION
263
pattern typical of ritsuryo planning. The scheme replicated the Heian
capital on a miniature scale, with the provincial headquarters standing in the place of the imperial palace buildings.57 The gun seats
(called gunke, or sometimes miyake) were also surrounded by walls
or moats. Besides the headquarters buildings and official residences,
they included a stable of post horses and official granaries, which, as
the document of 822 implies, required strict guarding.58
Although surviving official regulations provide scant information
about how all the mandated official work at the provincial level was
organized, it is reasonable to assume that it was thoroughly departmentalized. Some evidence appears in the directive of 822 where it
mentions the offices {tokoro) of clerks assigned to prepare the major
accounting report and the tax-grain report. Ancillary evidence of
departmentalization is found in the case of the Todaiji Construction
Office, the largest government bureau in the late Nara capital, which
was functionally divided into tokoro.59 It is thus possible that already
at the beginning of the ninth century most irregular appointees of
the provincial headquarters were attached to some such office. By
the middle of the tenth century, though, such was definitely the
case, and the offices had become permanent institutions, subordinate to the governors sent from the capital but also possessing a certain degree of autonomy. The irregular officials of the headquarters
were, more clearly than before, representatives of the locally privileged class.
One source describes the reception prescribed for a newly arrived
governor in the following way (the -sho in the document is the SinoJapanese allomorph of tokoro):
On this day, the irregular appointees of the offices come forward and pay
their respects (offices means the tax-grain office [zeisho], the major [accounting] report office [daichosho], the court report office [chdshusho], the
militia office [kondeisho], the provincial management office [kokushosho],
and the like). In this ceremony, those of the administrative office (mandokoro) lead the scribes and others standing in ranks in the courtyard. One
by one, each tells his office, rank, and full name, and after those statements,
[i.e., the governor] then commands,
all bow down again. The chief official
saying, "You are appointed . . ."6o
The concentration of local elites in the provincial headquarters as
irregular officials, and their extralegal division of fiscal and military
57 Fujioka Kenjiro, Kokufu (Tokyo:Yoshikawa kobunkan, 1974).
58 See Yamanaka Toshiji and Sato Koi, Kodai no yakusho (Tokyo: Iwanami shoten, 1984), for
a general description of gun headquarters, provincial temples, and barrier posts.
59 Kikuchi, "Tokoro no seiritsu to tenkai," pp. 100—54. 60 Choya gunsai, KT 29^520.
Cambridge Histories Online © Cambridge University Press, 2008
264
ADMINISTRATION AND LAND TENURE
functions into tokoro, show a degree of corporate autonomy quite antithetical to the original ritsuryo order. The balance between province
and gun established by the organic code was fatally upset, and the
gun and its officers merged into the corporate structure of the provincial headquarters. As members of the local elite rose to eminence
at the provincial level, there was a corresponding dilution of status
and function of the gun offices. During the eighth century, gun magistrates had to be examined by the Ministry of Ceremonial in the
capital before being appointed, reflecting the importance of the positions, but in 812 the government abandoned that procedure and allowed the headquarters of each province full authority to make the
evaluations. In 822 the government further stipulated that candidates for gun magistracies first be appointed provisionally for three
years of probationary service before becoming eligible for regular office. From that time onward, the orderly pattern of officeholding
prescribed for the gun by the organic code was a dead letter. Provisional magistrates tended to outnumber regular appointees, perhaps
because without formal tenure they were more firmly under the governors' control. The new system also permitted the total number of
gun magistrates to exceed the quotas of the code. Although provisional gun magistrates still ranked above the provincial irregular officials during the ninth century, both groups were merging into a
single dominant stratum of provincial gentry, overseers of tato for
whom gun boundaries meant very little.
The most striking structural change, however, took place on the
highest level of provincial government. The administration of provincial areas presented officers dispatched from the capital with
unique opportunities for private enrichment, and the original ritsuryo framework of rules soon proved inadequate to curb official rapacity. Early in the Nara period the government had sought to
strengthen its control of provincial governments by demanding, in
addition to the already staggering volume of correspondence required from provincial headquarters, the yearly tax-grain and great
accounting reports mentioned earlier. Those accounts were rigorously checked against accumulated prior reports at the capital, and
their acceptance by the relevant bureaus certified the provincial officers' good conduct. In the case of the tax-grain report, the government specified its acceptance by issuing a certificate of receipt
(hensho) indicating that the officers of both province and gun and the
province itself had discharged all tax-grain obligations. Both levels of
Cambridge Histories Online © Cambridge University Press, 2008
CUSTODIAL GOVERNORSHIP
265
officer could be held responsible for deficiencies and their own public-allowance rice {kugeto), confiscated to make up shortages.61
The organic code also provided for the occasional dispatch of very
high-ranking officials as circuit inspectors (junsatsushi) to make onthe-spot inspections of provincial administrations and report on the
conduct of the officers. The reports were basically personnel evaluations and could result in promotion, demotion, or dismissal for the
officers concerned.62 There were thus two kinds of sanctions that
could be used against unscrupulous officers. The government could
threaten their career interests by poor personnel evaluations, and it
could impose immediate financial penalties by forcing offenders to
make restitution for shortages in official stores.
THE ESTABLISHMENT OF CUSTODIAL GOVERNORSHIP
In 731 a new system of policing provincial officers was put into operation. Possibly inspired byT'ang-dynasty procedures, the new requirement obliged outgoing provincial officials to submit to an accounting by their newly appointed successors. New arrivals were to
incur responsibility for deficiencies they failed to detect, and officers
whose terms were expiring were ineligible for further appointments
until their replacements had issued discharge certificates (geyujo).
This innovation had profound effects on the ritsuryo system of
provincial office. It stressed the custodial aspects of office, forcing
incoming governors to seek out and take charge of all government
assets that were supposed to be on hand, particularly tax-grain. This
meant that some officers, usually the governor, had to assume complete responsibility for all such capital assets, leading in time to deterioration of the corporate integrity of the provincial-government
officer staff as the lower-level officers escaped all fiscal responsibility. The "discharge" (geyu) system also differed from earlier techniques of official oversight in its fundamentally adversarial, although
still bureaucratic, nature. It often resulted in lengthy disputes between incoming and outgoing officers of approximately equal rank
over whether or not a discharge certificate could legally be issued. By
the early ninth century, regulations governing this transfer of custo61 Sonoda, "Suiko," pp. 397-466. The term kuge occurs in the ryd;Zoryo, kuge no jo, Kiyowara
no Natsuno et al., comps., Ryo no gige, KT 228:340.
62 Nagayama Yasutaka, Ritsuryo futan taikei no kenkyu (Tokyo: Hanawa shobo, 1976), pp.
219-45; Abe, Heian zenki seijishi no kenkyu, pp. 47-62.
Cambridge Histories Online © Cambridge University Press, 2008
266
ADMINISTRATION AND LAND TENURE
dial responsibility had attained an amazing volume and complexity,
and every provincial headquarters of importance needed a clerk specializing in transfer procedures.63
These problems were already acute as early as 761, when legal doctors (myobo hakase), ordered to give an opinion on the criminal aspects of delays in transfer of office, and reasoning by analogy from
provisions of ritsu and ryo, concluded that if a new gubernatorial appointee could not obtain a satisfactory accounting from his predecessor within 120 days, he should bring his complaint to the Council of
State, or else forfeit his appointment and be held guilty of collusion
in misappropriation of official property. That opinion was enacted
into law by decree. A new appointee, under these rules, was now required to impeach his recalcitrant predecessor before the Council of
State, explaining to that body why he had refused to issue the discharge. The Council of State in its new supervisory role was thus
forced to duplicate the functions of the Tax Bureau {Shuzeiryo), the
Accounting Bureau (Shukeiryd), and other offices in the capital assigned to oversee the reports of the governors. The need for some way
to regularize the processing of the disputes arising from the discharge
system became increasingly clear in the latter half of the eighth century, when several ad hoc attempts were made to deal with them.64
At the center of many such disputes was the system of publicallowance rice (kugeto), first established in 74s.65 "Kuge rice" was in
fact a rice fund from which "seed rice" was lent to cultivators at an
annual rate of 30 percent, the proceeds of the loan being assigned as
stipends to provincial officials. Interest from such loans accounted for
many of the expenses of the provincial administration, as well as some
of the tax obligations owing to the central government. Called suiko,
the loans became an ever more important financial source for both
the imperial government and its local officials in the late Nara period,
each household within the jurisdiction of a provincial headquarters
being compelled to accept its share of the loans, regardless of need.
In 757 each type of provincial officer was allowed a specific number of shares in public-allowance rice as follows:66
63 For examples (the earliest from 959), see Choya gunsai, KT 2gA:529~33; See also Nagayama, Ritsuryo futan taikei no kenkyu, pp. 194-218.
64 Sugano no Mamichi and Fujiwara no Tsugutada, comps., Shoku Nihongi, Tempyo Hoji
2(758)^9/8, KT 2:255-56;Yoshioka Saneyuki, "Fuyogeyujo to kageyushi ni kansuru shiron,"
in Inoue Mitsusada hakushi kanreki kinenkai, ed., Kodaishi ronsd, vol. 3, pp. 87-120.
65 Shoku Nihongi, Tempyo 17(7455/11/27, KT 2:185. See Murai, Kodai kokka, pp. 11-36.
66 Shoku Nihongi, Tempyo Hoji i(757)/n/n, KT 2:243; Sugano no Mamichi et al., comps.,
Enryaku kotai shiki, KT 17:10-11; Sonoda, "Suiko," pp. 397-466.
Cambridge Histories Online © Cambridge University Press, 2008
CUSTODIAL GOVERNORSHIP
Governor
Vice-governor
Executive officer
Inspector
Clerk
Provincial professor
Physician
267
6
4
3
2
1
1
1
Although the scheme of apportionment changed slightly in Heian
times, the rights to public-allowance rice for provincial officers remained basically the same. They accrued both to those with substantive functions in the provincial government and also to provincial officers whose lack of real fiduciary responsibility under the
zuryo system had deprived them of any administrative role. Thus
arose a special class of nominal officers having a claim on revenue
from provincial stores.
Public-allowance rice had developed out of an earlier accounting
category called the provincial account (kokucho), established in 724
to pay certain local clerical expenses. The capital dedicated to this
account was reapportioned several times during the Nara period,
and part of the income was always reserved for kuge stipends. During Kammu's reign, in a reapportionment of 804, one-tenth the original amount was to be devoted as before to pay the local clerks and
the remainder disbursed to the higher officials. The province itself,
the edict illustrates, had become a major source of official emolument distinct from the central government. It is also clear that in the
period from 724 to 803, when the following order was issued, the category "provincial account" had become a subcategory of kuge (public-allowance rice). The original account now meant a subsidiary
portion, here one-tenth, of total kuge.
Determination of the portion of public-allowance rice to be reserved as
provincial account.
Great provinces: 12,000 sheaves. (In a general calculation of public-allowance-rice interests, from 10,000 sheaves, 1,000 sheaves should be set
aside for the provincial account. If the public-allowance rice is greater or
lesser in amount, always follow the same ratio. Upper, middle, and lower
provinces are also to follow this principle.)
Upper provinces: 9,000 sheaves.
Middle provinces: 5,000 sheaves.
Lower provinces: 3,000 sheaves. (The province of Shima and the three islands of Iki, Tsushima, and Tane are not within this rule.)
This matter has been examined. In the ordinance of the first year of Jingi
[724], third month, twentieth day, it is stated: "A portion of the tax-grain
Cambridge Histories Online © Cambridge University Press, 2008
268
ADMINISTRATION AND LAND TENURE
fund (shozei) should be set aside and let out for interest. This is to be designated provincial earnings and is to be allotted to the court-report messenger while he is away from the province, to unscheduled expenses, and
the clerks who proofread and copy reports, and also to the provision of supplies for bearers delivering articles other than tribute and commutation-tax
articles to the capital. The amounts used for lending are to be, for a great
province, 40,000 sheaves, and for a lower province, 10,000 sheaves."67
Note that the amounts of rice in ear authorized for each category
of province (12,000 sheaves, 9,000 sheaves, etc.) represent interest to
be distributed, rather than principal, and that they are all exactly 30
percent of the principal amounts given by the quoted order of 724.
By the standard specified in the document, an officer of a great
province could expect to enjoy a share in an annual income of 10,800
sheaves, the size of the share depending on his status.
That income, however, was not free and clear. According to the
regulations, any shortages in the main tax-grain fund (shozei) would
have to be made up out of the public-allowance fund before any distribution of it to local officers. Furthermore, provincial earnings
were to have priority over distributions to officers. Most interest income from the main tax-grain fund, meanwhile, was reserved as income for officials in the capital. The use of government rice for the
benefit of regional officers was originally regarded as a kind of interest-free loan to provincial officials out of the main tax-grain fund,
and logic as well as fiscal policy required that local public-allowance
beneficiaries receive last priority.68
Enforcing those priorities was, understandably, a major difficulty,
prompting the elaborate oversight mechanisms already described.
The discharge system was probably moderately effective in checking
misappropriations of public wealth. An incoming governor had an
incentive to see that the entire tax-grain fund was intact and that his
predecessor had not wasted any "government rice" (kanio). Assignment of primary responsibility for loss was a complex task. If there
was a shortage in a given granary, for example, under what circumstances should the loss be charged against the general publicallowance account, and when should the local officials in immediate
custody of the granary be charged? Such problems, repeatedly adjudicated, generated a substantial body of law.69
To expedite the litigation, the Kammu regime installed a new
67 Enryaku kotai shiki, KT 17:14-15; Murao Jiro, Ritsuryd zaiseishi no kenkyu (Tokyo:
Yoshikawa kobunkan, 1961), pp. 279-323; Abe, Heian zenki seijishi no kenkyu, pp. 49-51.
68 Murao, Ritsuryd zaiseishi no kenkyu, pp. 392-418.
69 Abe, Heian zenki seijishi no kenkyu, pp. 47-62.
Cambridge Histories Online © Cambridge University Press, 2008
CUSTODIAL GOVERNORSHIP
269
board, the Board of Discharge Examiners (Kageyushi), probably at
some time between 795 and 797. It was headed by Fujiwara no Uchimaro (756-812), then an official of Fourth Rank but destined later
to attain the title of Minister of the Right and Junior Second Rank.
In 803 he participated in the compilation of Enryaku kdtai shiki (Enryaku Regulations on the Transfer of Office), a compendium of rules
concerned with the administration of the discharge system by the
Board of Discharge Examiners. A copy was to be kept at every
provincial headquarters.
The new board was organized as a regular office with four levels
of officials: one chief, two assistant chiefs, three senior clerks, and
three junior clerks. Under such influential figures as Uchimaro and
later Sugano no Mamichi (738-811) the board was in a position to
exercise unchallenged authority over provincial officers. That authority took the form of compulsory arbitration and adjudication,
rather than administrative inspection of supervision.70
The establishment of the board appears to have been part of a
wholesale revision of the provincial administrations. During the 790s
the government forbade provincial officers the cultivation of any
local land whatever, and although it quickly modified its stand on
the issue, it took every conceivable measure to restrict the degree of
private control the officers could exercise over the use of labor or the
exchange of goods. That policy prompted edicts in 795 and 797 that
transformed the system of public-allowance rice. The first ruled that
deficiencies in tribute articles from a province would be a general
charge against the public-allowance fund, and the second set a limit
on the degree to which income from that fund could be impaired by
prior deficiencies, guaranteeing all provincial officials not personally
responsible for loss a minimum public-allowance income regardless
of the total deficit.
Although probably first established as a temporary expedient, the
Board of Discharge Examiners continued to function, not reaching
its full development until sometime after Kammu's reign. It was disbanded in 806 but reinstituted in 825 and was a permanent office
thereafter. During the nineteen-year hiatus in the Board's existence,
Circuit Inspectors were again appointed to oversee provincial administrations and their handling by the Controller's Office (benkari)
of the Council of State. A procedural change of 807 required a new
gubernatorial appointee first to present a charge of deficiency in the
70 Abe, Heian zenki seijishi no kenkyu, pp. 52—55.
Cambridge Histories Online © Cambridge University Press, 2008
27O
ADMINISTRATION AND LAND TENURE
provincial accounts to the incumbent governor before appealing to
the Controller's Office. That was intended to speed up the adjudication process by permitting the Office to disallow any new questions
not raised by the new appointee at the time of the original charge.
Only one such presentation of charges could be made, and the
incumbent was required to reply to each charge. The process resulted in a refusal-of-discharge statement (fuyogeyujo) containing all
changes and responses and signed by both parties. The discharge
process was also imposed on certain officers in the capital, resulting
in a work load too heavy for the Controllers' Office and its legal staff,
and the Board of Discharge Examiners was reestablished in 825, continuing the procedures instituted during its temporary demise.71
As noted earlier, the adoption of the discharge system gradually
accentuated the distinction within the provincial headquarters between officers with custodial responsibility (zuryo kokusht) and those
without, that is, between the custodial and the merely commissioned
provincial officials. The government of the late eighth and early
ninth centuries did not, to be sure, anticipate this strict differentiation between commissioned and custodial officials, and the collective responsibility of all officials serving in the same headquarters or
bureau was still stressed by the directives of the 790s empowering
the Board of Discharge Examiners. In early Heian times, moreover,
both the governor and the vice-governor of a province were considered responsible for official properties, and both needed a discharge
certificate for the properties when vacating office.
Yet power was quickly centered in the hands of a single zuryo governor. In 879 the governor of Bungo Province complained to the
Council of State that his commissioned assistants were obstructing
his administration, stating in part:
The welfare of a province always depends on the chief official, and the conduct of affairs is not ordered by assistants. Furthermore, as for crimes by
gun magistrates, the law has its provisions: there is reduction in rank, also
confiscation of office land, and in the case of the severest penalty, there is
deprivation of rank and office. But the commissioned appointees are not
the officials for this [i.e., the enforcement of the law]. They take their personal concerns into public affairs and express their resentment. Sometimes
trusting the word of lackeys, they wrongly judge gun officers, and sometimes opposing the will of their chief, they commit violent crimes against
the clerks. Because of this, people capable of doing service are all afraid to
71 Order of Enryaku i4(795)/7/27, Enryaku kdlai shiki, KT 17:16-17; order of 797, Enryaku
i6(797)/8/3, Enryaku kotai shiki, KT 17:13-14; Yoshioka, "Fuyogeyujo to kageyushi ni
kansuru shiron," pp. 87—120.
Cambridge Histories Online © Cambridge University Press, 2008
CUSTODIAL GOVERNORSHIP
271
work, and only a few unreliable people take office. Even if the governor is
an honest leader and leads with skill, if the gun magistrates are not aligned
with him, his authority will be unavailing. How much worse, then, when the
officials and the people are not at peace and the district is in turmoil. When
one cannot change the old ruts, one cannot expect a new government. I respectfully petition that commissioned appointees not be permitted to pass
judgment. If irregular appointees are at fault and must
be judged, let the
chief official pass judgment and afterwards enforce it.72
The petition was granted, marking an important step in the rise of
the zuryo governor. The Council of State in its order granting the petition, after exempting Fifth and higher ranking officials from the
governor's judicial monopoly, condemned the exercise of judicial authority by the commissioned officers as injurious to the prestige of
officials sent out from the capital. The activities of the commissioned
officers clearly were seen as a complicating factor in the often adversary relations between the "chief official" and the local peerage.
Another important consequence of the Council of State order, and
one that the original petitioner must have intended, was that governors could now feel at least somewhat justified in bringing their own
personal staffs of assistants, including military assistants, with them
to their posts, something that the organic code had prohibited. The
Council took a decisive step in 897 by ruling that mere commissioned officers were entirely unaccountable to the Board of Discharge Examiners.73
The concentration of functional authority in the zuryo governor,
and consequent abandonment of the provincial staff as a bureau of
the central government, led in the early tenth century to a revised
picture of the ideal good official. The emphasis shifted decisively
from magisterial benevolence toward the people at large to effective
negotiation of taxes with local elites, not excluding the use of force
where needed. These elites, in turn, often conflicted, sometimes violently, with the zuryo governor over the distribution of official and
unofficial benefices within the province.74
Provincial officers, whose appointments were mere conferrals of
public-allowance rice (kugeto), had existed ever since the Nara period, when acting or concurrent provincial posts were first awarded
as benefices to officials in the capital. To those "remote appointments" (yoniti) may be added, by way of contrast, the unstipended
72 Order of Gangyo 3(879)/g/4, Ruiju sandai kyaku, KT 25:318-19.
73 Kyaku of Gangyo 9(885)74/9, Ruiju sandai kyaku, KT 25:246-47.
74 Choya gunsai, KT 2gA:5i7-25, "Kokumu jojo no koto," esp. articles. 5, 6, 34, 39, and 40;
Sato, Heian zenki seijishi josetsu, pp. 149-59; Toda, Ryoshusei, pp. 115-65.
Cambridge Histories Online © Cambridge University Press, 2008
2/2
ADMINISTRATION AND LAND TENURE
assignments to distant provinces that served as a form of exile.75 A
famous instance is the posting of Sugawara no Michizane (845-903)
to the Kyushu Government Headquarters in 901. Except for such
exiles, supernumerary officials were forbidden by a decree of 766 to
visit their provinces.76 A ruling of 826, departing apologetically from
the established taboo against appointing high royalty to subordinate
positions in the bureaucracy, sanctioned the appointment of princes
of the blood to stipendary governorships of the eastern provinces of
Kazusa, Hitachi, and Kozuke, which remained prince-of-the-blood
provinces {shinno ninkoku or shinno koku) for the next century and a
half. The somewhat loftier title of supreme governor (taishu) and the
income from the governorship were awarded to a series of major imperial princes who, as before, were barred from leaving the capital
area. The vice-governor of a prince-of-the-blood province, called a
great vice-governor (psuke), was the custodial governor.77 This new
form of sharing provincial revenues, later to be extended to other
nobles in the capital, was among the more important outgrowths of
the zuryo institution.
LAND AND TAXES
Although the administration of stored tax-grain came first among
the financial concerns of provincial officers in the ninth century,
control over land use and revenues was undoubtedly a close second.
The ritsuryo system recognized a bewildering variety of land tenures,
but from the viewpoint of finance, there were three broad categories
of rice lands: taxable fields (yusoden), tax-exemptfields(fuyusoderi),
and rental fields (chishideri). The tax from taxable fields, called so,
was legally 1.5 sheaves of unthreshed rice ears per tan, which was
only 3 percent of the yield from a top-grade field, but perhaps 5 to 6
percent in the case of average land.78 When the registered "owner"
rented a field out, the 50 tax was always collected from the actual cultivator. Tax-exempt land included fields allocated to official temples
and high officials. Rental fields, from the standpoint of the provincial governments, were state land that had not yet been distributed
as allotment fields (kubunderi) to cultivators. Such land was let out
75 On yonin, see Yoshimura, Kokushi seido, pp. 350-81.
76 Sakamoto Tar6, Sugawara no Michizane (Tokyo: Yoshikawa kobunkan, 1962), pp. 109-15.
77 Ordinance (kyaku) of Tencho 3(826)/g/6, Ruiju sandai kyaku, KT 25:198; on shinnokoku
generally, Yoshimura, Kokushi seido, pp. 373-76.
78 On the so tax and Heian measures, see the appendix to this chapter, "Note on Heian
Measures."
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND TAXES
273
(in Nara-period terminology, "sold") on a yearly basis for a price
equivalent to one-fifth of the putative yield. That price or ground
rent was called chishi. In the terminology of the ritsuryo commentators, taxable fields of all kinds, including peasant allotment fields,
were "private." The allotment-field holdings of each household were
to be readjusted every six years in accordance with changes in
household population, yet they were considered in law as "owned,"
that is, managed, by the household members to whom they had been
allotted.™
There was one category of fully alienable rice land called konden
(reclaimed fields). In 743 the government ruled that rice paddy
opened to irrigation at the reclaimer's expense on land never before
registered as cultivated should, up to a maximum area depending on
the status of the reclaimer, become his private chattel {shizai) free
from the prospect of reallotment. By the early Heian period such
land was treated as freely heritable and alienable as long as kept
under cultivation. As with all other private and public lands, registration of reclaimed fields was required. The reclamation and transfer of fields had to be approved by both gun and province. The law
concerning reclaimed fields facilitated the opening of large tracts of
land under private auspices during the Nara and early Heian periods. Principal beneficiaries were rich and powerful local elites with
private rice stores to invest in reclamation. Small-scale "cleared
fields" called chiden (or harita) were an important development of
the early Heian period, but the massive projects typical of Nara
times were not to resume until the eleventh century, when they were
invariably carried out on the initiative of local, rather than central,
elites.80
Private suiko loans were perhaps the major bulwark of the rich and
powerful stratum. The interest rate on such loans was commonly 50
percent, significantly higher than the 30 percent for a public loan
from the tax-grain fund (shozei).81 Private loans were a potent means
of control, as is suggested by the following excerpt from an earlyninth-century tale of a gun magistrate's greedy wife:
Or, when she lent rice, she used a light-weighing scale, but when she collected it, she used a heavy-weighing scale. She did not show any mercy in
forcibly collecting interest, sometimes ten times and sometimes a hundred
79 ToraoToshiya, Handen shuju no ho no kenkyu (Tokyo: Yoshikawa kobunkan, 1961).
80 R/aku ofTempyo i5(743)/5/27, Ryo no shuge, Denryo, kohai no jo, KT 23:372; Shoku Nihongi,
KT 2:372; Ruiju sandai kyaku, KT 25:441.
81 Miyahara, Nihon kodai no kokka to nomin, pp. 137-48.
Cambridge Histories Online © Cambridge University Press, 2008
274
ADMINISTRATION AND LAND TENURE
times as much as the original loan. She was strict in collecting debts, never
being generous. Because of this, many people worried a great deal and
abandoned their homes to escape from her, wandering to other provinces.
There had never been anybody so greedy.82
Private suiko loans could be secured by the pawning of family
members, and creditors could distrain the property and the labor of
those in arrears. In 751, creditors' rights were severely restricted and
seizures of debtors' lands were forbidden, but the efficacy of these
legal restrictions seems to have been quite limited.83 Under such circumstances, the power of the rice lender was easily extended to
power over the land cultivated by the borrower.
Reclaimed fields and private suiko loans were among the major
factors leading to the breakdown of the land allotment system established by the organic code.84 Reclaimed fields did not, however,
preempt existing allotment land, as the regulations for the reclamation of land applied only to land never before cultivated, not to abandoned or ruined paddy.85 Loans, on the other hand, were a means of
de facto exploitation of every kind of land. That was recognized by
the central government, which occasionally forbade private loans.
The issue became even more acute when public loans from stored
tax-grain became a prime source of state revenue in late Nara times.
The conflict of local and central interests was most certainly a factor in the incidents of arson (shinka, literally "divine fire") that destroyed numerous tax-grain granaries in the late Nara and early
Heian periods.86
The collection of the ritsuryo tribute (cho) and labor-commutation
(yd) taxes could work to the advantage of proprietors of private granaries. According to a petition of 823,8/ peasants needing food in the
months immediately before harvest time obtained it from private
granaries in return for cloth and other commodities that would later
be needed for payments of those taxes. When the taxes were due and
82 Kyokai, Nihon ryoiki, pp. 392-97; trans Nakamura, Miraculous Stories, pp. 206-8. Cited in
Nakada, Shoen no kenkyu, pp. 276-83.
83 Kyaku of Tempyo shoho 3(75:0/9/4) Ruiju sandai kyaku, KT 25:403-4; Sakanoue Akikane
et al.j comps., Hossd shiyosho, in Hanawa Hokinoichi et al., eds., Shinko gunsho ruijit, vol. 4
(Tokyo: Naigai shoseki, 1983 reprint), pp. 164—211; Sato, Nihon no chusei kokka, pp. 52—55;
Murao, Ritsuryo zaiseishi no kenkyu, pp. 288-92. Ryo rules on suiko are found in Zoryo,
provisions 18, 19, 20, and 21, Ryo no gige, KT 226:336-37.
84 Nakada, Shoen no kenkyu, pp. 276-83.
85 Iyanaga Teizo, "Ritsuryoseiteki tochi shoyu," Kodai, vol. 3 of IK, pp. 33-78.
86 Murao, Ritsuryo zaiseishi no kenkyu, pp. 392-438, 49-257.
87 On these taxes, see Hojo, "Heian zenki chozei kiko no ichikosatsu," pp. 121-64; Izumiya,
"Choyosei no henshitsu ni tsuite," pp. 175-308. The petition of 823 is recorded in a kyaku
of Konin i4(823)/2/2, Ruijii sandai kyaku, KT 25:434-37.
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND TAXES
275
the harvest was in, the granary proprietors resold the tax commodities to the same peasants at much higher grain prices than they had
paid for them. The official system of acquiring tax commodities and
goods in exchange for grain afforded provincial officers comparable
opportunities for profit. The officers purchased the requisite items
from local producers at the low grain prices that prevailed before
harvest but forwarded them to the central government in quantities
equivalent to the higher postharvest values.
The close relationship between administration of rice lands and
rice lending prompted the Kyushu Government Headquarters in
823 to propose a novel way of replenishing dwindling tax-grain
stores in the nine provinces of Kyushu.88 The petition conveying the
proposal to the central government stated that the total area of allotment fields in Kyushu was about 65,700 cho, which under good
conditions produced field tax revenues (50), and that there were
about 10,900 cho of unallotted public fields (koderi), or extra fields
(joden), that were a potential source of ground rent. Those figures,
showing an approximate six-to-one ratio of allotted to unallotted
land, were typical of Japan as a whole. Extra fields, it appears, were
so firmly established as an element of the provincial economy that
their conversion to allotment fields was infeasible. From the total
landed resources of about 76,600 cho, the petitioners recommended
the expropriation of 12,100 cho of "good" fields not subject to flood
or drought and their establishment as "publicly operated fields"
(kueiden).
The amounts of allotment and extra fields to be expropriated were
about equal; all were to yield so revenues, as dictated by the codes,
of fifteen sheaves per cho, for a total of about 181,400 sheaves. The
term "operated" is explained by this statement in the petition:
Impressfivecorvee laborers to operate each cho, giving them compensation
and food, and just as is done among the people, allocate stored tax-grain
for operating expenses. After the autumn harvest, restore [the grain] to the
original granaries.
The clear inference here that grain stores were the source "among
the people" of labor power to work private fields is one of several indications of on what terms the rich and powerful had their land cultivated, and what local officials and magistrates did when they were
said to "cultivate" {den) their office lands directly. The same method
88 Kyaku of Konin 14(8235/2/2, Ruiju sandai kyaku, KT 25:434-37; Murai, Kodai kokka,
pp. 61-79.
Cambridge Histories Online © Cambridge University Press, 2008
276
ADMINISTRATION AND LAND TENURE
was clearly being employed at about the same period by the temple
proprietor Gangoji on parts of its Echi shoen in Omi, where overseers
from local ^wn-magistrate families supervised the operation.89
For the convenience of administering publicly operated fields in
the manner just described, small branch granaries (shoiri) were to be
located throughout the areas involved. The petitioners in 823 also integrated the collection of the tribute and labor-commutation taxes
into their plan, maintaining that the new system would ensure the
delivery of such taxes equal to the amount due from 60,240 ablebodied male subjects. Mollifying the exploitative "private" practices
then prevailing, the local administration under the proposed system
would offer a more reasonable price for the tax articles during the
growing season, namely, twenty sheaves for tribute tax items and fifteen sheaves for the labor-commutation tax items owed by each
able-bodied male. The projected annual budget was:
Total anticipated harvest
Expenses
Cultivators' compensation
Cultivators' food
Repair of facilities
Price of tax commodities
Field tax [50]
Total expenses
Surplus for storage as tax-grain
5,054,120 sheaves
1,451,400
723,084
110,000
1,507,790
181,425
3>973>699 sheaves
1,080,421 sheaves
The expected annual return of slightly over 21 percent, although
less than the 30 percent authorized for "public suiko," was undoubtedly less risky and also covered expenses for repairs of buildings and
irrigation facilities that were normally covered by a separate stored
tax-grain account called "miscellaneous rice" (zoio). The field tax
was not normally merged into the stored tax-grain account, and it
was therefore taken as a deduction from that account. It was destined for the "nonmoving" (fudo) stores of permanent reserves,
which were not to be lent out.
The Kyushu Government Headquarters' publicly-operated-fields
project required the labor of 60,257 corvee laborers, or aboutfiveper
cho, each man to work for thirty days, which was the limit set by the
89 The term den, meaning land under direct supervisory control, including rights to the entire harvest, first occurs in a document relating to Echi no sho; Murai, Kodai kokka, pp.
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND TAXES
277
code for the local corvee service called "irregular corvee" (zoyo). To
oversee the work, the proposal recommended:
selecting capable men from the villages, make each one a director. Assessing his capacities, entrust him with one cho or more of land, and insofar as
wind, infield work is concerned, leave it entirely to him. If damage from
sects, or hail occurs, excuse him in accordance with the facts.90
Requiring five workers per cho and one "director" (shocho) for a
"cho or more," the scheme reflected the scattered dispersal of the
publicly operated fields, which made necessary the services of several thousand corvee overseers. The capacities of the villagers to be
chosen as directors are nowhere explained, but almost certainly authority within the local community was a factor in determining the
amount of land to be left in a director's care.
The proposal recommended continuing the system of publicly
operated fields for thirty years, permitting a total accumulation of
over 32 million sheaves of stored tax-grain, but the Council of State,
while acknowledging the merits of the idea, permitted it to be put
into effect for four years only, remarking, "what has been done since
past times surely should not be changed abruptly." Among the likely
reasons for the Council's reluctance is the probability that the approximately 20 percent rental (chishi) paid by local lessees to use the
"extra" public fields had been going to the Kyushu Government
Headquarters, and merging that land into the publicly operated
fields would have the institutionally disruptive effect of diverting
this income to the several Kyushu provincial capitals. Furthermore,
the buying-in of tribute and labor-commutation tax commodities
under the system of publicly operated fields probably annoyed private lenders who had profited by taking those goods as security for
food loans.
Trading tax articles for grain was clearly a major source of income
for the petty gentry, a source that the new scheme was deliberately
designed to coopt.The radical reduction in the area of "extra" rental
lands, moreover, must have also resulted in substantial losses for
rent-paying tenants. A glance at the figures shows that the "good"
fields selected as publicly operated fields had an average expected
yield of about 421 sheaves per cho. Even if a 20 percent rental were
charged, a renter could have still profitably operated the land with
paid labor, that is, the same kind of direct cultivation the state was
90 Yoshida Takashi, "Ritsuryo ni okeru zoyo no kitei to sono kaishaku," in Sakamoto Taro
hakushi kanreki kinenkai, eds., Nikon kodaishi ronshu, vol. 2, pp. 223-62.
Cambridge Histories Online © Cambridge University Press, 2008
278
ADMINISTRATION AND LAND TENURE
now undertaking. In sum, the scheme of publicly operated fields
would have undoubtedly improved the accumulation of stored taxgrain at the provincial level, but it would have also made the utilization of private granaries for control of peasant labor that much more
difficult.
"Extra fields" (joderi) had been, in ritsuryo terms, publicly owned
but privately operated by the lessees. In the new system proposed in
823, publicly operated land was in fact to be publicly exploited land,
with all reasonably predictable revenues going to public stores. The
exploitation of land, as distinguished from labor, was probably not
at this time the major revenue source for either provincial or central
governments, but it was gradually becoming so as the government's
share of this labor power came under the control of the petty gentry.
The rationale behind the old field-allotment system of the ritsuryo
had been, as Miyoshi Kiyoyuki (847-918) stated in his famous sealed
memorial of 914, to enable the peasant to produce tribute and laborcommutation taxes and tax-grain for storage.91 Peasant land holdings were ultimately a form of compensation to the people for paying taxes. Publicly-operated-field projects of various kinds, including
provincial fields, continued throughout the ninth and tenth centuries. The commitment to land allotment as the best way to acquire
revenues was, by Kiyoyuki's time, all but abandoned.
A ruling of 801 decided that reallotment of fields would take place
once every twelve years rather than, as earlier, once every six, the excuse being the difficulty of surveying the land.92 Another attempt at
nationwide reallotment was made in 806, but thereafter reallotment
on that scale ceased, and each province followed a history of its own
in allotment matters. The major cause of the failure of land reallotment was very probably resistance on the part of the many local irregular officials (zdshiki), themselves members of the petty gentry,
on whose cooperation government surveyors had to depend.
More important, if the scattered small-scale operations contemplated under the publicly-operated-fields plan were, as stated in the
petition, patterned on the kind of management found on "private"
land, "among the people," it is likely that however often private allotment titles were reassigned, the actual distribution of labor over
91 Tokoro Isao, Miyoshi Kiyoyuki (Tokyo: Yoshikawa kobunkan, ig7o);Tokoro, "Ritsuryo jidai
ni okeru ikenfushin seido no jittai," in Kodaigaku kyokai, ed., Engi Tenryaku jidai no
kenkyu, pp. 162—97. For a text of Kiyoyuki's memorial with commentary, see Abe, Heian
zenki seijishi no kenkyu, pp. 228—48.
92 Kyaku of Enryaku 2o(8oi)/6/5, quoted in kyaku of Jogan 1(859)72/3, Ruiju sandai kyaku, KT
25:427; Torao, Handen shuju no ho no kenkyu, pp. 281-414.
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND TAXES
279
the land would not have been much affected. Census registers were
also taking on a fictional character that would have added considerably to the difficulties of reallotment. At the same time, publiclyoperated-field experiments demonstrated that rice fields cultivated
by peasants who were assured of food and seed supplies produced
more than allotment fields. Large-scale publicly-operated-fields projects, however, were likely to irritate the petty gentry, who were beginning to challenge the £wn-magistrate class.93
Varieties of publicly-operated-fields projects were carried out
from time to time in Kyushu and elsewhere, but as the ninth century
progressed those schemes became decidedly more accommodating
to local elites. In 879, 4,000 cho of rice paddy in the home provinces
were designated office fields (kanden), to be "publicly operated" and
the anticipated proceeds to be applied to stipends of certain minor
officials.94 This marked deviation from the ritsuryo order of things,
where the central treasury had been the designated source of all such
stipends, was part of a general effort by the capital government to divest itself of fiscal burdens by shifting them to specifically designated
sources of income. At first, shares in provincial suiko funds had been
awarded to the officials, but when that source proved unreliable,
stored provincial tax-grain was appropriated outright for the purpose, and, at last, the office fields were established.
Two years later, in 881, a new directive ordered a reduction in the
mandated rice revenue from the office fields. This reduction in
amounts collectible was entirely for the benefit of the local land managers. Retreating from the original plan of direct cultivation for the
entire bloc, the new plan instead provided for leasing, for the legally
stipulated rent of 20 percent of estimated yield, of half the area to the
managers and direct cultivation, through their agency, of the rest.
This concession to the petty gentry was one that, fifty-eight years
earlier, had not occurred to the architects of the Kyushu scheme. The
land managers, tato, were now, in a sense, sharing the proceeds of
cultivation with the government. More important, the price for the
management of publicly administered land was now the granting of
possessory interests in part of the land to be managed. The text of
the directive fully acknowledges mat a concession has been made:
93 Hirata Koji, "Heian jidai no koseki ni tsuite," inToyoda Takeshi kyoju kanreki kinenkai, ed.,
Nikon kodai chiiseishi no chihoteki tenkai (Tokyo: Yoshikawa kobunkan, 1973), pp. 59-96.
94 Kyaku of Gangyo 5(8gi)/2/8, Ruiju sandai kyaku, KT 25:439-41. Murai, Kodai kokka, pp.
129-44; Nagayama, Ritsuryo futan taikei no kenkyu, pp. 258-74; Sato, Heian zenki seijishi
josetsu, pp. 187-210.
Cambridge Histories Online © Cambridge University Press, 2008
280
ADMINISTRATION AND LAND TENURE
Although we are strongly desirous of direct cultivation, we are concerned
that it will be difficult for officials and people to bear, and yet if the whole
is rented out, we fear that the profit for the court officers will be but slight.
The land-managing tato was to become an increasingly significant
force over the next three centuries.95
Another, closely related change in the office-fields system in 881
was that the appointment of labor chiefs or directors was no longer
restricted to native residents of an area but could include migrants
(roniri) as well. Tato could come from anywhere. Although being a
migrant was not in itself illegal, their regular employment away from
home contributed to the weakening of the original ritsuryd household registration system. Nor was this the first time that migrants
had been employed as petty officials in a government project. In 873
orders for the establishment of publicly operated fields in Kyushu
for the support of local defense directed the governor to select capable chiefs regardless of whether they were natives or migrants.96
Although partially obscuring the importance of the rural elite,
the diffuse language of the sources ultimately confirms it. Gentry
land managers were occasionally termed rikiden no yakara, "those
who maintain the fields." Borrowed from T'ang China, this phrase
indicated commendable peasant worthies. It was used regularly for
persons who, having contributed their private wealth to public projects, had merited official recognition, usually accompanied by tax
remissions, along with other exemplary subjects like filial sons and
chaste widows. For rikiden farmers, the assumption of burdens was
the key to privilege, and, indeed, this was the underlying rationale of
all privileges enjoyed by the gentry. The use of grain wealth to make
private suiko loans and thereby command peasant labor power, as
may be seen here, was not always disapproved, and contributions of
wealth to agricultural projects or famine relief, even under compulsion, could sometime lead to rank status as well as tax exemptions.
Rikiden no yakara were, for purposes of rural administration, the indispensable allies of the officials among the "people." The term rich
and powerful (fugo no yakara), on the other hand, expressed disapproval of the same class of gentry when they displeased the provincial authorities above them. In the 881 ruling on office fields, the simple word "people" was inferentially applied to gentry managers when
95 Murai, Kodai kokka, pp. 217-39;Toda, Ryoshusei, pp. 18-87, 241-77.
96 Fujiwara noTokihira et al., comps., Sandai jitsuroku, Jogan 15(873)712/17, ^^4:333-34; on
the legal status of ronin, see Morita Tei, Nihon kodai ritsuryoho shi no kenkyii (Tokyo:
Bunken shuppan, 1986), pp. 274-96; Murai, Kodai kokka, pp. 246-55.
Cambridge Histories Online © Cambridge University Press, 2008
LAND AND TAXES
281
the government stated that direct cultivation of all the new officefield holdings would be "hard for officials and people to bear."97
The employment of local gentry as estate directors, and the consequent need to assure them a share in the revenues, was not limited
to publicly administered land. It was common to all landholdings of
absentee proprietors, that is to say, shoen in the broadest sense of the
term. In the tenth century the term for legally privatized rice fields
so managed was shoden (estate fields). This new category demonstrated the increasing importance of land as a form of wealth that accompanied the growth of the gentry.98 Some of the migrant land
managers were drawn from former low-ranking "commissioned"
(jiin'yo) provincial officers who had (illegally) taken up permanent
residence in their provinces on the expiration of their terms. Such resettlement of minor nobles from the capital in the countryside had
been noticed and condemned as early as 797 in an order to the
Kyushu Government Headquarters. That prohibition was repeated,
apparently without notable effect, at least nine more times during
the next hundred years.99
The authorities of the ritsuryo state generally discouraged private
linkages between central and local elites. In 744, and again in 868,
for example, provincial officers were forbidden to contract marriage
alliances with gun magistrates or other persons under their jurisdiction, and capital officials were repeatedly prohibited from traveling
privately to the provinces.100 Despite these prohibitions, powerful
migrants, many of them from the capital, were a generally acknowledged feature of the late-ninth-century countryside, and legally irregular transactions between local gentry and Kyoto aristocrats proliferated. A petition in 881 from the vice-governor of Hizen Province
complained about the conduct of "former provincial officers, sons or
grandsons of princes and ministers." The petitioner stated that such
rich and powerful migrants lived together, seized "the cultivation
97 On rikiden noyakara, see Abe Takeshi, "Sekkanki ni okeru chozei taikei to kokuga," in Kodaigaku kyokai, ed., Sekkan jidaishi no kenkyu (Tokyo: Yoshikawa kobunkan, 1965), pp.
Z
9~55> Kameda Takashi, Nihon kodai yosuishi no kenkyu (Tokyo: Yoshikawa kobunkan,
!973)> PP- 340-61; Sato, Heian zenki seijishi josetsu, pp. 130-34; Toda, Ryoshusei, pp. 14-32.
On privilege and assumption of burdens, "liturgy," see Max Rheinstein, Max Weber on Law
in Economy and Society (Cambridge, Mass.: Harvard University Press, 1954), p. 163.
98 See Takeuchi, Ritsurybsei to kizoku seiken, vol. 2, pp. 371—91.
99 Kyaku of Enryaku i6(797)/4/2g, quoted in kyaku of Saiko 2(855)76/25, Ruiju sandai kyaku,
KT 25:383-84; Yoshimura, Kokushi seido, pp. 338-39 and 654-56; Toda, Ryoshusei, pp.
139744100 On intermarriage with local elites, see kyaku of Tempyo i6(744)/io/i4, Ruiju sandai kyaku,
KT 25:302, and kyaku of Jogan io(868)/6/28, Ruiju sandai kyaku, KT 25:303; Yoshimura,
Kokushi seido, pp. 61, 219-28.
Cambridge Histories Online © Cambridge University Press, 2008
282
ADMINISTRATION AND LAND TENURE
rice funds" of the people, did not accept official rice loans, made private rice loans, and at harvest time obstructed public business. He
proposed the following remedy:
In accordance with the provincial precedent of Chikugo, without distinction between former officers and migrants, let stored tax-grain be distributed to both in proportion to the extent of the fields they operate, and let
them [also] be required like natives to cultivate publicly operated fields. If
the powerful among them do not comply with this decision, let them be expelled from the district and refused residence there.101
The management of proprietary or publicly operated fields, the
petitioner's argument implies, must be subject to the same burdens
in the case of both natives and migrants, and the presence in the area
of the latter should be conditional on acceptance of those burdens,
specifically: (1) the cultivation of publicly operated fields, as a sort of
compulsory public service due from all rikiden-type chiefs; and (2)
acceptance of stored-tax-grain loans, here viewed as a kind of surtax
on land, to be assessed in proportion to the area cultivated, a practice that was gradually undermining the old ritsuryd grain-banking
system. As loans from stored tax-grain came to be regularly distributed in proportion to land under cultivation without regard to the
need of the cultivator, the actual transfer of loan funds from the government was becoming a needless formality. In the late ninth century, the responsible cultivator merely paid the interest on the assigned tax-grain loan while actually receiving only about half the
principal.102
Rikiden were basically small-scale operators, but their resistance to
tax-grain loans was nevertheless having an effect. Custody of official
grain stores was now shared by the local gentry, who were also given
interest-free loans from provincial stores unlisted in the annual Tax
Grain Report. A related threat to the official loan system came from
the private grain stores of the noble households and great religious
institutions, also administered with the collaboration of regional
gentry. They could function as shelters for gentry holdings against
the demands of local authorities.103
101 Sandai jitsuroku, Gangyo 5 (881)73/14, KT 4:495—96, discussed in Toda, Rydshusei, pp.
1824.
102 Edict ofJogan 4(862)73/26, Sandai jitsuroku, KT 4:89-90, discussed in Murao, Ritsuryd
zaiseishi no kenkyu, pp. 297-99, 435-38; Murai, Kodai kokka, pp. 232-46.
103 Abe, Heian zenki seijishi no kenkyu, pp. 325—27; Takada Minoru, "Chiisei shoki no kokka
kenryoku to sonraku" Shicho 99 (June 1967): 6-25; Toda, Rydshusei, pp. 144-65. The sheltering of local grain wealth is attested by kyaku of Jogan io(868)/6/28, Ruijti sandai kyaku,
KT 25:603-4, kyaku of Kampyo 7(895)^9/27, Ruiju sandai kyaku, KT 25:604-5 and kyaku
of Engi 2(go2)/3/i3, Ruijlt sandai kyaku, KT 25:605.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
283
THE SURRENDER OF CENTRAL CONTROL TO
PROVINCIAL AUTHORITIES
The allegation in the Hizen petition of 881 that provincial precedent,
or practice, in the nearby province of Chikugo justified the use of
migrants and former officers as land managers, notwithstanding repeated prior edicts to the contrary, demonstrates the qualified withdrawal of the central authority from the field of local regulation in
late-ninth-century Japan. As a valid legal norm taking precedence in
individual cases over ritsuryo law and even recent imperial edicts,
provincial precedent first gained broad recognition from the Council of State in connection with local finances. An early instance occurred in 873, when the Kyushu Government Headquarters was ordered after a nineteen-year lapse to redistribute allotment fields in
the province of Chikuzen according to a new apportionment plan.
When the Headquarters reported its compliance with the order, it
also noted some surprising modifications it had made on its own authority. It had, to begin with, expropriated 950 did of good land for
publicly operated fields in order to achieve a more reliable source of
tribute taxes than allotment fields provided. Even though that meant
eliminating the ordered distribution of allotment shares to women,
the Headquarters explained, the allotment fields in the province
were still double those of other provinces. The Headquarters also established large blocks of rental fields for "miscellaneous expenses,"
and it provided for the appointment of migrants as field managers.
By the end of the ninth century, similar provincial precedents had
been recognized for the provinces of Shimosa, Mino, and Harima.104
The surrender to provincial precedent of Council of State authority to distribute tax burdens was to continue steadily in the following centuries. At bottom a concession to the interests of the provincial gentry, it contributed substantially to the discretionary powers
of the zuryo governors. Although land was important, especially to
the gun magistrates, who were permitted extensive office lands, the
main force of irregular officials in the provincial governments depended heavily on the income from loans of stored tax-grain and the
labor power it represented. Despite the many difficulties reported in
official petitions and edicts, tax-grain stores and suiko-loan revenues
probably grew during the ninth century, even as actual custody of
104 Ishimoda Sho, "Kodai ho," in Kodai, vol. 4 of IK, pp. 255-87. The term occurs frequently
throughout Ruiju sandai kyaku. For the allotment of 873 in Chikuzen, see Decree of Jogan
7» Sandai jitsuroku, KT 4:333-34.
Cambridge Histories Online © Cambridge University Press, 2008
284
ADMINISTRATION AND LAND TENURE
the stores fell more and more into private hands and official reports
became cluttered with legal fictions intended to reconcile official
standards with intractable fact. By the end of the century, conflicting demands on these grain stores from central government and
local elites had become acute. This partially explains the practice of
reporting rice not really in official custody as let out in suiko loan
when, in fact, "interest" payments were being made on receipt of a
fraction of the reported amount. The Tax Grain Report was becoming more a means for evaluating the incumbent local official than a
tool of actual financial supervision.1O5
The attempts of middle-status nobles and officials to augment
their dwindling incomes by exploiting local rice stores and cultivating provincial fields in absentia (i.e., by establishing shoeri) threatened the local officialdom and, in the final decade of the ninth century, prompted a flurry of prohibitory edicts intended to preserve
the authority of the provincial officials and of the ritsuryo order generally.106 One of the more troublesome problems confronting the reformist regime was that of imperial grant fields (chokushiden).
Mostly wasteland or abandoned paddy reclaimed with the use of
provincial tax-grain stores by authority of an imperial decree (chokushi), these lands were directed by the palace treasury and often
(but not necessarily) dedicated to the support of high-ranking imperial family members. Widely distributed throughout the country,
by the late ninth century they could be found in nearly every
province, mostly in blocks of one hundred cho or more. The force of
the decree gave such fields priority rights to irrigation water and immunity from the field tax (50). They were operated in the same general way as so-called publicly operated fields, but under the ultimate
protection of the palace authorities. For that reason, the dominant
Fujiwara leaders seem to have seen chokushiden as a threat to their
own power as well as an unwelcome intrusion into the government's
provincial base.107
In one of a famous series of edicts in 902 aimed at restricting intrusion into the provincial economies by great families and religious
105 On the critical increase in declared provincial suiko funds, see Murao, Ritsuryo zaiseishi no
kenkyu, pp. 232—46; on gunji benefices, Sakamoto, Nihon kodaishi no kisoteki kenkyu, vol.
2, pp. 142—51. On the evaluation olzuryo, Murai, Kodai kokka, pp. 70-86.
106 Sato, Heian zenki seijishi josetsu, pp. 264—73.
107 See Kochi Shosuke, "Chokushiden ni tsuite," in Tsuchida Naoshige sensei kanreki kinenkai, ed., Nara Heian jidaishi ronshu (Tokyo: Yoshikawa kobunkan, 1984), vol. 2, pp.
291-327; Miyamoto Tasuku, "Ritsuryoseiteki tochi seido," in Takeuchi Rizo et al., eds.,
Taikei Nihonshi sosho (Tokyo: Yamakawa shuppansha, 1973), vol. 6, pp. 49—138; Murai,
"Shoensei no hatten to kozo," pp. 41—88.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
285
institutions in the capital area, the government ordered the abolition
of imperial grant fields, the cessation of collusive sale or gift of lands
and dwellings by farmers to members of the imperial family or
upper-ranking courtiers for the implied purpose of establishing
shden, a halt to the occupation of vacant and abandoned land by the
same imperial and noble figures for similar purposes, and a return
of peasant lands held by temples and shrines to their original owners of record. The stated purpose of this order, to restrict the extent
of privately reserved land and granary holdings, was nevertheless
subject to one important qualification. The order exempted from its
scope any shden headquarters, or "estate house" {shoke), that would
otherwise come under its provisions if its head - that is, the shden
manager - had obtained his position through "transmission" (soderi)
from an ancestor. This concession conferred on the hereditary shoke
head, as distinguished from the self-established one, a quasi-proprietary right. Shoden management was becoming a protected household occupation (kagyd), in other words, a kind of estate.
Collusion between the capital elites and prosperous peasants in
the reclamation of rice lands and the establishment of shden was not
new, but government orders of the late ninth century show that the
scale of such activity was steadily increasing. Wealthy peasants could
more often than before choose to evade the provincial headquarters
and become shden managers for the nobility, thus removing their rice
wealth and the labor power at their command from provincial control. A shden manager's establishment was, as the order of 902 shows,
a place where harvested rice was stored, and also a depot for private
suiko loans and a source of payment for the labor costs of field work
and reclamation.108
There was a growing tendency in the ninth century for provincial
governments to use the owners of private storehouses as intermediaries in the operation of the suiko-loan system. Tax-grain was loaned
to such owners, who reloaned it at 50 percent interest to farmers in
the area, returning 30 percent in interest (the generally prevailing
public suiko-loan rate) to the provincial authorities. From the tenth
century on, the process was simplified: the tax-grain was paid in directly to the private storehouses, instead of going first to the provin108 Kyaku of Engi •2.(so-2.)ljji^, Ruiju sandai kyaku, KT 25:607-9. The other eight kyaku of this
series, all promulgated on Engi 2/3/12 or 2/3/13, are recorded in either Ruiju sandai kyaku
or Koremune no Tadasuke, comp., Seiji yoryaku, KT 28; Inagaki, "Ritsuryoseiteki tochi
seido no kaitai," pp. 139-72; Sato, Heian zenki seijishi josetsu, pp. 295-305;Takeuchi, Ritsuryosei to kizoku seiken, vol. 2, pp. 371-91; Toda, Ryoshusei, pp. 74-113.
Cambridge Histories Online © Cambridge University Press, 2008
286
ADMINISTRATION AND LAND TENURE
cial granaries, and the operators of the storehouses simply paid 30
percent interest on the rice to the province. The practice was generally, if tacitly, condoned, and the discharge system did not substantially inhibit it. The result was a diminution of direct involvement by
the provincial governments in the operation of the tax system that
provided it with much of its revenue.
A legal requirement that each province keep on hand a designated
amount of tax-grain, with specified quantities dedicated to the
stipends of provincial officers (public-allowance rice) and miscellaneous use (such as upkeep of provincial temples), may not have been
literally observed, but it did impose on the provincial governments
high minimum quotas of suiko-loan interest to be collected. That
burden could be spread over the inhabitants of a province in a number of ways. Early in the ninth century, the government had intended
that the compulsory suiko loans be made on a per capita basis, but
differences of wealth made that impractical. A natural response to
the difficulty was to apportion the loans in accordance with the
wealth of the borrowers, but that opened up too many opportunities
for abuse by the minor provincial officers who actually toured the
districts imposing the loans, and resulted in a level of suiko defaults
that prompted over half of the recorded ninth-century sales of privately reclaimed fields.
By the end of that century, the area of land under cultivation in
each province had, in the government's accounting, become highly
fictionalized. There was provincial precedent in most areas for the
imposition of suiko loans, but despite general agreement on that
standard, responsibility for suiko payments continued also to be imposed on any others who were able to pay, and the apportionment
of the burden still involved a degree of local discretion somewhat inconsistent with the ritsuryo model of a capital-centered economy.
A redistribution of the state's economic resources under the direction of central authority was undoubtedly a major ideal underlying
the nine reformist Council of State orders of 902, of which the example cited above is typical. Fujiwara noTokihira (871-909), then the
prevailing voice on the Council, was attempting through the reforms
to reinforce the function of the court treasury as the principal source
of income for the nobles and officials in the Heian capital, and to restore the ritsuryo structure generally. For the last time in history the
Council called for a nationwide distribution of allotment fields and
insisted that undeveloped areas be kept open for both public and private use. It revived the old emphasis on labor-commutation and tribCambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
287
ute taxes, commodities produced and delivered to the capital by
provincial farmers' households. These orders attest to a somewhat
belated recognition of the ritsuryo distributive system as essential to
the solidarity of the courtier class.
Fields designated as allotment land continued to make up a large
share of the arable in provincial accounts up to the year 1000. Such
fields were still monitored closely by both provincial and national
governments at the beginning of the ninth century, and periodic reallotment was merely one of several control devices employed. Cultivated fields were registered by owner or allottee and their location
and ownership indicated on official maps. They were to be surveyed
annually by provincial officers, who prepared the register of standing crops (seibyobo), listing each household eligible to receive an allotment of land, which of their fields were leased out and which cultivated directly, and what lands had been rented from others. The
register thus showed the actual cultivator of each plot who, regardless of ownership status, was the person responsible for the payment
of the field tax (50) to be reported later in the field-tax report (socho).
Basically a device meant to assure centralized control over all rice
cultivation, the register was also to report all cases of land and crop
damage, a justification for partial tax remission for the affected
households and reduction of tax receipts expected by the capital
from the province. An early concession to the local gentry may be
seen in an 845 Council of State ruling that the register, which was in
any case being neglected, would no longer be required by the capital except as evidence of crop damage requiring tax remission. The
government thereby declared an end to its policy of centralized
monitoring of all leasing arrangements. Such arrangements were,
however, crucial to the changes taking place in the countryside, as
the government acknowledged in its edicts of 902.109
The government sought to prevent high-ranking nonprovincials
from acquiring land in the provinces, where their local estate managers were said to have imposed harsh and cruel regimes on the
peasantry. The acquisition of legally transferrable land by capital
elites, moreover, was only a small part of an essentially local problem. The fields that were available for purchase, mostly reclaimed
fields (konderi), provided local magnates or shdden directors with a
109 Fujiwara no Tokihira et al., comps., Engi shiki, Shuzeishiki, Seibyobo no jo, KT 26:690-97;
kyaku of Jowa g(842)/6/9, Ruiju sandai kyaku, KT 25:327-28; kyaku of Jowa 12(845)79/10,
KT 22:328, Seiji yoryaku, KT 28:330-31; comments to Buyaku ryb, Suikan no jo in Ryo no
shitge, KT 23:392-403; Murai, Kodai kokka, pp. 327-36.
Cambridge Histories Online © Cambridge University Press, 2008
288
ADMINISTRATION AND LAND TENURE
base from which to exert economic dominance over the legally registered cultivators of inalienable fields such as kubunden. The legal
holders of kubunden were forbidden to transfer the land to others but
could transfer cultivation rights on a year-by-year basis. Rice-loan
indebtedness, a major reason for the sale of alienable land, could
lead to the loss of control over kubunden to the holders of local rice
storehouses, so that inalienability of land afforded little real protection. Official records of landholding took on an increasingly fictitious character as the ninth century wore on, and reallotment of rice
fields became even more difficult as the authorities in the capital
grew preoccupied with checking suspicious census data. With control of allotment field cultivation steadily passing from the capital to
the provinces, the central government's reviews of allotment data
submitted by zuryo governors became in effect negotiations between
countryside and capital about the amounts of taxes due. By the end
of the century the central government was accepting patently fictitious census data in which women and children vastly outnumbered
taxable males, data that justified the occupation of extensive areas of
allotment fields by households of record with very little accompanying liability for tribute-commodity or labor-commutation taxes.110
The original presumption of the architects of the ritsuryo system
seems to have been that the crop yields from allotment fields would
be almost entirely consumed as food by the allottees. The fields were
therefore left ungraded as to fertility (unlike the public extra fields
rented out by provincial governments) and taxed at a low uniform
rate. The government's chief revenue source was thus not so, the
field tax, but the tribute-commodities and labor taxes. That situation
changed, however, as the government began to impose additional
taxes on crop yields, first through the public suiko-loan system, and
occasionally by the same direct cultivation methods it condemned in
902 when applied by nobles of the capital to provincial lands.
Another step in that direction was attempted in 862 when the
Council ordered a general revision of the taxation system in the
home provinces. The three major features of the plan were:
1. An increase in sofromthe long-established rate of 1.5 to 3 sheaves
per tan for all allotment fields, and 2 sheaves per tan for most
other fields subject to the tax (excluding reclaimed fields).
2. A reduction in provincial corvee obligations from thirty to ten
110 On the fictionalization of census and tax reports, see Hirata, "Heian jidai no koseki ni
tsuite," pp. 59-96.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
289
days, and a complete exemption of households in the capital from
the obligation. Labor paid out of the increased so was to be used
instead.
3. The abolition of compulsory provincial 5M*&o-loan quotas, except
for provinces short of the land and rice funds needed for the support of large Buddhist temples.111
Although the scheme was intended to remain in effect for an experimental three years, it lasted only two. The two prevailing objections were (1) that most allotment fields were of "lower lower" quality, making increases in the so difficult to collect and causing land to
go uncultivated when increases of the tax were added to an existing
rental payment; and (2) that outlays of tax-grain could not defray the
increased labor costs without a depletion of official stores, shortages
in which would be made even more acute by the lowered grain revenues from suiko loans. The interrelatedness of 50, suiko-loan revenues from stored tax-grain, and provincial corvee presumed by
the experiment came about because, unlike the tribute and laborcommutation taxes, they were the foundations of the revenue system
of the provincial headquarters rather than of the central government. We may suspect, moreover, that the failure of the new system
resulted less from the infertility of allotment fields than from the reluctance of local gentry authorities to accept the plan. A century
later grain taxes against allotment fields had nearly quadrupled, and
the burden was being sustained by the cultivators, albeit reluctantly.
This could not have been due to a sudden surge in productivity. The
more plausible explanation is that land had become easier to tax
because of the abandonment of the allotment system, which left
the distribution of cultivation rights to those fields entirely in local
hands, as well as the absorption of large numbers of the local elites
into the provincial-headquarters structure.
The Council of State in 902 thus faced two major problems: first,
the maintenance of central control over allotment land and public
fields; and second, the restriction of shoen formation. At the time, the
two problems were fairly distinct. Allotment fields and other lands
subject to so and suiko fees (sozeideri) and the public or "extra lands"
leased out annually (chishideri) were not then in much danger of misappropriation, remaining firmly within the distributive control of the
provincial governments. Newly reclaimed rice paddy, however,
could be sold freely by the reclaimer or his successors. Reclamation
111 On these changes in the land tax, see Murai, Kodai kokka, pp. 312-27.
Cambridge Histories Online © Cambridge University Press, 2008
290
ADMINISTRATION AND LAND TENURE
projects and land so reclaimed diverted local labor and resources
away from provincial control and into the hands of noblesfromthe
capital and their agents. For this reason, the Council wished to prevent the further development oi shoen, which were a focus of private
reclamation efforts.112
The type of shoen that was the object of the Council of State's restrictions in 902 was not the same as the very large temple shoen established in Nara times, which declined precipitously in the ninth
century. The earlier shoen had been almost completely dependent on
the support of the central government, and as extensions of central
authority to the local scene, they were highly vulnerable, soon being
abandoned by their mostly nonresident cultivators. The new shoen
that evolved in the ninth century were more clearly private in origin,
representing typically a cooperative relationship between, on the one
hand, a high-ranking noble or member of the imperial family and,
on the other, a local magnate or official. Although they did not always enjoy formalized tax exemptions, shoen-based gentry could expect special consideration from the fiscal authorities. The increasing
prominence of the local gentry in the tenth century resulted in a new
sort of shoen that, like 6yama-no-sh5, could support an adversary relationship with the provincial headquarters.113
The ritsuryo system had never totally banned nobles or temples
from having special interests in local economies. Such interests,
however, were usually well differentiated from the ordinary holdings
of provincial farmers. Very high nobles and official temples, for example, were given "support households" (fuko) by the government.
The recipient of such households was entitled to the tribute and
labor-commutation taxes from them and to part of the field tax from
their allotment fields.114 Originally controlled entirely by the provincial officers on the recipient's behalf, some of the support household
grants evolved into shoen, but others seem simply to have reverted to
the provincial domain as the importance of commodity taxes in personal income declined.
The greatest number of new shoen probably originatedfromcollusive agreements between upper-level peasants and middle-ranking
nobles. In the countryside, peasant gentry sought to avoid forced
112 Nagayama, Ritsuryo futan taikei no kenkyit, pp. 304—9; Sato, Heian zenki seijishi joseisu, pp.
294-319113 Nakano, Ritsurydsei shakai kaitai katei no kenkyu, pp. 278-79; Sakamoto, Nihon ochd kokka
taisei ron, pp. 66-95.
114 On fuko, see note 25, above.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
2Q1
suiko loans and other burdens. In the capital, a shortage of available
posts and stipends forced downwardly mobile nobles to develop private estates of their own. Middle-ranking officials in the capital, who
did not receive grants of support households, had to rely entirely on
disbursements from the central treasury. The four thousand cho of
rice fields appropriated in 879 to sustain their stipends, while helpful, could not totally offset the steady reduction in commodity tax
revenues. Minor royals and nobles seeking new sources of income
could, however, offer the protection from forced suiko loans that the
upper peasantry needed. The consequence was a proliferation of regional shoen holdings, accompanied by the quasi-legal resettlement
of capital gentility in the provinces. Already alarmed by these developments, the central government in 902 made renewed efforts to
check estate growth.
Shoen holdings by individuals were not necessarily prohibited; if
the "documentation was clear," as the order of 902 put it, the tenure
was legal. A condemnation of the "private administration" of provincial land by capital nobles issued by the Council in 895 also recognized the legality of certain types of shoen holding, but prohibited
new acquisitions.
Officials of the Fifth Rank and above already have high position, their responsibilities are not unimportant and each of them has a stipend independent of cultivation. Why then should they covet the profits of the fields?
Accordingly, the various imperial, princely, and ministerial houses and persons of Fifth Rank and above are absolutely prohibitedfromcultivating any
land other than their
estatefields(skoderi), imperial rankfields,rank fields,
and office fields.115
The legality of estate fields was, as implied in the Council orders
of 902, certified jointly by representatives of both the central and the
provincial governments. The precise distinctions made in the Council's orders are far from clear, but the sho houses mentioned there
seem to have been an essential element of both legal and prohibited
shoen. The recognition later given to sho houses with a history of two
or more generations shows the hereditary patron-client relationship
that could develop between capital nobles and rural gentry. The sho
house was an extension of the noble house. Its chief, under the patron's protection, could disrupt the provincial government's control
of peasant agriculture, as the edict of 902 suggests, by lending rice to
neighboring farmers and thus reducing them to dependent-debtor
115 Kyaku of Kampyo 8(8g6)/g/2, Ruiju sandai kyaku, KT 25:444-45.
Cambridge Histories Online © Cambridge University Press, 2008
292
ADMINISTRATION AND LAND TENURE
status. This also led to control over the circulation of commodity-tax
items, and the government quite naturally saw this as a major threat
to its authority.116
Shd houses were substantial residences that usually included privately reclaimed fields as well as granaries. Whether old or new, the
houses reflected the emergence of a new and numerous stratum of
wealthier peasants, operators of private granaries and leaders of
reclamation projects both for themselves and for patrons in the capital. More than the purchase of provincial lands by nobles, it was the
emergence of that elite peasant stratum that made the taxes mandated by the codes harder to collect and goaded the Council of State
into attempts at reform.
The Heian government's policies toward the rich and powerful
were far from consistent. Some high officials, like Miyoshi Kiyoyuki,
regarded them with undisguised hostility, but the Council of State
often acted to protect their interests. In 896, for example, the Council modified the ordinance of 743 permitting permanent possession
of reclaimed fields in order to make the titles of smallholders of such
fields more secure against takeover by powerful nobles. The earlier
ordinance, while recognizing the permanent ownership of reclaimed
fields by their developers, also specified that if wasteland awarded by
a provincial government for reclamation was not in fact brought
under cultivation within three years, the award could be revoked and
reassigned to another petitioner. In 824 a similar restriction had
been placed on the reopening of permanently abandoned allotment
fields (jdkoden). The rule of noncultivation for three years (sannen
fuko) was abused by powerful figures in the capital in order to deprive smallholders of partially reclaimed fields. This was explained in
a complaint by gun magistrates addressed to a Circuit Inspector that
became the occasion for the ruling in 896:
The peasants of the villages petition for abandoned or unreclaimed land
and, following the ordinances, bring it into cultivation. Then later some
temple, shrine, prince, or ministerial house, claiming that the land has not
been cultivated for three years, notifies the provincial government and requests reassignment of that land. The provincial government, relying on the
wording of the ordinances, grants the requests and reassigns the land. The
nobles go into occupation with no interest in [further] development but
only in the profits of the land [from renting out already developed parts].
Having respectfully surveyed the situation, the gun magistrates suggest that
when a peasant opens three or four tan of a cho of land [that is, 30 or 40
116 Murai, Kodai kokka, pp. 232-38, 270-80.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
293
percent] that he has received [for reclamation], but is unable to open it
completely because of his poverty and weakness, it is a grievous thing to
award the land to another merely because of die terms of die ordinances.
The peasant is gready to be
pitied. We pray for a decision from the Inspector quickly granting relief.117
The Council, after reviewing the history of land reclamation law,
decided that the basic principle of promoting agriculture (kanrio) required the protection of local farmers. It accordingly ruled that as
long as one-fifth of the land claimed was under cultivation, the
three-year rule was not to be applied.
A survey in 859 of the Echi shoen, a holding of Gangoji Temple, revealed the same struggle between local and capital elites over "the
profits of the land" so strongly implied in the Council of State orders of 902. Located in Omi Province, the Echi shoen grew into a
fairly large domain by the eleventh century, but in 859 it probably
did not exceed ten cho of arable field, divided like most shoen in the
ninth century between fields let out for a fixed rent and directly cultivated fields (eideri). In addition to a local superintendent (betto),
there were on the domain two field managers (tato) who rented
fields in die shoen for cultivation. Contemporary documents from
the same area show that these managers were members of a local
elite, and had the status of irregular officers in the local government,
collecting taxes and witnessing land transfers in that capacity. They
had reclaimed-field holdings of their own, but it is clear that their
property was not very secure, and suiko-loan debts often required
them to sell off their land. Such sales account almost entirely for the
steady growth of the Echi shoen.
The field managers principally responsible for the cultivation of
Gangoji's fields were not part of the temple's administrative framework. Their status depended on their position in the local community, not on delegation from the shoen proprietor, whose fields simply happened to be located in the area. Their interests conflicted
with those of the temple, and they and the other renting cultivators
on the shoen took every possible opportunity to increase their own
holdings at the temple's expense and to minimize rent payments.
The survey report of 859, written by Empo, the temple's representative, records his efforts, beginning eleven years earlier in 848, to help
the superintendent vindicate the temple's proprietary claims.
One frequendy disputed issue was the amount of rent due from
117 Kyaku of Kampyo 8(896)74/2, Ruiju sandai kyaku, KT 25:486-87.
Cambridge Histories Online © Cambridge University Press, 2008
294
ADMINISTRATION AND LAND TENURE
leased fields. Legally, the amount of rent depended on the assessment of the fertility of each field in terms of upper, middle, or lower
grades, an assessment made by the provincial authorities. Even when
a field was part of a shoen, the rent to be paid by field managers or
other cultivators was fixed by the provincial government. That control, which lasted into the eleventh century, meant that in principle
leased shoen fields were the equivalent of leased public fields, and
both could be called rental fields {chishiden). Empo's efforts to increase rents was therefore a campaign to have official assessments
raised, as he reported in the case of two particular parcels.
The aforesaid two grid-parcels were originally classified as middle-quality
fields. At present, an on-site survey shows diat they are clearly upper-grade.
We accordingly summoned the field manager . . . Echi-no-Hata-no-Kimi
Yasuo for questioning, and the assessor said, "This is clearly upper-grade
field. Why do you render only middle-grade rent? How can that not be the
crime of violating goods of the Buddhist clergy?" He answered, "This was
decided long ago and is not a recent matter. There is no deliberate act of
offense, so how can there be a crime?" I, the representative, pressed him,
saying, "Even if the officials negligently fail to recognize the grade, why do
field managers not correct it? In accord with what is proper, the fields
should be made upper-grade." He answered, "It will be done in accord with
what is proper. How can there be any resistance?"118
Empo's disputes over land ownership illustrate the temple's lack
of control over vacant lands in the vicinity of its fields. Echi shoen at
the time consisted solely of buildings and arable fields either in or
out of cultivation. The temple had no firm prior option on reclaiming new rice paddies on undeveloped land and no legal right to enclose it, regardless of proximity to its own fields. The shoen was
merely a complex of estate fields {shoden) registered with the provincial government, which kept maps on which all fields were located in
a grid of one cho survey squares called grid squares (tsubo).
Renters of the temple's fields, on the other hand, were free to reclaim paddy land on their own account and were thus enabled to
hold property in reclaimed fields adjacent to those of the shoen. Ambiguities could arise, and Empo seems to have felt, probably correctly, that they were likely to be resolved by local officials in the cultivator's favor. Typical of the several ownership disputes summarized
in his survey report is one in which the issue at stake was the illegal
merger of temple fields into the adjacent reclaimed fields of a local
118 Survey report ofJogan i(859)/i2/25, HIB 1:107-10, doc. 128; on Echi-no-sho (the Gangoji
domain); Miyamoto, "Ritsuryoseiteki tochi seido," pp. 49-138, especially 131-36.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
295
cultivator. The original temple fields were registered as permanently
uncultivated and thus unproductive of revenue for the temple. The
document reads as follows (note that the field manager in the rental
dispute already mentioned appears again here to defend the cultivator against Empo's charges):
In the case of this grid square, the original notation is "permanently uncultivated." Now on viewing the land, we found that it had become the reclaimed field of . . . Echi-no-Hata-no-Kimi Otonaga. Whereupon I as representative disputed this, saying, "This grid square originally consisted of 1
tan 160 bu [1 bu = 3.3 sq. meters] of temple field and 60 bu of reclaimed
land. But now temple fields are claimed to be permanently uncultivated,
and reclaimed land, originally small in quantity, is presently cultivated in
large quantity. I surmise from this that the original fields of the temple have
wrongfully been designated reclaimed land." It was said in answer, "The
original fields of the temple are described as being to the east, but the present reclaimed land is in the center of the grid square. Since the direction
is not the same, how can you say it is temple field?" (The person making
this answer was . . . Yasuo.) I, the representative, disputed this, saying,
"There were originally in this grid square 1 tan 160 bu of temple field and
60 bu of reclaimed field. The meaning of'east' is that, as between the two,
the temple field is to the east and the reclaimed field to the west. It does
not mean that the temple fields are on the eastern edge of the grid square.
Furthermore, rice paddies are opened from the bottom land first. How can
the temple fields be on a hillside and the reclaimed land in the valley? Here
the owner of the reclaimed field is twisting reason."
The temple's shoen holdings here had originated partly from
alienable residence- or garden-land donated or sold by individual
owners. As in other such cases, there probably was an original core
of temple fields already established by donation from the government or imperial family. But the addition of new fields to the core
holdings was not perfected until the provincial authorities registered the acquisitions on its official maps, a process in which local
elders played a crucial role. Their testimony, moreover, was usually
decisive in cases where records were ambiguous. As Empo's report
illustrates, the official maps did not indicate the precise location of
any holding within a single square. Empo's investigations disclosed
three cases where fraudulently redesignated temple fields belonging to Gangoji were sold to third parties. He reported success in recovering not only that property but also other temple shoen land
that had been misrepresented as public fields owing rent to the
government. Although the net gain was a mere 3.86 cho, the temple's managers, by a vigorous policy of purchase and exchange,
consolidated the scattered holdings into a sold block, thus laying
Cambridge Histories Online © Cambridge University Press, 2008
296
ADMINISTRATION AND LAND TENURE
the foundation for the somewhat more extensive shoen revealed by
a document of 1051.119
This later document shows another significant change. Empo's
statement shows that the Gangoji holdings of 877 were being treated
by the officials of Omi Province as chishiden, to be rented at one-fifth
the putative yield, and for this purpose all such land was classified as
upper, middle, lower, or lowest. All public - in other words, unallotted - land was classified in this way under the ritsuryo system, but in
the tenth century such meticulous control over the land by the central government could no longer be maintained. In 1050, all fields
actually under cultivation in Echi-no-sho yielded a uniform chishi
rental of 3 to (1 to = 7.2 liters) of hulled rice per tan, minus 50, still
calculated in accord with the ritsuryo rate of 7.5 sho per tan.
Early shoen proprietors, as we see from this example, had very
weak support in the local community. Their fields were for the most
part let out for rent, a procedure that seems often to have required
annual written lease agreements registered with provincial authorities, with rates determined by provincial assessment. There was
some, probably not very extensive, directed cultivation on behalf of
the absentee proprietors of the shoen. In another Echi shoen, this one
a holding of Todaiji, only about two of the twelve end under cultivation were operated under direction of the proprietor in the late ninth
century.120 Early shoen, unlike those of the eleventh and twelfth centuries, mostly lacked their own proprietor-appointed operators, and
there was not a strong community of interest binding field-manager
lessees to proprietors. By 1060, however, when Echi shoen had grown
to more than sixty cho of rice paddy and was provided with a resident official staff, the shoen had become a domain in the true sense,
and the temple proprietor was threatening to expel cultivators who
resisted an increase in rental.121
Cultivation by field managers dominated agriculture in the tenth
century, gradually displacing allotment-field holdings as the single
major source of labor power from land. It was possession by field
managers that made land, rather than people, the major object of
taxation by both provincial headquarters and shoen proprietors. That
development meant that theritsuryosystem of allotment of land and
119 Dues assessment of Eisho 5(iO5i)/i/28, HIB 3:822, doc. 687; Morita, Heian jidai seijishi
kenkyu, p. 232.
120 Inagaki, "Ritsuryoseiteki tochi seido no kaitai," pp. 139-72, 154-56.
121 Shoen supervisor's petition of Kohei 3(1060)74/21, HIB 3:1005-8, doc. 931.
Cambridge Histories Online © Cambridge University Press, 2008
THE SURRENDER OF CENTRAL CONTROL
297
direct taxation of each household had failed. Tax revenues were no
longer available except through the rich and powerful, who by the
late tenth century were, as far as agriculture was concerned, the field
managers.
The same trend toward reliance on field managers was also affecting the system of local granaries and stored tax-grain. By the late
ninth century, not all of the tax-grain legally presumed to be in official granaries was actually there. A substantial portion of it consisted
of merely paper obligations, debts to the province assumed by consignees called "named obligees" {fumyo). Commenting on that situation in 891, the famous scholar-official Sugawara no Michizane, in
the course of opposing the dispatch of tax-grain auditors from the
capital to the provinces, wrote:
If, for example, a certain province has stored tax-grain to the amount of
1 million, in actuality an amount of 500,000 will be counted as lent back.
On the day for the collection of suiko loans, with respect to lent-back grain,
only the interest, not the principal, will be returned. The principal is allowed to remain in the custody of private people and will be lent back again
in the following year. 122
Precedents like this are long established and cannot
suddenly be changed.
Michizane argued that demands for strict accounting by the tax-grain
auditors (kenzeishi) would do more harm than good. His fear of disrupting provincial administrations and violating precedent shows the
development of a new relationship between provincial headquarters
and capital. The governor's formal accounting, as a tacit confirmation of private proprietorship over allegedly government grain stores,
had to be accepted as valid without authorization from above. Custody of official grain and the imposition of suiko loans had always
been sources of controversy, and in the early ninth century destruction of official stores by arson had become a serious problem. The
dispersal of tax-grain stores away from gun headquarters to branch
granaries and the appointment of village irregular officials to dispense tax-grain loans were intended to diffuse local resentments. The
lending-back system mentioned by Michizane, allowing local gentry
to hold and lend out tax grain as if it were their own, was an inevitable concession to gentry growth, rationalized in terms of provin122 Memorial of Kampyo 8(896)77/5, Sugawara no Michizane, comp., Kanke bunso, in
Kawaguchi Hisao, ed., Kanke bunso, Kanke koshu, vol. 72 of NKBT, pp. 569-70; Murai,
Kodai kokka, p. 235; Abe, Heian zenki seijishi no kenkyu, pp. 210-14; Robert Borgen, Sugawara no Michizane and the Early Heian Court (Cambridge, Mass.: Harvard University
Press, 1986), pp. 134-36, 208-10, 357.
Cambridge Histories Online © Cambridge University Press, 2008
298
ADMINISTRATION AND LAND TENURE
cial precedents.123 The need of provincial governors to come to terms
with local power structures, regardless of the prescriptions of the ritsuryo system, widened the cleavage between capital and provincial
regimes. Arguing against the dispatch of tax-grain inspectors, who
would have forced restitution of missing grain to official granaries,
Michizane insisted that the proper conduct of provincial affairs
sometimes demanded departure from the letter of the law. Provincial
precedent, he implied, did not always need to be confirmed by the
Council of State, and governors should be allowed to exercise considerable discretionary power in fiscal matters.
DISCRETIONARY TAXATION AND ELITE WEALTH
"Precedent" as used here referred not to the customary law of the
local people as such, but to the established practices of provincial
headquarters {kokuga). A Council order of 902, for example, acknowledged that because of the proliferation of personal tax exemptions among the provincials, it had been provincial precedent
(kokurei) since the Jogan era (859-77) f° r m e governors to impose
irregular levies {zoyaku) on them.124 Rinji zoyaku, the extraordinary
irregular levies that were to become one of the two principal categories of late Heian taxation, most probably originated in this way.
As the code-mandated tax structure collapsed, the discretionary
autonomy of the kokuga increased. It is important here not to be
misled by the word "extraordinary" {rinji). Extraordinary levies
were in fact routine, as implied by their justification by provincial
precedent.
Loosening regulatory supervision over the governors led to a series of legal fictions meant to establish limits beyond which they
were not to go. One such fiction, sanctioned by Council order on the
same day Kiyoyuki presented his memorial, was the "rule of sevenths" (shichibumpo), establishing artificial standards for tax remission claims based on crop damage. Only one-seventh of damaged
public land was to be deemed upper grade, and assessments of middle, lower and lowest were to be ascribed, regardless of actual fact,
in equal amounts to the remaining six-sevenths. Similarly, a stipulated grain value was assigned, province by province, to tax com123 Murai, Kodai kokka, pp. 47—59, and Nakano, Ritsuryosei shakai kaitai katei no kenkyu, pp.
252-71.
124 Kyaku of Engi 2(go2)/4/ii, Ruiju sandai kyaku, KT 25:635-36, cited inToda, Ryoshusei,
p. 30.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
299
modities purchased by the kokuga in exchange for tax-rice. In a
sense, the kokuga itself was becoming the principal object of taxation, and the central government was defining its minimum share.125
In his sealed memorial of 914, Miyoshi Kiyoyuki voiced the same
opinion as his former rival Michizane, declaring, "The administration of a provincial governor cannot in every instance be bound by
the formal law." Kiyoyuki's memorial stresses the importance of reinforcing the authority of the governor against unruly locals and
complains bitterly of the latter's numerous techniques of escaping
tribute and labor taxes, techniques that usually included the acquisition of such official or quasi-official titles as those of priests at official shrines and temples, constables, kebiishi, and palace guards,
posts that were often mere sinecures and could be obtained or renewed by purchase.
In recommending that the numbers of tax-exempt persons be limited to about 10 percent of those then existing, Kiyoyuki's aims were
not entirely fiscal. He was very much concerned that governors were
not being accorded proper respect. One article in his twelve-point
memorial begins:
Lately subordinate officers bearing a private grudge have brought false accusations against the chief official of their province; local people have also
lodged complaints against their governor under the pretense of public duty.
Sometimes the charge is misappropriation of public goods, sometimes illegal acts of administration.
Kiyoyuki, who clearly felt it outrageous to subject governors to
such abuse, proceeds to relate how oneTachibana no Mamiki, falsely
accused by an underling while he was governor of Awa, was subjected to investigation by an official sent out from the capital, a humiliation that thereafter rendered him a "cripple" without real authority in his province. "With the like of this, what gentleman of
honor will seek office?" Kiyoyuki asked. Only in cases of treason or
high crimes, he insisted, should a governor in office be embarrassed
in this way. Besides, he added, "the time is now one of decline, and
public duty is difficult to accomplish."
During Kiyoyuki's time, a provincial governor could be penalized
in three different ways. First, he could be reprimanded bureaucratically by central government agencies, often for failure to meet tax
quotas set by the accounting offices. Second, he could be brought up
125 Council order of Engi 14(914)78/8, Seijiyoryaku, KT 28:312-21, cited in Sato, Heian zenki
seijishi josetsu, pp. 312-16; on Kiyoyuki and his memorial, see note 91, above.
Cambridge Histories Online © Cambridge University Press, 2008
3OO
ADMINISTRATION AND LAND TENURE
before the Board of Discharge Examiners (kageyushi) by his successor, as already described. Finally, he could be impeached on the instance of his subordinate local officers; according to Kiyoyuki, that
was a common occurrence. In his memorial, he says:
The use of inspectors from the capital should be wholly discontinued in
this kind of impeachment procedure and the case left entirely to the newly
appointed officer, unless it is a matter of treason or sedition. If there are real
offenses, the charges can be set forth in a statement of nondischarge, and
after a finding by the Discharge Examiners, they can be submitted to the
original authorities for determination of crime and penalty.
Kiyoyuki, like Michizane, had experienced difficulties as a provincial governor. Both argued that good officials (rydri) were the only
guarantors of good provincial administration, that the governor with
Confucian virtue should be allowed broad discretionary authority as
the emperor's representative, and that continuous scrutiny from
above, as well as insubordination from below, could impair his effectiveness. Beneath the righteous Confucian tone of Kiyoyuki's recommendations lay the realization that the ritsuryo system of finance
could never be restored. His reliance on discharge proceedings as
the chief means of restraint on governors placed supreme importance on a single final accounting at the end of a gubernatorial term
and demonstrated his acceptance of the office of the accountable
custodial governor as the principal institution mediating between
capital and countryside.
The sealed memorial also showed its author's acquiescence in the
decline of the ritsuryo land system in at least two other ways. First, it
seems intentionally to minimize the problem of landholding by capital nobles, declaring the problem solved. According to Kiyoyuki, the
orders issued by the Council limiting the extent and development of
shoen more than ten years earlier had ended the difficulties for both
governors and people. Second, the memorial's attitude toward the
failed land allotment system was frankly acquiescent. In discussing
the subject, Kiyoyuki repeated the familiar complaint that many of
the taxable peasants listed in the annual major accounting report
were in fact dead or missing. He declared:
Over half of the peasants listed in the major reports from the various
provinces are fictitious. But the provincial administration, solely in accord
with the population report, assigns allotment fields and then parcels out
loan rice and imposes tribute and labor-commutation taxes. Where the field
allottee is an actual person, he cultivates a meager field and pays excessive
land and labor-commutation taxes. Where the allottee is dead or missing,
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
301
one of the household will privately lease the field, never tilling it himself,
and he succeeds in avoiding payment of the field tax, the loan-rice levy, and
the tribute- and labor-commutation taxes. Having inquired into this, I respectfully opine that the reason the court distributes allotment fields is to
collect tribute and labor-commutation taxes and to lend out tax-grain, but
now the fields have been misused, leading at last to deficiencies in the revenue offerings. Provincial governors vainly cling to useless land registers,
and the rich and powerful increasingly gather the profits of their accumulated land. This is not simply an injury to the government fisc but also an
obstruction to the conduct of administration. Now the various provinces
should makefieldallotments only to persons found to be actually present.
As for the remaining land, the provincial administration should take it back
and lease it out as publicfieldsat will. If land rent were collected, it could
be applied to the tax liabilities of thefictitiousallottees.
Kiyoyuki clearly believed that land-leasing by the province was a
more reliable means of collecting revenue than the allotment of
land. He was apparently quite willing to eliminate about half the
total allotment fields carried on provincial registers and convert
them to field-manager leaseholdings. In his informed judgment, improved knowledge of who actually controlled the "profits of the
land" and a more equitable distribution of tax burdens would result,
and actual revenues would not be reduced because rents from confiscated allotment fields could be used to purchase the equivalent of
the tax articles.
The problem of fictitious household registers and misappropriated
allotment fields had been addressed by the Council of State as early
as 864 when it accused governors of fraudulently increasing the census population but not the taxable population in order to take credit
for population increases while minimizing their obligations to produce revenue. In 875 the Council complained that failure to strike
dead persons from the registers had allowed some individuals to control the allotment fields of more than five hundred households. Kiyoyuki added a new dimension to the issue by blaming not the governors, but the central government itself. The local rich and powerful
could easily acquire immunities to personal tax liabilities from various authorities in the capital, he maintained, so that the governors
were constrained to "excuse the tax duties of actual able-bodied subjects who are rich and powerful and enter fictitious taxable subjects
on the accounting report." The rich and powerful, in other words,
did pay taxes, including some rice-loan interest, but what they paid
was attributed to allotment peasants whose liabilities they had assumed long before, and now the accounting report was being maCambridge Histories Online © Cambridge University Press, 2008
3O2
ADMINISTRATION AND LAND TENURE
nipulated in their favor. Taxation, then, was sometimes a matter of
unpleasant negotiations with the unofficial holders of supposedly allotted land. False reporting was often the official's best course, despite repeated threats from the capital. In 876, for example, provincial governors had to be cautioned against describing allotment fields
as being cultivated by the allottees, when in fact they had absconded,
leaving the land in the hands of "conniving migrants."126
By the third decade of the tenth century, the allotment system had
been tacitly, but clearly, abandoned forever. Some of the fields still
carried on official surveys as allotment land had been retaken by the
provincial governments and leased out as public fields. Under the
statutory code, unallotted surplus fields (joderi), regarded as public
land, were to be leased out for the direct use of the Council of State,
and by the late ninth century, the Council had established local stations throughout the country, called chuka (literally, "kitchen-garden
houses"), to collect the income. When households holding allotment
land became, at least for accounting purposes, "extinct" the land reverted to government control and was let out for rental (chishi), it did
not become part of the chuka system. The rice revenues were added
instead to the provincial stored tax-grain, and used to purchase the
dues or labor that actual allotment farmers would have paid.127
Under the leadership of Fujiwara noTadahira (880-949), the c e n "
tral government commenced more realistic efforts to prevent lapsed
allotment fields from escaping systematic taxation. In 925 the government acted on Kiyoyuki's recommendation by ordering that allotment fields registered to peasants who had died or moved away be
rented out and that the tribute- and labor-commutation tax quotas
for the absent peasants be met by applying the grain realized from
the rents to the purchase of the tax commodities.128 But Kiyoyuki's
advice on curbing the rich and powerful could not be taken. His own
experience as a governor had illustrated the difficulties of that.
The rich and powerful could menace and threaten provincial governors, Kiyoyuki noted, and it is very likely that he himself and his
126 Council order of Jogan i8(876)/6/3, Seijiyoryaku, KT 28:295.
127 See Abe, "Sekkanki ni okeru chozei taikei to kokuga," pp. 29-55, a t P- 39> Murai, Kodai
kokka, pp. 339—48; on Dajokan chuka, see Hashimoto, Heian kizoku shakai no kenkyu, pp.
11920.
128 Council order of Encho 3(g25)/i2/i4, Seiji yoryaku, KT 27:503; Nakano, Ritsurydsei shakai
kaitai katei no kenkyu, p. 163; Sato, Heian zenki seijishijosetsu, pp. 295-319; Kuroita Nobuo,
"Fujiwara noTadahira seiken ni taisuru ichikosatsu," in Kodaigaku kyokai, ed., EngiTenryakujidai no kenkyu, pp. 123-47.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
303
personal attendants had been subjected to such intimidation while
he was chief officer of Bitchii from 892 to 896. In any case, his biography indicates a confrontation of some sort with the Kaya family,
whose head was then chief magistrate of Kaya gun, where the provincial capital was located. The magistrate's brothers included another gun magistrate, a priest of the Kibitsu Shrine, the chief Shinto
shrine in the region, and a man who had purchased the office of Junior Secretary in the neighboring province of Bizen. There was, furthermore, a nephew of the chief magistrate who held a junior post in
the palace guards, and it is surely no coincidence that palace guards
(who were periodically on duty in the capital) and local priests were
so bitterly criticized in Kiyoyuki's sealed memorial. The memorial's
complaints about such guards and the local constabulary plainly reveal the limited ability of the governors to control the use of local
military force.
Kiyoyuki's memorial also castigated unlicensed Buddhist monks
and noted how, except for shaving their heads, they behaved like
other rich and powerful figures, controlling private wealth and even
attacking provincial offices. Such persons appear often in documents
of the period, sometimes as field managers. In 924, for example, a
communication from the Toji Temple in Kyoto to the provincial administration of Tamba demanding exemption from extraordinary irregular levies for the cultivators of the Oyama shoen lists a Priest
Heishu as superintendent of the shoen and three field managers with
monks' names.129 In 932 the provincial headquarters of Tamba complained to Toji about these individuals, referring to one of them as a
shoen custodian (sho azukari). Two were accused of withholding tribute silk, for which gun authorities had distrained their rice holdings.
A probable reference to these and other tato of Oyama-no-sho as officially "unlisted migrants" is one of the earliest acknowledgments of
corvee and produce dues as charges on landholding. Unlisted rbnin
were off the books, their names officially withdrawn from the rolls of
those liable for corvee and commodity-tax duties. Their licensed
presence on Oyama-no-sho was seen as an official endowment to the
proprietory temple, and to the Oyama estate in particular, of the
corvee services and produce normally at the command of the state.
Their argument was that since they had, with approval from the cap129 Communication of Toji of Shohei 5(935)710/25, HIB 1:360, doc. 245; on the role of the
gun here, see Abe, "Sekkanki ni okeru chozei taikei to kokuga," pp. 29-55.
Cambridge Histories Online © Cambridge University Press, 2008
304
ADMINISTRATION AND LAND TENURE
ital, been exempted from tax liability to the central government, they
should also be freed from the locally imposed burden of rinji
zoyaku. 13°
The document of 932, one of the earliest to show the assumption
of peasant tax liabilities by the rich and powerful, reads in part:
The inspector responsible for the village of Amaribe, Heki no Sadayoshi,
has reported, saying:
The aforesaid village has always been without land, its peasant allotment
grants being in other villages of the region. Accordingly, the tribute silk
for the village has by custom been levied on listed capable farmers of
those villages [where the allotment grants were located]. At present,
Heishu and [Seiho] are of the same capable status as laymen, and moreover in years past they have submitted the said tribute silk. The names of
Heishu and Seiho have then been entered on the original report for two
jo [6 m.] of silk each.
I went to the personal residences of Heishu and the others in order to make
them pay the said silk, but they ran away into the mountains and did not
pay. Accordingly, I have impounded two hundred sheaves of rice from each
man for the payment of the silk. After the silk has been forwarded, the rice
will be released.
"Capable farmers," kambyakusho, were those able to guarantee
the tax obligations of allotment-field peasants, whether the latter actually existed or not. They were, after the order of 925, the government's principal means of acquiring commodity taxes. The lists of
capable farmers alluded to in the document were far more important in revenue raising than the population-based tax-accounting reports, which were gradually losing all real fiscal significance, although they continued to be made until the century's end/ 31 The
old head taxes were, in effect, being farmed out to the more prosperous field managers.
In the meantime, Kiyoyuki's hopes to restore the ritsuryo census
and household taxing system had been totally abandoned under
Tadahira's new policy, which was aimed at preventing further erosion of the rural tax base, even if some concessions to the rich and
powerful had to be made. Judgments by the Board of Discharge Examiners in 933 and 941 made it clear that the provincial governor
was himself responsible for the production of revenue from aban130 Communication of Tamba Province of Shohei 2(932)79/22, HIB 1:354-55, doc. 240. See
also communication of Toji of Shohei 5(935)/io/25, HIB 1:360, doc. 245, listing the same
tato as cultivators of Oyama-no-sho.
131 Abe, "Sekkanki ni okeru chozei taikei to kokuga," pp. 29-55; Izumiya, "Choyosei no henshitsu ni tsuite," pp. 175-308.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
305
doned allotment fields and other such land, holding that a lack of capable farmers to tax was not a valid excuse for revenue shortages.
Recruiting and enrolling the necessary cultivators was ultimately
part of the governor's duty, and as the tenth century progressed,
kokushi and district officers (gunji) were held increasingly responsible for the desertion of public rice fields. In 933, for example, the
Board of Discharge Examiners refused to accept an alleged shortage
of capable cultivators as a justification for reporting an unacceptably
large area of land as not being worked.132
In the ritsuryo scheme, all arable paddy had to be registered and
all failures to cultivate it reported to the central government. One
standard exception to those requirements was damage to land or
crops resulting from storms, floods, insects, or other disasters, an exception that applied, however, only to fields in which a crop had already been planted. If no crop at all had been planted for three or
more years, the field was designated abandoned (kohai) or permanently out of cultivation (Joko). If, on the other hand, cultivation had
been discontinued more recently, the field was called uncultivatable
(fukanden). Fields could fall into that category for natural reasons
like flooding or infertility, but also for social reasons, such as the
flight of their assigned cultivators from the district or simply the inability of the cultivators to provide seed grain.
Provincial governors were required to report the extent of uncultivatable fields annually, the central government depending heavily
on such reports to assess the state of agriculture nationwide and also
to determine how much field tax (so) could be collected in each
province. Reported acreage totals of uncultivatable fields had a pronounced tendency to increase, not simply because of cupidity on the
part of the governors, who could pocket the revenues from uncultivatable fields that were in fact under cultivation, but also because
local communities abandoned registered fields of poor quality in
favor of unreported newly reclaimed lands. Governors were admonished to carry out inspections in person to detect undocumented
reclamation, and when uncultivatable-field and damage totals
seemed too high, special inspectors were dispatched from the capital
to check the accuracy of the governors' reports.133 Ever since Nara
times, provincial authorities had tended to overstate field damage
132 Kageyushi decision of Shohei 3(933)/n/2i, Seiji yoryaku, KT 28:328; see also decision of
Tengyo 4(941), Seiji yoryaku, KT 28:327-28.
133 On the meaning of the term fukandenden, see Sato, Heian zenki seijishi josetsu, pp. 321—35.
Cambridge Histories Online © Cambridge University Press, 2008
306
ADMINISTRATION AND LAND TENURE
and crop losses. This allowed them to take advantage of a legal provision that excused households that had suffered over 50 percent
crop damage from flood, drought, insects, or hail from the so, and
additionally canceled the tribute tax (cho) for households with 70
percent crop damage and both tribute and labor-commutation (yd)
taxes for those with 80 percent or more damage.134
The policies instituted by Tadahira were particularly concerned
with curbing, rather than preventing, falsification of fiscal reports by
accountable governors. Under his guidance, the Council of State
elaborated on the earlier measures of "standard damage" (reison), all
of which, it may be remarked, originated from administrative custom
rather than ritsuryo rules. The code rules regarding tax exemptions
and evasions were, after all, originally aimed at taxable subject
households, not governors. Government legal technicians were
nonetheless hard at work reinterpreting the old codes to make them
apply to the emergent real taxpayers, that is, the zuryo governors
themselves. Implicit recognition of this may be seen in a ruling of
915, giving the Board of Discharge Examiners, which was exceptionally well staffed with legal experts, authority to recommend rewards or penalties for zuryo, and to pronounce on their evaluation
generally.135 In 926 the Council ruled that whenever a province petitioned for tax remission on account of land damage, no more than
one-third of allegedly damaged households could be listed as over
the 50 percent damage bracket, and thereby eligible for remission of
taxes other than so. By the middle of the century each province had
its "standard damage" allotment, a kind of legal fiction that was beyond challenge. Similarly, uncultivatable fields became a kind of tax
deduction on the provincial account, having little relation to actual
conditions of arability.136
Tadahira's policies regarding damage reports also resulted in a revision of the system of accounting for public fields yielding rent. The
established practice of grading such fields into upper, middle, lower,
and lower-lower assessment categories was all but abandoned. The
rule had been that when damage to public fields under Council supervision was reported, only one-seventh could be upper field, while
134 Murao, Ritsuryo zaiseishi no kenkyu, pp. 87-101; Nagayama, Ritsuryo futan taikei no kenkyu,
pp. 219-45; Buyaku ryo, Suikan no jo, Ryd no gige, KT 226:119, Ry° "° shuge, KT 23:
392-403.
135 Ruling of Engi 15(915)712/8, quoted in ruling ofTentoku 3(959)712/4, Seiji yoryaku, KT
28:182.
136 Ruling of Encho 4(926)712/5, Seiji yoryaku, KT 28:504-5; Nakano, Ritsuryosei shakai kaitai
katei no kenkyu, pp. 99—107.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
307
two-sevenths were to be in each of the three other categories. The
"rule of sevenths" was the standard for delivery of land rents to the
Council's provincial collection stations until 928, when a "rule of
thirds" (sambumpo) was imposed. That rule completely eliminated
the "upper" category of land from damage reports and required that
for tax- and damage-estimation purposes public fields be presumed
to consist of equal parts of the three lower categories.137 The rule of
thirds meant for central government purposes that the land-grading
system was largely inoperative, since the average putative yield of
about three hundred sheaves per cho for the three lower assessment
categories eliminated all distinctions.
The virtual disappearance of the grading of fields as a topic of
concern in relations between the central and provincial governments
did not mean, however, that the subject ceased to be an important
issue in relations between the provincial governments and local cultivators. As Empo's land survey of the Echi shden shows, although
the grading of fields was the prerogative of the provincial headquarters, it was also subject to informal negotiation with field-manager
cultivators, and Tadahira's policies merely defined the area left open
to negotiation while making even more obvious the tax-farming aspect of the administrations of accountable governors. The use of mechanical formulas instead of factual surveys as a basis for taxation
was clearly an attempt to check the increased bargaining costs that
the enforcement of the old taxation rules against the governors had
entailed. It illustrates the sort of adversary situations in which legal
fictions develop within a context of formally codified law.
The administrative code provided that if paddy in a district increased by one-fifth or more during the term of a local officer of a
province or gun, his personnel evaluation was to be raised one grade
for every one-fifth increase; if the field area declined, the evaluation
was to be lowered one grade for every 10 percent lost. In an official
promotion system stressing Confucian values of character, diligence,
and talent, this mechanical, achievement-oriented standard was
somewhat incongruous, as contemporary legal commentators noted,
but conditions in the tenth century gave it heightened significance.
Renewed stress was also placed on an article of the criminal code
that provided similarly graded levels of punishment for officials who
allowed fields to drop out of cultivation. Allowing cultivated fields in
a district to decrease rated forth blows of the stick for the first one137 See note 124, above.
Cambridge Histories Online © Cambridge University Press, 2008
308
ADMINISTRATION AND LAND TENURE
tenth and an additional increment in the penalty for each further
tenth, up to a maximum of one year's imprisonment. 138 Those draconian penalties were not intended to be carried out in fact, however,
since the criminal code made a number of substitute punishments
available to the official classes. Loss of an office of Fifth-Rank status,
for example, was the equivalent of two years of penal servitude, and
loss of a lesser office, the equivalent of one year.139 Even the substitute penalties could be catastrophic for an accountable governor,
however, and the central government was usually reluctant to impose
criminal sanctions at all.
The rules rewarding local officials for increasing land under cultivation and punishing them for reductions reflected a major explicit
concern of the ritsuryo state: the promotion of agriculture, originally
regarded as more basic to regional administration than collecting
taxes. Misreporting of uncultivatable fields was not only a crime
under the penal code but also a violation of this basic policy, and yet
despite threats of dire sanctions, provincial and gun officers continued to falsify acreages, partly because the sanctions were not consistently applied. Tadahira's Council of State was the last seriously to
insist on accuracy. In an order of 918, the Council declared:
The penal code states:
When within a district damage occurs from drought, flood, frost, sleet,
worms, or locusts, and the chief official makes an exaggerated report, the
penalty is seventy blows of the heavy rod. Reexaminers who report falsely
are liable to the same penalty. If taxes are collected or excused in violation of the law, the crime is that of illicitly acquired goods, meriting an
additional seventy blows of the heavy rod. In assessing the penalty for illicitly acquired goods [which increases in proportion to the value of the
goods misappropriated], the maximum punishment is three years penal
servitude. As for improperly taken goods, the case is one of illicitly acquired goods if the goods have gone into government possession. If the
goods fall into private hands, the case is one of official extortion; in sentencing for that, if the amount extorted merits death, the actual sentence
will be life at forced labor in exile.
Now, however, when the various provinces send up pleas of crop damage
or uncultivatable fields, they regularly ignore actual facts, and when reexamination is made, the discrepancies are found to be excessive. In recent
138 Koka ryo, Kokugunji no jo, Ryo no gige, KT 226:157-58, Ryo no shuge, KT 23:591-595;
Kokon ritsu, Bunai denchu kohai no jo, Kuroita Katsumi and Kokushi taikei henshukai,
comps., Ritsu, KT 22A:ii5139 Ishii Ryosuke, Nihon hdseishi gaisetsu, 2nd ed. (Tokyo: Sobunsha, i960), pp. 144-46; Wallace Johnson, The T'ang Code; General Principles (Princeton, N.J.: Princeton University
Press, 1979).
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
309
years, nevertheless, offenders are merely made to restore the tax goods required, and the laws of criminal responsibility are never applied. Deception
has become the established rule by the growth of accumulated custom.
Now then, as ordinances have pointed out, the established penalties are severe, but forbearance has led to nonenforcement of the law.
The Great Minister of the Right [Tadahira] proclaims, announcing an
imperial edict:
Governance requires adapting to change, and its acts may be strict or lenient. From now on, on the day a reexamination is reported, the examining officer in his report shall divide the area of fields he finds unproductive by ten, and where the misrepresented area is not more than
one-tenth greater, criminal penalties will be specially remitted and the official ordered to restore the tax goods. If the limit is exceeded, let die
crime be punished according to die law, depending on whether die goods
have passed into public or private possession. If a reexaminer in his survey misrepresents the facts, he is always to be punished in accordance
with the law. Also, as for matters such as peasants dead from epidemic
disease, people requiring famine relief, or damage to official buildings or
dikes, when the report is within one-tenth of the actuality, punishment is
to be remitted and restitution only ordered.140
The allowance of a one-tenth margin of error in exaggerated uncultivatable-field reports and in various other certifications of provincial loss reveals a basic weakness of the ritsuryo system. The accountable governors were subject to forces far more compelling than
the incentives and sanctions of the codes. Those codes presumed
that promoting agriculture was primarily the responsibility of the
local administrators and that their direction of the agricultural enterprise, if properly conducted, would be unreservedly welcomed by
the populace at large. The fundamental assumption was that the
rulers would be in total control of basic resources and in no way dependent on the assets of the subject population. As the numerous irregular officers in each province, the more prosperous field managers, and other such local elites tightened their hold on granaries
and peasant labor, the accountable governors found it increasingly
difficult to carry out the letter of the law and meet their tax quotas
as well. As Miyoshi Kiyoyuki's memorial pointed out, such performance could not be expected of even the most scrupulous governor,
and he accordingly opposed direct government reexamination of
gubernatorial accounts. To function economically, the governors
needed more latitude than the organic code permitted, and the
140 Ruling of Engi i8(gi8)/6/2O, Seiji yoryaku, KT 28:493—94; SatD, Heian zenki seijishi josetsu,
pp. 326-33.
Cambridge Histories Online © Cambridge University Press, 2008
31O
ADMINISTRATION AND LAND TENURE
Council of State was forced to acknowledge the use of extralegal discretion in the performance of their duties. The 10 percent margin
permitted in the quoted document, however, proved to be an inadequate concession.
In 927 the Council tried once more to pressure the governors into
reducing reported losses of cultivatable fields. It imposed on newly
appointed governors a positive duty to reclaim lands vacated under
their predecessors, to increase the total area of land in production,
and to submit annual progress reports on gains made. Any further
net loss of arable land in a governor's jurisdiction was to be an absolute bar to future reappointment or promotion.141 The Council of
State seems by those actions to have tacitly acknowledged the difficulty of reasserting total power over reclamation policy, opting instead to establish a compulsory minimum amount of paddy under
cultivation to be reported from each province.
The measures taken in 927 failed to prevent further decline in officially registered paddy under cultivation, however, and the Council of State decided in 946 to resort once more to the criminal sanctions imposed in 916, which were still in force. Fourteen provincial
governors were impeached for presenting reports exaggerating by
more than 10 percent the area of uncultivated paddy in their provinces. 142 One surviving case report deals with the governor of Ise,
Tachibana no Korekaze, and two provincial clerks who cosigned a
false acreage report.143 Korekaze had reported about 2,070 cho of uncultivatable fields for his province, or about 12 percent of the 18,000
cho of registered paddy that other sources show for the province. The
inspector from the capital found only about 1,450 cho of land actually out of cultivation, and, following the established rule, recommended the signatories of the original report for criminal punishment. The 620 cho falsely reported out of cultivation was more than
30 percent above the maximum allowable error of 10 percent.
Asked to determine the proper legal penalty for the crime, the law
doctor Koremune no Kinkata responded that final dispensation required further inquiry into the facts. In the case of misappropriation
by provincial authorities of taxes from unreported fields under cultivation, the ruling of 918 and the criminal law both specified penal141 Dajokan order of Encho 5(g27)/i2/i3, quoted in the ruling of Tenryaku i(g47)/int. 7/23,
Seiji yoryaku, KT 28:271-72; see Abe, "Sekkanki ni okeru chozei taikei to kokuga," pp.
29-55; Sato, Heian zenki seijishi josetsu, pp. 330-38.
142 Dajokan order ofTenryaku i(947)/int. 7/23, Seiji yoryaku, KT28:271-72. One of these adjudications is that ofTenryaku i(g47)/int. 7/16, Seiji yoryaku, KT 28:278—80.
143 The case of Korekaze is outlined in Seiji yoryaku, KT 27:496-98.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
311
ties that differed depending on the use of the taxes. In the present
instance, Kinkata said, a finding on the matter of use was required
before a penalty could be assessed. If the revenues had gone to official agencies such as the provincial headquarters, the culprits were
to be sentenced for the crime of illicitly acquired goods, but if they
had been converted to the personal use of the provincial officials, the
much severer penalties prescribed for extortion were appropriate.
The authorities then discovered that taxes on the fields in question
had not been collected at all, and that there had therefore been no
misappropriation of revenues. The offense was found to be that of illicitly acquired goods, as if the tax-grain had gone to an inappropriate office. Although a general amnesty proclaimed late in the year
947 absolved Korekaze and his subordinates from criminal penalties
connected with the case, they were still required to make restitution
of the uncollected taxes.
The discovery that no taxes had been collected on the uncultivatable fields falsely reported by Korekaze cannot be interpreted to
mean that he was unaware of the cultivation of the fields or remiss
in the exercise of his duties. The absence of revenues from those
lands was the result more likely of the growing political power of the
local elite and the corresponding growth in its ability to resist tax
collection by provincial authorities. The adversarial relationship between accountable governors and provincials is as well established in
the sources for the period as the well-known ability of accountable
governors to use their office for personal gain. The strained legal reasoning in this case illustrates the need of the early Heian officials to
maintain the rhetoric of the ritsuryo order even as the command
economy embodied in it deteriorated.
Seeking to maintain the facade of the authority-intensive ritsuryo
system, the central government could not officially acknowledge
participation of the local elites in its control of the provinces, and it
was forced to resort to increasing use of legal fictions in order to
cope with them. The decision in Korekaze's case that the crime was
an instance of illicitly acquired goods, in other words, the misdirecting of revenue, rather than a failure to collect taxes or personal misappropriation of them, was in fact an indirect accommodation to the
power of the local elite through a fictionalized interpretation of penal
law. The imperial court was relinquishing its control over agriculture
and land use to other hands.
Although the reporting of uncultivatable fields continued for
some centuries, the system steadily weakened during the tenth cenCambridge Histories Online © Cambridge University Press, 2008
312
ADMINISTRATION AND LAND TENURE
tury, eventually becoming simply a heavily ritualized routine of the
central bureaucracy. The decline in the practical meaning of the reporting was marked by the discontinuance of the practice of canceling expensive court rituals in years when the area of uncultivatable
fields reported was unusually high, by the cessation of the dispatch
of special field investigators to determine actual acreages of uncultivatable fields, and by the establishment of reporting quotas of uncultivatable-field acreages for each province (typically pegged at 10
percent of the province's arable), observance of which exempted a
province's report from further scrutiny (the quotas, similar to the
standard damage allowances mentioned earlier, were called the
"standard uncultivatable," reifukan).The result was the complete fictionalization of the uncultivatable-field reports and the abandonment by the central government of an important device for overseeing its own tax base.
In the face of the growing power of rural elites and their increasingly successful quest for control over land and peasant labor, the
government had little choice but to relinquish authority over rural
production. The ritsuryo presumption of total peasant dependency
on the state could not be maintained against the interests of the rural
elites, and the central government's reins on the accountable governors themselves were slipping. The governors required not only
more freedom in their attempts to control their refractory populations but also their own private military and civil auxiliaries, paid for
from local revenues. Alarmed by the rebellious potential exhibited
by its subjects, most notably in the revolt of Taira no Masakado (d.
940), the central government was obliged to recognize the right of
governors to military escorts. It stopped short in this period, however, of recognizing the use of private retainers for that purpose.
Condoning greater scope for gubernatorial discretion and giving
less scrutiny to local land administration did not mean total loss of
control by the noble regime over the countryside, and revenues continued to flow in. The imperial court was, in part, merely confirming concessions already made over the years to the rich and powerful by the governors, concessions that had given local elites an
unofficial but profitable role in the revenue system. Perhaps the
clearest example of that trend was the system of tax-grain fund exchange (shozei kdeki), on which the government became increasingly
dependent during the ninth century. Exchange of tax-grain, for example, had been an important element of the scheme for publicly
operated fields proposed for Kyushu in 823. The exchange system
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
313
was expanded to all taxable lands in the tenth century. Provincial authorities used stored tax-grain funds to purchase the commodities
that ritsuryo law had required, with growing lack of success, of each
taxable household as tribute and labor commutation taxes.
The fiscal authority of provincial governments had never been
limited to control over land use and corvee labor. Provinces were required to supervise the manufacture of fine silks and other luxury
items, mining, and the gathering of natural food items for the court,
and the production of these rare or precious items was the responsibility of the provincial unit itself. In the distributive economy presumed by the ritsuryo, the kokuga stood directly below the capital in
the exchange system. By the tenth century most such activity was reduced to a system of official procurement, the purchase of the required goods with stored tax-grain. For example, the province of
Owari was required to provide the sovereign's Chamberlains' Office
(kurododokoro) with an annual supply of lacquer, almost certainly
purchased from local growers with tax-grain funds,144 and the province also bought silk, hemp fabric, karamushi fabric, vegetable oil,
paper bark, and a variety of other items demanded by the capital.
Disputes over prices and quotas could break out between provincial
authorities and local suppliers as each side attempted to maximize
its own share of the proceeds, and there was a similar area of friction
between the provinces and the central government.145
As the ritsuryo system of taxes became ever more fictionalized and
divorced from the actual productive yield of the countryside, the
central government began tapping into the provincial system of extraordinary irregular levies (jrinji zoyaku), at first justifying its new
demands by particular exigencies like the need to reconstruct capital buildings after fires or earthquakes. During the eleventh century,
the term rinji zbyaku acquired a different, far more general meaning.
It was applied to any of the various forms of corvee defined by the
old ritsuryo rules and, at times, to tribute and labor commutation
taxes. By the early twelfth century, it meant all taxes not perceived
as a direct charge on grain revenues, like so and suiko payments.
Taxes of this latter kind, by contrast, merged into a second major
category, kammotsu, a term formerly used for "official goods."146
Just how and why this transition occurred remains uncertain, and
144 Abe, Owari no kuni gebumi no kenkyu, pp. 156-60.
145 Nagayama, Ritsuryo futan taikei no kenkyu, pp. 219-45.
146 Abe, Heian zenki seijishi no kenkyu, pp. 309-10; Nagayama, Ritsuryo futan taikei no kenkyu,
pp. 285-319; Nakano, Ritsuryosei shakai kaitai katei no kenkyu, pp. 199-299.
Cambridge Histories Online © Cambridge University Press, 2008
314
ADMINISTRATION AND LAND TENURE
there is no exact agreement on which of die old ritsuryo taxes became rinji zoyaku and which kammotsu. Nevertheless, some aspects
of this long transition are clear. Most important was the steady advance of the tato as the major source of provincial revenues and the
chief consignee of resources, including labor resources, in every
province. The gradual sorting out of taxes into rinji zoyaku and kammotsu almost certainly reflected a distinction between labor and land
resources, in which control over labor power was, to an increasingly
greater extent, a function of control over land. The disposition of
stored grain, a major element in the exercise of governmental authority during the ninth century, gradually became a prerogative of
the local rich. During the ninth century, the government shifted the
burden of the imposts to the stored tax-grain of the provincial governments and used the revenue thus produced for purposes that had
once been met from the central treasury. This was an added burden
on the stored tax-grain, already strained by its use to purchase the
goods for commodity taxes owed the central government and by its
misuse on the part of provincial officials. The result was sequestration and hoarding of the stores, and the fictionalization of the suiko
loan system based on them.
The stored tax-grain on which so much depended was divided
into three parts: (1) official government grain funds, the "main
fund," or skozei in the narrow sense of the term; (2) public-allowance
rice (kugeto), reserved for the income of regular provincial officials,
including numbers of titular appointees; and (3) "miscellaneous
rice" (zoff), for the upkeep of official buildings, religious institutions,
and irrigation facilities, and the benefices of irregular officials, local
priests, and other rich and powerful provincials. Roughly speaking,
each fraction represented one of the three groups most interested in
provincial revenues: the nobles in the capital, the accountable governors and their retinues, and the rural elite.147
The first component of the stored tax-grain, the official government grain, was regularly tapped by the court during die tenth century as a source of rank- and support-stipends for the middle-ranking
nobility. Instead of receiving his stipend in silk and other commodities through the general treasury, a holder of court rank might be told
to collect the rice equivalent of the allowance from the stored taxgrain of a specific province, which was by rule always fairly distant
from the capital. Those "remote awards" (yoju) were frequently diffi147 See note 61, above.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
315
TABLE 4 . 3
Stipend grants from stored rice in Izumo
Year
Number of stipendiaries
939
957
1003
4
12
15
Total stipends
(in sheaves)
1
>°°5
^835
3,142
cult for a grantee to collect. It was his own responsibility to convert
the rice to shippable commodities and transport them to the capital,
and for that he was obliged at significant cost to employ the local accountable governor as his agent.148 The growth in the use of remote
awards during the tenth century may be illustrated by data available
from the province of Izumo for three different years.149
As the administrative and fiscal separation between the capital
and countryside became more pronounced, the difficulty of collecting such stipends forced middle-ranking nobles to seek help and
protection from individuals of the highest ranks, increasing the trend
toward clientage and estate partitioning.
The miscellaneous rice component of provincial stored tax-grain
had grown very rapidly during the ninth century, but the government in the tenth century, considerably less solicitous of local interests than its predecessors had been, sought to check that growth, insisting that official government rice was of prime importance and
that the miscellaneous rice account was of lowest priority. The government was understandably concerned that the governors might
conspire with local officials to accumulate hidden stores of unreported tax-grain under the guise of miscellaneous rice.15°
The Heian government's policy toward provincial administration
during the ninth and tenth centuries may best be seen as one of accommodation to changes beyond its power to control, especially the
reorganization of local resources by the regional elite. Direct taxation of land and peasant had to be replaced by a more indirect
method, in which a major object of taxation was the product that resulted from the power of the field-manager gentry over the peasants.
148 Murai, Kodai kokka, pp. 146-50; Yoshimura, Kokushi seido, pp. 362-64.
149 Murai Yasuhiko, Heian kizoku no sekai (Tokyo: Tokuma shoten, 1972), p. 44; Murai, Kodai
kokka, pp. 146-74.
150 Abe, Oviari no kuni gebumi no kenkyii, pp. 69-70; Murai, Kodai kokka, pp. 37-59.
Cambridge Histories Online © Cambridge University Press, 2008
316
ADMINISTRATION AND LAND TENURE
Close inspection of land and population resources had to give way
to a more flexible system within which the locally recruited resident
officials of the provincial headquarters were allowed a considerable
degree of autonomy, and it became necessary at the same time to
grant accountable governors broader discretionary authority in the
performance of their duties. In response to those developments, the
more powerful noble households and political bureaus of the capital
sought direct links with field-manager cultivators to gain independence of the gubernatorial stratum, as with the directly operated office fields. First established in 879 to provide stipends for centralgovernment officials, they were later parceled out among the central
ministries for their individual support.151
The ninth- and tenth-century restructuring of the provincial revenue system, with its heavy reliance on tax-grain accounts, remained
like the earlier tax system in that it was ultimately based on income
from land. But the emergence of a well-differentiated field-manager
stratum among the peasantry radically changed the system of linkage between the administration of taxes and the administration of
land that had been intended by the original ritsuryo planners. Standing between the economies of capital and province, the accountable
governors had to extract as much grain as possible from the field
managers in order to meet their legal tax obligations, but their success in that respect depended on collaboration with the resident officials in the evasion of ritsuryo law regarding the state's power to dispose of land and supervise agriculture. They were, in brief, forced to
resort to illegal actions in order to meet their legal obligations.
That development was an essential step in the formation of the
Heian period shden. By the mid-tenth century, the various enclaves
of shden and publicly operated fields constituted the only arable
fields that the capital nobility controlled directly. As the autonomy of
the provincial headquarters increased, and as provincial impositions
of irregular grain and commodity taxes on shden grew heavier, conflict between the headquarters and shden owners in the capital became inevitable. The conflict was reflected in the policies of Tadahira, apparently forcing even him, a staunch defender of public lands
against private encroachment, to retreat somewhat from the strict
policies of his predecessors regarding shden.
Tadahira's retreat can be inferred from a "communication" (cho)
151 Sakamoto, Nihonocho kokka taisei ron, pp. 127-94. See also, however, Morita, Heian jidai
seijishi kenkyit, pp. 32-233.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
317
issued in 920 by his household chancery to the governor of Tamba
instructing him to respect the tax immunities of the Oyama shden, a
holding of Toji. One of Kyoto's most powerful temples, Toji had acquired the land by purchase early in the ninth century. In 845 the
temple was awarded certificates by the Council of State and the
Ministry of Popular Affairs completely exempting the holding, as it
then was, from land taxes. The Popular Affairs certificate defined
the shden as consisting of slightly over nine cho of privately reclaimed paddy scattered over thirty-six grid squares, thirty-five cho
of woodland, and a large pond that may in fact have included some
rice land. The boundaries of the village-site area were noted, and
there was some "public field" within them. In 915 the provincial
headquarters of Tamba had recognized Toji's ownership of about
sixteen cho of additional paddy reclaimed within an already existing
irrigation system.152 But soon thereafter the headquarters began to
challenge Toji's claims, in two different ways. First, it refused to certify new paddy opened within the area as tax-exempt temple land.
Then, it refused to excuse the taw in charge from the various irregular and extraordinary levies just then being instituted throughout
the countryside.
The questions were referred to Tadahira's private household
chancery rather than to the Council of State, presumably because in
addition to being the senior minister in the Council, he was also official overseer of Toji. The first complaint was resolved in the temple's favor, at least for a time, by the communication of 920. The
document repeats the complaint received from Toji, stating in part:
As for the woodland, additional rice fields have been reclaimed [from it]
with each passing year. However, after the shden buildings have been
erected and the fields reclaimed, the provincial and gun officers have confiscated them and made them unallotted public land.
The chancery responded with a request to the provincial headquarters to look into the charge, and if it was true, to return the lands
to the temple's possession. Control over land use within the boundaries established in the shden certificates, the communication implied,
should rest with the temple rather than with the province or gun.
152 The "communication" {cho) of Engi i5(9is)/io/23, recognizing temple proprietorship
over specified additional fields appears as HIB 1:322, doc. 213. For the incident of 920,
see Toji petition of Engi 20(920)79/7, HIB 9.3473, doc. 4555, and Udaijin kecho of Engi
20/9/21, HIB 1:325—26, doc. 217; Sato, Heian zenki seijishi josetsu, pp. 270-83. On the history of Oyama-no-sho, see Miyagawa, Oyama sonshi, Hommon hen, pp. 69-76. For a good
account of this estate in middle and late Heian, see Elizabeth Sato, "Oyama Estate and
Insei Land Policies," Monumenta Nipponica 34 (Spring 1979): 73-99.
Cambridge Histories Online © Cambridge University Press, 2008
318
ADMINISTRATION AND LAND TENURE
Despite the ambiguities of the document, it seems likely that the
action of the local authorities in confiscating the reclaimed fields was
not illegal. The term "confiscate" as employed here, it must be
noted, did not indicate the extinction of Toji's proprietary rights. It
merely meant that the land in question, although unquestionably a
part of Toji's Oyama estate, was adjudged subject to all regular
charges of grain dues {kammotsu). Its cultivators, like all cultivators
of public fields, would also be subject to the newly established rinji
zoyaku imposts. Official maps, kept in the provincial headquarters,
designated specific areas in specific grid squares as exempt from
kammotsu. The governor had the power to confiscate the excess, and
would ordinarily do so except in grid squares where some exempt
cultivation was already established. Tadahira, through his chancery,
here preempted the largely discretionary power of the local authorities. The Toji complaint makes no claim that the affected land belonged to it by the well-established right of wasteland reclamation,
probably because it did not consist of reclaimed fields (konden) in
the legal sense of that term. It was very likely abandoned public-field
land that had been returned to production, a type of reclamation
that did not entitle the reclaimer to permanent ownership. If that
was the case, the local authorities would have been completely justified in seizing the fields as they did and returning them to publicfield status, which would have required that the land rent on them
be paid to the government, rather than to the temple.
In confiscating reclaimed land in the Oyama shoen, the Tamba authorities may have been acting in response to some now lost government order, but it is clear that the legal right to confiscate the land
did in fact rest with the governor. The registration and classification
of land had always been primarily the function of provincial governments. This power had been exercised earlier in the case of the
Oyama shoen when, as already noted, sixteen chd of newly reclaimed
paddy were confirmed in Toji's ownership. Tadahira through his
chancery clearly preempted the discretionary power of the governor
of the province to confirm an addition to the tax-exempt shoen land.
More important, however, is the gap seen here between the land
and taxation system mandated by official regulations and the actual
conduct of provincial governors responsible for collecting taxes. The
government was doing its best to allow the governors broad discretionary powers in dealing with local notables and shoen lords, while
at the same time maintaining full taxing authority over all land
legally specified as taxable. The result was the development of new
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
319
categories of taxation, like rinji zoyaku, permitting uneven application of the established rules. Acquiescence in gubernatorial discretion, furthermore, facilitated increased participation by local gentry
in the tax assessment process.
Land within the boundaries of a shoen newly brought under cultivation by the proprietor was subject to annual survey by local authorities whose recommendations reached the governor in the form
of an assessment by the "land office" (tadokoro). Tax exemptions
were, in effect, subject to ratification by the irregular officers of the
area concerned, as well as the governor. The assessment of shoen
land developed into a formalized process involving joint participation by local elites, agents of zuryb governors, and agents of estate
proprietors like Toji.
A document of 935 relating to Oyama-no-shd contains the earliest
specific information about rinji zoyaku, and illustrates the basically
routine nature of this tax. It is also clear from this text that the gentry managers of the estate and its tato cultivators were the principal
targets of tax gatherers. This document, a communication from Toji
to the kokuga of Tamba, expresses the temple's attempt to protect
them. It reads in part:
The said shofields,in accordance with the certificate of the Council of
State and the Ministry of Popular Affairs, of Jowa 12th year [845], 9th
month, 10th day, are fields for the support of transmitting Buddhist law,
and the use of their rice rent for teaching the law and copying the surras has
been established long since. The prospering of Buddhist law indeed rests on
this sho.
Accordingly, from the beginning the field tax {denso) and tax-grain funds
(shozei) were not assessed, and there were no impositions of extraordinary
irregular levies (rinji zoyaku). However, we have received a petition from
this slid saying:
The gun magistrates order us saying, and the kokuga orders us saying:
The irregular levy of official exchange silk thread, tribute tax sale silk,
rice in ear from provincially cultivated fields, cedar bark for the repair of
official buildings, and the [levy of] labor and horses, you are ordered to
render.
Because of this, we cannot rest at all either night or day. How can we perform our customary service to the s/wPWe respectfully pray that a communication be sent to the kokuga that we be exempted from special irregular
levies.
Our communication is as aforesaid. We pray the kokuga to examine the
matter and desire that, in accord with precedent, our sho custodians (sho
azukari) and sho retainers (shoshi, i.e., tato) be exempted from special irregular duties. Let the matter be in accord with goodwill, and without conCambridge Histories Online © Cambridge University Press, 2008
32O
ADMINISTRATION AND LAND TENURE
cealment. Accordingly, we append a copy of the former governor's order
granting exemption from special irregular duties.153
Rinji zoyaku, and immunity from it, were mentioned as early as
895 in a declaration made by the Usa Hachiman Shrine, and a petition from Oyama-no-sh5 dated 924 requested exemption from the
tax, but both documents fail to state what the levy actually was.154 As
the document here implies, rinji zoyaku simply did not exist when the
sho was established in the mid-ninth century, and the former governor's exemption of the sho from the levy was almost certainly a discretionary act. Especially noteworthy is the way ordinary provincial
taxes on tato cultivation are divided into (1) denso, the ritsuryo tax on
riceland; (2) skozei, a grain tax originally paid as interest on rice loans
from the fund called shozei but now assessed directly on management
of riceland; and (3) rinji zoyaku. The five particular types of rinji
zoyaku enumerated in the document here can be divided into two
categories: (1) revenue items procured by local purchase but destined for the capital, represented by "official exchange silk thread"
and tribute tax sale silk; and (2) revenue for purely provincial use,
here to be collected in such forms as cedar bark for roofing, horses
and men for the transport of goods, and, notably, an impost of rice
in ear presumably produced on provincially cultivated fields.
"Provincially cultivated fields" were, originally, maintained by
corvee labor for the benefit of the governor and other regular provincial officers.155 It is very possible that the fields of Oyama-no-s/zo had
been provincial fields of this kind before acquisition by Toji, but in
any case, the location of the fields supposedly dedicated to administrative expenses may not have been a major consideration. As we
have seen earlier, official direct cultivation of fields was by now a virtual impossibility. The cultivation of provincial fields had always
been a labor charge on the populace for the upkeep of official institutions (including the officials themselves). By the early tenth century, the rendering of such produce rice had to be mediated through
the local gentry, as did the cedar bark and transport power, human
and animal, also mentioned here. All these contributions to the ad153 Communication of T5ji of Shohei 5(935)/io/25, HIB 1:360, doc. 245, discussed in Murai,
Kodai kokka, pp. 97—110; Nakano, Ritsuryosei shakai kaitai katei no kenkyu, pp. 210—7;
Sato, Heian zenki seijishi josetsu, pp. 284—94.
154 Declaration of Usa Hachiman Shrine of Kampyo 7(8g5)/n/i7, HIB 9:3464-70, doc. 4549;
Toji communication of Encho 2(g24)/8/7, HIB 1:328, doc. 219.
155 On provincially managed fields, see Abe, Heian zenki seijishi no kenkyu, pp. 43-46, 134-40;
Abe, Owari no kuni gebumi no kenkyu, pp. 212—13; Murai, Kodai kokka, pp. 271—80; Nagayama, Ritsuryo futan taikei no kenkyu, pp. 288—89.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
321
ministrative overhead of the province that originally came under the
heading of "irregular corvee" (or zoyo) were now, when imposed on
the tato gentry, rinji zoyaku.
Rinji zoyaku, here as in all other documents of the period, was
treated as categorically distinct from imposts like denso and shozei,
grain taxes for which the kokuga was directly accountable to the central government. But rinji zoyaku was not simply a reconstituted
form of provincial corvee. As illustrated by the "official exchange silk
thread" and "tribute tax sale silk" mentioned in the document, it
also included the forced procurement of tax commodities for the
central government. This was the system of "exchange," the official
purchase of tax commodities at rates fixed by the kokuga. The local
elite had become crucial agents of procurement in an area where
local and capital budgets intersected.
The use of provincial rice revenues for commodity purchases was
a long-established practice, and the government regularly stipulated
conversion values for each item demanded. These evaluations were
controlling, however, only in transactions between kokuga and central government, and the zuryo governors had, within limits established by provincial precedent, the opportunity to purchase tax commodities from provincial grain at lower rates, and, in at least some
cases, to enrich themselves in the process. A petition lodged against
the governor of Owari Province in 988, in complaining about his
abuses of the official exchange system, reveals the existence of an
unofficial exchange network in which the grain value of silk was
much higher.
This system of "exchange" {koeki) often caused conflict between
governors and local elites. Forced purchases financed by suiko payments and land rents (chishi) from "surplus unallotted" fields, had
begun in Nara times. By the tenth century, each province had fixed
exchange quotas, and koeki had become a major source of tax commodities. The steady decline in directly collected commodity taxes,
cho and yd, prompted the central government to permit larger and
larger amounts of provincial grain to be used for exchange, and the
system sometimes faltered when zuryo governors pleaded that grain
reserves were insufficient to meet procurement quotas. In such
cases, the governors, with tacit acquiescence from the capital, procured the tax items at a forcibly reduced price {genjiki) .156 The
156 On the "exchange" system, see Murai, Kodai kokka, pp. 80-116; Murao, Ritsuryo zaiseishi
no kenkyu, pp. 331-51, 199-239; Nagayama, Ritsuryd futan taikei no kenkyu, pp. 288-90;
Nakano, Ritsurydsei shakai kaitai katei no kenkyu, pp. 199-239.
Cambridge Histories Online © Cambridge University Press, 2008
322
ADMINISTRATION AND LAND TENURE
growth of the exchange-procurement system probably contributed
substantially to the growth of gentry power in the countryside. "Exchange" was essentially a tax limited to the rich and powerful, the
only group that could provide the required commodities, and the expansion of the procurement system increased the kokuga's dependence on them. By the middle of the tenth century, this system of
procurement was often viewed by the increasingly vocal local elites
as a kind of illegitimate commandeering. Their growing command of
provincial wealth, confronted by the increasingly discretionary powers of the zuryo, led to an escalation of accusations and recriminations that the high nobility of Kyoto could not afford to ignore.
Even under the ritsuryo, provincial officials had considerable discretionary taxing authority. The code permitted them to demand up
to sixty days of extra labor annually from every able-bodied male in
their districts for work not explicitly required by the provision of law.
One consequence of this definition was, for example, that repairs or
maintenance of existing irrigation works, explicitly required by the
codes, were not counted as extra labor, while construction of new facilities was. This onerous requirement was the "irregular corvee,"
zoyo, already mentioned. Although reduced to thirty days during the
eighth century, it seems to have been regularly abused by provincial
officers. Zoyo, along with the obligations of the peasantry to provide
transport labor to and from the capital, accounted for a major fraction of provincial revenues. It is therefore reasonable to conclude
that rinji zoyaku was an outgrowth of this discretionary taxing authority. A further indication of this is that "households of the gods"
(kambe), originally exempt from zoyo, were also held exempt from
rinji zoyaku.157 Even in the eighth century, it should be noted, demands for materials as well as labor could be defined as zoyo, as was
the commandeering of pack horses.158
The fraction of special irregular duties paid to the central government, usually in response to orders issued annually to each province,
became increasingly important later in the tenth century. Shortages
in standard shdzei funds and the consequent difficulty in procuring
exchange items was one reason why the government in the capital
relied increasingly on emergency demands. Underlying the changing
relationships between capital and provinces were changes in the
zuryo governor's position vis-a-vis the local elite.
157 Nakano, Ritsuryosei shakai kaitai katei no kenkyu, pp. 212-16.
158 Nagayama, Ritsuryo futan taikei no kenkyu, pp. 82-113; Yoshida, "Ritsuryo ni okeru zoyo
no kitei to sono kaishaku," pp. 233-62.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
323
It is clear that, by the early tenth century, "capable farmers" (kambyakusho) like the Oyama tato Heishii and Seiho, were routinely assuming the tax burdens of the peasantry under their economic domination. Heishu and Seiho paid these redefined obligations to
£ww-based collection agents, like the "responsible inspector" for the
village of Amaribe in Tamba who, in 932, confiscated grain from
their residences when the required commodities were not forthcoming. The authority of such collection agents, too, was becoming a
kind of estate prerogative, and a higher order of gentry was taking
shape on the gun level.159
Heishu and Seiho were, the sources indicate, in control of fields
beyond Oyama-no-sho. Such capable farmers were in a sense purchasing the labor of state subjects by guaranteeing tax payment on
their behalf, or, more realistically, paying in taxes for the power they
had acquired over peasant labor. Moreover, the report that when
presented with the collection agent's demands, Heishu and Seiho
fled the district, cannot be taken too literally. In fact, their absence
was quite temporary, and the sources show that they were back a few
years later, in the same position as before. Such tato were indispensable to the operation of tenth-century provincial taxation. According
to a document of 988, for example, each official village of Owari
Province had four or five tato responsible for land dues, an estimate
suggesting that the entire province, of perhaps 55,000 souls, had
about 300 of this type of tax provider.160
Not coincidentally, rinji zoyaku came into being just at the time
when the kokuga was coming to rely on a limited number of tax
providers, at least some of whom produced goods and forced labor
as surrogates for the original peasant taxpayers. Taxes of the special
irregular type, listed together with demo and shozei by the Oyamano-sho custodians as potential encumbrances on shden revenues,
would in this context appear to be a land tax, but the documentary
history of Oyama and other holdings of the time does not support
this conclusion entirely. In the case of Oyama-no-sho, the arguments
for exemption from special irregular duties did not rest on the character of the land as such, but on the immunity of its tatofromcho
and yd, collectively termed kaeki. Earlier, in 924, Toji had requested
the Tamba provincial headquarters to exempt the sho custodians and
tato from rinji zoyaku.lSl The principal arguments advanced in that
159 Abe, "Sekkanki ni okeru chozei taikei to kokuga," pp. 29-55.
160 Abe, Owari no kuni gebumi no kenkyu, pp. 145-53.
161 T6ji communication of Encho 2(924)78/7, HIB 1:328, doc. 219.
Cambridge Histories Online © Cambridge University Press, 2008
324
ADMINISTRATION AND LAND TENURE
communication were: (1) that the sho was certified as exempt temple land by the Council of State and the Ministry of Popular Affairs;
(2) that in 920 Tadahira had caused the then incumbent governor to
recognize "ten unlisted migrants" for the benefit of the estate, thus
immunizing the tato involved from cho and yd; and that (3) all previous governors had acknowledged the holding's immunity, as set
forth in the prior governor's certificate exempting the custodians
and tato from rinji zoyaku. The term "unlisted migrants" (chdgai
ronin) refers to ronin deliberately taken off the ronin register (but undoubtedly listed elsewhere) and thus made exempt from all formal
head taxes (kaeki) otherwise due to the central government.l6z The
sequence of arguments strongly suggests that the nonlisted ronin
were in fact the tato themselves, and that their immunities were
predicated on their services toToji on its Oyama estate.This linkage
to the estate exonerated them from liability for what were in origin
personal obligations to the state. But for most gentry, who lacked
such advantages, the obligations were probably as much based on
command over labor power and grain as on occupation of land. The
Oyama documents also illustrate how migrants, anomalous and illegal under the original ritsuryd system, had become an integral element in local tax gathering structures.
The authority to impose rinji zoyaku, or to grant exemptions from
it, was a discretionary power of the incumbent zuryd governor. In
diis respect, the tax was exactly like the old provincial corvee, zoyd,
and fundamentally different from cho and yd, imposts under the control of the central government. Conversely, governmentally sanctioned exemptions from cho and yd, such as those enjoyed by the tato
of the Oyama estate, were no guarantee of exemption from rinji
zoyaku. The records show that each successive governor had to be
asked for the exemption, and, as in the case cited here, the tato could
often expect help from the proprietory temple in procuring a central
government order commanding the governor to desist.
The series of taxes collectively regarded as special irregular duties
could, most probably, be demanded of any nonexempt person, but
the principal objects of the levies were inevitably those rich and powerful who were able to bear the burden. These heterogenous imposts, imposed by the kokuga without the aid of elaborate census reports, were simply collected from those listed as able to pay. This
system of subscribed taxes, consisting of local arrangements made
162 Nakano, Ritsurydsei shakai kaitai kaiei no kenkyu, pp. 201-7.
Cambridge Histories Online © Cambridge University Press, 2008
TAXATION AND ELITE WEALTH
325
without monitoring from the capital, demonstrates the growing autonomy of the kokuga and its by now self-perpetuating stratum of
resident officials.
By the end of the tenth century, special irregular duties had become a major revenue source for the central government, particularly, as already noted, for occasional expenses such as court festivals
or palace construction. The court was thus taking a share in the
governor's discretionary power to tap into free floating resources. Increasing reliance on "occasional" demands, and on the zuryo governors' ability to extract extraordinary revenues, contributed substantially in the eleventh century to the granting of zuryo appointments
for the purpose of accomplishing special projects.163
The position of the listed, or "named," tax subscribers most certainly permitted some considerable degree of informal authority over
peasant labor, and was undoubtedly a manifestation of the growth of
patrimonial authority on the local level. The tax providers' economic
base was concentrated in stored rice holdings as much as in land,
and their personal power over their clients was not truly that of a domanial landlord. Their power, however, was domanial in a very special sense. The grain storehouses of the powerful taw were located in
his residential compound, and his dependents, reduced to clientage,
were considered to belong to his menage. This was the sort of establishment referred to as sho houses in the reformist edicts cited earlier, and was to develop into the local lordship (usually justified by
assertions of reclamation) of the twelfth century.1&* The Oyama estate did not in fact obtain unconditional immunity against rinji
zoyaku from Kyoto until the early twelfth century.165 This power of
the kokuga was most important for the later development of the
shden, as it enabled the governors to establish partially immune holdings in their provinces without permission from Kyoto.166
Despite the conflicts over fukandenden (uncultivatable fields) and
damage reports that so preoccupied the officials of the early Heian
era, rice land continued to be a most important object of taxation,
and the early tenth century regime quickly evolved a new method of
securing its share, while at the same time condoning expanded discretionary powers for the governors. With the allotment system aban163 Takeuchi Riz6, Jiryo shden no kenkyit (Tokyo: Yoshikawa kobunkan, 1983 reprint), pp. 1-77.
164 Toda, Rydshusei, pp. 5-8, 73-113.
165 Order of Eikyu 2(iii4)/n/6, HIB 5:1637-39, doc. 1181. See Abe, "Sekkanki ni okeru
chozei taikei to kokuga," pp. 29-55; Miyagawa, Oyama sonshi, Hombun hen, pp. 96-99.
166 On the so-called zbyakumen shden, see Murai, Kodai kokka, pp. 351-72.
Cambridge Histories Online © Cambridge University Press, 2008
326
ADMINISTRATION AND LAND TENURE
doned, and periodic compilations of allotment maps and registers
discontinued, the central government instead maintained a permanent set of standard land maps for every province, on which exempt
temple and sho fields were noted by grid square location. All other
rice paddy under cultivation, whether within formal shoen boundaries or not, was, if under cultivation, to be considered taxable. This
was the system applied to newly opened fields of Oyama-no-sho, as
seen above. Governors, or, more accurately, their surrogates, made a
complete survey of cultivated fields once during each new gubernatorial term, and the tokoro in charge of the actual assessment came to
be an important organ of negotiation between zuryo and local gentry. The governor's agents, with the participation of the local cultivators and resident officialdom, decided, within the quotas established
by map registration and damage allowances, how much kammotsu
was due from each holding.167 By the late tenth century, in fact, rawmanaged rice fields were being taxed at higher rates than ever before,
compensating the government, at least partly, for the loss of immediate power over peasant labor.
LOCAL ELITES AS A POLITICAL FORCE
The trend toward discretionary taxes negotiated with the local gentry led to marked increase of political activity on their part. A striking example of this may be seen in the "Petition of the Gun Magistrates and Farmers of Owari Province," an impeachment of the
incumbent zuryo governor presented to the Council of State in 988.
By this time, a regularized procedure for the presentation of such
complaints in the capital had been established. Meanwhile, direct allusions to the general population, "the people," had disappeared
from official documents since about 950, to be replaced by phrases
such as "gun magistrates and farmers (hyakusho)" unambiguous references to the local elite.
The appearance of a hundred or more demonstrators in front of
the Yomei palace gate, bearing a petition for reappointment of a locally esteemed governor, or the dismissal of an unpopular one, was
a commonplace event in Kyoto during the tenth and early eleventh
centuries. Usually but not always peaceful, this petitioning fre167 Abe, Heian zenki seijishi no kenkyit, pp. 305—13; Inagaki, "Ritsuryoseiteki tochi seido no
kaitai," pp. 139-72; Sakamoto, Nihon ocho kokka taisei ron, pp. 1—137; Toda, Ryoshusei,
p. 96.
Cambridge Histories Online © Cambridge University Press, 2008
LOCAL ELITES AS A POLITICAL FORCE
327
quently brought about the desired result. Confrontational negotiation could be effective because officials in the capital were mindful
of the increasing incidence of violence against governors in the
countryside. Demonstrators were not unequivocally hostile, but they
could be impressively tough-looking. On at least two such occasions,
in 987 and 1019, certain courtiers considered recruiting the more
menacing of the visiting provincials as sumo wrestlers.168
Owari was probably seen as an especially troublesome province
in 988. An earlier governor had been impeached in 974, and, in 939,
another had been murdered. The detailed, and floridly written,
thirty-one-part accusation of 988 stresses the illegality, only occasionally defined in ritsuryo terms, of the governor's behavior, especially his confiscatory tax policies. It asserts that the tato had been
required to pay a total of 13.2 sheaves per tan of paddy, in the following components:
Official goods {kammotsu)
Field-tax grain (sokoku)
Tax fund dues (shozei)
Total tax per tan
1 to 6 sho
3 to 6 sho
1 to 4 sho (2.8 sheaves)
6 to 6 sho (13.2 sheaves)
If, as the document clearly states, the governor collected 13.2
sheaves per tan, he had surpassed the highest land rent (chishi) permitted by the ritsuryo for government leasing of surplus unallotted
land. That was only 10 sheaves, payable for a year's cultivation of one
tan of upper-grade paddy, having an assessed yield of 50 sheaves (25
to of threshed grain). Although this rate of taxation was protested by
the petitioners as illegally high, the somewhat lower levels attributed
to prior governors was still far higher than those of the ninth century.
Article Five of the petition estimates total land-tax returns at 1 million sheaves, which, taken at the rate 132 sheaves per cho, would indicate a total paddy area of about 7,576 cho for the province, somewhat
more than the 6,280 given by another source, but not implausible,
particularly in the light of the petition's further charge that the governor had deliberately overestimated rice land under cultivation.169
Of the three categories of tax on rice fields listed in the petition,
168 General discussions of this document, and the whole process of petition and protest, can
be found in Abe, Owari no kuni gebumi no kenkyu; Morita, Heianjidai seijishi kenkyit, pp.
227-60; Nakano, Ritsuryosei shakai kaitai katei no kenkyu, pp. 247—71; Sakamoto, Nihon
ocho kokka taisei ron, pp. 203—22; Sato, Heian zenki seijishi josetsu, pp. 339-59.
169 Abe, Owari no kuni gebumi no kenkyu, p. 94.
Cambridge Histories Online © Cambridge University Press, 2008
328
ADMINISTRATION AND LAND TENURE
tax fund dues (shdzei) appears to have been a kind of ad hoc surtax
imposed in addition to normal shdzei quotas and without any distribution of principal, "without passing through any special demand
for court expenditures, but simply applied to his [the governor's]
privately planned uses, either to be used up in trading or hulled and
transported to his house in the capital." The second listed category,
field-tax grain (sokoku), was treated as not totally outrageous but
merely excessive. The ritsuryd rate for so was a modest 1.5 sheaves
per tan, or, in grain, 7 shd 5 go, while the 3 to 6 shd mentioned in the
complaint is nearly five times that amount. The text expressing this
complaint shows that the ritsuryd rule governing so was, as a practical matter, a dead letter:
As forfieldtax grain, the official law has set limits. That being the case, successive governors, although they plead standard damage (reisori), nonetheless assess the full amount. Some provincial governors exact one tofivesko,
and some demand two to or less, but the incumbent Motonaga no Ason
[Fujiwara no Motonaga,fl.985-98] has increased the impost to three to six
shd. Never has there been such an example before.
Although the "official law" or Council of State decree (kampo)
said to regulate so rates cannot now be identified, it is clear from the
text as a whole that the actual rate was established, within limits, by
the discretion of the governors. That would have been unthinkable
by ritsuryd standards, and the two to per tan given as a maximum was
still nearly three times the amount prescribed by the ryd. The governor's use of "standard damage" claims to divert so revenues further
demonstrates the increasing latitude the zuryd could exercise in
dealing with local resources.
The protest about the third category, official goods (kammotsu),
centered on the ritsuryd distinction between unallotted rice fields
rented out on a yearly basis {chishiden) and fields that had been distributed to subjects, including both kubunden and konden. In ritsuryd
parlance the latter type had been considered "private," that is, distributed by the government to its subjects for their private welfare.
Such fields, as they were subject to so, were also called "taxable
fields" (sozeideri). Chishiden was, in theory, land still in the hands of
the state, "public" land awaiting redistribution and let out for rental
in the meantime. As taw agriculture had become the general rule,
and allotment was no longer a possibility, this distinction had lost
much of its original force. Nevertheless, a difference in tax costs persisted. Chishiden, being in theory "public," was free of "tax," yielding
only chishi rent. While chishi rates remained stable, however, taxes
Cambridge Histories Online © Cambridge University Press, 2008
LOCAL ELITES AS A POLITICAL FORCE
329
imposed on sozeiden had increased, and chishiden had now become
the more advantageous tenure.170
In the words of the complaint, "Both types offieldshould be taxed
in accordance with the order of the Ministry of Popular Affairs, but,
committing an abuse of law [dbd], he increases taxes as if all were sozeiden and thus, for the tato farmers, there is no little distress." Although
the Ministry of Popular Affairs order mentioned here has not been
found, it seems, at least in the minds of the protesters, to have precluded discretionary tax increments on chishi-payingfields.The use of
the technical ritsu term "abuse of law," besides accentuating the general tone of impeachment, implies wrongful conversion of revenue.
In the ritsuryo order, allotment fields had been the typical form of
sozeiden. They were regarded as providing the peasant household
with a financial basis for the payment of produce and labor exacted
as head taxes. Chishiden was free, originally, from all association with
personal taxes, but in the tenth century it became ever move closely
integrated with the general revenue system. In 925, as we have seen,
the Council of State ruled that kubunden abandoned because of flight
or pestilence should be rented out for chishi and the proceeds used
to purchase cho and yd articles.171 This illustrates a more general shift
away from taxation of households to taxation of grain revenues and
the administration of land. As the management of agriculture, presumed under the ritsuryo system to be an official monopoly, passed
to the tato, all land so managed became the same. Motonaga's imposition of the added tax on chishiden was, in the light of these developments, far from illogical, and anticipated the eleventh century
abandonment of the old legal distinction between public and private
fields.
This distinction between chishiden and sozeiden, no longer grounded
on thefiscalrequirements of the ritsuryo state, embodied yet another
growing incongruity. The ritsuryo assumption was that a landholder
had to choose between transfer of possession for a promised rental of
about one-fifth the yield, or retention of possession and cultivation
with the aid of paid or household labor ("direct cultivation"), even
when the owner was the government or a government agency. That assumption was not at all consistent with local practice. As we have already seen, government landholding inevitably involved gentry cultivators with whom the product had to be shared. The ritsuryo rules
were not intended to accommodate for domanial possession.
170 Abe, Owari no kuni gebumi no kenkyu, pp. 77-78. 171 See note 94, above.
Cambridge Histories Online © Cambridge University Press, 2008
330
ADMINISTRATION AND LAND TENURE
Article Twenty-nine of the 988 Owari petition charges Motonaga's
"kindred and retainers" with "forcing all the gun magistrates and
farmers to cultivate possessory fields {tsukuda)" and then taking the
yield. While demonstrating the surprisingly large amounts of rice
land potentially available for immediate appropriation by incoming
governors and their personal staffs, the passage shows that these
fields remained under gentry management. The principal complaint
of this passage concerns the cultivation charges imposed on the local
gentry, and the failure to reimburse them from the yield. The operative portion of the complaint states:
No sooner did his kin and retainers arrive, on the day office was transferred, their possessory fields filled the province and not a single household
was overlooked in assigning them for cultivation. Especially the possessory
fields of his sonYorikata, four or five cho in some gun and seven or eight cho
in some villages, distributed for cultivation throughout the eight gun, are
extremely numerous. On the day for granting loan rice (suiko), management provisions are not assigned to those ordered to work the possessory
fields, but when the time for collection comes, with no regard for consent
or protest, the rice is taken. This is in fact a seizure of paid-in official goods
in the form of harvest ricefromforced cultivation. This is to say nothing of
the four or five to of rice per tan taken by the collection agents as local produce [taxes]. When these accumulations are added up, they reach twice the
amount of legitimate official tax goods.
The authors of this text seem to have felt no need to explain the legal
basis for this expropriation of land. Perhaps, as some have suggested,
the land was deserted kubunden, or, perhaps, land falling within standardized fukandenden exemptions.172 In any case, it is most probable
that Motonaga's kin and retainers were exercising usufruct powers
over lands already under the management of the several households
mentioned, and that the language implying direct cultivation should
not be taken too literally. Especially noteworthy in this connection is
the dispersal of Yorikata's "possessory fields" throughout the
province. There is, furthermore, another reference to Motonaga's efforts to provide for his retainers asserting that "he seizes the customary tillages (reisaku) of the gunji and makes them tillages of his
retainers." "Customary tillages" were abandoned fields that these
magistrates had been allowed to take over for an indefinite period.
Unlike the private acquisitions of zuryo class nobles, however, the
tsukuda distributed by Motonaga was unquestionably a holding of
the kokuga dedicated to the support of provincial officers. Since
172 Abe, Owari no kuni gebumi no kenkyii, pp. 210-13.
Cambridge Histories Online © Cambridge University Press, 2008
LOCAL ELITES AS A POLITICAL FORCE
331
other articles in the petition accuse Motonaga of failure to distribute
stipends due the other regular appointees, the issue of inequitable
distribution was clearly significant, but the main burden of this particular accusation is the grievance of the gun magistrates and "farmers" (meaning tax-providing farmers as elsewhere throughout the petition) assigned to carry out the cultivation of so-called possessory
fields.173 The lands controlled by each tato household fell within a variety of administrative categories, including "provincially managed
fields," that were vulnerable to expropriation by the governors.
The expropriations complained of here, however, could not have
entailed any wholesale shifts in actual management. Basically, the
tato managing this tsukuda were complaining of an insufficient share
in the proceeds. Providing management and labor costs theoretically
entitled the holder to take the entire harvest, but "management provisions" had become quite generous and, at least in the case given
here, the "possessor" was not expected to take all that was harvested.
Otherwise, how could Motonaga's kin take an additional "four or five
to per tan"? The "management provisions" (eiryo) in this case were
unquestionably distributed out of provincial shozei funds. This is
clear from the statements that eiryo was awarded in conjunction with
suiko loan rice and that seizure of harvested rice without deducting
the eiryo portion was in effect the theft of official tax goods (kammotsu). The gun magistrates and farmers, or at least some of them,
seem to have found, under more relaxed circumstances, the management of provincial lands quite profitable, even though the legal
system failed to account for the profits. Viewed from this perspective,
Motonaga was attempting to reassert control over sources of wealth
that were gradually being removed from the oversight of the capital.
Possessory lands, as indicated by the quoted passage, were benefice fields assigned by the kokuga to the upkeep of regular and irregular local officials, without close regard for ritsuryo quotas for
stipendary fields. The added complaint that Motonaga's followers
had usurped fields privately administered by gun magistrates points
up the conflicting interests of gunji, irregular kokuga officers, and
zuryo in the disposition of local land revenues. By the middle of the
eleventh century, clearly stipulated benefice fields were a prominent
feature of every provincial regime, constituting a kind of shoen administered by the kokuga for the benefit of its officials.174
173 Abe, Oviari no kuni gebumi no kenkyu, pp. 142-43.
174 Morita, "Heian chuki gunji ni tsuite no ichikosatsu," pp. 208-17.
Cambridge Histories Online © Cambridge University Press, 2008
332
ADMINISTRATION AND LAND TENURE
The growing, but as yet imprecisely defined, authority of the kokuga to control the distribution of free-floating resources such as
these benefices was a major factor in the repeated conflicts between
zuryo and provincials. One important effect of this was a gradual
consolidation of kokuga and gun networks into a consolidated
whole. The Owari petition alleges that provincial benefices and gunji
fields were appropriated by Motonaga's private staff, who then encroached on the authority of the gun magistrates and local officers
to preside over land surveys. Surveying and tax gathering were becoming estate prerogatives of the "resident officials" of the kokuga's
"offices" (tokoro).1™
Conflicts between zuryo and provincial elites over provincial resources of this type, removed as they were from the capital's immediate distributive power, were not direct challenges to the Kyoto
court's authority. The struggle between capital and countryside was
buffered by the zuryo and the opportunities he had for extralegal
gain. There was nevertheless considerable potential for future conflict, as the court tapped into the governor's discretionary prerogatives in the form of rinji zbyaku, the extraordinary irregular levies already discussed.
Even when acting on behalf of the state's proprietary interest in
official goods, the zuryo was in an equivocal position. The 988 Owari
petition's most striking illustration of this is undoubtedly in its
eighth article. This reveals that quantities of tax cloth, grain, and
other forms of official goods were
seized from the persons of the gunji on the claim that it is the accumulated
liability of their districts, and from the homesteads of the people. On the
rationale that it is their proportionate dues, they seize things fraudulently
and without cause. When one or two houses are broken into, devastation
reaches ten or twenty places. . . .
One of the more significant features of this item of complaint is that,
once again, despite its indignant tone, the actions it described were
not all violations of ritsuryo rules. Gun authorities, for example, had
often been held personally responsible for taxes not collected in their
districts. Complaints elsewhere in the petition that the governor "illegally" increased grain taxes or reduced the exchange price for tribute silk and other tax commodities describe actions that, however
unfair they may have been, did not actually violate the written law.
On the contrary, such measures can be seen as perfectly proper
175 Morita, Heian jidai seijishi kenkyu, pp. 239-41; Toda, Ryoshusei, pp. 360-61.
Cambridge Histories Online © Cambridge University Press, 2008
LOCAL ELITES AS A POLITICAL FORCE
333
means of achieving the revenue goals of the ritsuryo state, and, a century or so earlier, would probably have evoked little comment.176 But
the rise of local gentry to controlling positions in the kokuga and gun
meant the displacement of administrative regulation by estateoriented customary law.
The gunji and farmers, in asserting legal norms appropriate to
their position, cited provincial precedent in four instances, all of
them prescribing limitations on the quantity of items that could be
taken as tax, setting new boundaries that, in the eleventh century,
would be formalized in rules like "the rule of kammotsu apportionment," setting limits on the dues the landholder had to pay.177 The
property interests of the provincial rich and powerful of the provinces were thus awarded a much greater degree of legal protection
than the Owari gentry could have hoped for.178
In terms of legitimate authority, the most powerful of the local
elite before the eleventh century were probably still the gunji. Although new kinds of tax-gathering agents, with new titles, were appointed by the kokuga in the late tenth century, these seem mostly
to have been centered on the gun headquarters rather than the
provincial capital. The gun and its officials survived in somewhat altered form, and appointments were made largely in recognition of
the power they already possessed.179 Gunji were leading figures in
uprisings like Masakado's rebellion and numerous other less spectacular acts of armed resistance to the authority of provincial governors. Ninth-century gunji sometimes exhibited truly impressive
power. In 856, Kagami noYoshio, chief magistrate of Kagami gun in
Mino Province, and Kagami no Yoshimune, chief magistrate of
neighboring Atsumi gun in the same province, led a force of "over
seven hundred men" in an attack on a government-approved water
project. The Kiso (then called the Hirono) River, flowing between
the provinces of Mino and Owari over a marshy, flood-prone plain,
had changed its course, inundating land on the Owari side, and the
Owari governor had received permission to return it to its original
course, probably by digging out the blocked channel. Thus began a
short inter-provincial war, to which the Kyoto authorities seem to
have reacted rather mildly. The Mino governor, who shared legal responsibility for gunji misconduct, was simply transferred to another
176 Nakano, Ritsuryosei shakai kaitai katei no kenkyu, pp. 247—71. 177 See note 12, above.
178 Ishimoda ShS, "Kodai ho," in Kodai, IK, vol. 4, pp. 291-333.
179 Morita, "Heian chuki gunji ni tsuite no ichikosatsu," pp. 208-17.
Cambridge Histories Online © Cambridge University Press, 2008
334
ADMINISTRATION AND LAND TENURE
province, as was the governor of Owari. What, if anything, actually
happened to the offending gunji has escaped the records.180
The power to marshal such a numerous force, probably including
some resident palace guardsmen, could not have depended merely
on accumulated wealth, and was surely linked to the gunji's central
role in the collection and exchange of tax goods and particularly the
holding and distribution of skozei tax rice. Gunji, the immediate superiors of a large staff of "irregular officials," certified all land transfers for the kokuga's approval. They were also, by law, jointly responsible with the governor for conducting land surveys, a duty, as
we have seen, with considerable potential for conflict.
In addition to their official stipendary fields (six cho for the chief
magistrate) and possibly other possessory lands, higher-level gun officers occupied one or more parcels of land that constituted, as we
have seen, a mixed portfolio of private and public holdings. Such
parcels were made up of irrigated paddy of various kinds, often including proprietary reclaimed rice fields (konderi) and unirrigated
grain or garden plots scattered piecemeal around a central "house."
The house, from which cultivation was directed, was comparable to
the "sho houses" on lands of nobles and temples, and included granaries and storage facilities. Hata no Tametatsu, the kaihatsu ryoshu
of the late eleventh century mentioned earlier, held gunji office.181
The Owari petition provides good evidence that agricultural administration from houses or homesteads, typical of tato cultivation in
general, had saturated the entire province. It also shows that the
scale of agricultural management was usually small, a fact reflected
by the establishment of official branch granaries in local areas, partially replacing storage facilities concentrated at the gun headquarters. A prosperous house with its own granary, or a granary holding
consigned government rice, could extend its influence to neighboring lands through rice loans or assumption of commodity tax obligations. The gunji retained much of their importance in the tax exchange and distribution system well into the eleventh century, as
well as constabulary and land-surveying powers.
In Atsumi gun of Mino Province, the scene of the 856 disturbance
just outlined, the gunji remained as the major local power right into
the twelfth century. Opposition of the gunji and his "sycophant
180 Entries for Jogan 7(865)/i2/27, Sandai jitsuroku, KT 4:169; Jogan 8(866)/7/g, Sandai jitsuroku, KT 4:191; Jogan 8(866)/7/26, Sandai jitsuroku, KT 4:192; Sato, Heian zenki seijishi
josetsu, pp. 159-67.
181 Toda, Ryoshusei, pp. 116-32.
Cambridge Histories Online © Cambridge University Press, 2008
LOCAL ELITES AS A POLITICAL FORCE
335
party" was effective against attempts by Todaiji to expand Akanabeno-sho there, and the documents generated by the conflict indicate
that land surveys and maps were very largely under gunji control. Although the ritsuryd titles for gun offices had nearly disappeared in
favor of tokoro-like functional designations, gun-based authorities
were clearly the major powers to be reckoned with by outsiders.182
The loss of the old custodial monopoly over official goods in the
ninth century, and the later growth of domains within their territories, did not mean their extinction.1^
Gunji power did not, of course, always go unchallenged. As revenues received in the capital steadily decreased and a combination
of factionalism and hereditary prerogative restricted access to office
there, resettlement in the provinces became an attractive option for
the middle nobility. Transplanted nobility, very often the descendants of former governors, could sometimes rival longer-established
elites, and rivalry of this kind was an important factor in Masakado's
rebellion. Cooperation from zuryo governors, continuing patronage
from high noble households, and intermarriage with gunji families
were major elements in the establishment of immigrant nobles in the
countryside. During the ninth century, and especially during the last
decade when the Kyoto government was beginning its last serious
attempt to revive the old ritsuryd order, the Council of State did its
best to prohibit the resettlement of nobles as an unwarranted burden on provincial administration. By the middle of the tenth century, on the other hand, the court was sufficiently reconciled to this
new group to grant some of them regular appointments as vice-governors or lower in their provinces of residence. As a consequence of
this new deviation from Chinese bureaucratic norms of avoidance, a
new class of resident official, holding regular rather than irregular
provincial office, appeared in the kokuga structure. The Owari petition reveals that the commissioned (nin'yo) officers, now almost entirely powerless, had the right to participate in land surveys and, denied this by the governor's personal retinue, sided with the resident
officials against him.
The challenge to zuryo authority by local officers was strongest in
the frontier provinces of the northeast. Even during Miyoshi Kiyoyuki's tenure as governor of Bitchu, a few of the local notables he
found intimidating held posts on the lowest level of regular provin182 Koizumi, "Todaijiryo Akanabe-no-sho," pp. 430-548.
183 Murai, Kodai kokka, pp. 43-58, 347-50; Abe, Heian zenki seijishi no kenkyu, pp. 30-32.
Cambridge Histories Online © Cambridge University Press, 2008
336
ADMINISTRATION AND LAND TENURE
cial office,184 but the preemption olkokuga posts by transplanted nobility, besides adding new tensions to an already volatile situation in
the outlying provinces, was a particularly ominous sign for the future
of court authority.
Yet the Owari petition and numerous other sources demonstrate
that control over provincial wealth was largely in the hands of the
rich and powerful. Of these "capable farmers," some, especially
those on the gunji level, were far more powerful than others, and
were, like Hata no Tametatsu, destined to become the s/u&t-holding
regional lords of the twelfth and thirteenth centuries. Their local,
and largely unofficial, control of peasant labor and land, justified by
their role as tax subscribers, constituted the chief guarantee of revenue for the zuryo governors and, ultimately, the court in the Heian
capital.185 While the shden prohibitions of the late ninth and early
tenth centuries pointed up the court's hostility toward the rich and
powerful and condemned their efforts to seek protection from the
nobility, the government of the late ninth century was more inclined
to compromise with the rural elites who constituted its new tax base.
The gunji and farmers of the 988 Owari petition could express their
grievances in an indignant, self-confident tone.
The role of land in the newly developing system of taxation reinforced the importance of the grain-holding elite, who, as capable
farmers, paid taxes on land they dominated but did not technically
own. Taxes levied directly on irrigated rice fields were, as the 988 petition attests, higher than those prescribed by the ritsuryo, but there
were limits established by custom. In the early eleventh century
these limits were very precisely defined in each province. In Iga
Province, for example, the rate per tan was:186
Actual rice: up to 3 to (1 to delivered to capital granaries)
Oil 1 go
Rice equivalent: 1 to 7 sho 1 go
Actual rice in ear: 1 sheaf
Rice in ear: 2 sheaves
This formula, specifying exactly what the tato should pay, including
how much rice should be threshed and where it should be delivered,
184 Abe, Heian zenki seijishi no kenkyu, pp. 28-248; Tokoro, Miyoshi no Kiyoyuki, pp. 41—70.
185 Takada, "Chusei shoki no kokka kenryoku to sonraku," pp. 6-25.
186 Petition of resident officials of Iga, Hoan 3(ii22)/2, HIB 5:171-72 doc. 1958; see also tax
assessment of Ota-inumam myo, Eisho 1(1046), HIB 3:774-75, doc. 639, and Katsuyama,
"Koden kammotsu rippo no seiritsu to sono zentei," pp. 1-43; Murai, Kodai kokka, pp.
320-27.
Cambridge Histories Online © Cambridge University Press, 2008
LOCAL ELITES AS A POLITICAL FORCE
337
expressed a kind of compact between the capital and the provincials,
eliminating some of the uncertainties that had occasioned complaints against zuryo governors, although the imposition of extraordinary levies continued under various guises. More important than
the precision of the formula, however, was the generality with which
it was applied. These were the dues to be paid on all rice lands in the
disposal or taxing power of the kokuga. The older distinction between chishiden and sozeiden no longer mattered, as all such land was
regarded as public field (koden), as distinguished from fields enjoying some sort of exemption. Of equal importance is the way these
rules provided special rates for specific areas called betsumyo, literally, "special names." Like the settlement established by Hata no
Tametatsu in the late eleventh century, these entities may be regarded as embryonic forms of the medieval shden. Provincial certification of betsumyo, sometimes headed by more than one proprietor,
was one way the governors could come to terms with the more eminent of the rural gentry.
The definition of the term "public fields" changed accordingly.
Public fields were now fields within the provincial domain, whether
leased to tato from year to year or held as absolute property under
the rule of reclamation. The emergence of the provincial domain,
and the evolution of the kokuga into a kind of estate management
agency, was largely the result of the rise of the tato class.
The dichotomy between providing tax-rice and cultivating land is
one clue as to why local lordships were so slow to develop. Fumyo
owed their position in the first instance to their control over movable
wealth, official rice consignments, rather than land. This wealth enabled them to finance agriculture beyond their own households, giving them de facto control over land owned or occupied by others. As
the Owari petition put it, by despoiling one household, a rapacious official could ruin several others. The distinction between fumyo and
tato, fundamentally a distinction based on the kind of assets being administered, broke down in the late eleventh and early twelfth centuries, the same time that term myo ("name") came to be applied to
a landholding and the cultivators on it, viewed as a single unit. Only
then had the power of the tax providers over their clients matured into
dominion over the land they cultivated, and the exploitation of people become, in the emerging system of customary law, the exploitation of land. While showing how different the rural societies of the
tenth and twelfth centuries were, the Owari petition marks a crucial
stage in the transition from autocratic bureaucracy to estatist polity.
Cambridge Histories Online © Cambridge University Press, 2008
338
ADMINISTRATION AND LAND TENURE
APPENDIX: NOTE ON HEIAN MEASURES
Area
1 bu = 3.3 sq. meters, a square, 1.8 meters to a side
1 tan = 360 bu = 1,190 sq. meters, always a rectangular area, typically
54 X 21.6 meters (30 X 12 bu)
1 cho = 10 tan = 11,900 sq. meters, a square, 108 meters (60 bu) to a
side
Length (land or distance measures)
1 shaku = 10 sun = 29.7 cm
1 bu = 6 shaku = 1.8 meters
1 cho = 60 bu = 108 meters
For the measure of grain or metal, the codes mandated a standard
sho measure called a "large sho."
1 shaku = 7.2 cc
1 go = 10 shaku = 72 cc
1 s/a? (the "large sho") = 10 go = .72 liters
1 to = 10 5/zo = 7.2 liters
1 koku = 10 to = 72 liters
For measuring other commodities, the code required the use of the
"small shd," which was exactly one-third the volume of the standard
"large sho" used for grain.
Sheaf measure
1 grip, ha, yields approx. 5 go of hulled rice
1 sheaf, soku = 10 ha, yields approx. 5 shd of hulled rice
The most fundamental unit of volume applied to grain was the sho,
officially standardized by a measuring box, masu. During the eighth
century, when the ritsuryo system was at its height, one sho was about
720 cc. During the ensuing three centuries, volume measure was far
from uniform, and it is not often possible to give absolute equivalents for the units of volume that appear in the sources. It is clear
Cambridge Histories Online © Cambridge University Press, 2008
NOTE ON HEIAN MEASURES
339
that from the tenth century onward, units of volume grew larger.
One famous attempt at restandardization, the senji masu, or "decree
measure" promulgated by the court of Emperor Go-Sanjo in 1072,
set one sho at approximately 1.2 liters (one of the very few figures
still available), but by that time, both local authorities and estate
proprietors had established their own standard measures.
The ritsuryo text gives the so tax rate as 2.2 sheaves per tan. This
misleading number represents an unsuccessful attempt to redefine
grain quantities at the time the Taiho codes were promulgated (702)
and does not reflect any deviation from established practice.
In the old system of the seventh century, it was assumed (very optimistically) that one shiro, a rectangular area of approx. 23.8 sq. meters, and made up of five square segments called bu, could yield one
soku, or "sheaf" (actually a bundle of rice ears cut from the top of
the stalk). Three soku of rice in ear per 100 shiro of paddy, or 3% of
the putative harvest, regardless of actual yield, were payable as so.
Notwithstanding other changes, this was the actual rate maintained
throughout the ritsuryo period.
One soku was supposed to yield a volume of 1 to of unhulled rice,
momi, or half that volume, 5 sho, of hulled rice, kome. One bu was accordingly presumed sufficient to produce one sho (about 104 cc) of
hulled rice. When the rules of the Taiho codes required a reduction
in the area of the bu, units of grain measure were reduced proportionately in order to maintain this equivalency of bu and sho.
The new codes discarded the shiro but retained the tan, a much
larger rectangle, typically 10 by 25 pre-code bu, 119 square meters in
area and the equivalent of 50 of the old shiro, or 250 of the old bu.
Measured by previous standards, the so rate for this unit area would
have been 1 soku 5 ha, that is, 1.5 sheaves. The Taiho andYoro codes,
however, subdivided the tan into 360 bu rather than 250 as before. A
tan rectangle of 25 by 10 bu was now, under this new surveying standard, 30 by 12. The compilers of the code reduced soku and sho
quantities to correspond with the newly reduced bu. One sheaf, or
soku, in the old system became 1.44 soku (36/25) in the new. The
amount of so due from one tan, 1.5 soku by the old measure, became,
precisely calculated, 2.16 soku (1.44 X 1.5). The Taiho ryo text,
rounding out this amount at the first decimal place, called for the
payment of 2 soku 2 ha, i.e., 2.2 sheaves per tan. In explaining these
recalculations, commentaries to the codes show that there was never
any intention in the codes to change the actual amount of 50 to be
collected per unit area.
Cambridge Histories Online © Cambridge University Press, 2008
340
ADMINISTRATION AND LAND TENURE
The older, larger soku and shd units, abolished in law but not, apparently, in fact, were officially restored in 706, only four years after
the Taiho code went into force. The official formula for the so rate
was accordingly reestablished at 1 soku 5 ha per tan, and remained
so thereafter. The newly mandated bu unit, linked as it was to surveying and land allotment, became standard.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 5
CHINESE LEARNING AND
INTELLECTUAL LIFE
INTRODUCTION AND ASSIMILATION OF
CHINESE LEARNING
The subject of this chapter is the learning of the Heian upper
class - not all of it, by any means, but that portion of it which was
regarded as fundamental in the education of males and which, even
in times of decline, enjoyed the highest formal prestige. "Chinese
learning," for our purposes, can be defined as the reading and writing of Chinese and those kinds of knowledge most directly dependent on learned traditions which boasted Chinese roots. Our focus
will be on government, Confucian study, and belles lettres - overlapping categories in the Heian context. We will touch on those aspects of law mostly directly related to Chinese learning and its exponents, but we will largely leave aside some other forms of Chinese
knowledge such as medicine, various kinds of divination, the calendar, and explication of the sutras. Mathematics, though important to
the practice of government and included among the curricula of the
state Academy (Daigaku-ryo), ceased to prosper as a field of study
very early in the period. We will leave aside also the important topic
of the influence of Chinese learning on such native forms of literature as waka and monogatari,1 and on painting, music, and dance.
A view that is fortunately losing currency among Western students
of Japan holds that to the Japanese the Chinese language remained
permanently alien: if Heian males failed to produce masterpieces of
imaginative prose paralleling those of their womenfolk, it was owing
to the burden of having to compose in a medium in which they were
ill at ease; and if, for example, a Japanese literatus chose to write of
Taoist immortals in the belief that such beings also existed in his
i See e.g., Jin'ichi Konishi, A History of Japanese Literature, vol. 2: The Early Middle Ages,
trans. Aileen Gatten (Princeton, N.J.: Princeton University Press, 1986); Helen Craig McCullough, Brocade by Night: "Kokin wakashu"and the Court Style in Japanese Classical Poetry
(Stanford, Calif.: Stanford University Press, 1985); David Pollack, The Fracture of Meaning
(Princeton, N.J.: Princeton University Press, 1986).
341
Cambridge Histories Online © Cambridge University Press, 2008
342
CHINESE INTELLECTUAL LIFE
own land, he must have been playing with "exotic" ideas.2 It is true
that a note of something exotic and foreign clings to the idea of
China as it appears in the vernacular literature of the time, with
China as a place and the Chinese populace presenting the ultimate
challenge to Japanese self-esteem - a challenge that the hero of the
narrative always surmounts3 - and with certain types of Chinese
goods enjoying almost a magical cachet. But to the Heian Japanese,
Chinese culture and its products existed apart from national boundaries as requisite tools of civilization and, to a very high degree, as
the marks of civilization itself.
No precise date can be assigned to the beginnings of Chinese
learning in Japan. From the fifth century on, however, the importation of Chinese ideas and techniques expanded rapidly. The process
contributed to the growth in authority of the central government in
a variety of essential ways, from reinforcing the primacy of the emperor (tenrio) to providing means of record-keeping for his tax gatherers. Until the arrival of Chinese learning, Japanese society had
been unlettered, apparently lacking elaborated theories of government or social ethics, with no formal system of law, and without
recorded history or a religious canon. To say that Chinese learning
filled a vacuum would, of course, do injustice to the resources of oral
tradition, but whatever the power of native tradition, the importations often dealt with categories of thought and practice that had not
previously existed in Japan and for which the indigenous language
lacked words.
Chinese was the basis of government, both ideal and practical. A
bureaucratic system carefully modeled after that of the T'ang and
referred to by historians as the "statutory" (ritsuryo) regime had
reached its apogee in the eighth century. By the beginning of the
Heian era, the system had already begun to evolve in new directions;
by the beginning of the tenth century, although the conception and
rhetoric of Confucian government remained, as did its forms and
usages, many of its functions were being carried out by other means.
Aristocrats and their clients competed for office, empty or not,
within its bureaucracy. Chinese provided the medium for the memorials, decrees, codes, administrative regulations, ordinances, com2 See, e.g., Michele Marra, review of / / Dio incatenato: Honcho shinsenden di Oe no Masafusa,
trans. Silvio Calzolari, Monumenta Nipponica 41 (1986): 495-97, quoting a passage from that
book.
3 See, e.g., Thomas H. Rohlich, trans., A Tale of Eleventh-Century Japan: Hamamatsu Chunagon monogatari (Princeton, N.J.: Princeton University Press, 1983).
Cambridge Histories Online © Cambridge University Press, 2008
ASSIMILATION OF CHINESE LEARNING
343
mands, communications, and certificates by which the government
functioned. Chinese precedents governed the vocabulary and form
of these official writings. Chinese exempla and ethical teachings
were invoked to justify decisions of state. Above all, the ritual persona of the ruler, whether emperor or regent, was fashioned according to Confucian patterns.
The earliest bearers of Chinese culture in Japan were continental
immigrants (kikajin) and their descendants. According to the account in the Nihon shoki,* Chinese learning began about 400 in the
reign of Emperor Ojin with the coming of an emissary from Paekche
named Achiki, who was able to read the classics and who gave instruction to the crown prince. Achiki recommended that a more
learned man, Wani, be invited to Japan, and he arrived a year later.
Tradition says that Wani presented to the court the Analects of Confucius and the Thousand-Character Classic. Such accounts reflect the
prestige of the two works, not historical fact, if for no other reason
than that the Thousand-Character Classic was not composed until the
sixth century. Wani was said to have been the ancestor of a line of
scribes, the Kawachi no Fumi no Obito. A certain Achi no Omi, who
claimed to be a descendant of Emperor Ling (r. 168—89) of the Later
Han, came a few years later, accompanied by "the people of seventeen hsien (districts)," and became the ancestor of theYamato no Aya
no Atae, a large hereditary group of scribes and craftsmen who were
followers of the Soga. It was members of such groups, some recently
arrived, all with strong memories of their continental past, who carried on the profession of reading and writing.
Not until the beginning of the seventh century with ShotokuTaishi
did there appear the first writings of any length in Chinese composed
by a Japanese. By the early Heian, however, the kikajin had largely
been assimilated among the Japanese population, and their skills had
been acquired by Japanese. Although some families still boasted of
their continental ancestry, the government no longer felt obliged to
make special provision for employing them or educating their sons.
The ninth century in particular was a time when Chinese learning
thrived in Japan, and books in Chinese by Japanese authors included
histories, legal compendia and commentaries, manuals of court procedure, dictionaries, encyclopedias, religious tracts, travel diaries,
treatises on a variety of subjects, and poetic anthologies.
4 W. G. Aston, trans., Nihongi (Rutland, Vt., and Tokyo: Charles E.Tuttle, 1972 reprint), vol.
1, pp. 261-62.
Cambridge Histories Online © Cambridge University Press, 2008
344
CHINESE INTELLECTUAL LIFE
During the two hundred years that preceded the Heian era, the
Japanese actively sought Chinese knowledge through repeated embassies: ten such missions to Sui andT'ang are known to have taken
place or been attempted during the seventh century, and nine more
were begun between 702 and 779, of which seven were completed.
The cost was great, both in goods and in human lives. Many of the
Japanese who went abroad to study died before they could return
home; others, such as Buddhist monks, remained in China for
decades. Some - including monks who returned to secular life married Chinese wives and had children who aided in subsequent
missions. Of the men, ranging in number from about a hundred to
six hundred per mission, who embarked on the fragile ships, sometimes fewer than half would make their way back. Those who did
brought not only books but their personal experiences to shape the
statutory regime. The ambassadors themselves were usually of
the Fourth Rank, the top of the middle aristocracy, and their immediate subordinates were also men of prominent families. Though
none rose to be minister, a substantial proportion subsequently
achieved the office of Counselor or Consultant or became professors
at the state Academy. Accompanying them to China as students were
doctors of medicine, yin-yang masters, painters, sculptors, musicians, jewelers, metal casters, and other craftsmen.
Only two more embassies to China took place after 800. The first
set sail from Japan in 804 and returned in 806. The second departed
in 838, and the last of its four boats came back in 840. A final one
planned for 894 was abandoned under the urgings of the prospective
ambassador, Sugawara no Michizane (845-903), who cited reports
of civil disorder on the continent. The underlying reasons for the
cessation of the embassies are still a matter of debate. The government's declining ability to support expenditure on such a scale was
undoubtedly a factor, as was the growth of private commerce conducted by Chinese and Koreans, which, together with the frequent
arrival of embassies from the Manchurian kingdom of Po-hai, provided a safe and less costly means of maintaining contact.5
Japanese monks wishing to travel to China for study could do so
privately, while aristocrats were able to obtain through trade the luxury goods they prized. And the Japanese now believed that they had
acquired the essentials of Chinese knowledge. They had been seek5 Robert Borgen, Sugawara no Michizane and the Early Heian Court, Harvard East Asian
Monographs, 120 (Cambridge, Mass.: Council on East Asian Studies, Harvard University,
1986), pp. 246-50.
Cambridge Histories Online © Cambridge University Press, 2008
ASSIMILATION OF CHINESE LEARNING
345
ing selectively in China to serve Japanese needs and tastes. One of
the chief purposes of the pre-Heian embassies had been to deliver
and bring back long-term students (ryiigakusho) and monks (ryugakuso). In contrast, those who went abroad in the two ninth-century
embassies were shdyakusho ("students seeking to gain"), specialists
sent for a few months to resolve doubtful points of interpretation
(e.g., of law) or have specific questions answered; if monks (shoyakuso), to bring back doctrinal tracts lacking in Japan, receive initiation into orthodox lineages, and copy holy images.6 The selectivity
of the Japanese borrowings, whether of law or of styles in Chinese
poetry, is a theme that must be constantly kept in mind in any study
of Japan's cultural relations with the continent.
To what extent Chinese learning had been imported into Japan by
the end of the ninth century, as well as which aspects enjoyed the
most prestige, is suggested by a bibliography, Nihonkoku genzai
shomokuroku, compiled in 891 upon imperial order, to inventory the
books remaining in the country after the Reizeiin, a former imperial
mansion that housed a library collected over generations, was destroyed by fire in 875. The books listed in it totaled almost 19,000
scrolls (maki); and although the titles of native works are mixed in
among them, the number of books brought from China - most of
them as a result of the embassies - is impressive. The compiler, Fujiwara no Sukeyo (847-97), was a prominent literatus who enjoyed
the patronage of the regent Fujiwara no Mototsune (836-91).7
Sukeyo modeled his catalogue after the bibliographical essay in the
Sui shu, the official history of the Sui dynasty (581-618). Its forty
categories begin with the Confucian classics and include law, medicine, agriculture, warfare, astronomy, the calendar, the "five elements," and the works of the various philosophical schools. Especially large categories were those having to do with ritual and with
various kinds of historical writing. The extant catalogue lists almost
1,400 scrolls of Chinese dynastic histories and almost 2,000 scrolls of
books on rites and ceremonies. More emphasis is given to belles lettres than to the classics - reflecting a T'ang taste that was especially
6 Charlotte von Verschuer, Les Relations officielles du Japon avec la Chine aux viif et ixf siecles
(Paris: Iibrairie Droz, 1985); Edwin O. Reischauer, Ennin's Travels in T'ang China (New
York: Ronald Press, 1955); Robert Borgen, "The Case of the Plagiaristic Journal: A Curious
Passage from Jojin's Diary," in Aileen Gatten and Anthony Hood Chambers, eds., New
Leaves: Studies and Translations of Japanese Literature in Honor of Edward G. Seidensticker (Ann
Arbor: Center for Japanese Studies, University of Michigan, 1993), pp. 63-88.
7 Ohase Keikichi, Nihonkoku genzai shomokuroku kaisetsu kb (Tokyo: Komiyama shuppan,
1976). For Sukeyo, see Marian Ury, "The 6 e Conversations," Monumenta Nipponica 48
(1993): 37OCambridge Histories Online © Cambridge University Press, 2008
346
CHINESE INTELLECTUAL LIFE
congenial to the Japanese - and a number of books by Chinese authors appear in it that are not listed in continental bibliographies,
several of which have been discovered at Tun-huang. And probably
there were other books not deemed worthy of being entered into an
official bibliography: handbooks, practical books of various sorts,
and volumes to amuse and instruct the less educated, women, and
children.8
Chinese learning was both assimilated and appropriated into the
indigenous culture. As a separate cultural tradition or gathering of
traditions, however, Chinese learning became itself a Japanese tradition. Reified, it was possessed, transmitted, hoarded, honored, displayed, boasted of, cultivated, and lamented when it failed to flourish. It came to have its own patriarchs, a history, and curiosities. The
most appropriate metaphor might be that of a fund of intellectual
capital, consisting of the segment of Heian Japan's intellectual inheritance that depended directly on the use of the Chinese language
and education in Chinese books. Originally brought in from abroad,
it still received supplements from abroad, though in reduced number, but during the Heian era it can also be seen replenishing itself
from its own resources. The intent of this chapter is to describe this
Chinese intellectual capital, to portray those who made particular
claim to ownership of it, the ways in which they acquired it and the
environments in which they utilized it, and to suggest something of
its evolution during the first three centuries of the Heian period. The
texts it produced are very much more difficult than those of the vernacular traditions, while the aesthetic, spiritual, and moral needs it
served are generally alien to contemporaries, Japanese as well as
Westerners, in contrast with the more ingratiating and-perhaps
only superficially - more accessible intellectuality of the salons that
produced the vernacular literature, much of it written by women.
The study of the bunjin (literati) - virtually all of them men - and
their world has suffered comparative neglect, but acquaintance with
the subject suggests that the fund of Chinese learning supported a
weighty and on occasion lively intellectual life.
Whether or not he was destined to become a professional literatus, a little boy of the Heian upper class would typically begin to
learn written characters in his sixth or seventh year - about age five
by Western count. First would come the matriculation ceremony: the
child's teacher would read aloud the first few characters of the Clas8 Kawaguchi Hisao, Heiancho no kambungaku (Tokyo: Yoshikawa kobunkan, 1981), pp. 71—74.
Cambridge Histories Online © Cambridge University Press, 2008
ASSIMILATION OF CHINESE LEARNING
347
sic of Filial Piety (Hsiao ching) as annotated by the T'ang emperor
Hsiian-tsung (r. 712-56), and the child would pronounce them several times after him before the lesson was adjourned for the serious
business of congratulations and feasting. In learned families, the
teacher was likely to be some senior relative; among the highest nobility, an eminent literatus client. Crown princes were provided with
two Confucian tutors, and the post was a great honor. No one
thought it odd when a tutor was appointed for a crown prince who
was only six months old; as with many other Heian institutions, the
ceremonial function was considered to be of absolute value even in
the absence of practical activities. In addition to the Classic of Filial
Piety, beginners were taught the Thousand-Character Classic, Meng
ch'iu, and extracts from the poets.
Meng ch'iu consists of 592 four-character phrases, each recounting
the name and salient characteristic or deed of some famous person,
the phrases arranged into rhymed couplets for easy memorization.9
Commentaries, and presumably also the teacher, helped explain the
often cryptic text. Learning was by memorization, and memorization meant repeating aloud; students preparing for Academy examinations were expected to memorize commentaries as well as the
texts themselves. The student would be expected to be able to recite
the Chinese reading for each character, its equivalent in Japanese,
and the interpretation. Heian scholars had not yet devised the standardized method of reading Chinese texts known as kundoku, in
which words are rearranged and grammatical features added in
order to convert the Chinese original into pseudo-Japanese. Texts
were read aloud word by word as they appeared in the Chinese and
then read a second time rendered into Japanese, in a manner which
would have seemed free by later standards. Different schools, moreover, had different traditions for Japanizing their texts: that of the
Oe, for example, was freer and more Japanese-sounding than that of
the conservative Sugawara.
Chinese was not taught as a spoken language. At the beginning of
the Heian period there were Japanese - many of them of recent continental descent - who had facility in spoken Chinese, but the ability had all but disappeared by Michizane's time, so that by the end
of the period for a Japanese to be able to speak in Chinese at all was
cause for amazement. But the student's education would not have
9 Burton Watson, trans., Meng ch'iu: Famous Episodes from Chinese History and Legend (Tokyo:
Kodansha International, 1986).
Cambridge Histories Online © Cambridge University Press, 2008
348
CHINESE INTELLECTUAL LIFE
proceeded without reference to the sounds of Chinese. There were
teachers of pronunciation, of Chinese descent, at the Academy, at
least during its prosperous early days. A sign of the lack of contact
with oral Chinese was that in composing Chinese verse, the Japanese generally simplified the prosodic rules having to do with tone,
while observing those for rhyme.10 Both rhymes and tones might be
dealt with through the aid of handbooks and dictionaries. But the
Japanese were not themselves quite deaf to tone: Japanese lexicographers of the period recorded the pitch patterns of native words by
a method developed from that used for the notation of Chinese
tones. 11
Works such as Meng ch'iu provided Japanese learners with a fund
of historical and exemplary lore, a mainstay of elegant discourse.
Meng ch'iu, moreover, drew its material from a wide variety of canonical Chinese works, with which its male (or sometimes, as in the
case of Sei Shonagon, female) user could therefore become acquainted at second hand. In addition to primers from China, the
Heian beginner would also be given compendia of fundamental
knowledge written in the Chinese language by Japanese authors.
Such a work was Kuchizusami (the title might be loosely translated
"Fun with Recitation," or just "Fun with Learning"), written in 970
by Minamoto noTamenori (d. 1011) for Matsuo-gimi, Fujiwara no
Tamemitsu's oldest son, then in his seventh year. Tamenori was also
the author of a collection of maxims, of a handbook of government
offices, and of Sambo ekotoba, a collection of pious anecdotes to provide seemly amusement for a young princess turned nun.12 As a
nonpareil beginner's book, Kuchizusami was probably inspired by
Wamyo ruijti sho (comp. 935), a dictionary-encyclopedia of Japanese
words arranged by category of meaning, compiled (at the order of a
scholarly princess) by Tamenori's teacher Minamoto no Shitago
(911-83). (The categories in Wamyo ruiu sho, in turn, were inspired
by those of a Chinese encyclopedia compiled by the T'ang poet Po
Chu-i).The work became highly popular almost immediately after
its compilation and continued to be widely used, serving also as the
basis for late-Heian educational compendia.13 One of its virtues was
its accuracy: Oe no Masafusa (1041-1111), who admired Tamenori,
10 Konishij A History ofJapanese Literature, vol. 2, p. 49.
n Oka Kazuo, ed., Heiancho bungaku jiten (Tokyo: Toyodo, 1972), p. 332.
12 Translated by Edward Kamens in his The Three Jewels: A Study and Translation of Minamoto
Tamenori's Sanboe (Ann Arbor: Center for Japanese Studies, University of Michigan, 1988).
13 Kawase Kazuma, Kojisho no kenkyu (Tokyo: Kodansha, 1955), pp. 155-60.
Cambridge Histories Online © Cambridge University Press, 2008
ASSIMILATION OF CHINESE LEARNING
349
found only two lapses in it, each involving a single character.14
Tamenori's purpose of supplementing Meng ch'iu with material that
is "near" is mentioned in the preface in which he speaks of his young
charge, naturally endowed with intelligence but not yet used to
memorizing, fond of play and fond of singing. Now, if he will sing
these words out he will commit them to memory; he must constantly
keep this book in his hand as a plaything. . . . 15 The teacher's affection for the boy, easy to read between the lines, suggests a human dimension that historians too often overlook.
As a guide to what was considered essential knowledge for the
young Heian aristocrat, Kuchizusami well repays attention. The
work, about twenty-five pages in modern type with no commentary
except its author's own, presents its information under nineteen categories ("gates," mori), subdivided into "stanzas" (kyoku), here denoting memorizable units of various lengths. Many - almost certainly the majority-are lists of things and enumerations: the three
luminaries, the seven stars, the three great edifices, the twenty-five
great monasteries, the five tones, the seven tones, the eight tones,
the days of Buddhist observance, the gates of the palace, the bureaus
of the Ministry of Ceremonial. Under the category of "Geography"
are listed the three passes, the seven high mountains, the nine tumuli, and the three bridges. The knowledge that is taught is of two
principal kinds: that which is needed in public life, and that which
will most directly concern the personal well-being of the learner.
Science for its own sake does not exist, nor do abstractions. The
Heian mind was quite incurious about the natural world. Here and
there are mentioned some facts of Japanese history, but there is
rather more of Chinese. Much emphasis is placed on magical observances and recitations: the commentary in the opening section,
"Heavenly Phenomena," informs the learner of the locations of the
star deities - important because these were taboo directions16 - and
the category following, "Periods Within the Year," includes formulas
to be said to the deities of the four quarters and on Koshin night so
as to avoid contracting illness. Ommyodo, the Taoist divinatory science, is a major category, while the category of "Medicines," which,
significantly, immediately follows it, is largely concerned with lucky
14 Kawaguchi Hisao and Nara Shoichi, Godansho chu (Tokyo: Benseisha, 1984), pp. 1277-81;
for an introduction and partial translation of this important source, see Ury, "The Oe
Conversations," pp. 359-80.
15 Zoku Gunsho ruijit (Tokyo: Keizai zasshi sha, i960 reprint), vol. 32.1, p. 61 (doc. 930).
16 Bernard Frank, Kata-imi et Kata-tagae: Etude sur les interdits de direction a I'epoque Heian,
Bulletin de la Maison Franco-Japonaise, n.s., 5:2-4 (1958), passim.
Cambridge Histories Online © Cambridge University Press, 2008
350
CHINESE INTELLECTUAL LIFE
and unlucky days on which to dose oneself. The book includes a
rather thorough primer of Buddhism, emphasizing praxis and with a
number of hymns and gathas to memorize. Among "Animals" are
hawks (giving the dates when various provinces offer them as tribute
to the court), horses (including a charm to recit; that will cure diseases of horses' bellies), the shishimushi, a cicada-like insect whose
cry was thought to foretell death, and the nue, a thrush whose sad
song was associated with unrequited love. Units of measurement are
of the greatest importance to the future official as he is likely to be
concerned with tax assessments and tax collections, and thus appear
throughout the book under various categories, enumerated in great
detail and what seems to be every possible variation: two sets of
names for units of houses, from 5 to 2,500 (under "Dwellings");
measurements of length, area, and capacity (under "Agricultural
Land and Buildings"); how many years constitute a generation
(under "Periods of Years"). The final category in the book, headed
simply "Miscellaneous," consists of yet more terms of weight and
measurement and ends with a multiplication table, starting with
9 X 9 and going downward 1 X 1 , followed by a list of characters for
powers of ten; fourteen are given.
The activities of civilization, in the Heian scheme, had their focus
within the imperial compound. It is not surprising, then, that the
lengthiest category in the text, at least in its extant form, lists the
structures of the palace; it also describes the divisions of the capital
city, which would be viewed logically as an extension of the palace.
Another lengthy category lists government offices. The category
headed "Human Beings" typifies in miniature the mode of Heian
learning: it brings together lists of the five emperors of Chinese high
antiquity; the three dynasties of Hsia,Yin, and Chou; the eleven emperors of the Former and the twelve of the Later Han; the nine lesser
disciples of Confucius, and other lists of worthies of ancient and legendary antiquity; the names for the barbarians of the four directions;
the six kinds of portentous dreams; the three conditions under
which a man cannot spurn his wife; the seven grounds for divorcing
her; the three kinds of subordination which she owes ("to her father
before marriage, to her husband upon marriage, and to her son after
her husband's death"); the three persons to whom a man owes loyalty ("Father, teacher, and lord"); the organs and cavities of the
body; words for various degrees of direct ancestor and direct descendant; kinds of persons and transgressions to be forgiven; the
Cambridge Histories Online © Cambridge University Press, 2008
ASSIMILATION OF CHINESE LEARNING
351
dangerous years in human life ("13, 25, 37, 49, 61, 73, 85, 91"); a
charm to protect against a bad dream or make a good dream come
true, to be recited under a mulberry tree three times; and a verse to
chant should one come across a dead man in the road at night.
Readers of The Tale ofGenji will remember the skill with which the
hero, a paragon at least in his public behavior, dances and plays a
number of musical instruments. The lengthy section devoted to
"Music" accurately reflects the esteem that was accorded musical
performance in public and private life. The most essential of all fields
of knowledge, however, was surely that represented by the category
of "Writings" (shoseki). Heading its lists are those of the five, seven,
and eleven classics, the three histories (the Records of the Historian
[Shih chi] and the two Histories of the Han [Han shu and Hou Han
shu]), the eight dynastic histories from Wei through T'ang, the forms
of verse (bun) and the forms of literary prose (hitsu). Tamenori's
commentary supplies a succinct definition: "If it rhymes, it is bun; if
it lacks rhyme, it is hitsu. Bun is put together in units of two phrases
(ku); hitsu is made up of units of four phrases." There follow some
simple rules of prosody: a pattern each for five-syllable and sevensyllable regulated verse (Chinese, lii-shih); the seven "sicknesses" of
poetry; the seven kinds of antithesis to be used in the central couplets of regulated verse; the six kinds of poetry and principles of poetic rhetoric (those enunciated in the Greater Preface to the Book of
Odes [Shih ching]); the system of recording pronunciation of Chinese
words, known as fan-ch'ieh, used in Chinese dictionaries; the names
of the four tones; and the long lists of rhyme-word categories divided
by tone that one needed to know in order to consult a Chinese dictionary generally as well as, specifically, to compose Chinese verse.
It is interesting to note that Tamenori mixes current with antiquated
information. When he compiled his textbook, the T'ang history was
only twenty-five years old, a recent work by leisurely Heian standards. In the terms he uses to distinguish prose from poetry, however, he follows Six Dynasties usage; by the T'ang, bun had come to
mean prose (poetry was shi).
The remainder of Tamenori's category of "Writings" is given over
to Japanese; but whereas the Chinese items felt to be pertinent were
the classics and histories, along with aspects of Chinese rhetoric useful to Japanese who would have to compose in that language or at
least demonstrate discrimination in regard to the compositions of
others, Japanese books are of only two kinds, both, of course, fol-
Cambridge Histories Online © Cambridge University Press, 2008
352
CHINESE INTELLECTUAL LIFE
lowing Chinese models, both written in the Chinese language, and
both equally, from the Heian way of thinking, essential to civilization. These are the Six National Histories (Rikkokushi) and the legal
codes (writings to be discussed later in this chapter). The compilation of the codes is of some interest to Tamenori: one of his commentaries gives the dates over which this work took place, and a second commentary gives a brief history of the interest of the emperor
in legal compilation, beginning with the Seventeen Articles of
Sh5toku Taishi. The major divisions within the administrative code
(ryd) and the criminal code (ritsu) are enumerated, and this is followed by an abbreviated list of the eight kinds of oppression and six
of the eight "deliberations" as to personal status of the criminal that
require amelioration of punishment, a prominent feature of Chinese
law.17 Following, Tamenori offers a verse to help in remembering the
Japanese syllabary. Using the forty-seven syllables once each, it is
written in the full forms of Chinese characters in the manner of the
Man'yoshu. Tamenori comments that most people recite the more
familiar one that goes "Ame tsuchi. . . " but that his is better.
As an adult, Tamenori's young pupil would have been expected to
be able to compose a passable Chinese verse, with correct rhyme
words, in one of the Chinese shih patterns for arranging tones. That
might be as part of the entertainment of a particularly elegant banquet, with everyone pleasantly tipsy, or a formal outing. On more
ceremonious occasions, unless he were of a particularly scholarly
bent, and quite likely even then, he would present a poem composed
on his behalf by a professional literatus. If a composition in parallel
prose, the ornate Chinese style that later would come to be known
as p'ien wen, was required — such as the dedication {gammon) that
would accompany a gift to a Buddhist temple - he would almost certainly rely on the services of an eminent literatus, in whose personal
anthology it would later appear. Such writing was fraught with difficulty, and literati themselves would on occasion commission these
documents from colleagues of especial talent and reputation. In his
informal writing, the official would ordinarily use less formal Chinese with an admixture in various degrees of Japanese usages, the
style varying with the topic and the purpose of the document, as well
as the aptitude and education of the writer.
The advanced learner could also rely on compendia. Compiled in
17 See Wallace Johnson, trans., The T'ang Code, vol. 1: General Principles (Princeton, N.J.:
Princeton University Press, 1979), pp. 88-104.
Cambridge Histories Online © Cambridge University Press, 2008
ASSIMILATION OF CHINESE LEARNING
353
the year of its author's death for the youthful Fujiwara no Yorimichi
(992-1074), Tamenori's Sezoku gembun, the volume of maxims previously mentioned, is an assemblage of useful quotations from Chinese books, arranged by category. Only one of its three original
scrolls survives, but it is noteworthy for the range of its sources,
quoting no fewer than fifty-three different Chinese books, each
meticulously noted by the author.18 They include the Analects, the
Odes, the Spring and Autumn Annals with its commentaries, the Documents (Shu), the Rites (Li chi), the Classic of Filial Piety - all belonging to the Confucian canon - the Records of the Historian and the
two Histories of the Han, and to a lesser degree the histories of the
Wei and the Chin, and A New Account of Tales of the World (Shih shuo
hsin yu, a fifth-century collection of witty anecdotes about intellectuals of the centuries preceding). Works associated with the Taoist
tradition, in some cases tangentially, are also cited. Chuang tzu, Lao
tzu, Lieh tzu, Huai-nan tzu, Pao-p'u tzu, the fantastic geography
Shan-hai ching, and the legends of the Taoist immortals are quoted.
Study of such texts, the Chuang tzu especially, was popular with the
literati, even though these books were not included in the curricula
of the Academy. The Oe in particular interested themselves in the
Taoist sciences, several of them winning reputations as accomplished physiognomists. Of books composed in Japan, only the
Nihon shoki, first of the National Histories, and the ryo, the administrative code, are represented here.
Owing to the nature of the compendium, some other books that
figure prominently in Heian education are omitted. Missing are two
works that had to be memorized for one of the civil service examinations: the dictionary Erhya, part of the Confucian canon, and the
literary anthology Wen hsuan, which supplies numerous quotations
throughout Heian writing in Chinese. The poems of the beloved Po
Chii-i (772-846) and his friend Yuan Chen (779-831), popular
sources of couplets for decorating painted screens and singing in the
elegant style called roei, are also missing. Sezoku gembun falls short
of being fully representative of the learned culture of its time in its
omission too of the belletristic works in Chinese of the Japanese
themselves. Those of Sugawara no Michizane (or rather, the most
successful couplets from them) enjoyed great esteem. Japanese
literati in general took a lively interest in one another's works, so that
the author of an outstanding composition (which is to say, one that
18 Described in Kawase, Kojisho, pp. 164-66.
Cambridge Histories Online © Cambridge University Press, 2008
354
CHINESE INTELLECTUAL LIFE
contained a particularly fine couplet) might find that news of it had
traveled even to colleagues serving in the provinces.19
In a world in which a quotation from a Chinese classic constituted
the proof that would clinch any argument and in which the ascertainment of precedent was the first concern in matters ranging from
literary judgment to government edict, a collection such as that just
described fulfilled very practical needs. Serious scholars, of course,
would also study the texts in their entirety. One of the emperor's prescribed activities was to hear lectures on the Chinese classics, histories, and poetic anthologies; the occasions of such lectures, given by
the most eminent literati, are recorded in the National Histories.
Lectures and debates were part of the sekiten, the ceremony to Confucius performed twice annually at the Academy, in the second and
eighth months. Lectures were also held in the regent's mansion; their
occurrence often indicates genuine interest in the work itself, as well
as affirmation through ceremony, of the patron's status. On six
known occasions before 965, the court sponsored series of lectures
on the Nihon shoki, although - as Tamenori's selection suggests - no
such attention was given the subsequent histories of Japan. Among
popular texts for lectures were books that combined moral instruction with pragmatic advice on the conduct of life. Shuo yuan (late
first century B.C.E.), a collection of anecdotes teaching the behavior
proper to persons in various stations in life, and Yen-shih chia-hsuu
(late sixth century), written to provide guidance to its author's sons
and grandsons on personal decorum and conduct within the family,
were much studied. Both works are quoted in Sezoku gembun. Especially valued by the Japanese rulers was Chen-kuan cheng-yao, a collection of the conversations of the second emperor of theT'ang with
his ministers about the art of government. This text too was frequently lectured upon. Heian Japanese themselves, it might be
added, also wrote advice to their successors and descendants,
among the earliest and best-known examples of works of this genre
being Kampyo goyuikai, composed in 879 by Emperor Uda upon
ceding the throne, and Kujo Ushojo yuikai, by the Minister of the
Right Fujiwara no Morosuke (908-60);20 both of these works, which
19 Oe no Masafusa, "Bonen no ki," collected in Yamagishi Tokuhei, Takeuchi Rizo, Ienaga
Saburo, Osone Shosuke, eds., Kodai seiji shakai shiso, vol. 8 of Nihon shiso taikei (Tokyo:
Iwanami shoten, 1979)5 pp. 161-64. My translation of this work appears in "Oe no Masafusa and the Practice of Heian Autobiography," Monumenta Nipponica 51 (1996): 143-51.
20 In Kodai seiji shakai shiso, pp. 103-14, 115-22. For the former, see Borgen, Sugawam no Michizane, pp. 213-15; translation of the latter is in Inge-Lore Kluge, "Fujiwara Morosuke und
seine 'Hinterlassene Lehre'," Mitteilungen des Institots fur Orientforschung 1 (1953): 178-87.
Cambridge Histories Online © Cambridge University Press, 2008
IDEAL OF THE SAGE-KING
355
served as models for later ones, are of extremely practical bent, the
emperor commenting on the talents of certain individuals at court,
and the Minister of the Right even on such specific matters as the
lucky days for cutting finger- and toenails.
IDEAL OF THE SAGE-KING
At the center of Chinese learning and Confucian teaching in the
Heian scheme was an idea of the Chinese sage-king, superimposed
on the original sacerdotal and tribal role of the Japanese tenno . The
sage-king should be puissant but, as the ideal imagined him, able
to rule through moral force alone. The emperor could give evidence of his Confucian sagehood by meticulous attention to court
ceremonies and by selecting and rewarding good men. He should
request men of outstanding talent and rectitude to advise and admonish him. Whether or not the advice of these ministers had substantive content, the process of requesting and receiving the
memorials remained of value as ritual. The degree to which government was seen to inhere in observance of ritual proprieties cannot be overstated. They were simultaneously an essential element
in the processes of the state and the essential emblem of all civil
order.
According to the ideal, the emperor should conduct himself at all
times in the knowledge that Confucian histories, written by impartial observers, would hold him accountable for his actions, and he
should command the compilation of such histories of the reigns of
his predecessors as mirrors for future generations. He should honor
the aged and virtuous throughout the kingdom and exemplify filial
piety in his reverence toward his own parents. He should pay due attention to omens and portents reported to him and alleviate the lot
of the poor in time of drought or disaster by remission of their taxes,
himself setting examples of frugality and admonishing his courtiers
against luxury. He should encourage Confucian education and show
honor to learned men: if aged as well as learned, so much the better. He should promote Chinese literature. This last was one of the
most conspicuous activities of the sinified Japanese courts of the
first century and a half of the Heian period. Adult emperors and exemperors, whether in possession of some power to rule as well as
reign or, like Uda (r. 887-97) a ft er his abdication, largely impotent
in the political arena, presided over banquets at which high- and
middle-ranking courtiers assembled with professors from the AcadCambridge Histories Online © Cambridge University Press, 2008
356
CHINESE INTELLECTUAL LIFE
emy, who generally held more modest ranks, to compose verse in
Chinese on conventional subjects. The emperors themselves were
usually practicing poets, and the banquets were conducted in an atmosphere of great elegance.
Whether a given Japanese emperor was able to exercise the actual
powers of a ruler was largely beyond his control. The monarch
whose powers to govern were most like those of his Chinese counterparts was Kammu (r. 781-806). His court was hospitable to
literati, and he himself, before his accession, had held the post of
Head of the Academy. But he was too much occupied with practical
affairs to cultivate the literary graces himself, and the cost of his removal of the capital from Nara, first to Nagaoka and later to Heiankyd, was regarded by Confucians as extravagant. It was one of his
sons, Saga (r. 809-23), who most assiduously acted out the myth of
the Confucian sage-king guiding a Confucian state. From the time
of his abdication until his death in 842, Saga continued to dominate
the affairs of the court, and the tone that he set persisted through
the reigns of his successors, his brother Junna (r. 823-33) and his
own son Nimmyo (r. 833-50). Kdnin, the era name (nengo) of Saga's
reign often made the designation for the reigns of all three, is viewed
as a distinctive period in Japanese cultural history. It has been suggested that what finally ended the enthusiasms of the Konin court
was not intentional change of policy but rather the ill health of
Nimmyo's successor, Montoku (r. 850-58).
Saga and his ministers attempted practical measures to shore up
the statutory regime, ordering the reallotment of fields and, with
more success, the compilation of laws enacted since the promulgation of the original codes (the Konin kyaku and Konin shiki, discussed in a later section). He revived annual ceremonies allowed to
lapse by his predecessor Heizei and inaugurated others. He commanded the compilation of a manual of court ceremonial, the Dairi
shiki.21 He was a determined sinifier of the manners and customs of
his people. In 819 he issued a decree directing that the ceremonies
of the realm accord with those of T'ang, that men and women wear
T'ang costume, and that diplomas of rank for men of the Fifth Rank
and above follow T'ang usage. He decreed also that palace buildings,
cloisters, and gates were to bear tablets displaying names in Chinese
style. To these measures he was urged by a favorite Confucian, Suga21 Translated in Michael Charlier, Das Dairi-shiki: Eine Studie zu seiner Entstehung undWirkung
(Wiesbaden: Otto Harrassowitz, 1975).
Cambridge Histories Online © Cambridge University Press, 2008
IDEAL OF THE SAGE-KING
357
wara no Kiyokimi (770-842), a veteran of the embassy of 804 and
grandfather of the celebrated Michizane.
Under Saga, the Academy flourished as never before. As its history
will show, however, the age was fascinated more by Chinese belles lettres than by the classics. Frugality, one of the attributes of the ideal
sage-king, was neglected by Saga as it was by actual Chinese monarchs; personal rule was combined with a fondness for elegant amusement. The Shinsen'en ("Park of the Divine Spring"), a then spacious
garden immediately south of the Greater Imperial Palace enclosure,
was the site of many of his poetry banquets. Others were held at the
Reizeiin, where Saga inaugurated a new style in luxurious living for
retired emperors and installed his library. Characteristically, literature - which is to say, poetry - in Japanese was eclipsed by Chinese in
his court. Two of three anthologies of literature in Chinese compiled
by imperial command (chokusenshu) were produced in his reign:
Ryounshu (814), containing 90 poems, was compiled chiefly by Ono
no Minemori (778-830), himself a classic exemplar of the Confucian
"virtuous official" as well as a man of literary ability; Bunka shureishu
(818), with 140 poems, was compiled by a committee headed by Fujiwara no Fuyutsugu (775-826), the leading Fujiwara minister. The
third, Keikokushu (827), produced under Junna, included prose pieces
as well as verse; those heading the committee in charge included the
future emperor Nimmyo and one of Saga's half brothers, Yoshimine
noYasuyo. Poems by Sage himself, or written by his courtiers to "harmonize" with his poems, are dominant in all three. His sons, whose
education he personally supervised, and at least one of his daughters,
shared his enthusiasm for writing Chinese poetry. It is significant that
those closest to him in the work of governing - whether aristocrats or
bureaucrats - were members of his poetic circle.
Saga and his adherents took their cue from a phrase in the critical
treatise Lun wen by the Wei emperor Ts'ao P'i (187-226): "Literary
composition is a vital force in governing the state." This maxim they
interpreted in a more literal sense than could ever have been intended
by its author. It is quoted in the preface to Ryounshu and supplies the
title of Keikokushu, literally, "the anthology for governing the state."
In fact, the interests addressed in Saga's own poetry were increasingly
aesthetic rather than related to government. Some orthodox Confucians did protest the expenditure involved in the poetic gatherings
and revived ceremonies, but their voices were ignored.22
22 Goto Akio, Heiancho kambungaku ronko (Tokyo: Ofusha, 1981), pp. 7—53.
Cambridge Histories Online © Cambridge University Press, 2008
358
CHINESE INTELLECTUAL LIFE
It was not the time of Saga that was looked back upon in later
years as one of sage-reign but those of Daigo (r. 897-930) and Murakami (r. 946-67); together, their era names, Engi (901-23) and
Tenryaku (947-57), came to evoke visions of a golden age. But the
reputation of these two emperors is based less on accomplishment
than on the fact that each ruled without the aid of a Fujiwara regent.
This freedom from Fujiwara control was little more than nominal,
however, as Daigo dashed his father Uda's hopes for maintaining
some independence when Fujiwara leaders persuaded him to exile
Michizane, Uda's protege and the Fujiwara's presumed rival at
court. Daigo exerted his power as Confucian monarch with energy
in those spheres that remained open to it, while his ministers for
the final time attempted to revive the allotment system. The last of
the Confucian histories of Japan to be completed and the last, and
definitive, collection of administrative statutes, the Engi shiki, were
compiled under him. But - again, it is thought, in reaction to their
actual powerlessness - Daigo and Uda (at least during his long years
in retirement) devoted the major portion of their energy to the refined pleasures. In contrast to the Konin era, however, Engi saw a revival of native Japanese styles in the arts. Anthologies of poetry in
Chinese would continue to be compiled, but no longer by imperial
order; in their place were anthologies of poetry in Japanese, beginning with the classic Kokinshu in 905.
The reign of Murakami (946-67) looked back to Engi. Murakami
greatly admired his father, Daigo, and, at a time when central and
imperial authority had suffered even greater erosion, aspired to reign
in his style. In company with his elder brother the ex-emperor
Suzaku (r. 930-46), he too engaged in a full program of elegant pleasure. He revived the Chinese poetry banquets, which had been less
popular than Japanese ones under Daigo and more recently had
been allowed to lapse in favor of cockfights, and he inaugurated contests in Chinese poetry to match those in Japanese begun in the previous century. Such meetings were conducted in luxurious style; on
the occasion of the first Chinese poetry contest, for example, the
verses composed by both sides were written out by the eminent calligrapher Ono no Michikaze (Tofu) (894-966). But whereas the Chinese poets who gathered around Saga had been his close advisers
and those around Daigo included men whose opinions on practical
matters of government might still be solicited by the emperor, the
academicians whom Murakami summoned were elderly survivors of
Cambridge Histories Online © Cambridge University Press, 2008
SIX NATIONAL HISTORIES
359
his father's time, none of whom occupied a position of importance
in his government.
The men of Murakami's time felt tradtional rituals to be in danger of slipping out of use, and thus his court gloried in the production of manuals of ceremonies and court procedures. A handbook of
annual ceremonies was written by the Fujiwara regent Morosuke,
and Morosuke's successor Saneyori (900-70) was looked upon by
later times as the founder of a school of yusoku kojitsu, the study of
customs and precedents. But in the actual work of governing Murakami took little interest. His successor was mentally unbalanced
and was followed in turn by a child and then by a youth who reigned
only two years. Confucian teachings continued to be honored at
court; there were still learned men. The age of the sage-king, however, was over; it could not, in any event, have long survived the
changes in the statutory system and the weakening of its Confuciantrained bureaucracy.
SIX NATIONAL HISTORIES
The sage-king myth was cultivated in part through the writing of
histories in the Chinese manner, or at least in what Japanese had
come to regard as the Chinese manner. The Japanese already had
completed two such chronicles - Nihon shoki (720) and Shoku Nihongi (797) - and in the early Heian period added four more to
make up what came to be known as the "Six National Histories."
Prepared over a period of a little more than eighty years were Mhon koki, commanded by Saga in 819, presented in 840; Shoku
Nihon koki, commanded by Montoku in 855, presented in 869;
Nihon Montoku jitsuroku, commanded by Seiwa in 871 and presented in 879; and Nihon Sandai jitsuroku, commanded by Uda in
892 and presented in 901. The sage-king was not to rule alone, but
with the guidance and support of wise ministers, whom he rewarded; the writing of history was an enterprise involving both aristocrats and professional literati, requiring energy and expertise, and
serving the self-esteem of the compilers as well as those who commissioned it. According to a tenth-century protocol, the committee
for compilation of a National History should consist of the highest
minister of state, a Consultant to take practical charge, one of the
two Major Secretaries (daigeki) in the Council of State, and four or
five learned men selected from among the officials of the various
Cambridge Histories Online © Cambridge University Press, 2008
360
CHINESE INTELLECTUAL LIFE
bureaus.23 Thus, the original head of the Nihon koki committee was
Fujiwara no Fuyutsugu, that for Shoku Nihon koki Fujiwara no
Yoshifusa, and for Montoku jitsuroku, Fujiwara no Mototsune. Fuyutsugu, with some justification, enjoyed a reputation for Chinese
learning, and Yoshifusa seems to have participated actively in the
work of his committee. Since the committees also included members
too old or too young to be major contributors, it may be assumed
that they too were named to lend - or be lent - prestige.
Chinese histories were intended to promote wise government by
functioning as impartial mirrors of the successes, failures, virtues,
and vices of the successive emperors of a dynasty. If the activity of
writing orthodox history was based on the Chinese model, its products deviated from it in some significant ways. In form, Chinese dynastic histories contain tables, biographies, and treatises of various
kinds, as well as "basic annals" (pen-chi), whereas the extant versions
of the Japanese National Histories consist of annals only (Nihon shoki
originally included an imperial genealogy, now lost). Educated
Japanese were well acquainted with Chinese historical writing; in the
Academy, the study of the histories (kidendo) was a popular subject
that came to be amalgamated with the older literature curriculum.
The choice of a chronological arrangement was not made either
from ignorance or necessarily merely from a preference for simplicity. There are precedents in Chinese usage for such an arrangement,
most notably in the Spring and Autumn Annals. The Chinese genre
that the Japanese histories most resemble, however, is the shih-lu (jitsuroku, "veritable records"), as the titles of the two last examples acknowledge. Shih-lu were compiled after the death of an emperor on
the basis of diaries kept by his officials and constituted the stage in
the preparation of orthodox history immediately preceding the final
one of the dynastic history itself. They departed from strict chronological form only in order to include biographies of notable subjects
following the announcement of their deaths, a practice also followed
by the Japanese starting with Shoku Nihongi. Essentially full drafts,
the shih-lu awaited the demise of the dynasty to be transformed
through supplementation and revision into dynastic histories. Other
considerations aside, it was appropriate for the Japanese, already
proud of their dynastic continuity, to stop at the shih-lu level, as fuller
histories might have been taken to imply a break in the imperial line.
23 Sakamoto, Taro, The Six National Histories of Japan, trans. John S. Brownlee (Vancouver:
University of British Columbia Press, 1991), p. 97, quoting Shin gishiki.
Cambridge Histories Online © Cambridge University Press, 2008
SIX NATIONAL HISTORIES
361
It has been claimed that the Japanese had little understanding of
the monitory function that history was supposed to serve. Some of
the later National Histories do in fact venture criticism of such safe
imperial targets as the scapegrace Heizei and mention in their prefaces that praise and blame are to be impartially bestowed. Nevertheless, the purpose of the Japanese histories lay elsewhere: celebration of the imperial house, glorification of its ministers, and
promotion of the image of Japan as a Confucian state. To the last of
these ends they present examples of the practice by humble persons
of Confucian virtue on a heroic scale, inventing them, one suspects,
if necessary. Montoku jitsuroku, for example, records the award of the
lowest court rank to a woman who had spent thirty years in mourning her deceased husband; the same reward was given as well to a
woman who had mourned her husband, raised his children (presumably, though it is not stated, by another wife), and performed
works of Buddhist piety. Such exemplary instances, described in admiring detail, are included to suggest that they had been the result
of imperial virtue. Auspicious events too might be revered as the
consequence of beneficent rule. Upon discovery of a sweet spring in
Iwami province, the priests of the emperor's ancestral shrine at Ise,
as well as the owner of the property where the spring was found and
all officials of the district, were advanced in rank; and gifts of grain
were made to the aged (those over a hundred years of age got three
koku each; those over ninety, two; those over eighty, one) and the era
name changed.24
The compilers of the histories could draw on the archives of the
Bureau of Drawings and Books (Zushoryo) and the Ministry of Ceremonial (Shikibusho), which collected biographies of meritorious
subjects. The deaths of persons of the Fourth Rank and above were
supposed to be recorded, and biographies appended if the deceased
was of Third Rank or higher. Occasionally biographies also appear
even though the deceased was of lower rank, if only to explain why
the death of a person of humble station should be worth recording.
The largest proportion of such biographies are found in the last two
histories, which tend toward the anecdotal. Not all biographies are
hagiographical: Nihon koki, in particular, is distinguished by the acuity and occasionally the asperity of its biographies.
It is often possible to trace the influence of individual compilers.
24 Osamu Shimizu, "Nihon Montoku Tenno jitsuroku: An Annotated Translation, with a Survey of the Early Ninth Century in Japan," Ph.D. diss., Columbia University, 1951, Saiko 1
(854X3/9, Saiko 1/5/26, and Saiko 1/11/30.
Cambridge Histories Online © Cambridge University Press, 2008
362
CHINESE INTELLECTUAL LIFE
Haruzumi no Yoshitada (797-870) was the most renowned Confucian of his day, tutor to Crown Prince Tsunesada, afterward professor {hakase) at the Academy, ultimately Consultant. He owed his
success to the patronage of Yoshifusa, whose chief collaborator he
was on Shoku Nihon koki. His hand is seen not least in the frequency
with which Yoshifusa's name appears there, even where it need not.
Shoku Nihon koki, like Montoku jitsuroku after it, covers only one
reign, so that it comes even closer in form and spirit to the shih-lu
than do its predecessors. Nimmyo, its subject, is depicted as the perfect Confucian monarch, with special mention made of his filial respect toward his mother, while imperial children are praised for a
preternatural mastery of ceremonial deportment. This emphasis on
Confucian etiquette perhaps is also due to Yoshitada, as is the inclusion of many anecdotes of the supernatural, in which he was said to
take a keen interest. Miyako noYoshika (834-79), w n o was probably
the chief contributor to Montoku jitsuroku, is thought responsible for
including in that work an exceptionally large number of biographies
of members of the middle aristocracy; he himself achieved only Junior Fifth Rank Lower Grade.
When Uda commanded the compilation of Nihon sandai jitsuroku,
he appointed a committee that included his Ministers of the Right
and Left, Michizane and Fujiwara noTokihira, and in the spirit of opposition to Fujiwara power he made head of the committee Minamoto noYoshiari (845-97), a s o n of Montoku and the highest-ranking
Minamoto who was not too old for the task. When Daigo came to the
throne a few months after Yoshiari's death, no new members were
added. Michizane (also the author of the preface to Montoku jitsuroku) was exiled early in 901, and the work was not completed until
mid-autumn. It was Okura no Yoshiyuki whom the late-thirteenthcentury bibliography Honcho shojaku mokuroku lists as the chief author. The aged Yoshiyuki was a particular favorite ofTokihira in the
latter's role as Confucian patron. (Tokihira's chief achievement in
Chinese belles lettres is represented by Suisekitei shikan, the record of
a poetry party held, also in 901, to celebrate Yoshiyuki's seventieth
year.) With the aid of a disciple, Mimune no Masahira, who was
twenty-eight years younger, Yoshiyuki devoted a good part of his energies not only to his patron's but to his own glorification. Although
not even of the Fifth Rank, he caused his own name to appear in the
text a number of times. The committee itself, the processes by which
it was constituted and reconstituted, exemplified the relations between literatus and noble, of which more will be said.
Cambridge Histories Online © Cambridge University Press, 2008
SIX NATIONAL HISTORIES
363
Sandai jitsuroku is the most detailed of all the histories; the goal of
the compilers, as stated in its preface, was comprehensiveness. Like
its two predecessors it focuses narrowly on the court, recording in
detail all imperial rites of passage and also Tokihira's gempuku (coming-of-age) ceremony. All of the histories chronicle special rites and
festivals, but in addition, Sandai jitsuroku also records all of the ordinary seasonal festivals. Texts of decrees and memorials are reproduced in full. The compilers take great interest in Shinto shrines; the
new awareness of Japanese roots that in poetry produced the anthology Kokinshu is manifested here. Portents are recorded in detail - again, one may perceive a native fondness for anecdote. On the
thirtieth day of the fifth month of the fourteenth year of Jogan (872),
a great serpent appeared in one of the halls of the official provincial
Buddhist temple in Suruga. There were thirty-one copies of the
Heart Sutra wrapped around a single roller, and it ate them. The
monks who witnessed this tied a rope around its tail and hung it upside down from a tree. Shortly afterward, it disgorged the sacred
books and fell on the ground half dead, but then suddenly revived.
Again, on the twenty-ninth day of the seventh month of the second
year of Ninna (886), at the hour of the boar (roughly 9-11 P.M.), a
giant was seen strolling back and forth within the imperial enclosure, in front of the Shishinden. A page boy posting summonses saw
him and fainted from fright, and he was also glimpsed by a man who
was lighting torches in front of the station of the Right Bodyguards.
Subsequently, in the vicinity of the station of the Left Bodyguards a
cry was heard, as of a man strangling. People called the apparition
the "strangling ghost."25
Sandai jitsuroku was the last of the National Histories to be completed and the last to be preserved, but one more was attempted. In
936, the sixth year of his reign, Daigo's successor Suzaku appointed
officers to the History Compilation Bureau (Senkokushisho), which
had been set up in 880 to take charge of the daily records on which
histories were to be based. The first pair of Superintendents were a
Fujiwara Major Counselor and a Middle Counselor, with two men
of the scholarly Oe family their nominal assistants. Notices of appointments to the bureau appear in various sources over the next
thirty years, from the reign of Suzaku through Murakami's and into
Reizei's, but what exactly was produced is unclear. Possibly there
was some sort of draft, but it cannot have been more than that, and
25 Sakamoto, Six National Histories, pp. 178-79.
Cambridge Histories Online © Cambridge University Press, 2008
364
CHINESE INTELLECTUAL LIFE
only a few putative fragments remain. One specific reason for the
project's lack of success may have been the age and isolation of the
directors, so characteristic of the literati of Murakami's court. Oe no
Asatsuna was sixty-nine when finally made Superintendent in 954,
his brother Koretoki even older when he in turn became Superintendent in 957, and they lacked competent helpers.
The cessation of the National Histories was one aspect of the
general weakening of all the institutions of the statutory system. Education of the literary technicians who staffed the bureaucracy became formalized and impoverished, as will be seen, and standards
of written Chinese declined, except among a devoted few. The bureaucracy itself was becoming an empty shell. There was little financial support for writing history, and the bureaucratic archives on
which historians depended for their material were no longer maintained. The record-keeping function of the Senkokushisho was inherited by the Secretariat (Geki no cho) within the Council of State,
and for a time geki nikki, secretaries' diaries, were thought to furnish possibilities for the revival of national histories. In 1010, for instance, the Geki no cho was ordered to search precedents and report
to the throne, but the report must not have been encouraging.26
COMPILATION OF STATUTES
If the Six National Histories helped enact as well as record the fiction of a harmonious Confucian state, the compiling of official
statutes may well be the one substantive achievement of that state.
Initially, legal scholarship had consisted of compiling whole new
codes, culminating in the Yoro ritsuryo, drafted in 718 but not promulgated until 757. Attention then turned to explication and two
important commentaries were compiled early in the Heian period.
The first, Ryo no gige, written by an officially appointed committee
of twelve, was completed in 833 and authorized the following year;
the second, Ryo no shuge, was the private work of a single legal
scholar, Koremune no Naomoto, who completed it during the Jogan
era (859-77) • I* ls only in these commentaries that the texts of the
original codes are preserved. The codes themselves, however, were
not the only basis of early Japanese law. Over the years, many new
regulations were issued either modifying the codes or detailing how
they should be enforced. These were promulgated as various forms
26 Sakamoto, Histories, p. 191.
Cambridge Histories Online © Cambridge University Press, 2008
COMPILATION OF STATUTES
365
of imperial edicts or proclamations by the Council of State, and together they came to be known as kyaku and shiki. Although the precise distinction is unclear, the kyaku usually concerned modifications of the original codes and shiki, rules for their enforcement.
These statutes were the means by which Japanese court government
refined the codes, worked out details for their implementation, and
responded to changing circumstances. In the Heian, focus gradually
turned to the kyaku and shiki, which were collected and classified in
the Konin, Jogan, and Engi eras - the reigns of Sage, Seiwa, and
Daigo. Significantly, although in each case the compilation committee included legal specialists, many of its members were also members of the committees charged with preparation of the National
Histories; for example, of the eight men who received the imperial
command to revise and expand the Konin kyaku and shiki in the
Jogan era, three were eminent literati who were also compilers of
Montoku jitsuroku. Heian literati were generalists.
The Engi shiki is the major monument of Heian law as well as the
most notable achievement of the reign of Emperor Daigo - an expression, moreover, of the belief of his Fujiwara ministers that they
were repairing rather than undermining the foundations of the statutory system. Daigo's intention was that Japan be provided with a
complete code of ritsu, ryd, kyaku, and shiki. The command to compile kyaku and shiki was received in 905 by Tokihira. Among the
eleven other members of the original committee were three other
nobles, including Ki no Haseo (845-912), Michizane's friend and
follower who had managed to remain at peace with Michizane's enemies and been made Consultant in 902, and a certain elderly Taira
no Korenori, whom history records also as recipient of a command
from Emperor Uda to compile an anthology of Chinese verse composed since the K5nin era. Below them were eight literati who are
surmised to have done the actual work. Among them were the venerable Okura no Yoshiyuki and Miyoshi Kiyoyuki (847-918).2? The
lowest ranking member was Koremune no Yoshitsune, a former Professor of Law (jnyobo hakase).All of the literati were loyal to Tokihira,
and most were Yoshiyuki's disciples.
The kyaku, in twelve volumes, were completed first, in 907, and
authorized in 908. The compilation of the shiki, however, dragged on
through Daigo's reign and was not finished until 927, after the
27 Torao Toshiya, Engi shiki (Tokyo: Yoshikawa kobunkan, 1964), p. 58; see alsoTokoro Isao,
Miyoshi Kiyoyuki (Yoshikawa kobunkan, 1970), pp. 111-12.
Cambridge Histories Online © Cambridge University Press, 2008
366
CHINESE INTELLECTUAL LIFE
deaths of several members of the original committee, including its
head. One reason for the delay was surely the dimensions of the project, since the compilers incorporated the two previous collections of
shiki into their unified system, along with the great number of individual ordinances that had been issued since the Jogan period. The
resulting work, in fifty kan, consists of 3,300 individual articles. But
despite its apparent comprehensiveness, the code does not contain
shiki for such newly created government organs as the Chamberlains' Office (kurododokoro) and Imperial Police (kebiishi). The criterion for inclusion was not whether the office had existed in the original bureaucratic scheme of the statutory system; rather, it seems to
have been whether shiki for the office in question had appeared in
the two previous collections. The Engi shiki was therefore a summation of existing and well-established administrative theory and practice, not in any way an attempt to open up new ground, the final
stage in the systematization of past forms.
The Engi shiki was not authorized until 967, forty years after it was
first presented to the throne. Among the reasons for this added delay
was the lack of any real need to promulgate it: it merely systematized
edicts already in force. Another was that, increasingly, the governmental structure it was supposed to regulate functioned in name
only. Yet another, it has been suggested, was that as presented to
Daigo it was still incomplete and in need of revision. At least three
drafts were made, perhaps four. There is much evidence that Daigo
himself was consulted in regard to the second draft. Murakami, too,
was personally involved in revision. His aspiration to imitate Daigo's
style of rule has already been noted, and most likely it was his interest that caused work on the Engi shiki to be resumed. Even though
the decree enacting the code was formally issued by his successor,
Murakami had set the process in motion. As it had been for Daigo,
the Engi shiki was Murakami's one genuine claim to the reputation
of Confucian sage-king.
What parts of the Engi shiki were actually put into practice?
Clearly, regulations for the conduct of offices that no longer had real
functions to perform were of symbolic rather than practical importance. But its prescriptions continued to be followed in the activities
of the court itself and in a variety of religious observances and festivals. It was consulted as authoritative by the authors of manuals prescribing annual events in the court ritual calendar. Preserving correct ceremonial observance was to be a major concern of mid- and
late-Heian literati.
Cambridge Histories Online © Cambridge University Press, 2008
STATE ACADEMY
367
STATE ACADEMY
The institution that shaped the men who staffed the statutory system's bureaucracy, wrote the sage-kings' histories, and compiled
their laws was the Academy. Properly called the Daigaku-ryo (literally, "Bureau of the Great Learning"), the Academy was attached to
the Ministry of Ceremonial, the branch of government in charge of
evaluating candidates for office. It occupied spacious grounds immediately to the south of the imperial palace and west of the Park of
the Divine Spring. Like the civil service examination for which it
prepared its students, it was a simplified copy of a Chinese model.
T'ang China maintained six schools in its capital, all preparing students to take the civil service examination; the Daigaku, by contrast,
stood alone in the Heian capital and was much smaller.
The history of the Heian Academy is instructive for a number of
reasons. The changes within it illustrate the further Japanizing of a
Chinese ideal, first in its intellectual content, then in its institutional
and social forms. From one point of view, the commonsensical one,
it must be judged to have gone into terminal decline by the beginning of the tenth century; from another, it was being transformed,
like Chuang tzu's dying man, into something "all crookedy," assuming a new and in its own way equally valid shape in the course of disintegration. The early tenth century may be seen as a kind of watershed in the function, activities, and morale of the professional men
of learning. A description of the Academy is an essential preliminary
to that of its graduates before and after, and their place in society.28
The Academy at the beginning of the Heian period had already
undergone more than a century of development; it is thought to have
originated in the first official Confucian school, founded at Otsu in
the time of Emperor Tenji (r. 662-71). Nothing is known of its structure at this early time, but under the eighth-century civil code it had
an administrative staff consisting of a Head (kami), an Assistant
(suke), two secretaries (jo), and a small complement of clerks and
watchmen; a little later it also acquired a Superintendent (Jbetto) to
supervise practical affairs. The faculty at the start of the eighth century consisted of a Professor (hakase) with two Assistant Professors
(suke hakase) who together provided instruction in the Chinese clas28 Standard sources for the Heian academic world include Momo Hiroyuki, Jbdai gakusei no
kenkyu (Tokyo: Meguro shoten, 1947), and Hisaki Yukio, Nikon kodai gakko no kenkyu
(Tokyo: Tamagawa Daigaku shuppambu, 1990). See Borgen, Sugawara no Michizane, pp.
69112, 12440.
Cambridge Histories Online © Cambridge University Press, 2008
368
CHINESE INTELLECTUAL LIFE
sics. In addition, there were two Professors of Chinese Pronunciation (era hakase), two Professors of Calligraphy (fumi no hakase), and
also a Professor of Mathematics {san no hakase), whose status was
slightly below that of the others and who was in charge of a separate,
less highly regarded curriculum. In addition, positions for two Professors of Law (myobo hakase), one Professor of Literature (rnonjo
hakase), and three lecturers (chokko) in classics were created in 728.
Study of the Chinese classics (later called mydgyodo) was the central
curriculum, and its Professor stood above the others. His rank, however, was only Senior Sixth Lower - below the line of aristocratic
privilege. His assistants and, after them, the other Professors were
ranged in various degrees of the Seventh Rank. Like other lowranking functionaries, these teachers received their income from the
seasonal stipends appropriate to their rank, and they also received
fees from their students. The statutes set the number of students at
four hundred, with an additional thirty in the mathematics course,
and they too were to receive stipends. From 757 on, the revenues of
certain lands, known as kangakuden ("fields for the promotion of
learning"), were set aside for support of the Academy; in the early
Heian their number was increased, but by the tenth century almost
all had been reappropriated.
If this was the matrix provided by the codes, the school as it
emerged in the Heian period, briefly flourished, and then declined
into a ceremonial, hereditary institution of peculiarly Japanese aspect that was in actuality much different. The ninth century saw the
rise of the curriculum in letters (monjddo) to the highest status and
prosperity. The subject of this curriculum was the art of elegant literary composition, its textbooks Wen hsuan and three historiesthose of the Records of the Historian and of the Former and Later
Han. Originally, it had been of distinctly inferior status, its Professor
holding the same modest rank as the Assistant Professor of Classics.
In 730, when places were created at the Academy for students of literature (and law), they were to be selected from among men holding the lowest, nonaristocratic, positions in the government or true
commoners. By the early Heian, this had all changed, and literature
came to be the most esteemed subject at the Academy. The prestige
of this curriculum corresponded not only to Saga's convictions but
to the demands of a mode of government that required preparation
of memorials and edicts in a refined style. So elegant indeed was the
parallel prose of memorials that many were prized for their beauty
and even included in literary anthologies, quite apart from any relaCambridge Histories Online © Cambridge University Press, 2008
STATE ACADEMY
369
tionship to their original context. In 813 Saga formally placed the
curriculum in letters at the head of the others. Its twenty students
constituted an elite group who were trained for the most challenging of the civil service examinations. A decree issued by Saga in 820
gave its Professor - then Sugawara no Kiyokimi - Fifth Rank Junior
Grade, thus making him a member of the middle aristocracy. When
the separate curriculum in the histories, created in 808, was abolished in 834, a second professorship of letters was created. Only a
monjo hakase would be given the post of Head of the Academy, an
appointment of great honor.
Whereas China had the ideal of universal education, however imperfectly realized, in Japan not until the Edo period did Confucian
education become available to members of diverse classes of society.
Provincial academies (kokugaku), greatly simplified analogues to the
academies in the Chinese provinces, were intended to educate the
sons of district officials, but they foundered early, with the exception
of the school in Dazaifu. Admission to the Academy in the capital
was originally based on rank. The earliest rule guaranteed places for
sons and grandsons of men of the Fifth Rank and above and accepted the sons of men of the Six through Eighth on special petition.
Another group eligible was the sons and grandsons of government
scribes descended from Korean immigrants, although by the early
Heian they no longer appear among the extant scattered references
to students of the Academy. Although students of both literature and
law originally were supposed to be of more humble background, in
practice, commoners were few at the Academy. One family, the
Nakahara, which came to share dominance in the classics curriculum with the Kiyohara, is thought to have been of common origin
despite its claim of descent from an ancient and surely legendary
emperor.29 It, however, was the exception.
A decree by Emperor Heizei set the lower age for enrollment, formerly thirteen, at ten; in 824 the upper, formerly sixteen, was set at
twenty. The celebrated monk Kukai was in his eighteenth year when
he entered, and Miyoshi Kiyoyuki was most likely in his seventeenth.
There was no division of students by age or grade. As in China, a
student might remain as long as nine years; after that he would be
expelled if he had failed to pass the Ministry's examination. Within
the Academy, students were examined annually and also every ten
days, those who did poorly being expelled. The Professors lectured
29 Hisaki, Nihon kodai gakko, pp. 290-92.
Cambridge Histories Online © Cambridge University Press, 2008
370
CHINESE INTELLECTUAL LIFE
on a text from beginning to end - lecturing, it may be surmised, consisting of reading the text aloud and explicating the words and
phrases. The Engi shiki stipulates the number of days allotted to the
exegesis of each book: the Li chi and Tso chuan at 770 days each, Shih
ching at 480 days, and so on.3°There was a holiday following each examination day, and there were two vacations, a month each, during
the year. Students were enjoined to be serious and forbidden such
amusements as archery and playing the koto (zither).
As literary studies became increasingly popular in the ninth century, an elaborate system for screening students evolved. In its final
form, a student who wished to enter the curriculum in literature
would first have to study the classics. If judged outstanding, he would
be allowed to take an examination administered by the Daigaku-ryo,
the ryoshi (bureau examination), and if he passed, he would become
a Provisional Scholar of Letters (gimonjdsho) .The next step for the aspirant was an examination administered by the Ministry of Ceremonial, the shoshi (ministry examination); by passing this, he became a
Scholar of Letters {monjosho) and bore the honorary title of shinshi,
which, strictly speaking, was supposed to be awarded to those who
had passed the civil service examination modeled on China's chin
shih civil service examination but lacking the prestige of the original
and hence rarely taken. The stages in this progress could be long or
short: Michizane became a monjosho at eighteen, Haruzumi noYoshitada at twenty-eight. Michizane preserved his ministry examination.
It consists of six short (sixteen-character) verses praising recent auspicious portents.
The two best monjosho were selected by recommendation or occasionally by examination to become Distinguished Scholars of Letters (monjo tokugosho) and receive a special stipend; they bore the
title shusai (Chinese, hsiu-ts'ai), which again was supposed to indicate success on a civil service examination, in this case the most
challenging one. Tokugosho were also given nominal provincial offices; and some did, in fact, become provincial officials and cease for
a time to be students before returning for their examinations. There
was, again, no set period of preparation; Kiyoyuki remained a
tokugosho for seven years. The highest examination, the one that had
originally qualified candidates for the title shusai, was the horyakushi,
administered by imperial command and of such difficulty that in the
30 Felicia G. Bock, Classical Learning and Taoist Practices in Early Japan: With a Translation of
Books XVI and XX of the Engi-shiki (Tempe: Arizona State University Center for Asian
Studies, Occasional Paper No. 17, 1985), p. 70.
Cambridge Histories Online © Cambridge University Press, 2008
STATE ACADEMY
371
more than two hundred years from 704 to 931, only sixty-five men
passed it. The candidate was obliged to compose two essays in ornate Chinese parallel prose treating problems in such areas as morality, philosophy, and Chinese history. The answers, called taisaku,
were included in Heian anthologies of Chinese belles lettres, their
outstanding couplets much praised. According to later anecdotes,
any means was thought justified in passing this examination: Miyako
noYoshika, who in his turn was to be Michizane's examiner, was reputed to have obtained the discarded draft of an examination question by seducing a maid of Yoshitada, his examiner; the answer that
he then composed included a passage describing the islands of the
immortals later anthologized in Wakan roeishu, that lovers of the arts
might sing them or use them in the decoration of screens. Although
the anecdote mentions only one passage, Wakan roei shu actually
contains two fromYoshika's examination. Michizane himself refused
to stay shut up in his examination shed, rambled about the Ministry
grounds, and when puzzled for an answer at one point sent a friend
galloping off to a certain mysterious "Recluse Gentleman of Saga"
for advice.31
The rewards of passing the examinations were nevertheless too
low to make the Academy attractive to those who had other means
of attaining advancement. First in 739, the Council of State ordered
all aristocratic youths to study at the Academy. Again, in 806, Heizei
attempted to make enrollment compulsory for all imperial princes as
well as for the children and grandchildren of aristocrats, but the attempt failed and six years later the edict was rescinded on the
grounds that ignorant minds are not easily improved and some had
wasted many years without mastering a single subject; better to leave
academic work to those who were interested. Readers of the Tale of
Genji will remember the young Yugiri's misery at being singled out
from his companions to be made a student. Sons of men holding the
Fifth Rank and above and grandsons of men holding the Third and
above automatically received rank upon reaching their twenty-first
year. By this system of "shadow ranks" (on'i), the heir of a man who
held Third Rank, for example, would receive Junior Sixth Rank
Lower Grade, while a younger son of a man who held Junior Fifth
Rank, at the bottom of the scale of those eligible, would receive Junior Eighth Rank Lower Grade. By comparison, a candidate who
31 Ozawa Masao, Goto Shigeo, Shimazu Tadao, and Higuchi Yoshimaro, Fukuro sdshi
chiishaku, vol. 1 (Tokyo: Hanawa shobo, 1974), pp. 298-303; Kawaguchi and Nara,
Godanshb chit, pp. 1117-23.
Cambridge Histories Online © Cambridge University Press, 2008
372
CHINESE INTELLECTUAL LIFE
was not entitled to shadow rank could obtain no better than Senior
Eighth Rank Upper Grade by passing the highest examination, no
matter how brilliant his performance. And if he possessed shadow
rank, passing the examination would raise his rank only one step.
The gains from the other examinations were even smaller. Those
who aspired to them were youths from modestly placed, ambitious
families to whom success on the examinations offered the possibility of rising in the world or, increasingly as time passed, scions of
families whose hereditary profession was Chinese letters. In the
eighth century one hears of provincial families exhausting their resources to send a promising son to the Academy, but all those who
passed the highest examination after 889 were sons or grandsons of
Confucian officials.
By Michizane's day, the academic profession was gradually becoming the exclusive possession of certain families who guarded
their prerogatives as jealously as the regental Fujiwara did theirs.
Eventually, the monjodo would belong to the Sugawara, the Oe, and
the Hino, Ceremonial, and Southern branches of the Fujiwara comparative newcomers whose entry, achieved with the support of
their influential relations, was resented - while only members of the
Nakahara and Kiyohara might become mydgyo hakase. The Monjoin,
a collegium founded by Oe no Otondo (811-77) and Sugawara no
Kiyokimi to minister to the needs of students in the belles-lettres
curriculum, was divided into an East House and a West House. The
East was dominated by the Oe, Ki, and Takashina; the West by the
Sugawara, Tachibana, and Fujiwara, and the spirit of rivalry between
their adherents was often intense. One of the complaints against
Michizane, and according to some reckonings a major factor in his
ultimate fall, was his relentless promotion of his own disciples,
pupils or former pupils of the Sugawara family private school, Kanke
Roka, to the exclusion of others.
If the first part of the ninth century, and in particular the Konin
era, was a period of prosperity for the Academy, the abandonment of
embassies was followed by a rapid and irreversible decline in its fortunes. An early-tenth-century document complains of favoritism in
examinations, the decrepitude of the buildings, and the skimpiness of
the students' rations. Some students might be able and industrious,
but a majority were not, and many of these stayed on year after year,
into destitute middle age.32To forestall favoritism, examinations were
32 David John Lu, Sources of Japanese History, vol. 1 (New York: McGraw-Hill, 1974)) pp.
Cambridge Histories Online © Cambridge University Press, 2008
STATE ACADEMY
373
supposed to be administered by a professor from the house other
than the candidate's, but as degree-taking became increasingly a matter of hereditary privilege, factional resentments became ever more
conspicuous. In 997 a factional dispute over the grading of an examination provoked a scandal. Soon, however, the examination of candidates from influential families became an almost purely ceremonial
undertaking, a form of high-toned entertainment. In 1090, for instance, ex-emperor Shirakawa and Emperor Horikawa summoned
literati for a gimonjosho "examination" which was in reality a Chinese
poetry competition on the topic "The clothes of the dancers flutter
in the palace garden."33 Within the actual Academy, fathers would
often resign their posts to their sons as soon as the latter reached majority. Since the new Professor would have undergone an examination that was a formality only and had little inclination to memorize
the books on which he was to lecture, it became customary to note
down the readings and interpretations traditionally taught by each
house in copies of the texts passed down from father to son. This was
the origin of o-koto ten, an early system for recording Japanese readings of Chinese texts and a precursor of the marks used in modern
kundoku. To that extent, if no other, kagaku ("the learning of the
houses") ultimately contributed to the spread of education. Physically, too, the Academy fell into ruins. There were frequent fires. After
a fire in 960, the buildings were reconstructed, but in 1135 a petition
bemoans the fact that where lecture halls once stood weeds now
grow. What remained of the halls was destroyed in the great fire of
1177 and no attempt was made to rebuild.
As was the case with other organs of the statutory system, the relevant functions of the Academy were taken over by extrastatutory institutions. Attached to the Academy and gradually eclipsing it in importance were a number of institutions known as besso, established
and supported by the individual clans. The best-known and most
prosperous was the Kangakuin, founded by Fujiwara no Fuyutsugu
in 821. The Kangakuin is often inaccurately described as the private
school of the Fujiwara. Rather than an independent school, however,
it is argued to have been originally a combination dormitory and research institute, providing housing for poor Fujiwara boys while they
attended the Daigaku - in Western terms, a collegium rather than a
college.34 Just as the Daigaku-ryo itself was charged equally with
33 Kawaguchi Hisao, Oe no Masafusa (Tokyo: Yoshikawa kobunkan, 1968), p. 153.
34 Omori Kingoro, "Ocho jidai no shigakko ni tsuite," Rekishi chiri, 53 (1929): 330-37; for an
opposing view, see Takahashi Toshinori, Nikon kyoiku bunka shi, vol. 1 (Tokyo: Kodansha
[i933l. 1978), PP- 97-99Cambridge Histories Online © Cambridge University Press, 2008
374
CHINESE INTELLECTUAL LIFE
conducting the semiannual ceremonies to Confucius and with instructing students, so the Kangakuin also served diverse needs. An
obligation of its students was to go in procession to offer their respects upon the birth of a prince to a Fujiwara imperial consort. The
Kangakuin also served administrative functions for the Fujiwara
family temple, Kofukuji. After the fire of 1177, it was rebuilt; by that
time there were lecture halls in it, proof that it had undergone the
transition to college.
Other important besso were the Gakkan'in, founded in the Jowa
era (834-47) by Saga's Tachibana empress on behalf of her clan; and
the Kobun'in, founded ca. 820 on behalf of the Wake clan by the then
Superintendent of the Academy, the renowned literatus and physician Wake no Hiroyo, and noteworthy for housing the library of a
thousand volumes that had belonged to the founder's father, Kiyomaro. In the Monjoin, the East and West houses came to function as
besso for the Oe and Sugawara clans respectively. The Shogakuin,
founded in 881 by Ariwara no Yukihira for the education of princes
and located in the southern part of the Academy grounds, had an interesting fate: it declined as an educational institution during the late
Heian, but its superintendency became hereditary in a branch of the
Minamoto clan and, ultimately, in the house of the Tokugawa shogun, persisting as a title until the Meiji Restoration.
In the late Heian, the main locus of education changed from these
besso to the private academies of individual literati. There is evidence
that private schools had existed as early as the seventh century, and
it can easily be surmised that at all times they played an important
role. For example, the Sugawara school at its peak under Michizane
is said to have had several hundred students and of them, Michizane
boasted, nearly a hundred had gained admission to the Academy.
The structure of such schools must have been extremely simple,
consisting in the main of a teacher - typically a Professor at the
Academy - who lectured, aided perhaps by a chief disciple who was
likely to be his own son or grandson. The student who was also enrolled at the Academy could expect the advantage of his master's
favor in his career at the official institution. Those of the nobility
who wished to acquire learning studied, as they always had, with private tutors, generally senior relatives or literati clients.
The subject of schooling cannot be left without some mention of
Kukai's Shugei Shuchiin ("Academy of Arts and Sciences"), typical
of an idealism found only at the beginning of the Heian period.
Kukai's intention was to establish a school that would make educaCambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
375
tion available to poor boys as well as the well-to-do, and the comprehensive curriculum he envisioned was to combine instruction in
Exoteric and Esoteric Buddhism with Confucian teaching. The protocols of the school were written in 828, and construction was
started at a site near the Toji, but the school lasted - if it was ever
properly in operation-no more than seven years. After Kukai's
death in 835, it fell under the supervision of his disciple Jitsue
(786-847), who was not a powerful man. There were financial troubles, and such buildings as had come into existence already needed
repair. The site was soon sold to obtain money to expand the Toji's
estates in Tamba, and by mid-century only Kukai's eloquent proposal remained of this visionary scheme.35
SCHOLARS AND THEIR ACCOMPLISHMENTS
Even a cursory acquaintance with the lives of the most prominent
literati leaves an impression of their individuality; familiarity makes
it hard to credit the lack of a sense of "selfhood" said to be common
to Japanese. Perhaps the great difficulty of the course of study, the
encountering and surmounting of disappointments, praise - or the
hope of it - for possessing "talent" (a key word in traditional biographies), and a conviction of the centrality of the knowledge acquired,
together with pride at the ranks and offices attained or, in the vast
majority of cases, years of resentment over the paltriness of the official reward, all contributed toward making men of distinctive character. Literati of Murakami's time and onward tended to exceptional
piety; those active from the middle of the eleventh century were
likely to see themselves as isolated or as members of a small fraternity, each member of which was to be valued. None of the men
whose work or character is discussed below is fully representative of
the others of his time, but each is of interest in representing some of
the currents of his age and class, and some of the possibilities of temperament. Behind each one may be seen the figures of other literati,
some of them equally or almost equally eminent in the eyes of their
colleagues, and perhaps others, less successful, even more humbly
employed, whose identities are lost to history.
If there are traits common to many of the prominent literati, they
35 Wm. Theodore de Bary, ed., The Buddhist Tradition: In China, India and Japan (New York:
The Modern Library, 1969), pp. 309-13;Yoshito S. Hakeda: Kukai:MajorWorks (NewYork:
Columbia University Press, 1972), pp. 56-58. Kukai's Shingon sect alluded to its patriarch's
school when it gave the name Shuchiin Daigaku to its Kyoto seminary in 1949.
Cambridge Histories Online © Cambridge University Press, 2008
376
CHINESE INTELLECTUAL LIFE
are energy and the sense of a mission to educate, expressed in the
voluminous production of works intended to supply the practical
needs, as the Heian period defined them, of cultivated people. Kukai, discussed at length in Chapter 7, deserves further mention
here. In his youth a student at the Academy (before the removal of
the capital to Heian-kyo), he brought new secular as well as religious learning with him on his return from China. He owed Saga's
favor more to the emperor's appreciation of his artistic gifts than to
his religious eminence. He was ranked as one of the three great calligraphers of his day, along with another former visitor to China,
Tachibana no Hayanari (d. 842), and Saga himself. He was a frequent partner in the emperor's poetic exchanges, accompanying
him on his outings to the Park of the Divine Spring, composing
verses to be copied on screens in the palace and inscribing characters on tablets for the palace gates. He wrote dedications on behalf
of Saga's nobles - a task, despite the religious content of the compositions, more typical of literati than of monks. His most celebrated secular work, however, is Bunkyo hifu ron, a digest of a number of Six Dynasties and T'ang treatises on the rules for poetic
composition, presented to the throne in 819, its subjects ranging
from rhyme, tone, and diction to the sources of poetic inspiration.
Many of the works that he excerpted are no longer extant and are
now known only because he quotes them. If none of Bunkyo hifu ron
represents Kukai's original thinking - as would scarcely have been
expected - Japanese scholars have nevertheless found occasion to
note his independence of judgment in selecting and arranging his
sources.36 The demand for such a work is testified to by the fact
that, a year later, he prepared an abridged version, under the title of
Bumpitsu ganshin sho. Kukai was also responsible for Tenrei bansho
myogi, an edition of the popular Six Dynasties character dictionary
Yu p'ien. The work boasts some special annotation for Japanese
users and is the oldest surviving (though not the oldest) character
dictionary produced in Japan. The Heian period in general was fertile in dictionaries and glossaries of many kinds compiled to serve
religious as well as secular needs.
Whereas Kukai, born in Shikoku, was the scion of a branch of the
ancient Otomo clan, Sugawara no Michizane was the fourth in a line
of professional scholars ennobled by Kammu. His father Koreyoshi
(812-80) had a career typical of the successful literatus, occupying,
36 Abe Akio, Chuko Nihon bungaku gaisetsu (Tokyo: Shuei shuppan, 1977), p. 23.
Cambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
377
among other posts, those of tutor to the crown prince. Head of the
Academy, and provincial governor, lecturing at court, becoming a
senior noble in his old age, and leaving behind him a great quantity
of writings, which included a thesaurus, Togu setsuin, based on thirteen Chinese rhyming dictionaries, and an anthology of his own
Chinese poems in the usual form of six chapters.
Michizane himself was the most admired of all classical Japanese
poets in Chinese. The quantity of his productions and the number
and nature of those that he produced on behalf of others deserve to
be noted, as a suggestion of what was expected of the court scholar;
he was prolific, but perhaps not very exceptionally so. Michizane
composed his first Chinese poem in his eleventh year, his first couplet worthy of being quoted in an anthology in his fourteenth, his
first prose work on behalf of a patron in his fifteenth. He completed
his Academy studies in 870, his twenty-sixth year - an early age at
the time. The following year, his compositions included a dedication
to accompany a donation by an imperial consort and a series of three
memorials submitted to the throne by the regentYoshifusa, in which,
as etiquette demanded, the regent attempted to decline an addition
to his emoluments. The year after that, he was assigned to the entertainment of an embassy from Po-hai. A major part of the entertainment consisted of exchanges of Chinese poems with the foreign
visitors. (In this instance, Michizane's activities were cut short by his
mother's death, but he was to officiate in receiving two subsequent
embassies.)37 His compositions on behalf of others this year included a resignation proffered by an Assistant Professor at the Academy and a memorial from the new Minister of the Right, Fujiwara
no Mototsune, giving thanks for his appointment, as well as a letter
in parallel prose from Emperor Seiwa to the king of Po-hai. In the
next year, 873, he wrote a dedication for an official of the Treasury
Ministry, presenting house and lands to the temple Urin'in, and a
dedication for Mototsune to accompany a gift of rice fields to the
Fujiwara clan temple Kofukuji.
These writings represent only a portion of his output during a few
typical years early in his career: Kanke bunso, the principal collection
of Michizane's belles lettres, records eleven prose pieces attributed
to the two years 872-73, and there must have been other works,
thought not worth preserving. In all, the anthology contains more
than 150 compositions in formal prose, including rhyme-prose (fit),
37 Borgen, Sugawam no Michizane, pp. 231-40.
Cambridge Histories Online © Cambridge University Press, 2008
378
CHINESE INTELLECTUAL LIFE
funeral inscriptions and eulogies, prefaces (Jo), examination questions and answers, edicts written on behalf of emperors, and a variety of memorials. Of more interest to present-day readers are the
more than 500 Chinese poems in the anthology. Many are conventional verses composed for banquets, but the best are expressions of
their author's deepest feelings. The turning point in his evolution as
a poet came with his unwilling posting as governor to the province
of Sanuki in Shikoku, in 886. The poems written after Michizane's
exile to Kyushu in 901 (preserved in a supplementary anthology) are
especially prized for their poignancy.38
Among Michizane's compositions from the year 873 is a formal
preface to Chiyo shakuen, a handbook of government he was compiling to aid students preparing for their civil service examinations.
The handbook itself was intended to be of monumental dimensions.
Although it remained unfinished, it prepared the way for his later
major historical work, Ruiju kokushi, in which the contents of the Six
National Histories are rearranged by topic. Ruiju kokushi was begun
at the command of Emperor Uda and presented to the throne in
892. (How exactly it came to include the contents of Sandai jitsuroku
is not known.) The thoughtfulness of its arrangement is noteworthy,
for it makes use of a kind of cross-referencing, and it deals meticulously with its sources, unlike the histories compiled privately later
in the Heian period. The categories were devised by Michizane on
the general model of the Chinese lei-shu but conforming to Japanese
priorities. They tell us a good deal about the concerns of the court.
First is the Way of the Gods, followed by emperors, imperial women,
and the rest of humanity; then "Calendar," "Music," "Banquets,"
"Memorials," "Government Bureaus," "Literature" (bun, i.e., literature in Chinese), "Agriculture," "Felicitous Omens," "Disasters."
The relegation of Buddhism to a place near the end of the list is due
to a sense of the proprieties of Confucian history, not a lack of piety.
Unfortunately, less than a third of the 200-chapter original is extant.
Michizane's rise to power and sudden fall belong to political history and are discussed in Chapter 1, but a few additional points are
relevant to the concerns of the present chapter. One has to do with
the lineaments, discerned in later anecdotes about him, of a strong
but unstable character. He inspired devotion, but he was also arrogant and had violent outbursts of temper. Another is the suggestion
38 Burton Watson, trans., Japanese Literature in Chinese, vol. 1: Poetry and Prose in Chinese by
Japanese Writers of the Early Period (New York: Columbia University Press, 1975), pp.
73-13OCambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
379
that the coup against him was brought about through the animosity
of rival literati. This cannot be the whole story, but there may be
much truth in it, as one act can serve a number of motives. Not only
all desirable Academy posts but half or more of all bureaucratic positions were said to be held by Sugawara disciples, so that his exile
might have been expected to open the way to literati clients of the
rival party. Michizane was habitually outspoken in his scorn of those
he considered his intellectual inferiors. One who had good reason to
resent him was Miyoshi Kiyoyuki, of whom he was intemperately
contemptuous and whom he initially failed in the doctoral examination. Perhaps with reason, traditional accounts of Michizane's fall
make Kiyoyuki one of the chief villains of the affair.
Kiyoyuki's rise was more gradual than Michizane's; his final rank,
Junior Fourth Lower Grade, still left him below the senior nobility,
and his highest post, that of Consultant, though a considerable
prize, was achieved only when he was past seventy. Michizane, by
contrast, became Consultant in his forty-ninth year. Rather than descending from a family of scholars, he was the son of an obscure
provincial governor. He seems to have owed his success at the Academy, over Michizane's opposition, to Kose no Fumio (824-92),
whose disciple he was; his promotion to monjosho coincided with the
elevation of his teacher to Head of the Academy. In 883, two years
after the original examination, Michizane changed his grade to a
passing one and Kiyoyuki was appointed Assistant Professor under
Fumio. In the Ako Controversy, Kiyoyuki ranged himself with the
regent's party against Emperor Uda, and just as Michizane was rewarded for his loyalty to the emperor after the regent's death, Kiyoyuki, among others of Mototsune's clients, found himself rusticated
in an unwanted provincial governorship. Even after his return in 897,
his way to further advancement was blocked, for not only did
Michizane's followers occupy all available academic posts, but in 897
Michizane's son Takami was Head of the Academy. Kiyoyuki did
succeed in being named Professor of Literature in 900 - after Uda,
by taking the tonsure, had effectively eliminated himself as Michizane's protector. Following the exile of Michizane and his sons,
Kiyoyuki became Head of the Academy as well, a triumph soon
compounded with the addition of the office of Vice-Minister of the
Ministry of Ceremonial.
Factions were rife among Heian literati. Personal loyalties, of
course, were an important consideration, but in addition, temperament, talent, and ideology distinguished the pragmatists among
Cambridge Histories Online © Cambridge University Press, 2008
380
CHINESE INTELLECTUAL LIFE
them from those of more purely literary and scholarly inclinations.39
If Michizane was an example of the latter, Kiyoyuki is an example of
the former. Scarcely the poet that Michizane was (although he accumulated enough verses to form a six-chapter collection), Kiyoyuki
was an outstanding writer of ordinary expository prose. His work is
noted for breadth of subject matter and clarity of expression - the
latter a quality neither common in Heian Chinese nor admired by
writers of the age. Like many other Heian literati-for example,
Haruzumi no Yoshitada and Oe no Otondo before him and Oe no
Masafusa after him - Kiyoyuki was strongly interested in the Taoist
art of prolonging life and in the occult. For men of the Heian era, the
occult was an aspect of reality, and it was the man of inquiring mind
who garnered anecdotes about it, as Kiyoyuki and Masafusa both
did. Kiyoyuki was especially devoted to the study of Chinese calendrical lore. A learned memorial, the Kakumei kammon, which he submitted in 900, stressed that the following year, the fifty-eighth in the
cycle of sixty and a fateful year in the larger cycles as well, would be
unlucky and bound to bring "revolution" (in the sense of change in
leadership) unless a new era name was adopted. This may have been
intended as part of his campaign against Michizane. His argument is
marred by some faulty arithmetic, but there is no reason to believe
that he was insincere in the science that he invoked. The court was
indeed persuaded to change the era name in 901, and the memorial
enjoys independent fame for having originated a custom followed
through subsequent sixty-year cycles as well as providing the model
for the compositions appropriate to such occasions.
Where Michizane's unwelcome posting to a provincial office
brought about his maturation as a poet, Kiyoyuki's undoubtedly
deepened his appreciation of the practical difficulties of governing.
The fruits of his experience are recorded in a celebrated document,
the Iken fuji submitted in 914 in response to a request made by
Daigo in 909. (Iken fuji were statements of opinion on the successes
and failures of government, prepared after a call was issued by the
emperor.) The document noted real problems and expounded concrete remedies. It can be read not only as evidence of general conditions at the time but as a sign of the sympathies and anxieties of Kiyoyuki's class. The introduction deplores past extravagances, praises
Kiyoyuki's patron, the late regent Mototsune, and depicts a depop39 Goto, Heiancho kambungaku, pp. 79-93; Hayashi Rokuro, Jodai seiji shakai no kenkyu
(Yoshikawa kobunkan, 1969), pp. 381—509.
Cambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
381
ulated, impoverished countryside. The topics of the individual articles are as follows:
1. Preventing floods and droughts and obtaining rich harvests:
Kiyoyuki blames unsatisfactory conditions on the fact that
priests who carry out the great national rituals are not men of
pure conduct.
2. Luxury should be forbidden.
3. Redistribution of government allocation fields should be carried
out in the provinces in proportion to the actual population. (A
redistribution had been ordered in 902. Kiyoyuki's motive is less
fairness to the peasantry than the hope of increasing revenues
through confiscation of excess land.)
4. Funds allotted for support of students at the Academy should be
increased.
5. The number of dancers at the Gosechi Festival should be reduced; at present, Kiyoyuki says, parents of the dancers compete
in expenditure.
6. The number of judges should be increased to that of former
times. Provincial governors should be exempt from punishment
for minor crimes or because of unfounded accusations.
7. All officials should be paid their half-yearly stipends without exception; in recent years, only the senior nobles have received
them regularly.
8. There must be an end to the practice of discharging provincial
officials upon accusation by the local gentry and under-officials.
9. A limit should be placed on the number of men in the provinces
exempt from taxation as nominal officials of the lowest class.
10. Sale of appointments to the Imperial Police and Guards should
be stopped. Soldiers should be trained in the use of the catapult.
11. The disorderliness of imperial guards and monks in the
provinces must be forbidden. Over half of the two to three hundred men who become monks each year are rascals. Provincial
householders take the tonsure to avoid taxation but do not
abandon secular ways; some even join gangs of thieves. Guards
sent to the provinces do not return to the capital but stay and
terrorize the populace.
12. A renewed plea that the port of Uozumi in Harima Province be
reopened.40
40 Complete German translation in Inge-Lore Kluge, Miyoshi Kiyoyuki: sein Leben und seine
Zeit (Berlin: Instirut fur Orientforschung, 1958), pp. 40-70; partial English translation in
Lu, Sources of Japanese History, vol. 1, pp. 60-65.
Cambridge Histories Online © Cambridge University Press, 2008
382
CHINESE INTELLECTUAL LIFE
Kiyoyuki believed strongly in the Confucian ideal of the provincial
governor who was frugal, compassionate, loyal, and honest to the
point of impoverishing himself. When he argues that provincial governors should be excused their failings, his assumption is not that
the men are invariably upright but that to some degree corruption,
quarrels with local magnates, and failure to meet the central government's expectations are inevitable. He is keenly aware also of the
chronic dissatisfactions of students and lesser officials, which he no
doubt shared. Where he complains against the evils of luxury, his
modern biographer sees, besides Confucian idealism, the resentment of the literati, as a class, against the senior nobles.41 Kiyoyuki's
memorial contains eloquent testimony to hard times.
In 954 Murakami issued another call for Ikenfuji. Three and a half
years later, Michizane's eminent grandson Sugawara no Fumitoki
(899-981), known especially as a master of poetic rhetoric, responded by producing a memorial in three articles. He urged (1) the
prohibition of luxury, (2) stopping the sale of offices, (3) supporting
the Academy and the literati. Not only was Fumitoki's fuji shorter,
but it represented the viewpoint of aristocrats and officials in the
capital and was extremely abstract. Each of its articles is related to
one of Kiyoyuki's and includes expressions that may be adaptations
of Kiyoyuki's.42 Following the correct model for composition was
more important for Fumitoki than giving advice wrought out of his
personal experience - advice that in any event the emperor could
scarcely have implemented. By his day, most of the central government offices that literati had been trained to fill existed in name
only; the primary relationship of the literatus was with the nobleman
who was his patron. The literatus was a specialist from whom the
noble received such instruction as he desired but whose professional
skills he need not personally aspire to - except in the way that,
through some quirk of talent or circumstance, he might aspire to
mastery of a particular musical instrument. He demonstrated his
Confucian commitment by his beneficence to his literati clients, entertaining them suitably and sponsoring them for court appointments and provincial governorships. They were of use to him also as
stewards (keishi) in his household office and administrators of his
provincial estates. When the court held Chinese poetry competitions, each team would consist of literati who composed the poems
offered for judgment and nobles who presented the poems with due
41 Tokoro, Miyoshi Kiyoyuki, pp. 168-69. 4 2 Tokoro, Miyoshi Kiyoyuki, pp. 182-83.
Cambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
383
ceremony and defrayed the often considerable expenses of the trappings and refreshments.
When a mid-Heian noble of high rank was himself learned in Chinese things it was often because he was an artistic polymath. An example is Fujiwara no Kinto (966-1041), a second cousin of Michinaga and brother of one of Emperor En'yu's consorts, a man skilled
not only in Chinese poetry but in Japanese waka and in calligraphy
and music and an authority, characteristically, on ceremonial customs and precedents. Tastes in Chinese learning were also cultivated
by persons who might have been in power but were not. Daigo's sixteenth son, Kaneakira (914-98), a distinguished poet known as the
Saki no chusho-o ("Former Archivist Prince"), and Murakami's seventh son,Tomohira (964-1009), the Nochi no chusho-o ("Later Archivist Prince"), presided over literary salons. Another active patron
was Fujiwara no Michinaga's unfortunate nephew Korechika, who
was remembered by later literati as a poetic arbiter worthy of respect, although he composed no Chinese poems after returning
from exile. The role of Confucian patron was also flattering to the
grandiose style of Fujiwara no Michinaga. He amassed a large Chinese library, and Chinese learning as a whole underwent a modest
revival (which seemed a great one a generation later), thanks to the
encouragement he gave it. But all of the Fujiwara regents were Confucian patrons, if only because their position as uji no choja (clan
chief) of the Fujiwara made it incumbent upon them to promote the
moral well-being of their clan and descendants.
There was no slackening at court of Confucian ceremonies, nor of
Chinese poetic entertainments. In the fifth month of 1003, for example, on the sixth day, there was a poetic gathering in the palace.
Among those attending, in addition to Michinaga, were such aristocratic literary lights as Kinto, Fujiwara no Arikuni (943-1011), and
Fujiwara noTadanobu (967-1035). The topic (dai) given the versifiers was "the thin, solitary voice of the first cicada." On the twentyseventh day, Michinaga held an entertainment at Uji, with Chinese
and Japanese poetry and instrumental music. The topic for Chinese
poems was "the freshness of mountains and streams after the sky has
cleared." It was typical of the age that men with talent for composing Chinese poetry were unlikely to confine themselves to that language; many of those who distinguished themselves at these gatherings, like Tadanobu, are better known for their work in Japanese.
Among those who frequented the literary gatherings at Prince Tomohira's mansion were two disciples of Fumitoki who, though men
Cambridge Histories Online © Cambridge University Press, 2008
384
CHINESE INTELLECTUAL LIFE
of ability, had to be content with modest rewards in their official careers. Fujiwara noTametoki (947?-iO2i?), father of the great writer
Murasaki Shikibu, owed his initial progress in office to a roundabout marital connection with the fifth son of a former regent. With
the abdication of Kazan, he lost his posts and, except for a brief
term as provincial governor, went many years without official employment. Yoshishige noYasutane (d. 1002) exemplifies the difficulties encountered by the literatus who was not of a family of hereditary scholars. That he would pass the highest civil service
examination was not expected, and after he did so, for all his brilliance, he was unable to obtain any significant promotion. He ended
his career almost at the rank at which be began. His notable works
include Chiteiki43 - modeled in part on a work of Po Chu-i and itself a predecessor in Chinese of Kamo no Chomei's classic Hojoki and Nihon djo gokuraku ki, which inaugurates a genre of hagiographical collections in Chinese, celebrating the lives of Japanese
Buddhist holy men, monks and laymen alike, the humble as well as
the highly placed. A follower of Kuya, revered in turn by Genshin
(who saw to it that his devotional poetry was sent to China),44 he
was the founder of a characteristic Heian religious institution, the
kangaku-e, a two-day semi-annual gathering at which twenty Tendai
monks and twenty literati laymen would hear lectures on the Lotus
Sutra and compose Chinese poems on topics chosen from the sutra.
When Yasutane formally entered religion he took the name Jakushin, and as a monk, his deeply felt, literal-minded piety, which took
the Buddhist doctrine of compassion for all living beings to an extreme, became the subject of legend. A number of Oe literati became his disciples in religion and, like their master, are themselves
commemorated in the hagiographical literature; one went to China
and died there.45
Other names might be mentioned: enumerating in admiration the
eminent men - from ministers to warriors - who added luster to the
court of Emperor Ichij5 (r. 986-1011), Oe no Masafusa listed ten
names of literati, and placed his great-grandfather Oe no Masahira
(952-1012) at the head.46 Masahira, tutor to the crown prince and
lecturer to Ichijo and Sanjo, had die advantages of his ancestry, his
43 Translation in Watson, Japanese Literature in Chinese, pp. 57-64.
44 Kawasaki Tsuneyuki, Jimbutsu Nihon no rekishi, vol. 3. Ocho no rakujitsu (Tokyo: Yomiuri
shimbunsha, 1966), pp. 112—15.
45 "Zoku honcho ojoden," in Inoue Mitsusada and Osone Shosuke, eds., Ojoden, Hokke genki,
Nihon shisd taikei, vol. 7 (Tokyo: Iwanami shoten, 1974), pp. 247-48.
46 Inoue and Osone, eds., Ojoden, p. 224.
Cambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
385
skill in poetic rhetoric, and the patronage of Michinaga in achieving
a highly respectable, though not exalted, Fourth Rank. He too was a
pious man. He remained throughout his life a subordinate, as a
modern biographer has pointed out, with neither influence nor aspirations to influence in the practical conduct of government. The
same biographer also notes the extent to which he was professionally occupied with words: criticizing the use of two characters in an
edict, proposing era names (two of those he recommended were
adopted), and advising Michinaga on names for imperial princes.47
A similar observation, of course, might be made about almost any
other literatus serving at court throughout the period.
The most elegant of the writings of these men were collected by a
Professor of Literature, Fujiwara no Akihira (989-1066). His Honcho
monzui, in fourteen chapters, took its name from a Chinese anthology, T'ang wen-ts'ui, and its form and ambitions from Wen hsiian. He
is thought to have completed it during the Kohei era (1058-65).
Again, as with Tamenori's inclusion of the T'ang history in his
Kuchizusami, Akihira's choice of title demonstrates that books from
China continued to appear in Japan, for T'ang wen-ts'ui was completed in 1011 and first printed in 1039. Although there were a number of other privately compiled anthologies of belles lettres, Honcho
monzui became the most prestigious, providing models for later writers. The work seems to have been conceived as a successor to the
three imperial anthologies compiled in Saga's time. The periods
most represented in it are those of Daigo and Murakami, but the individual writer whose compositions appear most often is Masahira.
Akihira succeeded to Masahira's eminence and honors, and he
came to regard himself, no doubt with justice, as the sole preserver
of the glories of Masahira's time. He was additionally, however, a
man of independent and original mind, with a keen eye for the contemporary scene. He wrote to educate, and his Shin sarugaku ki and
Unshu. shdsoku are as remarkable for their liveliness as for their informativeness. Both works are accounted ancestors of the later genre
of textbooks called orai mono.
Shin sarugaku ki presents the principal occupations of the time
and their vocabulary with encyclopedic thoroughness. The work is a
kind of fiction: to a performance of the entertainment called sarugaku come a captain of the Gate Guards, his three wives - one old
47 Francine Herail, "Un lettre a la cour de l'empereur Ichijo: 6e no Masahira (953-1012),"
in Melanges offert a M. Charles Haguenauer en I'honneur de son quatre-vingtieme anniversaire.
Etudes japonaises (Paris: L'Asianque, 1980), pp. 369-87, esp. 383-84.
Cambridge Histories Online © Cambridge University Press, 2008
386
CHINESE INTELLECTUAL LIFE
and jealous, one housewifely and maternal, one young and adored his sixteen daughters and nine sons. Each of his daughters' husbands, some of his daughters, and each of his sons is a paragon of a
different occupation or example of a social type, with the symmetries
and hyperbole of Chinese rhetoric contributing to the reader's education. There is a master gambler, a bowman, an exemplary provincial official skilled in agriculture, a metalsmith, a sumo wrestler,
a teamster, a physician and acupuncturist, a yin-yang master, a
painter, a calligrapher, a maker of Buddhist images, a yamabushi,
and aTendai monk as impressive in appearance as he is learned. One
daughter is a courtesan, another a pious widow, yet another, extremely rich, a shamaness. One daughter is a beauty "not inferior to
Yang Kuei-fei," while an ugly sister, whose defects are described, has
formed a liaison with a charcoal burner. Another ugly sister is married to a criminal the list of whose transgressions enriches the
reader's vocabulary. There are suitors: a Japanese poet much fancied
by the nobility, and a boy musician beloved by all the monks. Perhaps of greatest interest in the present context are a trader who deals
in Chinese imports as well as native goods and a scholar learned in
all four curricula of the Academy, who has read Wen hsu'an, Shih chi,
Han shu, and all the classics; the ritsu and the ryo, the kyaku and the
shiki; has mastered the varieties of calculation; and is skilled in the
composition of Chinese verse and rhyme prose, prefaces, edicts of all
sorts, commands, memorials, dedications, prayers, letters and responses, interbureaucratic memos, requests, and diaries. Akihira
gives him the evenhandedly concocted name of "Sugawara no Masafumi," the "masa" suggesting Masahira, and the "fumi," Fumitoki;
the two words in combination meaning "correct writing."
Unshu shosoku is a collection of more than two hundred model letters, many of them, it is thought, supplied from Akihira's personal
and official correspondence. As with Shin sarugaku ki, its Chinese is
notably influenced by Japanese usage.48 Written very much from the
point of view of Akihira's own class, the work gives invaluable insight
into the activities, mores, and needs of the middle-ranking official,
as well as into the etiquette of polite address. There are invitations to poetry parties, athletic contests, scenic excursions, a moonviewing party - congratulations, and requests for information. A
complaint about the slow receipt of tax revenues is accompanied by
a hint that the writer would like to be promoted. There are com48 Konishi, A History of Japanese Literature, vol. 2, pp. 185-86.
Cambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
387
plaints about the difficulty of promotion, which can best be accomplished if the aspirants will donate a building to their lordships. A
writer protests that he is unable to provide a feast requested for an
observance at the regent's mansion. A former governor writes to a
monastic official asking for instruction on dyeing priests' garments
and preparing offerings. Someone who is housebound because of a
directional taboo is given a description of a festival he has missed; a
near riot, mountebanks cavorting obscenely - great fun in the letter
writer's opinion. An Academy student facing an examination is advised to borrow a certain crib book. There is a great deal about borrowing and lending, suggesting to a modern commentator that middle officials were evidently obliged to live beyond their means. A
series of dunning letters is also provided. Akihira, who served as governor of Izumo, offers a number of letters dealing with provincial affairs. It has been suggested that the work was intended as a textbook
for an advanced course for students of present-day middle-school
age, and that its contents grew in response to very specific needs.49
Its provision of polished phrases with which the individual could
meet the exigencies of his economic and social life makes it an interesting complement to Tamenori's primer.
Oe no Masafusa (1041-1111), in his twenty-sixth year when Akihira died, succeeded to his lonely eminence. A child prodigy (precocity is a common element in Chinese literary hagiography, both
Japanese and continental), in old age he composed a brief memoir
reciting the names of the great men, now long dead, who had petted
and praised him; the memoir is no less poignant for following,
phrase by phrase, a poetic preface composed by Michizane's disciple
Ki no Haseo for the rather different purpose of bemoaning the factionalism of his day.50 It is characteristic of the Japanese literati to
have modeled their compositions on the admired examples of Japanese predecessors. Masafusa is an intriguing figure for a number of
reasons, one being that he sums up so many of the qualities and tendencies of his class. He was both a conservative and something of a
new man. The proceeds of a term as governor of Mimasaka Province
enabled him to construct a building to house the extensive Oe family library; his most important work is a monumental ceremonial
compendium, Goke shidai. He achieved some distinction as a poet in
49 MihoTadao and Miho Satoko, eds., Unshii orai (Osaka: Izumi shoin, 1982), p. 480. Partial
translation of the text in Clemens Scharschmidt, " Unshu orai oder die Briefsammlung des
Unshu," Ostasiatische Studien 20 (1917): 20—114.
50 Goto, Heiancho kambungaku, pp. 282-309.
Cambridge Histories Online © Cambridge University Press, 2008
388
CHINESE INTELLECTUAL LIFE
Japanese as well as Chinese and was wont to stress the value of
Japaneseness, on one notable occasion instructing a regent that yamato-damashii (Japanese spirit) was more important for the regent's
young son than Chinese book-learning. Like Akihira, he wrote about
the lively scene, taking as his subjects itinerant entertainers, the
prostitutes of Eguchi and Kanzaki, and the dancing mania of
1096 - although it is impossible for the modern reader of the compositions in question not to feel the pressure of his learned Chinese
style against the accurate expression of what he saw. He was a pious
Buddhist, compiling a continuation of Yasutane's Nihonojo gokuraku
ki; unlike its predecessor, however, Masafusa's collection combines
political with religious concerns, in keeping with its author's own
character. As was traditional among the Oe, he was adept in Taoist
lore, and he also compiled a collection of lives of Japanese Taoist
"immortals," beginning with the legendary Yamato Takeru and
Prince Shotoku. He had a reputation as a polymath - although the
story that he gave instruction in the art of warfare to Minamoto no
Yoshiie is almost certainly apocryphal. His reputation as a greedy
and dishonest governor is more likely well deserved. His fondness
for collecting anecdotes in his old age provoked the disapproval of
more dignified contemporaries. He took great pride in inheriting the
traditions of the Oe and pride also in surpassing his forefathers in attaining Second Rank (Junior, 1094; Senior, 1102).
In his pragmatic and inquisitive temperament, accompanied by
his fascination with the magical and supernatural, Masafusa resembles Kiyoyuki, but his attainment of high position suggests a comparison with Michizane, for whom Masafusa's own reverence went
deep.51 Masafusa owed his success to the determination of emperors
Go-Sanj5 (r. 1068-73) a n d Shirakawa (r. 1073-87) to rule for themselves, but whereas Michizane in power was isolated, Masafusa was
one of a number of Confucian advisers to the emperor, appointed to
relatively inconspicuous posts that kept him in close contact with his
master. His association with Go-Sanjo began while the future emperor was crown prince. Appointed tutor to the crown prince in
1067, the final year before Go-Sanjo's succession, he remained tutor
to successive crown princes until 1085. Confucianism was the storehouse to which every aspiring ruler would go for his intellectual and
ideological stock, the more conspicuously the better, and Go-Sanjo
51 Robert Borgen, "Oe no Masafusa and the Spirit of Michizane," Monumenta Nipponica 50
(1995):
Cambridge Histories Online © Cambridge University Press, 2008
SCHOLARS
389
was an attentive student. Asked about his abilities, Masafusa replied
enthusiastically that the emperor was "as good a scholar as Oe no
Sukekuni."52 Sukekuni, an older contemporary and a great-grandson of the famed 6 e no Asatsuna (886-957), w a s o n e °f th e most
distinguished literati of the time. Go-Sanjo was said to have taken
Masafusa's advice on appointments, and Masafusa is credited with
having been the architect of one of Go-Sanj5 dearest projects, that
of regulating the skoen. Masafusa was also a confidant of Shirakawa;
when Shirakawa abdicated, he was one of the five inaugural superintendents (betio) of the new ex-emperor's household office. Another
of Masafusa's patrons was the Fujiwara regent Moromichi (106299), who disapproved of Shirakawa's activity as ex-emperor, and it
may well have been the tension between the two, as well as the desire to enrich himself, that prompted Masafusa's ready absence from
the capital for a term as governor general of Kyushu. In Kyushu,
Masafusa endowed shrines and instituted additional ceremonies to
Michizane.
Masafusa, both as individual and as representative of the interests
of his class, is revealed most fully in Godan ("The Oe Conversations"; also, Gddansho), a record of his talks transcribed by a youthful disciple along with, perhaps, some of the anecdotes he himself
had recorded. Masafusa's subjects in it are men, Chinese poetry,
public affairs - meaning ceremonial, not economics or the science of
governing, despite his actual activities - the supernatural, and learning as a virtuoso accomplishment. The men are those of the Japanese literati past, and he is interested in them as individuals. He portrays them as cantankerous and competitive. In poetry, he is
interested in words and phrases. A particularly fine couplet may be
the achievement that justifies a life - as he said of himself.53 The tone
of a particular Chinese word may be the subject of a lengthy discussion. An occasion on which an allusion has been identified as inappropriate is worthy of note, in part because it suggests the superior
subtlety of the literatus who has so identified it. China itself is mentioned as the source of accolades: the works of a poet are said to be
worthy to be sent to China. But Masafusa is not interested in the
Chinese poets as individuals, nor in their poetry for its own sake.
Chinese learning in Japan has become a self-contained tradition,
with its roots and objects in its own past.
52 Takeuchi Rizo, Bushi no top, vol. 6 oiNihon no rekishi (Tokyo: Chuo koron sha, 1972), pp.
144-45, from Zoku kqjidan. 53 Kawaguchi and Nara, Gddansho chu, pp. 1154-60.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 6
ARISTOCRATIC CULTURE
The term aristocratic culture is here used to mean a style of social
and artistic expression characteristic of the Japanese court at Heiankyo and limited primarily to its members. The small, isolated, tightly
knit court community shared traits that conditioned the development of the culture: a strong sense of status and a firm subordination of the individual; an emphasis on order, decorum, and conformity; a greater interest in immediate solutions to practical problems
than in ethical questions, philosophical speculation, or scientific inquiry; a general tendency toward emotionalism in preference to intellectualism; a pervasive, melancholy concern with the changes
wrought by the passing of time; a high esteem for literature, calligraphy, and music; an acute sensitivity to beauty and to the moods of
nature; and an unwavering belief in the importance of taste as an
index of character and breeding. As a group, those characteristics
seem to have crystallized during the last decades of the ninth century and the first half of the tenth - a period, symbolized by the cancellation in 894 of the court's last projected official mission to China,
during which the Japanese repudiated earlier efforts to make their
court a mirror image of the one at Ch'ang-an, and moved instead toward the amalgamation of foreign and native elements into a civilization distinctively their own.
The preoccupation with beauty, one of the most conspicuous aspects of the new culture, influenced attitudes toward nature, standards of judgment in the arts, appraisals of human worth, and norms
of social behavior. It also powerfully affected almost every facet of
ordinary life, both public and private, as may be seen from a survey
of upper-class living accommodations, dress, and dietary customs.
DOMESTIC ARCHITECTURE AND FURNISHINGS
From around 950 on, the typical aristocratic residence consisted of
a group of buildings situated in a large urban estate, its stands of
390
Cambridge Histories Online © Cambridge University Press, 2008
DOMESTIC ARCHITECTURE
3Q1
pine and maple trees, artificial hills and streams, and island-studded
lake imitating the natural landscape of the Kyoto basin, or echoing
the features of a famous scenic spot in the provinces. In architecture
of this type, known as shindenzukuri (main hall construction), the
buildings were carefully designed to harmonize with the setting. The
main hall (shinden) was a simple, graceful, rectangular one-story
structure made of unpainted wood, with a gabled, shingled roof and
a wooden floor raised a few feet off the ground on posts. Its large
central room (moya) was divisible into smaller areas by movable
curtains and screens. Secondary rooms, known as eave chambers
(hisashi no ma), gave onto open verandas, from which they were separated by removable wooden shutters (shitomi) and bamboo or reed
blinds (sudare). Well suited to the hot, humid summers of the region,
the dark, drafty, high-ceilinged rooms offered little comfort in winter, when they were heated only by inefficient charcoal braziers.
Since the shinden was ordinarily used as a residence by the master or mistress of the establishment, it commanded the best view of
the main sanded courtyard, the lake, and the gardens with their
rocks, shrubs, and seasonal displays of flowering plum, cherry, wisteria, yamabuki Japan globeflower), white chrysanthemums, autumn leaves, and plume grass (susuki), all of which adjoined it on
the south. It was connected by covered galleries to lesser buildings
of similar architecture, which varied in size and number according
to the owner's wealth and social importance, but which normally included at least one or two wings (tai) assigned to the reception of
callers, the accommodation of family members and attendants, and
other uses. Great estates were provided with libraries, Buddhist
chapels, racetracks, dance platforms, and fishing pavilions; and the
waters, hills, trees, and flowering plants of their extensive gardens
were among the happiest expressions of Heian sensitivity to visual
beauty.1
The simplicity of the architecture was complemented by the deliberate sparseness of the interior furnishings. Even the bestequipped room seldom contained more than screens and curtains;
one or two double-tiered cabinets or chests; boxes for writing equipment, poetic anthologies, small toilet articles and mirrors; an incense burner; an armrest; a lampstand; some mats and cushions;
and a small curtained alcove, the chodai (curtain-dais), which did
double duty as bedchamber and sitting room.
1 For an exhaustive study of the shindenzukuri style, see Ota Seiroku, Shindenzukuri no kenkyii
(Tokyo: Yoshikawa kobunkan, 1987).
Cambridge Histories Online © Cambridge University Press, 2008
392
ARISTOCRATIC CULTURE
Wood, lacquered for decorative purposes and as a protection
against insects and rot, was the preferred material for furniture and
articles of everyday use. The many lacquer techniques and designs
introduced from China in the eighth century continued to be employed to some extent in the ninth and early tenth, but thereafter
they were supplanted by a new, purely native style.The technique on
which the style was based, known as makie, consisted in applying
metal powder to a drawing traced in the lacquer base, or directly
onto the wet lacquer. Motifs, usually inspired by the natural surroundings, included flowers, trees, birds, butterflies, and stylized
waves. By the end of the period, superbly decorative romanticized
landscapes had appeared, such as the one on a small Chinese-type
chest for Buddhist objects, a national treasure preserved at Mount
Koya, where the artist has depicted marsh irises and plovers in a setting reminiscent of a famous episode from he monogatari {Tales of
he), a tenth-century book of tales about poems. The perfectly balanced use of mother-of-pearl inlay, gold, silver, and pale gold (a mixture of gold and silver powders) complements the chest's rounded
lines, producing an effect of quiet opulence and soft, delicate beauty
that can be said to typify Heian aristocratic taste at its best.2
Metals and ceramics figured less prominently than lacquer in the
Heian house - in part, it seems, because it was difficult to achieve the
desired effect of softness when working with those materials. As in
earlier periods, bronze was the principal metal used, although silver
also became popular for dinnerware, boxes, and other articles. Instead of attempting to develop new metalworking techniques, craftsmen concentrated on creating simple, elegant forms and flowing designs using natural motifs. The heavy octagonal Chinese-style mirrors
of the eighth century, for example, were replaced by thin, plain circular forms, with stylized flowers for knobs and delicate decorative
designs of cranes, pine branches, butterflies, and autumn grasses.
T'ang-style three-color ceramics (sansai), a notable development
of the Nara period, were no longer produced, probably because they
lacked the subtlety demanded by the new era. The traditional
unglazed sueki, a grey, wheel-turned pottery that had been introduced from Korea around 400 c.E., continued in production, as did
a new off-white variant of the same ware, which was often glazed in
pale green and incised with graceful floral designs. But aristocratic
2 For an illustration, see YamanobeTomoyuki, Okada Jo, and Kurata Osamu, Senshoku shikko,
kinko, vol. 20 of Genshoku Nihon no bijutsu (Tokyo: Shogakukan, 1969), p. 63.
Cambridge Histories Online © Cambridge University Press, 2008
DOMESTIC ARCHITECTURE
393
taste preferred lacquered wood for objects of everyday use and imported Sung porcelains for special treasures; and the Heian period
was consequently one of general decline in the ceramic arts.
The largest, most conspicuous furnishings of a room were blinds,
screens (both folding and rigid), and curtains (both fixed and portable), which were used to partition spaces, prevent drafts, and
guard privacy. Eiga monogatari {A Tale of Flowering Fortunes), an
anonymous eleventh-century chronicle, describes articles of this
type assembled in 1023 for the coming-of-age ceremony of Princess
Teishi (1013-94), Fujiwara no Michinaga's granddaughter.
The bombycine3 lavender curtains, shading to purple toward the bottom,
were decorated with embroidered branch designs, and their Chinesebraided streamers were cluster-dyed in purple. The curtain-dais was decorated in the same manner... .The curtain-stands and folding-screen frames
were inlaid with mother-of-pearl and gold lacquer. On thefive-footscreens
there were quotations from Chinese works, inscribed in elegant Chinese
characters by Major Counselor Yukinari; and on the four-foot Chinese
damask screens, their colored-paper sections lightly tinged with purple,
there were texts in Yukinari's cursive Japanese script, the calligraphy and
underlying designs together producing an effect of indescribable brilliance
and taste. The edgings were of Chinese brocade.... The blinds were edged
in green bombycine with a large figure.4
This passage bears witness to the care with which colors and designs were coordinated, and to the continuing attraction of Chinese
material culture. It also reflects the importance attached to literature
and calligraphy, the two arts which - together with music and its
handmaid, dance - had traditionally been esteemed by Chinese
Confucianists as instruments of self-cultivation and government,
and which were regarded by Heian Japanese as supremely desirable
social accomplishments. Ladies and gentlemen were expected to
write a good hand, recognize a literary allusion, compose a creditable poem when called on, perform with skill on one or more musical instruments, be enough of a connoisseur to judge the efforts of
others, and, in the case of men, master the steps of certain dances
and sing familiar songs in a subtly original manner. So seriously
were the requirements taken - so great was the prestige of Chinese
example and so active the sponsorship in high circles - that the arts
3 Orimono, a term thought to have designated a luxurious silk fabric, possibly a changeable
damask. See William H. and Helen Craig McCullough, A Tale of Flowering Fortunes, 2 vols.
(Stanford, Calif.: Stanford University Press, 1980), p. 140, n. 26.
4 Matsumura Hiroji and Yamanaka Yutaka, eds., Eiga monogatari, vols. 75-76 of Nihon koten
bungaku taikei (Tokyo: Iwanami shoten, 1964-65), vol. 2, pp. 104-5; translation after McCullough and McCullough, Flowering Fortunes, p. 584.
Cambridge Histories Online © Cambridge University Press, 2008
394
ARISTOCRATIC CULTURE
permeated everyday life to an extent seldom, if ever, matched elsewhere. Little Princess Teishi's domestic furniture, with its inscriptions in the hand of one of the world's greatest calligraphers, may indeed be said to call into question the utility of attempting to
distinguish between the fine and the applied arts for the Heian period. We do so primarily for convenience.
TEXTILES AND COSTUMES
Damasks, brocades, and bombycines, mentioned in the description of
Princess Teishi's screens and curtains, were among the principal fabrics worn by members of the upper classes, who dressed almost exclusively in silks. Except for a brief early period, Heian weaving and
dyeing techniques were less varied than those of the eighth century,
during which the Japanese had mastered tie-dyeing, stenciling, and
batik techniques and skillfully imitated a wide assortment of continental weaves, including rich damasks, many types of brocades with
striking designs, and intricately woven chiffon-like gauzes. Heian brocades employed fewer colors and smaller designs, the gauzes were less
elaborate, and the damasks fell below Nara standards. Throughout
the Heian period, long after the cessation of official intercourse, Chinese fabrics were imported by the wealthy because of their superior
quality. But it would be a mistake to infer that Heian craftsmen were
incapable of first-class work, or that their patrons were less interested
in dress than their Nara-period predecessors. The changes represented, rather, a response to a changing aesthetic - a new interest in
subdued effects, achieved not by flamboyant polychromes but by the
graceful, flowing lines and subtle monochromes of wide-sleeved
robes, woven of luxurious but unassertive fabrics and worn in voluminous layers. The demand was for large-scale production, and the
weavers consequently avoided difficult, time-consuming methods.
Far from dismissing personal attire as unimportant, society regarded it as a major indicator of taste. Much anxious thought was
devoted to the subject, especially by women, whose beauty was
judged primarily by their dress and by the length, thickness, and lustre of their hair, rather than by their faces, which were made up according to a convention calling for chalky skin, rouged cheeks and
lips, thick painted eyebrows, and blackened teeth.5 Interest could be
5 For short discussions of Heian cosmetics, see Ivan Morris, The Wbrld of the Shining Prince
(New York: Alfred A. Knopf, 1964), pp. 203—4, a n ^ Ema Tsutomu, Yusoku kojitsu (Kyoto:
Kawara shoten, 1965), pp. 75-77.
Cambridge Histories Online © Cambridge University Press, 2008
TEXTILES AND COSTUMES
395
lent to a garment by artful shading of a color, or by the combination
of robe and lining in such a way as to suggest an aspect of nature appropriate to the season. There were, for example, wisteria robes of
pale purple lined with green, and fallen-leaf robes of yellowish
brown lined with yellow. Another alternative was to use discreet
woven or embroidered designs, such as lozenges and stylized birds
and animals; another was to accentuate texture, as with the redplum robe, in which the warp thread was purple and the woof red,
or the grape robe, which had a warp thread of pale red and a woof
of pale purple.
In each such choice, the wearer's sensitivity was at stake, but the
supreme test came when a complete costume was assembled - an
ensemble that might include, in addition to the short outer jacket
and long train prescribed for women on formal occasions, as many
as twenty or more identically cut robes, their softly blending colors
visible at the collar, sleeve openings, and hemline. As with individual
lined garments, layers were expected to harmonize with the season.
Eiga monogatari describes spring combinations worn by both sexes
at a New Year banquet held in 1025 by Grand Empress Kenshi,
Michinaga's second daughter.
After the ceremonial obeisances, Yorimichi, as Minister of the Left, led a
stately procession up the east steps of the main hall. He occupied the seat
of honor east of the south steps. The Ononomiya Minister of the Right was
next to him, and then cameTadanobu and all the others. They sat on square
cushions facing north, with the tails of their under-jackets draped over the
balustrades behind them. The color combinations of the jackets were [red]
glossed silk, willow, cherry, grape, and, in the case of the younger men, red
plum - a most delightful and glittering array.
Once seated, the gentlemen inspected the edges of the rows of blinds in
front of them. By mutual consent, each of the ladies on the other side was
wearing three of the same five color combinations - willow, cherry, yellow,
red plum, and yellowish green. Some had on five robes in each of their
three combinations, a total of fifteen; others, six or seven, amounting to
eighteen or twenty-one in all. Some were wearing Chinese damasks; others
seemed to have on bombycines that were either bound- or float-patterned,
the difference being determined by the color combination. Some of the
mantles were five-layered; others seemed to be glossed unlined garments
dyed pale green and other such colors. The jacket colors were chosen from
among the same five combinations, and the trains were decorated with
seashore patterns. The stand curtains were in red plum, yellowish green,
and cherry colors, deepening toward the bottom, and were decorated with
paintings and brilliant green streamers.6
6 Matsumura and Yamanaka, Eiga, vol. 2, pp. 177-78; translation from McCullough and McCullough, Flowering Fortunes, pp. 651—52.
Cambridge Histories Online © Cambridge University Press, 2008
396
ARISTOCRATIC CULTURE
As this passage intimates, the wearer's freedom of choice was circumscribed by the necessity of dressing in a manner recognized as
appropriate to his or her age.There were also official regulations correlating colors and fabrics with court rank. Speaking earlier of Kenshi's appointment as empress in 1012, Eiga monogatari comments:
In the past, it had been almost impossible to distinguish the various ranks
of Kenshi's attendants, who had all dressed as they pleased. Some of them
had disapproved of such laxity, but many of diose very ladies found their
prescribed costumes a source of embarrassment on the day of their mistress's elevation. The timid and conservative were obliged by the regulations
to put on bombycine jackets, whereas otfiers who had prided themselves on
dieir elegance were suddenly confronted by the devastating necessity of appearing in plain silk.7
Similarly, in Genji monogatari (The Tale of Genji), the fiction writer
Murasaki Shikibu describes the plight of Genji's young son, Yugiri,
who continues to wear the ignominious blue of the Sixth Rank after
his contemporaries have moved up to more prestigious colors.8 Such
regulations are further evidence of the great importance attached to
dress, as are the sumptuary edicts that were repeatedly promulgated
during the Heian period in a vain effort to curb extravagant displays
of the type just described.
Incense
To complete a costume successfully, it was necessary to scent the
garments. Incense for the purpose was compounded from aromatic
substances such as spikenard, sandalwood, musk, herbs, and cloves,
which were pulverized, bound together with honey or some other
sweet agent, and kneaded into hard balls. Incense making was considered an art, and one of the marks of taste was the ability to discriminate between traditional blends (which bore such names as fragrant-robe, gentleman-in-waiting, and plum-blossom), to create
original variations, and to judge the products of others.
Fans
Careful attention was also devoted to the costume's principal accessory, the folding fan, a Japanese invention dating from around the
7 Matsumura and Yamanaka, Eiga, vol. 1, p. 324; translation from McCullough and McCullough, Flowering Fortunes, p. 332.
8 Yamagishi Tokuhei, ed., Genji monogatari, vols. 14-18 of Nihon koten bungaku taikei
(1958-63), vol. 2, p. 277; Edward G. Seidensticker, trans., The Tale of Genji, 2 vols. (New
York: Alfred A. Knopf, 1976), pp. 361-62.
Cambridge Histories Online © Cambridge University Press, 2008
TEXTILES AND COSTUMES
397
tenth century. As was recognized by the Chinese, who were already
importing them in the Northern Sung period (960-1126), many of
these objects were works of art. They were of two types, one for winter use and one for summer. The former was made of joined strips
of wood, the material, color, and decoration of which were determined by the user's sex, age, and social status; the latter had paper
on one side and exposed ribs on the other, and was bound by no restrictions as to style or use. Most contemporaneous descriptions are
of the summer variety, like the one below, which appears in Okagami
{The Great Mirror), a historical tale dating from around the beginning of the twelfth century.
Most of their creations had ribs of gold lacquer, or of carved or inlaid silver, gold, aloeswood, or sandalwood; and their gorgeous paper surfaces
were inscribed with unfamiliar Chinese and Japanese poems, or adorned
with pictures of famous places mentioned in poetic handbooks. But Yukinari, with his usual flair, chose plain, tasteful lacquered ribs and yellow Chinese paper decorated with intriguingly faint pictures. On thefronthe wrote
a Chinese song in elegant formal script, and on the back a few graceful cursive lines. The emperor examined the fan many times and then treasured it
carefully in his handbox. The others he forgot after a brief show of interest.
Say what you will, nothing is better than a sovereign's approbation.9
This author's bias in favor of elegant simplicity, in support of
which he musters the highest possible social authority, reminds us
that luxurious ornamentation and elaborate conceits were not sufficient in themselves to meet the standards of exacting connoisseurs. It can perhaps be said, by way of conclusion to the foregoing
discussion of the major applied arts, that in this sphere, as in others, fastidiousness and restraint played a fundamental role in shaping taste, and, further, that those traits fostered a concern with total
effect - with the achievement and appreciation of a sophisticated
balance between the rich, glowing beauty of gold-lacquered cabinets, crimson fulled robes, and purple gossamer curtains, on the
one hand, and, on the other, the spartan bareness, clean lines, and
cool, dim interiors of the rooms in which they were displayed. A
sure sense of visual effect produced sumptuousness without vulgarity; and fresh, original designs attest to the fact that in this
realm, at least, the individual enjoyed considerable freedom of
expression.
9 Matsumura Hiroji, ed., Okagami, vol. 21 of Nihon koten bungaku taikei (i960), pp. 144—45;
translation from Helen Craig McCullough, Okagami, The Great Mirror (Princeton, N.J.:
Princeton University Press, 1980), p. 147.
Cambridge Histories Online © Cambridge University Press, 2008
398
ARISTOCRATIC CULTURE
DIET
Surrounded though they were by extraordinary natural and manmade beauty, Heian aristocrats lived far from sybaritic lives. Caftanlike robes with huge sleeves provided inadequate protection against
the winter cold of unheated houses; sanitary facilities were primitive; and mosquitoes, fleas, lice, and other insects were no respecters
of rank. Members of society sat or reclined on the floor during the
daytime and lay on the floor at night, covered by a robe or two.
Worst of all was the poverty of their diet, which caused boils and
other medical problems stemming from malnutrition, and which, in
the view of some writers, may have been at least partially responsible for the passivity and pessimism that bulk so large in Heian history and literature.10
The upper classes consumed two main meals a day, probably
around ten A.M. and four P.M., supplementing them with occasional
snacks. After brief experimentation with Chinese cuisine at the beginning of the ninth century, court society had returned almost entirely to native foods and native methods of preparation. Most of
the calories in the diet came from polished rice, which was served
in a number of ways - boiled with water, steamed and dried for
travel fare, combined with other grains and vegetables in cakes, or
thinned with water to produce a gruel, which was sometimes mixed
with red beans, chestnuts, or the like. A standard banquet dish,
sweet potato gruel, consisted of rice cooked with thin slices of sweet
potato and flavored with amazura, a sweet liquid obtained from a
vine.
The principal source of vitamins was a fairly wide variety of
vegetables - eggplant, bamboo shoots, cucumbers, miscellaneous
greens, burdock, onions, long scallions, daikon radishes, various
legumes, and so on - which seem to have been consumed chiefly in
pickled, boiled, or steamed form. Ample iodine was supplied by several kinds of seaweed.
Protein came primarily from such fish and shellfish as bonito, sea
bream, eel, carp, sea bass, mackerel, sardines, trout, whitebait,
prawns, squid, jellyfish, crabs, and clams. Because of transportation
and storage difficulties, most fish had to be dried unless they were
10 See Ishimura Teikichi, Yusoku kojitsu kenkyu, 3 vols. (Tokyo: Yusoku kojitsu kenkyu kankokai, 1958), vol. 2, p. 353; Watanabe Minoru, Nihon shokuseikatsu shi (Tokyo: Yoshikawa
kobunkan, 1964), pp. 109, 113.
Cambridge Histories Online © Cambridge University Press, 2008
DIET
399
obtained from local waters. Buddhist-inspired official edicts prohibited consumption of the flesh of such animals as dogs, monkeys,
horses, and oxen, but custom sanctioned the use of boar meat and
venison, as well as pheasants (a particular favorite) and other game
birds and their eggs. By the end of the twelfth century, however, most
members of the nobility seem to have been subsisting on largely vegetable and grain diets, supplemented by inadequate amounts of fish.
Supply problems had long since ended the early Heian consumption
of milk and two butter-like dairy products, with the result that the
average diet lacked sufficient fat.
This simple fare was rendered more palatable by the use of such
condiments and seasonings as salt, onion salt, ginger, garlic, vinegar,
miso, and fish broth, and of three sweeteners - honey, amazura, and
glutinous rice jelly. Imported sugar, when obtainable, served almost
exclusively as a drug. For special occasions, there were a number of
kinds of elaborately shaped "Chinese cakes" (karagashi), made of
one or more types of flour, often stuffed with bean jam, rice jelly, bits
of vegetable, or duck egg, and fried in sesame or walnut oil. Pears,
tangerines, persimmons, loquats, plums, pomegranates, peaches, apples, strawberries, pine nuts, chestnuts, and other fruits and nuts
were special delicacies, used, for example, to tempt the appetites of
invalids. Another rare treat was shaved ice, obtained from blocks
kept in storage chambers during the summer.
Tea, the seeds of which had been brought to Japan in 805, was
grown in and around the capital, but almost entirely for medicinal
purposes. (It was prepared for drinking by pounding the leaves,
combining them with amazura or ginger in a ball, and steeping in
hot water.) The only common beverage other than water was rice
wine (sake), which was produced in a number of unrefined varieties.
Although its alcohol content was relatively low, the wine seems to
have been highly intoxicating, possibly because of the absence of fat
in the diet; and we may assume, on the basis of much contemporary
evidence, that it and the host's presents constituted the two main attractions of the many formal banquets on the official calendaraffairs at which dishes of negligible gastronomic interest, presented
with exquisite attention to visual effect, were eaten in ceremonious
silence by richly attired gentlemen, each of whom consciously contributed to the elegance of the scene by his appearance and behavior, relaxing into informality only after the red lacquered plates and
silver cups had disappeared and the festive bowl had begun to circulate freely.
Cambridge Histories Online © Cambridge University Press, 2008
4OO
ARISTOCRATIC CULTURE
ARISTOCRATIC OCCUPATIONS AND PASTIMES
Although works like The Tale of Genji may convey a different impression, it is probably safe to say that the overriding concern of most
male Heian aristocrats was the maintenance and enhancement of
status. In economic terms, that aim meant securing a generous patron, seeking lucrative provincial governorships, and/or husbanding
and increasing private sources of income, such as shoen (rural estates), in order that the individual might house and clothe himself
and his many dependents in the discreetly luxurious manner already
described. Politically, it meant using talents and connections in an
unceasing effort to rise in the bureaucracy, and, in particular, to
achieve multiple offices of high ceremonial visibility. Socially, it
meant actively pursuing advantageous alliances with other families
and exhibiting such esteemed personal qualities as beauty, skill in
poetry, music, and calligraphy, and the ability to rise to an occasion
with a witticism or a comment showing sensitivity to the evanescence of worldly things.
Members of minor court families, barred by birth from high position, clung to modest official niches by developing expertise in, and
hereditary claims to, such specialities as precedents, ceremonies,
law, mathematics, the preparation and processing of documents, and
the rudiments of astronomy and medicine, and by assuming responsibility for routine governmental operations. It was they who kept
the bureaucracy running - and, as will be seen later, it was from their
ranks that many quasi-professional poets emerged to achieve a degree of recognition that would otherwise have been unthinkable. In
the status-bound world of the Heian court, outstanding literary ability was virtually the only avenue to relative prominence for men
whose ambitions were frustrated by the accident of birth, or by the
superior manipulative skills of their peers, but it was not a substitute
for rank and office, as can be seen by the regularity with which lowranking poets lamented their inability to climb the official ladder. As
a class, men from minor families lacked prestige. It was their superiors who set the tone of Heian culture.
For the highborn noble, whose success in life depended on circumstances only tangentially connected to his competence and diligence in office, the pursuit of status assumed forms not always readily distinguishable from the pursuit of pleasure, with the result that
he is sometimes portrayed as a carefree dilettante, enjoying himself
with whatever came to hand in the daytime, and flitting from flower
Cambridge Histories Online © Cambridge University Press, 2008
OCCUPATIONS AND PASTIMES
4OI
to flower at night. Such descriptions overlook the demands and restrictions imposed on the individual by his censorious, gossipy
peers, who expected him above all to function as a smooth cog in the
social machinery, and who had a correspondingly low tolerance for
unorthodoxy. The line between public and private behavior was ambiguous - or perhaps it would be more accurate to say that very little of a man's behavior was private in the sense of being outside the
purview of rules, whether explicit or unwritten. It was, of course, of
the utmost importance that he contribute by his attire, demeanor,
and knowledge of protocol to the successful execution of the annual
ceremonies (nenju gydji) around which court life revolved - those
great Chinese-inspired civil pageants, themselves a supreme affirmation of status, which symbolized the values and preoccupations of
the pacific Heian court as surely as tournaments of arms represented
those of its medieval Western counterparts. But the scrutiny of society was equally intense, and the demand for conformity equally insistent, on lesser occasions, including those that might impress an
observer as having been designed purely for amusement.
Excursions
Participation in the quasi-official pleasure excursions of imperial
personages and Fujiwara leaders was both an honor and an obligation, both an occasion for enjoyment and an opportunity for a man
to distinguish himself by his appearance, horsemanship, knowledge
of precedent, wit, or literary proficiency. Such events were usually
linked to the seasons, focusing in the spring on blossom-viewing at
suburban sites like Kitano and the Urin'in cloister, or in autumn on
enjoyment of the foliage at Arashiyama or some other favorite spot.
The Oi River, at the base of Arashiyama in what is now Ukyo-ku,
Kyoto, was the scene of many elegant entertainments, including an
autumn river excursion arranged by Michinaga, at which the talented Kint5 (966-1041) received the signal honor of being invited to
choose from among the three boats - "one for guests who were
skilled in the composition of Chinese verse, another for expert musicians, and a third for outstanding waka [Japanese verse] poets."11
The usual winter objective was falconry, the only form of hunting
sanctioned by the court, which was apparently prompted by the
sport's popularity to disregard the Buddhist prohibition against the
11 Matsumura, Okagami, p. 94; translation from McCullough, Okagami, p. 113.
Cambridge Histories Online © Cambridge University Press, 2008
402
ARISTOCRATIC CULTURE
taking of life. Falconry was a favorite pastime of at least eleven emperors, from Kammu in the ninth century to Shirakawa in the
twelfth] and there are many records of festive outings at which royal
spectators and their courtiers watched the activities of falconers and
dog handlers, who gradually came to be members of specialist families, versed in secret traditions and masters of elaborately ritualized
techniques. Whatever the occasion, the excursions of leading court
figures usually began with a procession through the city streets and
ended with food, wine, music, poetry, and gifts from the host.
Smaller, more private outings for similar purposes were somewhat
less formal but followed the same general pattern.
Horsemanship
Most of the retinue traveled on horseback during such pleasure
jaunts, which sometimes extended over several days but more typically were arranged so as to avoid spending the night away from the
capital. The ability to ride was also necessary if a man was to play his
assigned role in great state events like the Imperial Purification, held
on the Kamo River beach at the beginning of a new reign, or the
Kamo Festival, which was preceded by the most magnificent procession of the year. In earlier periods, riding had been closely associated with military prowess, a tradition that persisted in the warrior
class throughout the Heian centuries. Numerous members of the
upper aristocracy are also praised for their horsemanship in contemporary records - for example, Minamoto no Makoto (810-69),
who met his death in a racing accident, and such prominent Fujiwara nobles as Uchimaro (756-812), Michinaga, Tadazane (10781162), andTadamichi (1097-1164). But the connection with warlike
activities is seldom, if ever, made in such cases. Rather, there is increasing emphasis on the individual's contribution to a total visual
effect, a task requiring not only skillful horsemanship but also careful attention to the horse's appearance, to the rider's costume, and
to the burgeoning rules that prescribed the manner in which the bridle was to be held, the angle at which the whip was to be applied,
and the like.12
Racing, which had claimed Makoto's life in the ninth century,
12 See Matsumura, Okagami, p. 221 (McCullough, Okagami, p. i96);YamadaYoshio,Yamada
Tadao, Yamada Hideo, and Yamada Toshio, eds., Konjaku monogatari shu, vols. 22—26 of
Nihon koten bungaku taikei (1959-63), vol. 4, pp. 271—72; Bernard Frank, Histoires qui sont
maintenant du passe (Paris: Gallimard, 1968), pp. 156—57.
Cambridge Histories Online © Cambridge University Press, 2008
OCCUPATIONS AND PASTIMES
403
continued in favor, partly because it was associated with gambling,
but, like hawking, it became primarily a spectator sport for the upper
classes. The main races of the year were offical functions, held on the
fifth and sixth of the fifth month. Two teams of men from the guards
divisions competed on mounts provided by princes and other notables, the winners presented a musical program, and in the evening
there was feasting, with music and imperial gifts. A similar pattern
was followed on private occasions, as when an emperor or retired
emperor visited the home of a minister of state.13
Archery
Archery, another sport with strong roots in the martial past, retained
its popularity throughout the Heian period. Participation in the
mounted variety was left primarily to warriors and members of the
lower nobility, although emperors and regents demonstrated their interest by rewarding proficiency. Foot archery, which claimed the
prestige of Chinese endorsement, was considered an appropriate exercise for a gentleman. (It was one of only two forms of recreation
sanctioned for university students, the other being playing the koto.)
Numerous ninth-century princes and other personages were renowned for their marksmanship, and the sport was further recognized by the inclusion of an official contest, the Archery Ceremony
(Jarai), in the court calendar. Men of every rank were expected to
make themselves available for the two competing Archery Ceremony
teams, which performed on the seventeenth of the first month. As
time went on, however, senior nobles showed themselves reluctant to
participate, and a subsidiary contest on the eighteenth, the Bowman's
Wager (noriyumi), established itself as the focal point of attention, apparently because the higher nobles were able to watch guardsman
teams compete without being obliged to demonstrate their own skill.
There were also less formal matches, both at court and in private circles, between teams of wellborn young men. On the lighter side, such
events furnished an excuse for drinking and betting, but they were
also serious matters for the families involved, because they offered a
youth an opportunity to attract favorable notice by his dress, mastery
of techniques (and of a growing number of complicated rules), and
performance in the later musical entertainment.
13 See "An Imperial Visit to the Horse Races," McCullough and McCullough, Flowering Fortunes, pp. 631-39.
Cambridge Histories Online © Cambridge University Press, 2008
404
ARISTOCRATIC CULTURE
Wrestling
Another type of physical prowess was celebrated in the seventhmonth wrestling (sumai), a lavish summer spectacle featuring competition between two teams of champions, recruited from the provinces and trained by members of the imperial bodyguards. After the
ninth century, wrestling was almost entirely a spectator sport, but
the official matches were one of the highlights of the year, and it was
probably economic difficulties, rather than flagging interest, that led
to their disappearance from the court calendar in 1174.14
Kemari
The tendency to shun participation in strenuous sports may have
stemmed in part from the inadequacy of the upper-class diet, but the
principal explanation is doubtless to be sought in Chinese-inspired
notions of decorum, and in the increasing disposition to treat sport
as tableau. The strong emphasis on social harmony seems to have
played a role in the preference for team competition. It is significant
that the element of competition was almost completely lacking in kemari (kickball), the one notably active sport in which male aristocrats consistently indulged. Kemari was a game played with a small
deerskin ball on a hard-surfaced square court, approximately seven
meters long on each side. Eight men, of whom two were stationed
under each of four trees at the court's corners, attempted to keep the
ball in the air for as many counts as possible, using only their feet.
Counting, which began after the fiftieth kick, continued in theory to
1,000, the perfect score, but the usual score seems to have been
below 300. Every noble family had its kemari court, and the game's
enthusiasts included such prominent figures as Emperor Go-Toba
(1180-1239), the poet Saigyo (1118-90), and Saigyo's teacher, Fujiwara no Narimichi (1097-1159), the greatest of all kemari masters,
who held the high court rank of Major Counselor. Even this innocuous pastime, however, was censured by Murasaki Shikibu and
Sei Shonagon, those supremely articulate spokeswomen for naturalized Chinese values (see section entitled "Literature: Narrative
Prose"), who found it indecorous and inelegant for gentlemen to
rush around in pursuit of a ball. Given the strength of the attitude
14 For a description of the sumai festivities, see McCullough and McCullough, Flowering Fortunes, pp. 391-92Cambridge Histories Online © Cambridge University Press, 2008
OCCUPATIONS AND PASTIMES
405
the two represented, it is not surprising that upper-class amusements tended to be sedentary, and that we hear nothing of such
highly competitive activities as swordsmanship.
Feminine occupations
Most aristocratic women had little contact with, or knowledge of,
the official activities of men, or of such masculine recreations as
hawking, archery, and wrestling. They led quiet, very private lives at
home, practicing calligraphy, studying the poetic anthologies, improving their musicianship, reading or listening to stories, assembling costumes, caring for children, and keeping up a poetic correspondence with men and others in the outside world. The monotony
of their existence was broken by the companionship (often chiefly
nocturnal) of husbands or lovers, by domestic ceremonies and religious activities, and by occasional pilgrimages and sightseeing excursions, made behind the curtains of ox-drawn carriages - vehicles
used by both men and women, which accurately reflected the wealth
and social status of their owners in their size, construction, fittings,
and ornamentation.
Members of both sexes played go, character guessing, backgammon (sugoroku), and other parlor games, among which backgammon
seems to have been a particular masculine favorite. "Whenever
Michinaga and Korechika settled down to gamble [at backgammon] "Okagami says, describing what may have been a fairly typical
case, "they bared their torsos, bundled up their robes around their
waists, and kept at it until midnight or beyond. .. . Some remarkable
and very tasteful stakes changed hands."15
Monoawase
Much more to the liking of fastidious feminine writers - and of
women in general, we may suppose - were "matchings of things"
{monoawase), of which Sei Shonagon wrote in "Things That Bring
Happiness," "How could anyone help feeling happy after winning
one of those contests in which various things are compared?"16
The monoawase, ranked by a modern authority as one of the three
15 Matsumura, Okagami, p. 184; translation from McCullough, Okagami, p. 173.
16 Ikeda Kikan, Kishigami Shinji, and Akiyama Ken, eds., Makura no soshi, Murasaki Shikibu
nikki, vol. 19 of Nihon koten bungaku taikei (1958), p. 281.
Cambridge Histories Online © Cambridge University Press, 2008
406
ARISTOCRATIC CULTURE
favorite Heian pastimes (the others being falconry and kemari),17
was in many respects the amusement most characteristic of the period and the society. With the exception of the ancient sport of cockfighting, which might be called a special case, such contests were
group enterprises, in which the victory earned by one of the two
sides - identified as the Left (the socially superior) and the Right was scarcely more important than the preparation of costumes and
accessories, the scene when the entries were presented, the attendant
music, poetry, drinking, and other festivities, and the general opportunity to display taste, ingenuity, and wealth in an atmosphere of
well-bred harmony. Sometimes these elegant battles were fought
with man-made weapons, such as fans, incense balls, small boxes,
musical instruments, pictures, poems, and romances. On other occasions, the combatants turned to nature, comparing the plaintive
cries of insects displayed in dainty bamboo cages; matching the
plumage or songs of ducks, quail, warblers, or doves; arranging
sprays of spring blossoms along the borders of a garden stream or
pond; or presenting autumn plants, such as chrysanthemums or colored leaves, on "sandbar-beach tables" (suhama), so called because
of their gracefully curving tops, and because the entries were often
incorporated in a seacoast setting.
The account below conveys some of the flavor of such occasions:
the attention to symbolism and visual effect, the prominence assigned to music and poetry, and the preoccupation with status,
shown in the care with which ranks and titles were recorded and
seating arrangements noted. It describes a sweet-flag root contest
held in the middle of the eleventh century - that is, during the heyday of Fujiwara opulence and splendor, which coincided with the regency of Michinaga's son Yorimichi (992—1074; regent 1017-67).
Yorimichi was the moving spirit behind a succession of extravagant
contests, many of them nominally sponsored by one or another of
the regent's female relatives or, as here, by the emperor. The leaves
and roots of the sweet flag, or calamus, were considered to possess
medicinal properties. They figured in a number of customs associated with the Sweet-Flag Festival, held on the fifth of the fifth month
as a protection against summer diseases; and it was therefore appropriate to associate them, as the Left did, with pines, cranes, and tortoises, all symbols of longevity. Of the two empresses mentioned,
Kanshi was Yorimichi's daughter and Shoshi the daughter of his
17 Ishimura, Yusoku kojitsu kenkyu, vol. 2, p. 487.
Cambridge Histories Online © Cambridge University Press, 2008
OCCUPATIONS AND PASTIMES
407
nephew, the late emperor Go-Ichijo. We learn from another sovirce
that Kanshi's ladies-in-waiting were "brilliantly attired in sweet-flag,
China-tree, wild pink, and azalea combinations, with bordered
sleeves and gold and silver flower and bird designs," and that Shoshi's wore identical glossed silk sweet-flag robes, wild pink bombycine cloaks, mugwort Chinese jackets, and China-tea trains.18
There was a sweet-flag root contest in the imperial palace on the fifth of
the fifth month in the sixth year of Eisho [1051]. His Majesty had summoned one or two proficient senior nobles and a number of other courtiers
for an archery contest on the last day of the third month, and there had also
been cockfights, but no clearcut victory had been won, and so it had been
determined that the decision should rest on the outcome of a sweet-flag
contest.
The preparations in the apartments and grounds were the same as for the
poetry contest in the tenth month of the fourth year of Eisho [1049]. Both
empresses were present. Among those who attended were the Palace Minister Yorimune, the Minister of Popular Affairs Nagaie, the Inspector Major
Counselor Nobuie, the Ononomiya Middle Counselor Kaneyori, the Commander of the Left Gate Guards Takakuni, the Chamberlain Middle Counselor Nobunaga, the Nijo Middle Counselor Toshiie, the Master of the Empress's Household Tsunesuke, the Consultant Middle Captain of the Left
Yoshinaga, the Middle Captain of Third Rank Toshifusa, and the Lesser
Captain of Third Rank Tadaie. The members of the Left and Right teams
arrived in the evening.
First, oil was provided. Then it was time to produce the suhama prepared
by the Left and the Right. The suhama of the Left, which was four feet high,
was carried in and deposited east of the door leading to the east bay of the
south eavechamber. It depicted a seaside scene, with silver pine trees, silver
cranes and tortoises, and a silver stream flowing among aloeswood rocks.
There was a scroll on a stand in front. On the scroll paper, which was decorated with delicately edged designs, there were five colored squares, each
containing a poem. The green wrapper was decorated with silver, the roller
was of amber, and the cord was of silver. There was a green gossamer cloth
with a wave design to go under the suhama. Five long roots, twisted into
circles, were arranged on the pine trees and beside the shore.... Five medicinal balls with long multicolored streamers were arranged in circles on the
beach.
The members of the side seated themselves on the east veranda. Next,
the scorekeeper's suhama was presented by chamberlains, who carried it in
and put it down east of the principal suhama. It contained rocks and tiny
pine trees, and there were artificial sweet flags to be used as tallies.
Next, other chamberlains carried in the suhama of the Right, which held
a drumstand on a pedestal about two feet square, surmounted by a drum.
In front of the drum there were dolls, representing children performing the
18 Matsumura and Yamanaka, Eiga, vol. 2, p. 449. For the Sweet-Flag Festival, see McCullough and McCullough, Flowering Fortunes, p. 412.
Cambridge Histories Online © Cambridge University Press, 2008
408
ARISTOCRATIC CULTURE
butterfly dance; and on each root a poem had been inscribed. Everything
was made of silver. The long streamers of the gold and silver medicinal balls
were arranged in circles near the shore.
The members of the side seated themselves on the west veranda. Next,
the scorekeeper's suhama was presented. A single Chamberlain carried it in
and put it down west of the principal suhama. It held an imitation of one
of the bamboo clumps outside the imperial residence, and its tallies were
bamboo stalks.
Next, in compliance with an imperial command, the senior nobles divided up into Left and Right sides. The senior nobles of the Left withdrew
from their places, crossed to the east by way of the veranda in front of the
emperor, and seated themselves. They were the Palace Minister, Lord Morokata, Lord Kaneyori, Lord Nobunaga, Lord Tsunesuke, and Lord Toshifusa. The captains of the Left and Right, Head Chamberlain Controller
Tsuneie and Head Chamberlain Middle Captain Suketsuna, came forward
and took their places below their suhama. Meanwhile, two child scorekeepers took their places, one for each side. They were sons of LordTakakuni who were in service at the Courtiers' Hall.
Tsuneie summoned Yoshimoto and Suketsuna summoned Motoie. . . .
Tsuneie picked up a long root and handed it to Yoshimoto, who stretched
it out under the south eaves. The Right followed the same procedure, and
then the lengths were compared. The Left's root was eleven feet long
and the Right's twelve; hence the Right won. A second and a third round
followed. In each, both roots measured ten feet, but the Right's was slightly
longer, so the Right was adjudged the winner. It was decided that the contest would end with the third round.
Next, the five poems [of the Left and the five of the Right] were read. The
reciters and their assistants were Nagakata and Tsuneie for the Left, and
Takatoshi and Suketsuna for the Right. The Palace Minister was the judge.
The topics were "Sweet Flags," "The Cuckoo," "Rice Seedlings," "Love,"
and "Felicitations." Everyone returned to his original seat after the readings.
Next, his Majesty gave the command for music. The Japanese koto was
played by the Minister of Popular Affairs, the thirteen-stringed koto by the
Middle Counselor of Second Rank, the lute by Tsunenobu, the mouth
organ by Sadanaga, the flute by [missing], and the oboe by Takatoshi. The
singer was Sukenaka. After an oboe solo, the Palace Minister . . . presented
a flute to the emperor. His Majesty took it and told the Minister to use the
clappers. The Minister assented and returned to his seat. Then the song
"Ah! How August!" was sung. At the end of the song in the ritsu scale, His
Majesty presented gift robes to the senior nobles, who then withdrew. I believe I have heard that there were no imperial gifts for the other courtiers
on that occasion.19
Murasaki Shikibu, in The Tale of Genji, devotes a chapter to a contest in which two teams of ladies, representing rival imperial con19 Nagazumi Yasuaki and Shimada Isao, eds., Kokon chomonju, vol. 84 of Nihon koten bungaku taikei (1966), pp. 498-500.
Cambridge Histories Online © Cambridge University Press, 2008
SECULAR PAINTING
409
sorts, offered paintings for judgment in the emperor's presence. The
preliminary maneuverings of the consorts' supporters as they
sought to outwit the competition, and the anxious attention devoted
to costumes, boxes, mountings, and cords, remind us that the winner of such a contest was considered to have gained a significant advantage in the incessant jockeying for favor that went on among
powerful families with daughters in the palace harem. It was not
only Prince Genji's enthusiasm for art, but also his desire to protect
the interests of his protegee, Akikonomu, that impelled him to the
exertions Murasaki describes.20 Even when the stakes were lower,
any monoawase - and particularly any public one - had its serious
side, because it exposed the taste and sensitivity of the participants
to exacting scrutiny. Like the other major pastimes we have reviewed, this one was seldom taken lightly.
SECULAR PAINTING
Toward the end of Murasaki Shikibu's "Picture Competition" chapter, Prince Hotaru remarks to his brother, Prince Genji, "Our father
used to say, 'It goes without saying that Genji has mastered the art of
poetic composition. As regards the other major accomplishments, he
is best at playing the seven-stringed koto; then come the flute, lute,
and thirteen-stringed koto.' Everyone else thought the same, so I assumed painting was merely something you did for amusement."21
That rather ambiguous comment might imply that painting was regarded as an important aristocratic accomplishment - a proposition
for which there would appear to be support in The Tale of Genji itself,
where members of the imperial family are depicted as zealous wielders of the brush; and also in other works, such as Okagami and Eiga
monogatari, which describe the proficiency of leading court personages. We will probably be closer to the truth, however, if we assume
that painting was acknowledged to be a skill requiring native ability (a
point Prince Hotaru makes elsewhere in the same Genji passage), and
that members of society were under no compulsion to try to master
it. It seems to have been viewed in somewhat the same light as cooking, another hobby in which eminent gentlemen often dabbled. Like
the applied arts, it was essentially the province of professionals - men
20 See H. Richard Okada, Figures of Resistance: Language, Poetry, and Narrating in 'The Tale of
Genji' and Other Mid-Heian Texts (Durham, N.C.: Duke University Press, 1991), pp.
232-38, for a discussion of the interplay between politics and aesthetics on this occasion.
21 Yamagishi, Genji, vol. 2, p. 186.
Cambridge Histories Online © Cambridge University Press, 2008
41O
ARISTOCRATIC CULTURE
of low social status who worked in the court Painting Office, or in private or temple ateliers. Although there was some overlapping, secular
art was produced primarily by court and other lay painters; Buddhist
art, by painter-monks associated with the big temples.
Only a few Heian painters' names have survived. Their Naraperiod predecessors had been anonymous artisans, and the status of
the profession remained much the same, even after the ninthcentury rise of secular painting had brought new opportunities for
the display of individual talent. Those of whom we hear remain
shadowy figures, like Kudara no Kawanari (782-853), the first artist
to enter the historical record, who was celebrated for his realism, and
Kose no Kanaoka (fl. ca. 980), an expert painter of horses and the
founder of the long-lived Kose school. Mentions of Kawanari, Kanaoka, and others recur from time to time in Heian texts, but we get
little sense of their accomplishments, because their work, unlike that
of the professional poets, no longer survives. Of the handful of extant Heian secular paintings, none can be attributed with confidence
to a known artist.
We must likewise rely almost exclusively on literary sources for the
history of Heian secular art. A great many paintings are known to
have been executed on screens and panels (which were needed in
large numbers after the adoption of shindenzukuri architecture), and
also in small picture books and scrolls, designed primarily for feminine enjoyment. During the early ninth century, when Chinese influence was all-pervasive, styles and subjects were usually Chinese,
but Japanese themes became increasingly popular by the tenth century, and a distinction was then made between karae (pictures with
Chinese subjects) andyamatoe (pictures with Japanese subjects). As
the Heian period advanced, karae continued to be produced. The
only extant Heian landscape screen, for example, a work dating from
around 1050, is a karae preserved at the Toji Temple, depicting a
young court noble visiting an elderly recluse who is probably to be
identified with the Chinese poet Po Chu-i (772-846).Z2 But the
yamatoe became the predominant form, evolving from the karae through a process that cannot be traced with certainty - into the mature Japanese style we find in the oldest remaining examples, most
of which date from the twelfth century.
22 Dlustrated in Saburo Ienaga, Painting in theYamato Style (New York and Tokyo: John Weatherhill and Heibonsha, 1973)) plate 26; also in AkiyamaTerukazu, Emakimono, vol. 8 of Genshoku Nihon no bijutsu (1968), p. 164.
Cambridge Histories Online © Cambridge University Press, 2008
SECULAR PAINTING
411
Rise ofYamatoe
In the second half of the ninth century, the Japanese were actively
naturalizing or discarding many of the imported elements in their
everyday life, in areas ranging from food and dress to music, art,
and literature. The pivotal event of those decades was the emergence of a workable native syllabary (kana). By freeing the Japanese language from total dependence on an alien logographic writing system, kana paved the way for the remarkable social rise of
Japanese poetry - that is, of the thirty-one-syllable waka, which can
be said to have constituted the dominant element in aristocratic
culture from the tenth century on. Paradoxically, this seeming
move away from China proved to be a major step toward a truly
Chinese-style society, because it made poetic expression an integral
part of upper-class existence, instead of a self-conscious exercise in
a foreign tongue. The evolution of kana also resulted in the development of narrative prose fiction, which, if it lacked Confucian
sanction, nevertheless contributed to the literary atmosphere at
court. As we shall see, it also influenced the direction taken by
Japanese calligraphy. And it profoundly affected the content and
form of the yamatoe.
Screen and panel pictures, the earliest form assumed by the yamatoe, were not regarded as independent works of art, but, rather, as
companions to poems and as vehicles for the evocation of bittersweet
emotion. The overwhelming majority were landscape paintings with
added genre elements. They focused on the passing of time, as illustrated by the natural phenomena and human activities conventionally associated with the four seasons. Figures gathering young greens
indicated spring, as did hazy hills dotted with flowering trees; a
cuckoo pointed unmistakably to summer; deer or colored leaves to
autumn; snow or falconry to winter. Colored-paper squares, positioned to enhance the total design, contained graceful poems (or, less
frequently, prose), inscribed in flowing script, which endowed the
pictures with specific connotations.
Although the pictures have long since vanished, many of the
poems remain, helping us to visualize the content of the paintings
and to understand the manner in which man and nature were
linked. The five waka below, which may be considered typical, were
all composed in 875 for a screen behind the guest of honor at a
longevity celebration.
Cambridge Histories Online © Cambridge University Press, 2008
412
ARISTOCRATIC CULTURE
kasugano ni
wakana tsumitsutsu
yorozuyo o
iwau kokoro wa
kami zo shiruran
yama takami
kumoi ni miyuru
sakurabana
kokoro no yukite
oranu hi zo naki
mezurashiki
koe naranaku ni
hototogisu
kokora no toshi o
akazu mo aru ka na
chidori naku
sao no kawagiri
tachinurashi
yama no ko no ha mo
iro masariyuku
shirayuki no
furishiku toki wa
miyoshino no
yamashitakaze ni
hana zo chirikeru
The gods must know well
The feelings with which I pray,
"Ten thousand years"
As I pluck the tender shoots
On the plain of Kasuga.
So high the mountains
They seem to float in the sky Those cherry blossoms
My spirit visits daily.
Longing to break off a bough.
Yours is not, cuckoo,
A song we hear but rarely Why, then, should it be
That listening through the years,
We never weary of you?
Mists must be hovering
Above the Sao River
Where plovers call out,
For now the mountain foliage
Takes on ever deeper hues.
When white flakes of snow
Flutter thick and fast toward earth,
Flowers indeed scatter
Before the gale sweeping down
From fairYoshino's mountains.23
In order to form a notion of the actual appearance of these fourseasons screens and panels, we must examine the few scraps of evidence remaining from the Heian period itself, as well as comparable
examples of the later yamatoe style. Naturalized landscape backgrounds dating from the mid-eleventh century are to be seen in the
Toji karae screen and in religious door paintings at the Byodoin in
Uji; and there are numerous small-scale landscapes, including representations of screens and panels, in twelfth-century scrolls. The
pictorial designs on lacquer, ceramic, and metal objects also offer
useful hints. It seems safe to conclude, after such a review, that most
Heian yamatoe landscapes modified Chinese techniques in order to
achieve soft, delicate, romantic effects, and that the total impression
was one of elegant refinement.
23 Saeki Umetomo, ed., Kokin waka shil, vol. 8 oiNihon koten bungaku taikei (1959)) nos. 35759, 361-62. Here and below, Kokinshu translations, sometimes slightly altered, are from
Helen Craig McCullough, Kokin Wakashu: With 'Tosa Nikki'and 'Shinsen Waka' (Stanford,
Calif.: Stanford University Press, 1985). Poem numbers in Saeki are identical with those
in Shimpen kokka taikan.
Cambridge Histories Online © Cambridge University Press, 2008
SECULAR PAINTING
413
Emakimono
The appearance of the yamatoe coincided with the development of
such vernacular narrative prose forms as the romance, the diary, and
the poem tale (see section entitled "Literature: Narrative Prose")- It
was not long before romances and similar works began to inspire
paintings, which sometimes took the form of screen decorations but
most often appeared as booklets (soshi) or horizontal, hand scrolls
(emaki[mono]) - small treasures for highborn ladies, who gazed at
them while attendants read from related texts or told stories of their
own invention. The horizontal scrolls were made of sheets of paper
pasted together and attached to a mounting at one end and a roller
at the other. Quite apart from the aesthetic value of their paintings,
they were objects of art in their own right, with braided silk cords,
richly colored mountings, rollers made of jade, crystal, or precious
wood, and textual passages inscribed in exquisite calligraphy on
paper flecked with silver and gold. That they were favored over the
plainer soshi is suggested by the frequency of their mention in works
like The Tale of Genji, and by the fact that all the principal surviving
Heian yamatoe are in emakimono form.
Two main types of Heian secular narrative emakimono exist today,
known respectively as onnae (feminine pictures) and otokoe (masculine pictures). Both probably derive from Chinese antecedents,
through a process that cannot be reconstructed, and both are closely
associated with native literary genres - onnae with romances, kana
diaries, and poem tales; otokoe with an anecdotal, supposedly factual, ultimately oral genre of very short stories called setsuwa. The
oldest extant set of onnae scrolls, the Genji monogatari emaki {The Tale
of Genji Picture Scrolls), is a masterpiece that obviously represents
the culmination of a long line of development. There is also ample
literary evidence to show that onnae were probably being produced
by around the middle of the tenth century, and that they were extremely numerous and popular from at least the eleventh century
on. No comparable information exists for otokoe, but it does not necessarily follow that the otokoe was a much later phenomenon, as has
sometimes been maintained.24 Rather, in view of the high artistic
24 Tokyo National Museum, Painting 6th-14th Centuries, vol. 1 of Pageant of Japanese Art
(Tokyo: Toto Shuppan, 1957), p. 37; Seidensticker, Genji, pp. 307-17; Hideo Okudaira,
Narrative Picture Scrolls (New York and Tokyo: Weatherhill and Shibundo, 1973), p. 29;
Dietrich Seckel, Emakimono: The Art of the Japanese Painted Hand-Scroll (London: Jonathan
Cape, 1959), p. 25.
Cambridge Histories Online © Cambridge University Press, 2008
414
ARISTOCRATIC CULTURE
level of the oldest known examples, it seems best to postulate an ancestry of considerable antiquity. We may conjecture that the lack of
references to otokoe in the writings of court women, our best sources
in such matters, is due not to their absence from the society but to
a feminine preference for pictures illustrating literature of a different
kind.
Very soon after the appearance of The Tale of Genji in the early
eleventh century, that great work of fiction was canonized as the
supreme embodiment of the Heian spirit, and as a magisterial exposition, in particular, of the central aesthetic concept known as mono
no aware, which may be roughly denned as deep but controlled emotional sensitivity, especially to beauty and to the tyranny of time. The
book was read, re-read, quoted, imitated, explicated, and illustrated
by generations of admirers; and it is surely no coincidence that it furnished the subject matter for what is not only the earliest surviving
set of onnae scrolls but also, in the view of some scholars, the finest
achievement in the history of Japanese painting. As is true of most
emakimono, the four Genji monogatari emaki scrolls contain both
pictures and textual passages. Traditionally ascribed to the court
painter Fujiwara noTakayoshi (fl. ca. 1147), they are now recognized
to have been produced by different painting ateliers and different
calligraphers, probably around the 1120s or 1130s. In its present incomplete state, the set includes paintings of nineteen separate
scenes, almost all of which are laid in shindenzukuri apartments and
adjoining verandas. The strong, dramatic parallel lines of railings,
partitions, and lintels contrast magnificently with the richly colored
robes, screens, curtains, and blinds; and the stylized human figures,
with slit eyes and hooks for noses, are represented in static, pensive
poses, perfectly attuned to the majestic pace and melancholy tone of
Murasaki's work, and to the planes and masses of the total composition. As in the society of which it is a microcosm, there is small
place in this romantic, dreamlike world for ill-bred assertions of individuality or violent outpourings of emotion. Prince Genji's face is
impassive when he holds his wife's infant son by another man, and
only the tilt of his head hints at his feelings.25
Two outstanding examples of the otokoe technique survive, both
dating from around the middle of the twelfth century. The three
scrolls of the first, the Shigisan engi emaki (Shigisan Legend Picture
25 Akiyama, Emakimono, plate 2; also Ivan Morris, The Tale of Genji Scroll (Tokyo: Kodansha
International, 1971), facing p. 54.
Cambridge Histories Online © Cambridge University Press, 2008
CALLIGRAPHY AND PAPER
415
Scrolls), depict miraculous events associated with the monk Myoren
(fl. tenth century); the subject of the second, the Ban dainagon emaki
(Ban Major Counselor Picture Scrolls, 3 scrolls), is the Otemmon
palace gate fire of 866, said to have been set by the court noble Tomo
(or Ban) no Yoshio (809-68) in an attempt to discredit a political
rival. The Shigisan engi emaki stories unfold in a swift, cinematic
style, characterized by robust realism and freely flowing brushwork.
Colors appear only in thin washes, and the human figures are almost
all members of the lower classes, who reveal their feelings of alarm,
amazement, or joy through exaggerated facial expressions and hand
and foot movements. Most of the action takes place outdoors. The
Ban dainagon emaki makes more conspicuous use of color and devotes more attention to the upper classes, but its vigorous realism,
boisterous crowds of uninhibited commoners, and arson and fisticuffs remove it, too, very far from the feminine milieu of the Genji
monogatari emaki.26
Of other emakimono remaining from the late Heian period, the
most important are the first two scrolls of the Choju jimbutsu giga
(Bird, Animal, and Human Caricatures), a work consisting of four
scrolls in all. The two Heian scrolls (mid-twelfth century?), which
may be from a single hand, contain drawings of monkeys, rabbits,
and frogs mimicking humans (scroll 1) and sketches of horses, oxen,
roosters, lions, dragons, and other real and imaginary creatures
(scroll 2); the other two (early thirteenth century) repeat the subjectmatter of scroll 1, with additional scenes of monks and laymen gambling (scroll 3), and depict monks and laymen engaged in frequently
enigmatic activities (scroll 4). The symbolic intent, if any, is not understood. All four are executed almost exclusively in ink, using a
technique derived from Buddhist copybooks; and the first and second are particularly noteworthy for their skillful composition, lively
realism, and fluent, powerful brushwork.27
CALLIGRAPHY AND PAPER
The Choju jimbutsu giga scrolls, unusual in so many other respects,
are also among the few Heian emakimono without calligraphic sections. All the other major narrative scroll sets contain brief textual
passages, which were probably intended not so much to inform the
26 For illustrations, see Akiyama, Emakimono, plates 22-30 (Shigisan engi emaki) and plates
31-35 (Ban dainagon emaki).
27 Illustrated in Akiyama, Emakimono, plates 36-40.
Cambridge Histories Online © Cambridge University Press, 2008
416
ARISTOCRATIC CULTURE
viewer, who could have been expected to know the stories, as to enhance the visual effect and reinforce the connection between painting and the prestigious arts of literature and calligraphy.
Although Chinese writing styles had been competently copied
earlier, Japanese study of calligraphy as an art began with the great
religious leader Kukai (or Kobo Daishi, 774-835), who absorbed the
major T'ang styles during his fourteen months in China. The square
(kaisho), running (gyosho), grass (sosho), and other styles sponsored
or introduced by Kukai after his return were essentially those perfected by the legendary Wang Hsi-chih (32i?-7i?) and his son Wang
Hsien-chih (344-88). They provided the foundation for what was
later known as the Chinese style (karayd); and Rukai and two of his
contemporaries, Emperor Saga (786-842) and Tachibana no Hayanari (d. 842) - the so-called Three Brushes (sampitsu) - were recognized as the style's best early practitioners.
Meanwhile, the Japanese were continuing the experimentation
that was to lead ultimately to the modern hiragana syllabary. Their
first step had been to adopt a bewildering variety of Chinese characters for the phonetic rendering of proper nouns, poems, and the
like. Such characters, called man'ydgana because of their prominence in the eighth-century poetic anthology Man'yoshu {Collection
of Ten Thousand Generations), continued to be used by early Heian
writers, who often set them down in the grass style, producing a
form called sogana. By around the second half of the ninth century,
a relatively small number of sogana were being further streamlined,
a process that led to the creation, by the early eleventh century if not
before, of a syllabary {kana) similar to the modern one except for the
survival of numerous alternative forms for most sounds. It was primarily in that syllabary, known also as the woman's hand (onnade),
that poetry, romances, and women's diaries were written. {Katakana, the other modern syllabary, which dates from the same general period, developed as a utilitarian system of notation outside the
mainstream of Heian culture.)
There is evidence that kana were used to some extent by men during the period of gestation, but most scholars assume that the lead
was taken by women, for whom the study of Chinese characters was
considered unsuitable, and who consequently needed a script for the
letters and poems that bulked so large in their daily lives. It would
be a mistake, however, to conclude that the suppleness and elegance
of the script thus developed were qualities that appealed only to feminine taste. The first great calligrapher of the tenth century, Ono no
Cambridge Histories Online © Cambridge University Press, 2008
CALLIGRAPHY AND PAPER
417
Michikaze (orTofu, 894-966), wrote Chinese characters in a graceful, simple manner that was already significantly different from the
dignity and vigor of the orthodox style followed by the Three
Brushes. The naturalization process was carried further by Fujiwara
no Sukemasa (944-98), who was born about fifty years after Michikaze; and it reached maturity a generation later with Fujiwara no
Yukinari (972-1027). Although Yukinari is bracketed with his two
predecessors as one of the Three Calligraphers (sanseki) in the Japanese style (wayo), he is by far the most notable figure of the three,
because it was his gentle, smooth brushwork that became the classic
model, transmitted for generations by the immensely influential
Sesonji school.
Yukinari's name is also associated with the classic kana style,
which achieves an effect of great fluidity and elegance by linking individual kana in a long series of graceful loops and curves. The best
examples of kana calligraphy in its golden age are the Koyagire (Koya
fragments, ca. 1100?), a group of scrolls and fragments containing
portions of the first imperial poetic anthology, Kokinshu (Collection of
Early and Modern Times).^ Two later works, the Genji monogatari
emaki and Sanjurokunin shu (Collection of Thirty-Six Poets, ca. 1110-
20?), are celebrated for the superbly decorative manner in which
they combine the arts of calligraphy, literature, painting, and paper
making.
Sanjurokunin shii, in particular, utilizes many different kinds of fine
paper-heavy, white domestic michinoku; Chinese rosen, decorated
with wax designs; numerous kinds of domestic and imported karakami (Chinese paper), the kana paper par excellence, decorated with
mica paste designs and gold and silver dust - juxtaposing different
colors and textures with great verve and originality. One white sheet,
for example, is decorated with an overall silver wave pattern and a
picture showing an island, a boat, and wild geese in flight. An irregular brown band of differently textured paper has been added in the
approximate center, with its own design of gold dust and plume
grass, and there is a dark brown accent at the bottom of the page,
studded with bits of gold leaf.29
Similarly opulent paper was used for religious purposes, as in the
case of the well-known Heike nokyo (Taira Family Dedicatory Sutra).
28 Illustrated in Ozawa Masao, ed., Kokin waka shu, vol. 7 of Nihon koten bungaku zenshu
(Tokyo: Shogakukan, 1971), p. 6.
29 For illustrations of Genji monogatari emaki and Sanjurokunin shu paper and calligraphy, see
Ienaga, Painting, plates 95-97; also Morris, Genji Scroll, passim.
Cambridge Histories Online © Cambridge University Press, 2008
418
ARISTOCRATIC CULTURE
The thirty-three scrolls in this set of the Lotus Sutra, presented to
Itsukushima Shrine in 1164 by the powerful Taira family, are inscribed on paper lavishly decorated with gold and silver, dainty floral patterns, and under-paintings; and the karae and yamatoe illustrations, gold and silver fittings, mother-of-pearl inlays, and crystal
rollers make the viewer feel that "their splendor and sumptuousness
[suggest] collections of elegant verses rather than sutras," as Eiga
monogatari comments of a similar set.30
It is not surprising that Heian aristocrats should have copied sutras in much the same spirit as poems, with equal attention to calligraphy, paper, and general artistic effect. To a considerable extent,
religion itself was regarded as an aesthetic experience - a source
of material and spiritual benefits, to be sure, but also an opportunity
to delight the senses with the gorgeous ecclesiastical vestments, the
solemn massed chants, the clouds of fragrant incense, and the
pageantry that were associated with esoteric rituals, in particular. Sei
Shonagon expressed what must have been a common opinion when
she said, "A preacher ought to be handsome. Otherwise, his ugliness
leads us into sin by encouraging us to let our attention wander."31
And since members of court society were the principal patrons of
the great temples, their tastes inevitably affected the development of
Buddhist painting and sculpture.
BUDDHIST ART
As with secular screen paintings and emakimono, we know from literary sources that Buddhist art was produced in huge quantities
during the Heian period. The two esoteric sects, Shingon and
Tendai, which dominated the early religious scene, required at least
one new icon for each of the innumerable special rituals their monks
performed day in and day out for aristocratic patrons; and groups of
as many as a thousand paintings or statues were commissioned repeatedly by wealthy believers, especially in the eleventh and twelfth
centuries.
Most of those works are gone, but a substantial number remains - fifty or sixty statues from the ninth century alone. Often pre30 Matsumura and Yamanaka, Eiga, vol. 2, pp. 43-44; McCullough and McCullough, Flowering Fortunes, p. 532. For Heike nokyd illustrations, see Ienaga, Painting, plate 98; also Robert
Treat Paine and Alexander Soper, The Art and Architecture of Japan, Pelican History of Art
(Baltimore: Penguin Books, 1955), plate 57; and H. Minamoto, An Illustrated History of
Japanese Art (Kyoto: K. Hoshino, 1935), plate 79.
31 Paraphrased from Ikeda, Makura no soshi, pp. 73-74.
Cambridge Histories Online © Cambridge University Press, 2008
BUDDHIST ART
419
served in isolated temples, where they were relatively safe from
spreading fires, warfare, and other dangers, such statues and paintings enable us to trace broad stylistic developments with comparative assurance. Art historians have usually treated them in terms of
two periods, Jogan (or Konin, or Konin Jogan, after ninth-century
era names) and Fujiwara, with Fujiwara representing the kind of
courtly taste we have been discussing. In the view of some scholars,
the Fujiwara period begins as early as 894; in that of others, as late
as 980. To avoid ambiguity, we shall speak here of centuries, or of
early Heian (ninth century), mid-Heian (tenth and eleventh centuries), and late Heian (twelfth century).
Sculpture
For Buddhist art, as for other aspects of Heian culture, Chinese influence predominated at the beginning of the ninth century. Sculpture imitated mid- and late-T'ang models, and through them the art
of southern India, continuing a late eighth-century trend away from
the realism and classic repose of the Nara masterpieces. The material was almost invariably a single block of wood, frequently embellished with polychrome decoration. The style, copied from imported
statues and pattern books, was characterized by somber facial expressions; stout, almost corpulent bodies; formal, stylized poses;
powerfully carved drapery swirling in abstract designs; and a suggestion of sensuous languor, imparted by half-closed eyes, full lips,
and swelling flesh.
Although it is possible to detect a certain degree of naturalization
well before the end of the ninth century, all of the above traits are
present not only in the JingojiYakushi Nyorai, which probably dates
from around 800, but also in works attributed to the mid- and lateninth century, such as the Kanshinji Nyoirin Kannon and the Hokkeji Eleven-Headed Kannon.32 Moreover, they are still to be found
in statues of the early eleventh century. This continental style persisted, in short, well after the eclipse of Chinese poetry and Chinese
fashions in calligraphy and secular art-partly because glowering
faces and powerful torsos were suitable attributes for the Mystic
Kings and other fierce deities who figured prominently in esoteric
32 For the JingojiYakushi Nyorai, see Minamoto, Japanese Art, plate 41, and Kurata Bunsaku,
Mikkyo jiin to Jogan chokoku, vol. 5 of Genshoku Nihon no bijutsu (1967), plate 27; for the
Kanshinji Nyoirin Kannon, Kurata, Mikkyo, plates 63-64; for the Hokkeji Eleven-Headed
Kannon, Minamoto, Japanese An, plate 48, and Kurata, Mikkyo, plates 72-73.
Cambridge Histories Online © Cambridge University Press, 2008
42O
ARISTOCRATIC CULTURE
rituals; partly because of the conservatism nurtured by iconographic
pattern-books and by the strictly formulaic approach of the Shingon
sect. In the end, however, it failed to withstand the challenge of innovations catering to the aristocratic preference for refinement,
mildness, and luxurious decorative effects.
When members of the court circle were in need of immediate divine assistance - for example, to secure a promotion, ensure a safe
journey or an uneventful childbirth, resolve a land dispute, or recover from an illness - they turned to the esoteric sects, the native
gods, the yin-yang prognosticators, and their own special protective
buddhas and bodhisattvas. When they contemplated the afterlife (a
frequent practice in a society preoccupied with ephemerality), they
found solace chiefly in the hope of rebirth in Amida's Pure Land
paradise. Worship of Amida, designed to secure forgiveness for sins
and assistance for the dead, had existed in Japan for centuries, leaving its artistic mark most notably in the Golden Hall frescoes at the
Horyuji, where Amida's Pure Land, among others, was depicted in a
style whose Chinese antecedents can be seen at Tun-huang.
In 985, the Tendai monk Genshin (942-1017) gave new prominence to Amidism with a treatise, Ojd yoshu (Anthology on Rebirth in
Pure Land), which made the cult an alternative, rather than an adjunct, to other forms of Buddhism; and which contained vivid descriptions of the beauty and compassion of the Buddha and his attendants, and of the pleasures of the Pure Land.33 As depicted in Ojo
yoshu, the Pure Land was a place where the ear was delighted with
music, the nose with fragrance, and the eye with drifting blossoms,
crystal pools, golden palaces, and exquisite raiment - an idealized
counterpart, we might say, of the Heian capital.
Around the beginning of the eleventh century, thanks in large part
to Genshin, Amidism began to move religious art and architecture
in a new direction. The esoteric and other sects continued to produce works to meet their own needs, but lay believers increasingly
commissioned serenely benevolent images of Amida and his bodhisattvas, which they installed in tile-roofed halls decorated with
gold, jewels, lacquer, and mother-of-pearl. The most magnificent of
many such undertakings was Michinaga's Hojoji Temple, where the
grounds and buildings were consciously designed to suggest the glories of the Pure Land, and where the central icon, a sixteen-foot
33 For a partial translation, see A. K. Reischauer, "Genshin's Ojo Yoshu: Collected Essays on
Birth into Paradise," Transactions of the Asiatic Society of Japan, second series, 7 (1930):
16-97.
Cambridge Histories Online © Cambridge University Press, 2008
BUDDHIST ART
421
statue of Amida, "shining with peerless holy radiance," conformed in
every feasible respect to the glittering description of the Buddha in
Ojoyoshu. s4 Another was the Byodoin at Uji, founded by Michinaga's
sonYorimichi, where the main image, seated Amida, still survives in
the original airy, elegant Phoenix Hall (H56do), a building that combines shindenzukuri architectural features with those of edifices in
Tun-huang Pure Land paintings. Executed in 1053 by Jocho (d.
1057), a sculptor who had won acclaim earlier for his contributions
to the Hojoji, the statue is a noble work, its graceful, well-proportioned body, soft drapery, and tranquil, dignified face combining
with a sumptuously decorated golden halo to produce an effect of
the utmost beauty and refinement.35
As we might expect, the Phoenix Hall Amida did not spring fullblown from Jocho's inventive genius. Its ancestry can be traced back
through at least one hundred years of evolution. Nor can we assume
that it was hailed as the outstanding masterpiece of its day, since we
hear much more about the vanished splendors of the Hojoji from
contemporaneous writers. Nevertheless, it is a work of unique importance, not only for its intrinsic quality, but also for its effect on
the subsequent course of Heian sculpture. Admiration assumed the
form of imitation, with the result that the last hundred and fifty years
of the Heian period witnessed no new developments, but merely the
gradual debasement of Jocho's style into weak conventionalism.
There was mass production of bland images, assembled from many
small wooden parts (a technique Jochd had perfected), and decorated with bright colors, intricate patterns, and cut gold in a manner
that occasionally bordered on vulgarity, as in the Joruriji Kichijoten,
with its elaborately simulated textile designs and its multifarious
streamers, bangles, and hair ornaments.36
Painting
Before the rise of Amidism, there were two main categories of Heian
Buddhist painting, both intimately associated with the Shingon and
Tendai sects. One was the mandala, a symbolic representation of the
34 Quotation from McCullough and McCullough, Flowering Fortunes, p. 567. For the Ojo
yoshu description, see Allan A. Andrews, The Teachings Essential for Rebirth: A Study of Genshin's Ojbyoshu (Tokyo: Sophia University, 1973), pp. 13-14.
35 Illustrated in Paine and Soper, Art and Architecture, plate 37; also in Kudo Yoshiaki and
Nishikawa Shinji, Amidado to Fujiivara chokoku, vol. 6 of Genshoku Nihon no bijutsu (1969),
plates 1, 4-5.
36 Illustrated in Paine and Soper, Art and Architecture, plate 35B.
Cambridge Histories Online © Cambridge University Press, 2008
422
ARISTOCRATIC CULTURE
universe. In the most common type of mandala, scores of miniature
divinities were grouped hierarchically in geometric patterns around
a central image of Vairocana (Dainichi), of whom they were regarded as manifestations. A variety of decorative effects was achieved
by executing the figures in fine gold or silver lines on dark blue
paper, or in red outlines complemented by a palette of bright colors;
and there were also variations in the types of circles and rectangles
employed to make up the total composition. But a basic uniformity,
the result both of the tiny sizes of the figures and of detailed iconographic regulations, may be said to limit the aesthetic interest of the
mandala.37
Awesomely energetic fierce deities are the representative subjects
of paintings belonging to the other category, which is characterized
by an iconic, expansive, forbidding style paralleling that of contemporaneous sculpture. With the possible exception of the massive
Myooin Red Fudo (which art historians classify merely as "early"),
no major ninth-century work in this vein survives, but excellent examples remain from the next century and a half- for example, the
Boston Museum's Daiitoku and the Sh5ren'in Blue Fudo - showing
that the style, protected by strong conservative forces, was able to
maintain its integrity until well after the first manifestations of
heightened interest in Amida and the Pure Land.38
Meanwhile, at temples like the H5joji and the Byodoin, painters
were exploring a new subject, the descent of Amida and his heavenly
host to escort the dying believer to paradise. Pictures of this kind,
called raigo ("coming to welcome"), were not unknown in China,
but it was in Japan that they became a major element in Pure Land
art. Relatively unencumbered by iconographic considerations,
painters created compositions reflecting aristocratic taste: richly attired bodhisattva musicians riding on purple clouds, with a golden
Amida in the central position and a yamatoe landscape at the bottom. The oldest extant raigo paintings, a group executed in 1053 on
walls and doors at the Byodoin, are now badly worn and faded, but
protected sections, uncovered in the course of twentieth-century repairs, have revealed bright yellows, oranges, reds, blues, purples, and
37 For a short discussion of mandalas, see Akiyama Terukazu, Japanese Painting (New York:
Rizzoli International Publications, 1977), pp. 37-40; for illustrations, Takada Osamu and
Yanagisawa Taka, Butsuga, vol. 7 of Genshoku Nihon no bijutsu (1969), plates 38-42, 45-53.
38 For the Myooin Red Fudo, see Takada and Yanagisawa, Butsuga, plate 59; for the Boston
Museum Daiitoku, Takada and Yanagisawa, Butsuga, plate 68, and Akiyama, Japanese
Painting, p. 54; for the Shoren'in Blue Fudo, Takada and Yanagisawa, Butsuga, plates 56-57,
and Akiyama, Japanese Painting, p. 55.
Cambridge Histories Online © Cambridge University Press, 2008
BUDDHIST ART
423
greens, as well as ornamental designs produced by the application of
cut gold leaf (kirikane), a new technique that was to assume increasing prominence during the remainder of the Heian period. As
noted earlier, the pines and hills in the Byodoin raigo are among the
oldest extant fragments of yamatoe landscape painting.39
The greatest Heian raigo, known as the Koyasan triptych, consists
of three hanging scrolls, which appear to have originated as a single
painting. Brownish clouds, believed to have been lavender, shape the
composition into an ellipse, indented at the lower left to admit an
autumnal mountain scene. The bodhisattvas, outlined in clean red
lines, have skins delicately flushed with pink or tan, and their graceful white streamers repeat the undulating motions of the clouds, as
do the bands of red, blue, green, orange, and purple formed by their
patterned robes. Extensive and extremely sophisticated use of gold,
especially for the central figure and its halo, supports the attribution
of this work to the late Heian period.40
Mildness, delicacy, grace, and decorativeness, qualities ideally
suited to the content of raigo paintings, played an increasing role in
Buddhist art as a whole during the last hundred and fifty Heian
years. One of the finest works of the late mid-period is a hanging
scroll at Mount Koya, dating from 1086, which represents the death
ofSakyamuni, the historical Buddha. The background is occupied by
hills and water in the yamatoe style. In the center, the Buddha lies in
tranquil grandeur, many times life-size, framed by graceful sala trees
in full bloom, and surrounded by mourners, who spill over into the
foreground. Like the aristocrats in onnae emakimono, the divine personages show their grief only through slight gestures, if at all; like the
commoners in otokoe, the human figures shriek and weep without
inhibition. A sorrowing lion writhes in the lower right-hand corner;
a calm Queen Maya, the Buddha's beautiful mother, hovers in the
upper right. Painstaking attention has been devoted to color contrasts and harmonies, and to the patterns in the mourners' luxurious
robes, which closely resemble mid-Heian descriptions of upper-class
attire, but the commanding presence of the Buddha and the grief of
the mourners serve as a counterpoise to the decorative elements,
and the total effect is one of great vitality and textural richness.41
39 Illustrated in Kudo and Nishikawa, Amidado, plate 7.
40 Illustrated in Akiyama, Japanese Painting, pp. 46-47; also in Takada and Yanagisawa, Butsuga, plates 115, 118.
41 Illustrated in Takada and Yanagisawa, Butsuga, plates 1-5; see also the discussion in
Akiyama, Japanese Painting, p. 49.
Cambridge Histories Online © Cambridge University Press, 2008
424
ARISTOCRATIC CULTURE
In many other cases, the modern viewer feels that dignity, compositional strength, and religious feeling have been sacrificed to surface glitter and fussy detail. Images of the bodhisattva Fugen, a cult
figure worshiped particularly by women, appear against a background of falling blossoms, wear elaborate polychrome costumes
and innumerable bracelets, necklaces, and pendants, and ride elephants whose saddles are embellished with designs resembling the
intricate patterning of Persian miniatures.42 The once powerful bodies of fierce deities are swallowed up in near-abstract designs, where
the focus is on kirikane, tiny patterns, and the interplay of planes of
reds and browns.43 The Peacock King, another esoteric divinity, is
festooned with dozens of green and gold loops, and the feathers of
his mount become layers of contrasting colors, each with its own
complex pattern traced in gold.44
If we are tempted to accuse some of these paintings of excessive
ornamentation, or even of garishness, we may remind ourselves that
they were intended to be seen in dusky surroundings, not in brightly
lit museums or on the glossy pages of art books. But impressions of
insipidity and sentimentality are harder to dismiss. Without wishing
to deny the many attractive qualities of late Buddhist painting, we
must agree with those who find that here, as in the sculpture of the
same period, conservatism and overrefinement have resulted in
decadence. It is not to the temples that we can profitably turn for the
best in twelfth-century Japanese art.
MUSIC
In the four centuries immediately preceding the Heian period, the
introduction of many kinds of foreign instrumental music and
dance brought revolutionary new aesthetic experiences to the Japanese upper classes, for whom music had previously meant simple
vocal performances and dances, with or without flutes, bells, drums,
and six-stringed kotos by way of accompaniment. The native tradition survived, thanks to an intimate association with tenacious
magico-religious beliefs and practices, but it was the importations
that bulked largest in the official and private lives of the nobility by
the start of the ninth century.
The earliest arrivals had been sankangaku (music of the three Ko42 SeeTakada and Yanagisawa, Butsuga, plates 14-16.
43 Takada and Yanagisawa, Butsuga, plate 29.
44 Takada and Yanagisawa, Butsuga, plates 73-74.
Cambridge Histories Online © Cambridge University Press, 2008
MUSIC
425
reas), mixed Chinese and Korean styles from the Korean states of the
Three Kingdoms period (313-668).Then came gigaku, introduced in
612, which seems to have consisted mainly of satirical dances, performed with masks suggestive of a Central Asian or Western origin.
Dora-gaku and rin'yugaku, from short-lived unidentified countries
that may have been located in Southeast Asia, appeared in the eighth
century, as did Bokkai-gaku, the music of the Tungusic state of Pohai, which occupied parts of eastern Manchuria, the Russian Maritime Province, and northern Korea between 700 and 926. And from
around 685 on, there was piecemeal importation of togaku, the flourishing, cosmopolitan music and dance ofT'ang China.
Gagaku
Of three principal types of T'ang music, two found their way to
Japan. One was yen-yueh (Japanese, engaku), "banquet music," a formal, dignified amalgam of many elements, including folk songs and
dances and the music of Central Asia, India, and other areas, which
the Chinese used for court entertainments. The other was san-yueh
(Japanese, sangaku), "scattered music," a popular, relaxed form of
entertainment, which was accompanied by juggling, acrobatics, stiltwalking, and the like.45 The third, ya-yueh (Japanese, gagaku), "elegant music" used at rituals and ceremonies to ensure the harmonious functioning of the Confucian state, was not imported - it
being felt, apparently, that sufficient resources for such purposes
existed - but it gave its name to the Gagakuryo (Bureau of Elegant
Music), a government office established in 701. Although the original mandate of the Gagakuryo covered all forms of court music, native music was transferred to a new office, the Outadokoro (Folk
Music Office), around the end of the eighth century; and we shall
therefore use the common but ill-defined term gagaku to designate
only foreign or foreign-style music and dance of the type supervised
and performed by the Gagakuryo after the separation.
By the beginning of the ninth century, the Chinese engaku (usually called wgaku, "T'ang music") had become by far the most important of the gagaku genres. Rin'yiigaku, which was prized as representative of Indian music, occupied second place; the music of the
three Koreas and Po-hai followed at a considerable distance; and gi45 For details, see James T. Araki, The Ballad-Drama of Medieval Japan (Berkeley and Los
Angeles: University of California Press, 1964), pp. 50-54.
Cambridge Histories Online © Cambridge University Press, 2008
426
ARISTOCRATIC CULTURE
gaku, doragaku, and sangaku were well on the way to their subsequent disappearance from court life. The native music was still flourishing because of its unique acceptability to the Shinto gods.
Around 835 the sinophile court, dissatisfied with the chaotic state
of the gagaku corpus, set out to make music and dance more "correct." Retired Emperor Saga, an enthusiastic amateur who is said to
have been an expert at several Chinese instruments, supplied the initial leadership in what became one of the major musical developments of the Heian period, a process of gradual consolidation and
reorganization lasting for more than a century. Non-Chinese pieces
were recast in the Chinese mold; fragmentary compositions were
fleshed out; large numbers of exotic instruments were discarded;
scores were rewritten to eliminate all but two structural types and six
main modes; and new works were composed, both by professionals
and by prominent members of the court.
As the decades passed, systematization shaded imperceptibly into
naturalization. We can seldom be sure when a given change took
place, but it is apparent that gagaku was much more reflective of
Heian aristocratic taste in the late tenth century, at the end of the reorganization process, than it had been in 800. By that time, in addition to the changes just mentioned, the gagaku repertoire, including
the dance repertoire (bugaku), had been divided into two paired categories. The first, tdgaku or sagaku (Music of the Left), consisted of
the old tdgaku and rin'yugaku pieces, plus Japanese compositions in
the Chinese vein; the second, komagaku (Korean music) or ugaku
(Music of the Right), of drastically revised pieces from Korea and
Manchuria. The categories were the provinces of two troupes of professional musicians, who played approximately the same kinds and
numbers of instruments - the transverse flute, oboe, mouth organ,
lute, thirteen-stringed koto, and drums for the Left; Korean flute,
oboe, and drums for the Right - and who dressed in sumptuous harmonizing costumes, red for the Left and green for the Right. Dances
were performed in pairs, with, for example, a Left bird dance matching a Right butterfly dance, or a Left masked warrior a Right masked
warrior.
Another aspect of the reorganization was the establishment of regulations assigning music an integral role in all the main court ceremonies and rituals. The two orchestras and their dancers also figured
prominently in the entertainments accompanying wrestling matches,
horse races, poetry contests, and other competitions, in which the
contesting sides were likewise designated as Left and Right. And, as
Cambridge Histories Online © Cambridge University Press, 2008
MUSIC
427
noted below, gagaku was an essential element in the elaborate Buddhist services sponsored by members of the nobility.
This kind of emphasis, reinforced by the keen personal interest of
successive sovereigns, created a musical atmosphere at court and led,
early in the ninth century, to the inauguration of the custom of gyoyu
(or on-asobi, "august music"), woodwind and string gagaku music
played for recreation at the imperial residence by the emperor and
other gentlemen of the court. There is much literary evidence to
show that the performances of talented amateurs were considered
fresher and more elegant than stereotyped renditions of the same
compositions by professional musicians, and that musical competence was expected of every member of society. The sons of high nobles, taught by court musicians to dance and play the major gagaku
instruments, were given ample opportunity to display their skills,
both in childhood, when they were the featured performers at
longevity celebrations for older members of the family, and on innumerable later formal and informal occasions. Among the lower nobility, musical expertise came to be especially vital for members of
the palace guards {efu), who found themselves constant participants
in gyoyu because of their proximity to the imperial person.
Ironically, the popularity of gyoyu, a type of performance that permitted at least a degree of spontaneity and originality, seems to have
contributed materially to the ultimate fossilization of gagaku. In the
tenth century, the musical preeminence of the guardsmen was recognized by the creation of a new organ, the Gakuso (or Gakusho,
Music Office), staffed with efu members, which took over the functions of the Gagakuryo. By the early eleventh century, Gakuso posts
were hereditary, and by the twelfth the office was a bastion of conservatism, with each family jealously guarding its secret lore and all
alike insisting on the inviolability of tradition and on the status of
music as a quasi-mystic Way - an art that sanctioned subtle refinements but strictly prohibited innovations like new compositions. It
may be said, in short, that gagaku, the principal Heian musical form,
followed a course similar to the one we have already traced for Buddhist art: an initial phase of sinitic vigor, an intermediate stage of elegant refinement, and a final period of decline.
Shinto and Secular Vocal Music
At the Heian court, the traditional Japanese preference for vocal
music was partially satisfied by the performance of Shinto sacred
Cambridge Histories Online © Cambridge University Press, 2008
428
ARISTOCRATIC CULTURE
songs - notably the long cycle known as mi-kagura, a carefully structured combination of poetry, music, and dance, perfected at the
start of the eleventh century, which focused on the performance of
two choruses, singing to welcome, entertain, and send off a divine
visitor.46
A number of new secular song forms also appeared. One was the
saibara (a name of uncertain meaning), which entered aristocratic
circles around the beginning of the Heian period and enjoyed its
greatest vogue in the early eleventh century. Simple Japanese lyrics often of folk origin, and characterized by the inclusion of meaningless syllables {hayashikotoba) to adjust the rhythm - were set to gagaku melodies and sung in a drawn-out style to the accompaniment
of gagaku instruments.
A second genre, the rdei (recitation), also flourished around the
beginning of the eleventh century. As the name suggests, it was more
recitation than song. Rdei lyrics usually consisted of a pair of sevenword lines from a familiar Chinese poem, rendered in a combination
of Chinese and Japanese; and the performance style, which resembled round singing (with or without musical accompaniment), made
the form particularly appropriate for social occasions. Two wellknown collections of lyrics attest to the popularity of the rdei: Wakan
rdei shu {Collection of Japanese and Chinese Rdei, ca. 1011?), by Fujiwara no Kinto, and Shinsen rdei shu {Newly Selected Collection of Rdei,
ca. 1107-23?), by Fujiwara no Mototoshi (1056-1142).
The last major secular vocal genre to appear was the imayd[uta]
(modern style [song]), which seems to have originated around the
last quarter of the tenth century, and to have been fairly well known
by the early eleventh. The melodies were taken from a few favorite
gagaku compositions, and many of the lyrics were adaptations of simple Japanese Buddhist liturgical pieces {wasan). The heyday of the
imayd was the second half of the twelfth century, when the songs
were sung both by courtiers and by female professional entertainers,
such as dancers, puppeteers, and courtesans, who sometimes served
as music teachers to high-born students. The greatest aristocratic aficionado was Retired Emperor Go-Shirakawa (1127-92; r. 1155-58),
who compiled an imayd collection, Rydjin hisho {Secret Selection of
Songs), held imayd contests, including one that lasted for fifteen
nights, and personally consoled an ailing eighty-three-year-old entertainer by singing imayd and reciting the Lotus Sutra at her bedside.
46 For details, see McCullough and McCullough, Flowering Fortunes, pp. 410-11.
Cambridge Histories Online © Cambridge University Press, 2008
MUSIC
429
Buddhist Music
The recognition achieved by female imayo singers presaged the approach of the medieval era, when Japanese culture and aristocratic
culture would no longer be virtually synonymous terms. Among the
many factors responsible for the new order, the most conspicuous
were the decline of the nobility and the ascension of the military, but
an almost equally important part was played by Buddhism, which,
throughout the Heian centuries, quietly lent its support to types of
popular entertainment that were destined to exert a profound influence on the mainstream of medieval music and literature. For example, temple patronage helped to preserve sangaku, one of the ancestors of the no drama, after its banishment from court; and it also
sustained the itinerant blind reciters known as biwa hoshi (lute
monks), who developed the great body of medieval oral literature
known to us today as Heike monogatari {The Tale of the Heike). Since
the biwa hoshi remained a subterranean plebeian element in Heian
culture, they do not fall within the purview of this chapter, but it
may be noted that they, like the imayo performers, were harbingers
of the medieval renaissance of vocal music (and the concomitant
eclipse of gagaku), and that their chanting style was deeply indebted
to Buddhist vocal forms, which, like Buddhist art, constituted one
aspect of aristocratic culture.
Heian Buddhist vocal music, known as shomyo, consisted primarily of liturgical music, sacred texts, and eulogies, all of which were
sung or chanted by monks - at first in Chinese styles introduced by
the patriarchs of the esoteric sects, and later in naturalized styles,
perfected especially by the Pure Land monk Genshin. The chanted
forms, which influenced the later Heike recitations and no, were relatively simple; the songs, like the secular saibara, decorated a single
syllable with many notes. Both were complemented by instrumental
music and dances from the gagaku repertoire, performed as an integral part of the services and also afterward as entertainment.
Contemporaneous writers have little to say about the musical aspects of Buddhist rituals before the last quarter of the tenth century,
but the rise of Amidism seems to have nurtured a new interest in the
relationship between music and salvation. It became common practice to call in small groups of musicians to perform compositions of
the kind the believer expected to hear in the Pure Land; and also to
send off the dying with woodwinds, strings, and song - partly, perhaps, in the hope of summoning Amida and his attendants, as the
Cambridge Histories Online © Cambridge University Press, 2008
43°
ARISTOCRATIC CULTURE
shaman's koto had summoned the gods in ancient times.47 The intimate connection perceived by the middle and late Heian mind
between music and the Pure Land is further apparent in the prominence assigned to musicians and their instruments in raigo paintings,
and in the comments of writers like Murasaki Shikibu, who compares Prince Genji's singing voice (on a purely secular occasion) to
the warbling of a kalavihka bird in paradise.48 We have already noticed that the layouts, architectural features, images, and furnishings
of temple compounds like Michinaga's Hojoji were intended as
earthly replicas of sights to be seen in the Pure Land; and the same
spirit is discernible in the magnificent rituals staged at such religious
institutions, in which music played a central role. The point is made
quite explicitly in the famous seventeenth chapter of Eiga monogatari, an account of the H5joji Golden Hall dedication, which bears
the title "Music":
Five or six imposing monks, dressed in red and green robes and surplices, began to clear people out of the way with a great show of vigor. Marshals arrived, and then came the Lecturer and Reader riding in litters, with
Censors, officers of the Bureau of Buddhism and Aliens, and others from
the two ministries walking before them on the left and right, as though for
a Golden Light Sutra lecture. Heralded by a tremendous burst of fast music
from the Music Office [Gakusho] orchestra, a lion danced out leading a
cub. The spectacle as all awaited the Emperor seemed part of another
world.
Next the monks filed in from the south gallery, forming lines on the left
and right; and tears came to the eyes of the speechless spectators at the
sight of that great multitude of holy men moving forward in unison, each
group headed by a marshal. The monks' costumes varied in accordance
with their offices - Clear-tone Singers, Tin-staff Chanters, and the like.
Those who wore patchwork surplices had imported them from China especially for the dedication, and the colors shone widi all the vivid freshness
of ropes of gems, creating an effect of great dignity and splendor. Incense
smoldered in silver and gold censers, filling the compound with the scents
of sandalwood and aloeswood, and blossoms of many hues scattered from
the sky. . . .
Innumerable bodhisattva dances were presented on the platform, and
children performed butterfly and bird dances so beautifully that one could
only suppose paradise to be little different - a reflection that added to the
auspiciousness of the occasion by evoking mental images of the Pure Land.
There were peacocks, parrots, mandarin ducks, and kalavihka birds, and
the harmonies produced by the Music Office were utterly delightful - true
voices of the dharma. Those who listened felt as though the singers were
47 Ogi Mitsuo, Nikon kodai ongaku shiron (Tokyo: Yoshikawa kobunkan, 1977), pp. 144-45.
48 Seidensticker, Genji, p. 132.
Cambridge Histories Online © Cambridge University Press, 2008
POETRY
431
the celestial beings and the holy multitude, lifting their voices in praise of
the Buddha's teachings.49
(On the construction and dedication of the H6j5ji, see the section
"New Imperial and Fujiwara Buildings" in Chapter 2.)
The collapse of court power in the late twelfth century ended aristocratic sponsorship of such lavish religious spectacles. It also radically altered the way of life in which music had figured so prominently. Of all the major Heian musical forms, the Buddhist shomyd
alone weathered the passage into the new era, surviving to become
the ancestor of almost every new native secular vocal genre. The others, vocal and instrumental alike, receded into obscurity.
LITERATURE: POETRY
In poetry, as in other aspects of court life, the first half of the ninth
century was a period of direct imitation of Chinese models.50 The
composition of Chinese poems {kanshi) was an integral part of
many official functions, symbolizing the authors' cultivation and the
emperor's status as foremost patron of the arts; and three imperial
kanshi anthologies, compiled between 814 and 827, attest to the importance attached to such activities. Meanwhile, Japanese poetry
disappeared almost completely from public life. If there were gifted
waka poets in the first Heian decades, their names have not been
preserved. We know, however, that the waka was undergoing significant changes behind the scenes.
During the seventh and eighth centuries, native poets had produced impressive long compositions (choka) of great lyric intensity
and technical complexity, notably the elegies and encomiums of
Kakinomoto no Hitomaro (fl. late seventh century), who seems to
have served as an official bard. But there was no longer a demand for
the services of a Hitomaro by the beginning of the ninth century.
Moreover, although the waka survived in private - protected by the
inertia of custom and, we may conjecture, by the desire of women to
demonstrate at least one type of literary competence - it tended to
degenerate into little more than a pawn in the half-joking game of romantic intrigue, "nothing but empty verses and empty words . . . the
province of the amorous," as the poet Ki no Tsurayuki (873?-945?)
49 Matsumura and Yamanaka, Eiga, vol. 2, pp. 71-72; translation from McCullough and
McCullough, Flowering Fortunes, pp. 556-57.
50 See Helen Craig McCullough, Brocade by Night: 'Kokin Wakashu' and the Court Style in
Japanese Classical Poetry (Stanford, Calif.: Stanford University Press, 1985), chap. 3.
Cambridge Histories Online © Cambridge University Press, 2008
432
ARISTOCRATIC CULTURE
complained.51 With the development of kana, it also became an instrument of casual social intercourse, functioning as the main element in a note of invitation, thanks, or condolence, or as a well-bred
response to a moving incident, or as a witty escape from an awkward
situation. And because the choka was obviously less well adapted to
such uses than the thirty-one-syllable tanka (short poem), the tanka
became the standard waka form.
When the waka reemerged after its so-called dark age, therefore,
serious writers like Tsurayuki confronted difficulties unknown to
Hitomaro and his contemporaries. The tanka was, in effect, the sole
approved form, and the frivolous function of poetry as social grace
and pastime was firmly established. As will be shown, there were
other formidable handicaps as well. It is easy to perceive ways in
which Heian poets fell short of their Man'yoshu predecessors; less
easy, perhaps, to appreciate the skill, ingenuity, and dedication with
which the best of them met the challenge of a new era, or to do full
justice to the remarkable edifice they erected on an unprepossessing
foundation.
The waka began to reappear in public around the middle of the
ninth century.52 Its return was associated with a number of other
developments, among them the general revival and refinement of traditional interests and values; the resurgence of the hereditary principle, which diminished the utility of a Chinese education; the perfection of kana; and the increasing tendency of the great families to
seek power through their feminine representatives in the imperial
harem, each of whom was the potential mother of a malleable child
sovereign. In particular, the buildings where the consorts lived,
known collectively as the rear palace (kokyu), were becoming centers
of musical, artistic, and literary activity. It seems to have been from
their luxuriously furnished apartments that folding screens decorated with Japanese poems spread to the public parts of the palace
(ca. 850-900), and it was their mistresses' interest in elegant competitions, noted earlier, that probably contributed most significantly to
the birth of the poetry contest, one of the major cultural phenomena
of the Heian period.
There was a sharp spurt in the demand for formal waka during
the last fifteen years of the ninth century. The Japanese screen poem
(bydbu uta) entered its century-long heyday, poetry contests were held
51 Saeki, Kokin viaka shu, p. 97; McCullough, Kokin Wakashu, p. 5.
52 McCullough, Brocade by Night, chap. 4.
Cambridge Histories Online © Cambridge University Press, 2008
POETRY
433
with increasing frequency, and the waka began to supplant the kanshi
at banquets and other official functions. Quasi-professional poets, of
whom Tsurayuki became the most highly esteemed, arose from the
ranks of the lesser nobility, using their literary prowess to forge ties
with the great, and working in many ways to elevate the status of native verse. Early in the tenth century, this activity culminated in the
compilation of the first imperial waka anthology, Kokin[waka]shu,
which was edited by Tsurayuki and three other minor-bureaucrat
poets and submitted to the throne in or around 905.
Tsurayuki supplied the anthology with a preface that compared
the waka to a plant "with the human heart as its seed and innumerable words as its leaves," reviewed the form's illustrious history, discussed its social and political virtues, and otherwise sought to establish it on an equal footing with Chinese poetry.53 "Thanks to this
collection," he ended in a burst of rhetoric, "poetry will survive as
eternally as water flows at the foot of a mountain; thanks to the assembling of these poems in numbers rivalling the sands of a beach,
there will be no complaints of the art's declining as pools in the
Asuka River dwindle into shallows; there will be rejoicing for as long
as it takes a pebble to grow into a mighty boulder."54 His predictions
erred on the side of optimism, but the basic aims of the compilers
were more fully realized than they could have dared to hope.The exalted sponsorship of Kokinshu, together with its precedent-setting
status, made the waka the nation's supreme literary form and established compositional norms that endured well into the nineteenth
century. The occasional radical challenge was turned back, and other
new developments remained little more than variations on a theme.
The Kokinshu style became the orthodox style in a society where orthodoxy was almost all.
Some of the 1,111 Kokinshu poems, like the anonymous composition below, are simple, straightforward expressions of emotion, similar in construction and tone to the many folk-influenced tanka preserved in Man'yoshu.
sugaru naku
aki no hagihara
asa tachite
tabi yuku hito o
itsu to ka matan
How long must I wait
To see again the traveler
Who leaves this morning,
Journeying where wild bees hum
In autumn bush-clover fields?55
53 Saeki, Kokin waka shu, p. 93; McCullough, Kokin Wakashu, p. 3.
54 Saeki, Kokin waka shu, p. 103; McCullough, Kokin Wakashu, p. 8.
55 Saeki, Kokin waka shu, no. 366.
Cambridge Histories Online © Cambridge University Press, 2008
434
ARISTOCRATIC CULTURE
The style usually identified with the anthology's name is more complex, however, and much more aristocratic in tone. As modern
scholarship has shown, it reveals strong influence from the ninthcentury kanshi, andfromthe kanshi's Chinese models.56 It is natural
to suppose that masculine members of the court circle, who were
constantly endeavoring to compose kanshi, would have begun to use
continental conceits and techniques in the waka they exchanged
with women, and that the women would have responded in kind particularly because the kanshi style in favor was a witty, indirect
one, suited to "the fencing about the truth that characterized the romantic intrigues of the day."57
Early-ninth-century kanshi were imitations of sixth-century Chinese southern-court poetry - that is, of the style of the latter part of
the Six Dynasties period (220-589), which feigned inability to distinguish between such natural phenomena as plum blossoms and snow,
made extensive use of figurative language, and delighted in drawing
rational conclusions, which were often based on cause-and-effect relationships. The examples below are from two typical poems by Yii
Hsin (513-81), "In a Boat, Gazing at the Moon" and "The Mirror."58
The mountains are bright; one wonders if snow might have fallen.
The shore is white, but not because of sand.
A moon appears in which there is no cinnamon tree;
Afloweropens, but it will not follow spring.
Many compositions of the same kind appear in Japanese kanshi
collections dating from the early ninth century, and there is evidence
to show that the waka was developing along similar lines. Just as Yii
Hsin pretends to confuse moonlight with snow and sand, so Ono no
Takamura (802-53), who, significantly, is best known as a kanshi
poet, professes himself at a loss to tell plum blossoms from snow:
hana no iro wa
Although your blossoms
yuki ni majirite
Elude our gaze, their color lost
miezu tomo
Amidflakesof snow,
ka o dani nioe
Send forth, at least, your fragrance,59
hito no shirubeku
That men may know you are here.
And just as Yti Hsin makes moon and flower metaphors for a
beautiful woman, so Ariwara no Narihira (825-80) uses flowers
(meaning cherry blossoms) in an indirect accusation of fickleness:
56 The definitive study is Jin'ichi Konishi, "The Genesis of the Kokinshu Style," trans. Helen
C. McCullough, Harvard Journal of Asiatic Studies 38, 1 (June 1978): 61-170.
57 Konishi, "Genesis," p. 164. 58 Konishi, "Genesis," pp. 85, 102.
59 Saeki, Kokin waka shu, no. 335.
Cambridge Histories Online © Cambridge University Press, 2008
POETRY
kyo kozu wa
asu wa yuki to zo
furinamashi
kiezu wa ari tomo
hana to mimashi ya
435
Had I not come today
They would have fallen tomorrow
like drifting snowflakes.
Though they have not yet melted,
They are scarcely true flowers.60
Six Dynasties techniques are omnipresent in the Kokinshu
"Spring," "Summer," "Autumn," and "Winter" books, a large subcollection in which the natural phenomena of the four seasons are
closely associated with human emotion - especially with the feeling
of sadness evoked by the passing of time. Such seasonal poems,
which tended to be formal and public in nature, must have been
available to the compilers in comparatively large numbers, and preference was no doubt given to compositions with an up-to-date Chinese tone. But it should be noticed, in order to view the Chinese contribution in perspective, that the influence of the imported style is
weaker in "Love," the second of the two major categories in the anthology. We may surmise that authentic love poems of the kind we
have postulated, which were private in nature, were harder to come
by than seasonal verses, and, further, that many of them were too
personal for inclusion in a public anthology, where individualism and
strong expressions of emotion were no more permissible than in an
onnae painting. Like the examples that follow, most "Love" compositions are elegant, gracefully subdued statements of longing, with
the speaker's feelings closely linked to nature, as in seasonal poems,
but with the "reasoning" style less apparent, if it is present at all.
michinoku no
asaka no numa no
hanakatsumi
katsu mini hito ni
koi ya wataran
Shall I always love
Someone I have scarcely met A girl as pretty
As an Asaka marsh iris
Blooming in Michinoku?61
awanu yo no
furu shirayuki to
tsumorinaba
ware sae tomo ni
kenubeki mono o
If, like these snowflakes,
The nights when we fail to meet
Should accumulate,
I, too, must surely perish
Along with the melting drifts.62
Although the source of these two compositions is unknown, their
plaintive, passive, somewhat impersonal quality is duplicated in
many others that can be shown to have originated as screen poems verses in which the sentiments belong not to the authors but to fig60 Saeki, Kokin waka shu, no. 63. 61 Saeki, Kokin viaka shu, no. 677.
62 Saeki, Kokin waka shit, no. 621.
Cambridge Histories Online © Cambridge University Press, 2008
436
ARISTOCRATIC CULTURE
ures in the paintings, and where the intent is not to intrigue the
reader with clever reasoning but to hint at a romantic situation. A
screen poem serves, in effect, as a clarification of the picture it
accompanies, a reversal of the process by which an emakimono amplifies a tale (rnonogatari). Kokinshu poems of the byobu uta type,
closely associated with women and their interests, are among the
least public in the anthology. Nevertheless, they are clearly formal in
nature, adhering to the conventions of a genre that discouraged
originality and circumscribed artistic freedom by predetermining
topics, themes, and imagery, and by requiring a tone of quiet harmony and moderation, in keeping with the idealized beauty of the
pictures and the tastes of the screens' owners.
There is also a strong resemblance between Kokinshu poems and
compositions preserved in the more than 470 surviving Heian contest records. From its inception around 885 until the late eleventh
century, the poetry contest (uta awase) was not a literary event but a
social function, comparable in structure and emphasis to the monoawase described earlier, in which the aim was to spend a pleasant
evening in an atmosphere of opulence, taste, and friendly rivalry.
The judge's comments were bland, inoffensive, and as objective as
possible. The first poem of the Left, the side of superior status, was
invariably the winner. Neither side ever lost by an embarrasing margin, and the contest never ended in a tie, which would have made it
impossible for winners and losers to present appropriate musical entertainments. A poem mentioning the gods was assured of a win, as
was one into which the poet had managed to introduce an auspicious sentiment. Unconventional or inadequate treatment of the assigned topic constituted grounds for defeat, as did illogicalities,
flights of fancy, unorthodox imagery, and indecorous, inelegant, or
inauspicious language. Personalism was so assiduously avoided that
even phrases like waga yado, "my dwelling," came to be proscribed.
From around 960 on, if not before, acceptable precedents were required for departures from customary practice-for example, the
use of archaic or unfamiliar diction, or the failure to include a word
like "water" or "bank" in the same context with fujinami, "wisteria
wave," a term for cascading clusters of wisteria blossoms.
It is natural that the conception of poetry as game should have
made its influence felt with particular strength in the area of contest
rules. What may seem less understandable is thatTsurayuki, the first
Heian writer to advocate the reinstatement of the waka as serious
literature, should have given tacit approval to essentially the same
Cambridge Histories Online © Cambridge University Press, 2008
POETRY
437
criteria. There is no discernible difference in tone, subject matter, or
form between the ninety-five known contest poems in Kokinshu,
which are representative of their genre, and the remaining 1,016.
The contest judges, it appears, were merely expressing a consensus
about the nature of the public poem. Just as a screen poem was required to harmonize in topic and sentiment with the delicate beauty
of a yamatoe painting, so a contest poem struck the right note at a
well-bred gathering only if it was graceful, conventional, and readily
comprehensible.
If a high price was paid for the public rehabilitation of the waka,
the many admirable poems in Kokinshu and later anthologies show
how much of value was obtained in return.63 Moreover, literary considerations were never ignored altogether, even by the social contest
judges, who commented regularly on diction, conception, freshness,
and similar matters, often with considerable acumen. Particular
weight was attached to auditory effect, because contest poems were
apprehended through the recitations of the competing spokesmen.
Compositions were criticized for sounding awkward, crabbed, unpolished, and "unintelligible" (iishirenu, used of archaisms, colloquialisms, and complicated word plays, all of which were also said to
"grate on the ears"); they were praised for elegant cadences, refined,
flowing effects, and clear, serene, rhythmic beauty. That the same
considerations were present in the minds of Tsurayuki and his colleagues is apparent from the melodic flow and intricate sound patterns of Kokinshu compositions like the two below - attributes that
unfortunately disappear in translation.
yo no naka ni
taete sakura no
nakariseba
haru no kokoro wa
nodokekaramashi
hisakata no
hikari nodokeki
haru no hi ni
shizugokoro naku
hana no chiruramu
If ours were a world
Where blossoming cherry trees
Were not to be found,
What tranquillity would bless 64
The human heart in springtime!
On this springtime day
When the celestial orb
Diffuses mild light,
Why should the cherry blossoms
Scatter with unquiet hearts?65
There are Kokinshu poems richer in imagery, more original in
conception and treatment, more complex, and more moving than
63 The best study in English on this subject is Robert H. Brower and Earl Miner, Japanese
Court Poetry (Stanford, Calif.: Stanford University Press, 1961).
64 Saeki, Kokin waka shu, no. 53. 65 Saeki, Kokin waka shu, no. 84.
Cambridge Histories Online © Cambridge University Press, 2008
438
ARISTOCRATIC CULTURE
these two by Ariwara no Narihira and Ki no Tomonori (fl. ca. 890),
but none falls on the ear more pleasingly, and none is more representative of the main elements in the style - the intellectualizing, the
formal, public authorial stance, the association of man with nature,
the expression of elegant conventional sentiments about a topic with
strong overtones of romantic beauty, the sensitivity to the passing of
time, and the attention to sound.
In all, seven imperial anthologies were compiled in the Heian
period: Kokinshu (ca. 905), Gosenshu (Later Selection, ca. 951),
Shuishu (Collection of Gleanings, ca. 1006), Goshiiishu (Later Collection
of Gleanings, 1087), Kin'ydshu (Collection of Golden Leaves, ca. 1126),
Shikashu (Collection of Verbal Flowers, ca. 1151), and Senzaishu (Collection of a ThousandYears, ca. 1188).
The second and third, Gosenshu and Shuishu, which closely resemble Kokinshu in style, were both commissioned during the high
noon of aristocratic culture. The other four, which incorporate some
noteworthy innovations, coincided in time with the long twilight of
the court, the hundred-year period during which the emperors and
their Fujiwara regents struggled with ever diminishing success to
maintain the traditional way of life. The principal poetic development of the century was the emergence of a number of gifted writers who regarded composition as a high calling, and who developed
independent ideas about the manner in which the potentialities of
the tanka form could best be realized.
The fourth anthology, Goshiiishu, departed from tradition by introducing a new style of descriptive lyricism, typically focusing on
bleak autumn and winter landscapes, as in this poem:
mishi yori mo
Still more desolate
are zo shinikeru
Than when I saw it of old:
isonokami
Isonokami
aki wa shigure no
Where the rains of late autumn
furimasaritsutsu
Shower and shower again.66
Kin'ydshu and Shikashu included many poems of the same general
type, among them outstanding examples of what was coming to be
known as the lofty style, a compositional mode in which the objective was to convey an impression of power and dignity, usually
through the presentation of a panoramic scene in which no human
presence except that of the poet-observer intruded. The Kin'ydshu
66 Shimpen kokka taikan, ed. Shimpen kokka taikan henshu iinkai, 10 sections, each composed of text and index volumes (Tokyo: Kadokawa shoten, 1983-92). Goshiiishu, no. 367.
Cambridge Histories Online © Cambridge University Press, 2008
POETRY
439
compiler, Minamoto no Shunrai (or Toshiyori, io57?-ii29), and his
father, Tsunenobu (1016-97), were outstanding practitioners of the
lofty style, as may be seen from the two poems below, the first by
Tsunenobu and the second by Shunrai.67
yu sareba
As night settles down
kadota no inaba
There comes a rustling of leaves
otozurete
In home paddy fields
ashi no maroya ni
And the whistling autumn wind
akikaze zo fuku
Visits huts with reed-thatched roofs.
uzura naku
In the autumn dusk
mano-no-irie no
Waves of plume grass come rippling
hamakaze ni
Blown by the shore wind
obana namiyoru
At Mano-no-irie
aki no yugure
Inlet of quail's plaintive cries.
Tsunenobu and Shunrai were also prominent disputants in hot arguments about poetic practice - controversies pursued with especial
vigor at the literary poetry contests by which the old social competitions had been replaced around 1085. Tsunenobu tried to broaden
the scope of the waka, principally by introducing themes from rural
life; Shunrai advocated a closer connection between poetry and
everyday experience, as well as enlargement of the tiny Kokinshu
word hoard (about 2,000 items) through acceptance of colloquialisms and Man'yoshu vocabulary. Kin'yoshu and Shikashu contain
numerous poems reflecting their views and those of like thinkers,
but the imperial patrons of the two anthologies apparently balked at
admitting what is perhaps the most famous "progressive" poem of
its type, a "Love" tanka in which Shunrai uses two conspicuously
crude images - dog and crow, symbolic in popular lore of fidelity
and infidelity, respectively. The poem has been preserved in Shunrai's private collection.
iisomeshi
Do your feelings now
kotoba to nochi no
Match the speeches you uttered
kokoro to wa
When our love began?
sore ka aranu ka
Or might they be different?
inu ka karasu ka
Are you a dog or a crow?68
Fujiwara no Shunzei (or Toshinari, 1114-1204), the Senzaishu
compiler, was an outstanding poet and contest judge who wielded
tremendous authority during the last decades of his long life. He refused to countenance liberal views like Shunrai's, insisting that the
67 Shimpen kokka taikan, Kin'yoshu, no. 173 and no. 239.
68 Shimpen kokka taikan, Samboku kikashii, no >"—
Cambridge Histories Online © Cambridge University Press, 2008
440
ARISTOCRATIC CULTURE
abundant overtones and delicate nuances of the traditional vocabulary constituted an indispensable resource for poets seeking to make
significant statements in thirty-one syllables. At the same time, he
rejected the preciousness, mannerisms, and shallow cleverness that
marred the Kokinshu legacy, espousing instead the strongest qualities of the evolving descriptive style. In the best Senzaishu poems, a
melancholy, lonely, austere tone is combined with the use of compression and association to achieve rich complexity and pregnant
ambiguity. One of Shunzei's own compositions from the anthology
can be said to exemplify what might be called the Senzaishu compromise - the blending of "old words and new feeling," as Shunzei
put it, which constituted the last great poetic achievement of the
Heian period.
yu sareba
As evening falls,
nobe no akikaze
From along the moors the autumn wind
mi ni shimite
Blows chill into the heart,
uzura naku nari
And the quails raise their plaintive cry 69
fukakusa no sato
In the deep grass of secluded Fukakusa.
Here Shunzei uses familiar diction to express "awareness of mutability" (mujokari), the most familiar of all Japanese poetic sentiments. But the brooding, nostalgic tone of his poem (heightened by
the reference to Fukakusa, a place apart from the capital and its life,
and known moreover as a burial ground) differs strikingly from the
romanticism and facile grace of the two poems by Narihira and
Tomonori on the same theme.
Without pretending to explain a complicated literary development
in a sentence, we may venture the statement that objective circumstances were far more hospitable to lighthearted poses at the beginning of the tenth century, when not a cloud obscured the brilliance
of the court's prosperity, than at the gloomy end of the twelfth. In the
same connection, it should be noted that the decline of the court had
destroyed the cultural influence of the salons maintained by imperial consorts and princesses during the Fujiwara hegemony - in other
words, from around 850 until the last quarter of the eleventh century.
The tastes of women like Michinaga's daughters Shoshi, Kenshi, and
Ishi, who married three successive sovereigns between 999 and 1018,
had helped to shape almost every aspect of the culture we have been
reviewing, and thus to nurture an environment to which composition
69 Shimpen kokka taikan, Senzaishu, no. 259; translation from Brower and Miner, Japanese
Court Poetry, p. 17.
Cambridge Histories Online © Cambridge University Press, 2008
NARRATIVE PROSE
441
in the light, witty, social Kokinshu style was peculiarly appropriate. It
is unlikely to have been an accident that Goshuishu, the first imperial anthology to depart from that style, appeared at just the moment
when the salons were fading from the scene.
LITERATURE: NARRATIVE PROSE
The salons made an equally significant contribution to the history of
narrative prose, the only other noteworthy Heian belletristic form.
One genre, the setsuwa, mentioned earlier, remained outside their
purview, but all the others were affected in varying degrees: the
poem tale, the literary diary, the miscellany, the romance, and the
historical tale.
As with secular art, relatively little survives of what must have
been a corpus of impressive dimensions. (Some eighty titles of romances have been preserved from the ninth and tenth centuries
alone.) The poem tale (uta monogatari), a brief, elegant anecdote
about aristocratic life, centering on one or more poems, is represented by three short anonymous tenth-century collections, the best
of which is the oldest, he monogatari (Tales of he) - a classic expression of Heian aesthetic ideals, a poetic handbook, and a compendium of types of courtly love, written in simple, chaste language.70
All three appear to have been compiled by men.
The author of the oldest extant literary diary, Tosa nikki (Tosa
Journal, 935), was also a man, the poet Ki no Tsurayuki. Perhaps because the usual masculine diary of the day was a nonliterary Chinese
record of matters useful for bureaucratic and family reference,
Tsurayuki assumed the persona of a woman in setting down his account, which is a brief history of a fifty-five-day journey from
Shikoku to the capital, dominated by sixty poems.71 Of several surviving later Heian works in this rather misleadingly named genre, by
far the best is Kagero nikki (Gossamer Journal, ca. 982?), by "Michitsuna's mother," a secondary consort of the regent Kaneie (929-90),
which describes, with powerful realism and tight thematic unity, the
misery and resentment of a neglected wife.72 Izumi Shikibu nikki
70 For a translation of Ise monogatari, see Helen Craig McCullough, Tales of Ise: Lyrical
Episodes from Tenth-Century Japan (Stanford, Calif.: Stanford University Press, 1968).
71 Tosa nikki is translated in McCullough, Kokin Wakashu, pp. 263-91.
72 Translated in Edward Seidensticker, The Gossamer Yean: The Diary of a Noblewoman of
Heian Japan (Rutland, Vt., and Tokyo: Charles E. Turtle, 1964). See also the partial translation in Helen Craig McCullough, ed., Classical Japanese Prose: An Anthology (Stanford,
Calif.: Stanford University Press, 1990), pp. 102-55.
Cambridge Histories Online © Cambridge University Press, 2008
442
ARISTOCRATIC CULTURE
(Diary of Izumi Shikibu, ca. 1004?), ascribed to the poet Izumi Shikibu (b. 976?), is a lyrical, vaguely melancholy, poem-studded account
of the protagonist's love affair with an imperial prince.73 Murasaki
Shikibu nikki (Diary of Murasaki Shikibu, ca. 1010), a short memoir
by the author of The Tale of Genji, focuses on the activities surrounding the birth of an heir to the throne.74 Sarashina nikki (Sarashina Diary, ca. 1060?), by "Sugawara no Takasue's daughter," portrays the uneventful life of a romantic, religious woman.75
As a group, works in these two genres - and also Eiga monogatari,
the first historical tale, and Sei Shonagon's Makura no sdshi (Pillow
Book, ca. 993-1001?), the sole miscellany, discussed below-contained material of obvious interest to women: poetry, gossipy stories
about well-known people, courtly anecdotes, talk of dress and babies
and domestic problems, and, above all, accounts of relations between
the sexes. Romances ([tsukuri-]monogatari) offered most of the same
features, and, in addition, were often considerably longer. Whereas
the typical he monogatari story occupies less than a page in a modern Japanese printing, even the shortest extant Heian romance, Taketori monogatari (The Tale of the Bamboo Cutter, ca. 900?), uses thirty-
nine, and three volumes are required to accommodate the longest,
Utsuho monogatari (The Tale of the Hollow Tree, ca. 970-1000?). For
women with time on their hands, such works must have been especially attractive. Sei Shonagon brackets romances with go and backgammon as among the three best cures for boredom, and Takasue's
daughter tells us how eagerly they were sought - and how hard they
were for the ordinary woman to come by, even though hundreds appear to have been committed to writing.76
The scarcity of paper, a luxury item reserved for the wealthy, was
chiefly responsible for the difficulties encountered by would-be
readers, and by putative authors as well. Sei Sh5nagon, a member of
73 Translated in Edwin A. Cranston, The Izumi Shikibu Diary: A Romance of the Heian Court
(Cambridge, Mass.: Harvard University Press, 1969).
74 Translated in Richard Bowring, Murasaki Shikibu: Her Diary and Poetic Memoirs (Princeton, N.J.: Princeton University Press, 1982); see also Edward Seidensticker, "Murasaki
Shikibu and Her Diary and Her Other Writings," Literature East and West 18, 1 (March
1974): i7
75 Translated in Ivan Morris, As I Crossed a Bridge of Dreams: Recollections of a Woman in
Eleventh-Century Japan (New York: Dial Press, 1971).
76 For Sei Shonagon's comment, see Ikeda, Makura no sdshi, p. 195; and Ivan Morris, trans.,
The Pillow Book of Sei Shonagon, 2 vols. (New York: Columbia University Press, 1967), vol.
1, p. 145. For Takasue's daughter, see Suzuki Tomotaro, Kawaguchi Hisao, Endo Yoshimoto, and Nishishita Kyoichi, eds., Tosa nikki, Kagero nikki, Izumi Shikibu nikki, Sarashina
nikki, vol. 20 of Nihon koten bungaku taikei (1957), pp. 490-93, and Morris, Bridge of
Dreams, pp. 53-55.
Cambridge Histories Online © Cambridge University Press, 2008
NARRATIVE PROSE
443
the middling aristocracy like most of her fellow writers, explains that
she wrote Makura no soshi because her mistress happened to make
her a present of a large quantity of paper; otherwise, presumably, she
could not have afforded to do so.77 Similarly, the average noble lady
must have had to while away innumerable hours with her attendants' gossip, anecdotes, and oral tales for every one spent with a
written monogatari. But imperial consorts and princesses were able
to make use of their social and economic advantages to collect, reproduce, and order literary works with relative freedom. Manuscript
copying was one of their attendants' tasks, and women with literary
gifts were regularly brought into their entourages, both to add luster
to the social life centering on their salons, as Sei Shonagon did for
Empress Teishi, and to increase their store of romances.
So few romances remain today that it is easy to underestimate the
scope of such activities. The major poetry of the period has survived,
thanks to its brevity, public nature, and imperial sponsorship, but
most of the romances, like the paintings that illustrated them, vanished with the old life. Examination of the ones still available suggests that standards were not exacting, and might lead to the conclusion that the salons were of dubious literary significance. Of the
three known early works in the genre, the first, Taketori monogatari,
is a charming fairy tale about a moon maiden and her earthly suitors; the second, Utsuho monogatari, is a rambling story of upper-class
life, unified to some extent by recurrent illustrations of the miraculous power of music; and the third, Ochikubo monogatari (Tale of the
Sunken Room, ca. 985?), presents a Japanese version of the worldwide Cinderella theme, told with realistic detail and occasional
earthy humor.78 A handful of others, all from the late period, vary in
length and quality, the longer ones tending to indulge in improbabilities while attempting unsuccessfully to imitate The Tale ofGenji.79
77 Ikeda, Makura no sdshi, p. 331; Morris, Pillow Book, vol. 1, p. 267. Morris's book contains
a full translation of Sei Shonagon's miscellany; McCullough, Classical Japanese Prose, translates extended excerpts (pp. 158-99).
78 Taketori monogatari has been translated by Donald Keene in J. Thomas Rimer, Modern
Japanese Fiction and Its Tradition (Princeton, N.J.: Princeton University Press, 1978), pp.
275-301; Ochikubo monogatari, translated by Wilfred Whitehouse and Eizo Yanagisawa in
Ochikubo Monogatari or The Tale of the Lady Ochikubo: A Tenth Century Japanese Novel
(Tokyo: Hokuseido Press, rev. ed. 1965). For partial translations of Utsubh monogatari, see
Edwin A. Cranston, "Atemiya: ATranslation from the Utsubo Monogatari,"Monumenta Nipponica 24, 3 (1969): 289-314; Wayne P. Lammers, "The Succession {Kuniyuzuri): A Translation from Utsubo Monogatari," Monumenta Nipponica 37, 2 (1982): 139—78; and Ziro
Uraki, The Tale of the Cavern (Tokyo: Shinozaki Shorin, 1984).
79 For translations of post-Genji romances, see Robert L. Backus, The Riverside Counselor's
Stories: Vernacular Fiction of Late Heian Japan (Stanford, Calif.: Stanford University Press,
Cambridge Histories Online © Cambridge University Press, 2008
444
ARISTOCRATIC CULTURE
More important than the general level of quality that may have
prevailed, however, are certain other considerations. There were, in
effect, two kinds of literature at the Heian court: poetry, which was
the recognized public genre, suitable for both masculine and feminine attention; and narrative prose, which was largely relegated to
the rear palace (unless, like he monogatari, the work functioned as a
poetic manual). The imperial ladies thus presided over the only
catholic literary centers of the day, embracing both poetry and prose
in their interests, and encouraging literary activity both indirectly, by
setting a prestigious example, and directly, by patronizing individual
writers. The antecedents of the literary diary, for example-the
genre exploited so brilliantly by Michitsuna's mother - can be traced
to the tenth-century salon of Empress Onshi (885-954). Michitsuna's mother herself never served at court, but every other known
major diary author except Tsurayuki was a lady-in-waiting to a consort or princess at some time in her life. It may well be, indeed, that
we would not possess either Kageto nikki or any of the others if it had
not been for the custodial function assumed by the salons.
Feminine sponsorship at court also helped he monogatari to win
recognition as a classic, and feminine interest in the romance paved
the way for the single most impressive accomplishment of Heian civilization, Murasaki Shikibu's The Tale of Genji {Genji monogatari,
early eleventh century). In that magnificent work, which must rank
high on any list of the world's great psychological novels, Murasaki
echoes the concerns and adopts some of the techniques of the waka
poet, developing two major themes - the tyranny of time and the inescapable sorrow of romantic love - within the context of man's relationship to nature. But whereas the poet seeks to distill mono no
aware into the briefest of lyric expressions, Murasaki uses both poetry and prose to explore the concept in evocative, leisurely detail,
weaving a fabric of infinite richness and complexity. Whereas the anthology poet avoids indecorous personalism, Murasaki fills her stage
with more than five hundred characters, each a recognizable individual, and makes her long story develop logically from their
thoughts and feelings, and from the interplay of their personalities.
1985); Carol Hochstedler, The Tale of Nezame: Part Three ofYovia no Nezame Monogatari
(Ithaca, N.Y.: Cornell University East Asia Papers 22,1979); Thomas H. Rohlich, A Tale of
Eleventh-Century Japan: Hamamatsu Chunagon Monogatari (Princeton, N.J.: Princeton
University Press, 1983); Rosette E.Willig, The Changelings: A Classical Japanese Court Tale
(Stanford, Calif.: Stanford University Press, 1983); and Wayne P. Lammers, The Tale of
Matsura: Fujiwara Teika's Experiment in Fiction (Ann Arbor: Center for Japanese Studies,
University of Michigan, 1992).
Cambridge Histories Online © Cambridge University Press, 2008
NARRATIVE PROSE
445
She is both the quintessential representative of a unique society and
a writer who speaks to universal human concerns with a timeless
voice. Japan has not seen another such genius.80
Murasaki appears to have written the oldest parts of The Tale of
Genji away from court around the beginning of the eleventh century
(possibly for a private patron), to have been taken into Empress
Shdshi's service after word of the work spread, and to have completed it during the next decade or so. Her celebrated older contemporary, Sei Shonagon, who became a lady-in-waiting to Sh5shi's
rival, Empress Teishi (g76?-iooi), around 993, probably finished
Makura no soshi shortly after Teishi's death in 1001. Makura no soshi,
classified by Japanese scholars as a miscellany (zuihitsu), is a jumble
of reminiscences about the author's days at court, random observations on people, nature, and life in general, and innumerable lists of
things - waterfalls, mountains, bridges, musical instruments, games,
"Adorable Things," "Things That Make One Impatient," "Things
That Give a Vulgar Impression" - all set down in a sparkling, witty
style with fastidious sensitivity and small tolerance for human shortcomings. Under "Things That Are Unpleasant to See," she lists "a
lean, hirsute man taking a nap in the daytime. Does it occur to him
what a spectacle he is making of himself? Ugly men should sleep
only at night, for they cannot be seen in the dark and, besides, most
people are in bed themselves. But they should get up at the crack of
dawn, so that no one has to see them lying down."81
Sei Shonagon has a trenchant wit and a marvelously observant
eye. If her gaze is less serious and penetrating than Murasaki Shikibu's, she does us the service of calling attention to the gay, high-spirited, somewhat feckless side of court life, which Murasaki's preoccupation with weightier matters tends to obscure. We must read both
if we are to capture the flavor of aristocratic society, just as we must
80 Seidensticker, Genji, contains a complete translation of Murasaki's work. There is also a
virtually complete translation in Arthur Waley, The Tale of Genji: A Novel in Six Parts, 2 vols.
(Boston and New York: Houghton Mifflin, 1935) • Helen Craig McCullough, Genji and
Heike (Stanford, Calif.: Stanford University Press, 1994), pp. 25-242, translates many of
the chapters dealing with Prince Genji. For studies and bibliographies, see Haruo Shirane,
The Bridge of Dreams: A Poetics of "The Tale of Genji' (Stanford, Calif.: Stanford University
Press, 1987); Norma Field, The Splendor of Longing in The Tale of Genji (Princeton, N.J.:
Princeton University Press, 1987); Richard Bowring, Murasaki Shikibu: The Tale of Genji
(Cambridge: Cambridge University Press, 1988); Okada, Figures of Resistance; and Andrew
Pekarik, ed., Ukifune: Love in The Tale of Genji (New York: Columbia University Press,
1982). See also the discussions in Donald Keene, Seeds in the Heart (New York: Henry Holt,
1993), pp. 477-513; and Jin'ichi Konishi, A History ofJapanese Literature, vol. 2 (Princeton,
N.J.: Princeton University Press, 1986), pp. 317-46.
81 Ikeda, Makura no soshi, pp. 168-69; translation from Morris, Pillow Book, vol. 1, p. 266.
Cambridge Histories Online © Cambridge University Press, 2008
446
ARISTOCRATIC CULTURE
take both Six Dynasties obliquity and screen-poem romanticism
into account when we seek to understand the Kokinshu style.
However favorable the circumstances, a tiny court circle, isolated
from outside influences, is unlikely to produce a steady stream of
important writers. Only one other woman need be noticed here: the
anonymous author - probably Akazome Emon (b. ca. 960?), a poet
in the service of Michinaga's wife - who created the historical tale
(rekishi monogatari) genre by writing Eiga monogatari {A Tale of Flowering Fortunes, started ca. 1030?) .82 Akazome, as we may identify her
for convenience, was her country's first vernacular historian, and the
first Japanese to treat historical materials in the monogatari style. Her
book is a naively enthusiastic celebration of Michinaga, his family,
and his times. As literature, it falls far short of The Tale of Genji, its
apparent model, but its detailed descriptions of everyday aristocratic
life make it an invaluable source for the social historian. We are also
indebted to Akazome for inspiring another writer to produce Okagami {The Great Mirror, ca. 1086-1125?), m e second and most important work in the genre she invented.83
We return to the world of masculine letters with Okagami, a book
written by a man for what must have been primarily a male readership, judging from its contents. The anonymous author had had a
number of male predecessors in the general area of narrative prose.
Strangely, all three of the surviving tenth-century romances and at
least one from the eleventh century appear to have been written by
men; and Tsurayuki was not quite the only male diarist. Men also
monopolized the tenth-century poem tale genre, probably because it
was, in effect, a special type of setsuwa collection.
The compilation of anecdotes and short factual tales, which enjoyed the prestige of Chinese example, seems to have been a fairly
common occupation for educated men throughout the Heian period.84 Many collections were put together for utilitarian purposes.
Of the ones that survive, the bulk are Buddhist, assembled by monks
to enliven sermons; and most of the others are twelfth-century compendiums of scraps of information about musical instruments, poetry, official ceremonies, and the like - further evidence of the late
Heian desire to embalm the golden past. Nostalgia probably also
played a part in the compilation of the greatest monument of setsuwa
literature, Konjaku monogatari[shu] ([A Collection of] Tales of Times
82 Major portion translated in McCullough and McCullough, Flowering Fortunes.
83 Translated in McCullough, Okagami. 84 See McCullough, Okagami, pp. 1-14.
Cambridge Histories Online © Cambridge University Press, 2008
NARRATIVE PROSE
447
Now Past, ca. nao?), an anonymous collection of more than a thousand tales, dealing with a great variety of Buddhist and secular subjects, and with many social classes.85 Certainly, the conservatorial
spirit was in the air. But the inelegant concerns of Konjaku monogatari are a far cry from the world of Genji, the Shining Prince, and
it may be best to think of the collection as primarily a preachers'
manual - and as a forerunner, like the imayo and the biwa koshi, of
the medieval rise of new elements in the Japanese cultural mix.
Okagami, like he monogatari, can be considered a special kind of
setsuwa collection. The author, looking back on Michinaga's spectacular career, finds himself dissatisfied with Akazome's treatment of
it, and determines, as he tells us at the outset, not only to describe
but also to explain.86 He retells Akazome's story, forsaking the
chronological approach for the anecdotal, and selecting his materials to support the argument that Michinaga's success arose from a
combination of luck, family connections, and favorable personal
qualities. To readers of The Tale of Genji and Makura no soshi, it
comes as no surprise that Michinaga is praised for physical beauty,
taste in dress, and poetic ability, and that he is depicted as sponsoring literary entertainments like the Oi River excursion described earlier in this chapter. But nothing in the works of feminine authors
prepares us for the discovery that equal space is devoted to his
prowess as an archer and a horseman. Even more unexpected is the
extensive documentation, in story after story, of his courage, coolness, prudence, and resourcefulness in public life - characteristics,
noted approvingly in comments on other men as well, of which we
hear virtually nothing from women writers (whose concerns, as has
been seen, are basically private), but which are also singled out for
praise in Konjaku monogatari and other collections of anecdotes.87
Okagami provides an important corrective to the notion that the
Heian court produced remarkable aesthetes but bestowed little esteem on practical men of affairs. The truth is surely that the virtues
85 For partial translations, see Robert Hopkins Brower, "The Konzyaku monogatarisyii: An
Historical and Critical Introduction, with Annotated Translations of Seventy-Eight Tales,"
Ph.D. diss., University of Michigan, Ann Arbor, 1952; Frank, Histoires; Marian Ury, Tales
of Times Now Past: Sixty-Two Stories from a Medieval Japanese Collection (Berkeley and Los
Angeles: University of California Press, 1979); William Ritchie Wilson, "The Way of the
Bow and Arrow: The Japanese Warrior in Konjaku Monogatari," Monumenta Nipponica 28,
2 (Summer 1973): 177-233.
86 McCullough, Okagami, p. 68.
87 McCullough, Okagami, pp. 48-53; Ury, Tales, p. 17; Hiroko Kobayashi, The Human Comedy of Heian Japan: A Study of the Secular Stories in the Twelfth-Century Collection of Tales,
Konjaku Monogatarishu (Tokyo: Centre for East Asian Cultural Studies, 1979), pp. 159-60.
Cambridge Histories Online © Cambridge University Press, 2008
448
ARISTOCRATIC CULTURE
of both were recognized. Living as they did in a stable society almost
devoid of foreign or internal threats, Heian aristocrats were under no
compulsion to practice the martial arts as their ancestors had done,
but they did not wholly reject the heritage of the turbulent era before
the advent of Chinese culture. Rather, they performed the exceptional feat of reconciling the most useful virtues of their barbaric
forebears with a revolutionary new conception of civilized behavior;
and the synthesis proved enduring enough to gratify the most
conservative of Heian hearts. Although the establishment of the Kamakura shogunate brought ancient attitudes into renewed prominence, the Japanese never turned their backs on the ideals represented by the lacquerer of the Koyasan marsh-iris chest, the Genji
monogatari emaki painters, the calligrapher Yukinari, the sculptor
Jocho, the poet Shunzei, and the novelist Murasaki Shikibu. Remote
as Heian society may seem to us today, many of its essential characteristics survive in modern Japan. Practical ability continues to be
highly valued. Great importance is still attached to status within a hierarchy, to group solidarity, to decision by consensus, and to peer approval. There is less individual freedom than in Western countries of
comparable international stature. And there remains a persistent
feeling that a true Japanese, whatever his walk in life, ought to be able
to compose a verse, judge a specimen of calligraphy or an artistic
performance, and savor beauty with a proper appreciation of its
ephemerality.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 7
ARISTOCRATIC BUDDHISM
THE PRELUDE TO HEIAN BUDDHISM
Since its introduction into Japan in the middle of the sixth century,
the Buddhist religion experienced steady growth under the patronage
of the imperial family and powerful clans who sponsored the founding of magnificent temples and monasteries, which they generously
endowed with gifts of agricultural land to provide an economic base
for the upkeep of these institutions. The rapid expansion of the
church is evident in the following figures taken from eighth-century
sources. A census conducted in the year 624 - less than a century
after Buddhism made its appearance at the court - revealed that
there were forty-six functioning monasteries in Japan that accommodated a total of 816 monks and 569 nuns.1 By the year 681 the number of monasteries in the capital region alone had grown to twentyfour.2 Larger monasteries capable of accommodating hundreds of
monks appeared during the course of the seventh century. Thus an
entry in the Nihon shoki for the year 690 notes that gifts of cloth were
presented to some 3,363 monks residing in seven monasteries.3
Although we lack reliable figures for the total number of monasteries in Japan at this time, it is possible to confirm on the basis of
archaeological evidence, mainly tiles and foundation stones, the existence of at least two hundred temple sites dating from the AsukaHakuho period (593-710).4 As might be expected, most of these sites
1 Nihon shoki (720), kan 22, in Sakamoto Taro, Ienaga Saburo, Inoue Mitsusada, and 6no
Susumu, eds., Nihon shoki, vols. 67-68 of Nihon koten bungaku taikei (hereafter NKBT), 100
vols. (Tokyo: Iwanami shoten, 1957-67), vol. 68, p. 210. Other abbreviations used in this
chapter are: BZ for Dai Nihon Bukkyo zensho, 151 vols. (Tokyo: Bussho kankokai, 1912-22);
DDZ for Dengyo Daishi zenshu, 5 vols. (Sakamoto: Hieizan tosho kankosho, 1926-27); KT
for Kuroita Katsumi, ed., Shintei zbho: kokushi taikei, 62 vols. in 66 (Tokyo: Yoshikawa
kobunkan, 1929-64); T for Takakusu Junjiro, ed., Taisho shinshu daizbkyb, 100 vols. (Tokyo:
Taisho issaikyo kankokai, 1924-34).
2 Nihon shoki, kan 29,Temmu 9 (68o)/5/i, NKBT, vol. 68, p. 440.
3 Nihon shoki, kan 30, Jito 4 (690)77/14, NKBT, vol. 68, p. 504.
4 Summary of two studies by Ishida Mosaku, cited in note in Nihon shoki, NKBT, vol. 68, p.
563, note 9.
449
Cambridge Histories Online © Cambridge University Press, 2008
45°
ARISTOCRATIC BUDDHISM
are in Yamato Province, although a few are found as far west as
Bitchu and as far east as Owari, which gives some indication of the
diffusion of Buddhism in this early period. Although the government
supported the Buddhist religion in the expectation that its rituals
would provide protection for the state, the authorities became increasingly concerned about the unrelenting growth in the size of the
monastic establishments and the tax-exempt agricultural lands that
they held.
The position that the church was to occupy in the new centralized
state was made abundantly clear by the Taiho ritsuryo, the legal code
that was promulgated in 702, and the Yoro ritsuryo, which replaced
the former in 757. Both the Taiho ritsuryo and Yoro ritsuryo contained a section entitled Soniryo (Administrative Laws Pertaining to
Monks and Nuns) that consisted of twenty-seven articles regulating
the behavior of the Buddhist clergy.5 While some articles basically
restated prohibitions already enunciated either in the Vinaya (the
section of the Buddhist canon containing ecclesiastical law) or in
certain sutras - namely, the prohibitions against murder, fornication, theft, gambling, the consumption of meat and wine - the majority of the articles sought to define the position, responsibilities,
and scope of the activities of the clergy within Japanese society.
A primary concern of the Soniryo was to prevent monks and nuns
from taking advantage of their respected religious status to interfere
in political affairs, as can be seen from the first article, which specifically forbids clerics to "speak falsely about misfortunes or blessings
based on interpretations of mysterious natural phenomena, discuss
matters relating to the state, delude the common people, or read
military books."6 Fearing that monks and nuns might exploit their
charisma to gain a large personal following, the Soniryo banned the
establishment of private chapels (dojo), proselytizing among the
masses, itinerant begging without a permit (Art. 5), religious training in the mountains (Art. 13), the practice of divination, and the
use of magic to cure illness (Art. 2). There was also a prohibition
against wearing clerical garb while pleading a case in civil court
(Art. 17).
It is apparent from the Soniryo that the government was much
troubled by the abuse of the privileges accorded the clergy, the foremost being the exemption from all taxes. Members of the clergy
5 For the reconstructed texts of the Taiho ritsuryo soniryo and the Yorv ritsuryo soniryo, see
Futaba Kenko, Kodai Bukkyo shiso-shi kenkyu (Kyoto: Nagata bunshodo, 196a), pp. 167—76.
6 Futaba, Kodai Bukkyo p. 167.
Cambridge Histories Online © Cambridge University Press, 2008
PRELUDE TO HEIAN BUDDHISM
451
were forbidden to "lend their names" to others (Art. 16), which was
a technique commonly employed in tax evasion. Although monks
were free to accept devout young boys as their personal attendants,
the law stipulated that these boys must be sent back to their homes
before reaching the age of seventeen so that their names could be entered in the tax registers (Art. 6). Monastic officers were required to
notify the authorities whenever a monk died so that his identity
could not be assumed by another (Art. 20). It was further decreed
that persons privately ordained (shido), that is, ordained without official authorization, could not be accorded clerical status even
though they might have received all the formal training expected of
a monk (Art. 24). Finally, members of the clergy were denied the
right to private ownership of land, buildings, and valuables, or to engage in trade or collect interest on loans (Art. 18). In addition,
monks and nuns were prohibited from accepting personal gifts of
slaves, cattle, or weapons (Art. 26).
Although it is clear that some attempt was made during the early
years of the Nara period to enforce the Soniryo, the clergy became
increasingly independent of state control after the accession of the
strongly pro-Buddhist emperor Shomu (r. 724-49). The chronicles
covering the latter half of Shomu's reign abound in references to the
illegal acquisition of land by monasteries, faulty record-keeping by
temple officers, and fraudulent practices by monks.7 In open defiance of the Soniryo, not to mention the Vinaya, some monks took
wives and engaged in usury. Monasteries practiced pawnbroking to
augment their income, charging rates as high as 180 percent per
annum.8
Efforts to limit the growth of the Buddhist church in the Nara period proved largely unsuccessful because of the enthusiastic patronage it received from Shomu and his daughter Kdken (reigned 749-58
and again, under the name of Shotoku, 764-70), who succeeded him
on the throne. When the imperial court moved to Nara in 710, major
monasteries such as the Daikan daiji (renamed Daianji in 729),
Yakushiji, Gangoji, and Kofukuji that had been scattered throughout
the Asuka region were dismantled and reconstructed in the newly
established capital on an even grander scale. In addition to these
older transplanted Asuka monasteries, magnificent new ones such as
7 Shoku Nihongi (797), kan i6,Tempyo 18 (746)73/16, KT, vol. 2, p. i86;Tempy6 18/5/9, KT>
vol. 2, pp. 187-88; kan 22,Tempyo hoji 3 (759)/6/22, KT, vol. 2, p. 264.
8 See the Shosbin documents dated Hoki 4-6 (773—75) inTakeuchi Rizo, Nara-chb jidai ni
okerujiin keizai no kenkyu (Tokyo: Ookayama shoten, 1932), pp. 205-6.
Cambridge Histories Online © Cambridge University Press, 2008
452
ARISTOCRATIC BUDDHISM
the Todaiji, Toshodaiji, and Saidaiji were built in Nara throughout
the course of the eighth century.
Just as the government failed in its effort to restrict the number
and size of monastic establishments, so too was it unsuccessful in
limiting the flow of people joining the clergy. In 696 a decree was issued stipulating that henceforth ten monks should be ordained on
the last day of each year.9 The authorities, however, were unable to
keep the annual ordinations at this modest level for long. An earlyninth-century document reveals, for example, that 773 monks were
admitted to the priesthood in the annual ordination ceremony held
at the imperial palace in the twelfth month of 741.10 In a similar ceremony conducted in 747 at Emperor Sh5mu's palace at Naniwa,
6,563 monks were ordained.11 A tabulation of the various records
shows that at least twenty-seven separate ordination ceremonies were
conducted during the Nara period, at which a total of 18,520 persons
were granted clerical status.12 Individual monasteries by this time
were housing large numbers of monks: the Horyuji had 176 resident
monks and 87 novices in the year 747;13 the Daianji during the same
year accommodated 473 monks and 414 novices.14
With the rapid expansion of the Buddhist church in Nara and the
great respect accorded to learned or charismatic clerics, it was perhaps inevitable that monks would become deeply involved in the affairs of state, the strictures of the Soniryo notwithstanding. Typical of
such political monks was Gembo, who had spent seventeen years in
China, where, in recognition of his scholarship, he was appointed by
Emperor Hsiian-tsung to the Third Rank (san-p'iri) and granted the
purple robe, the highest honor that can be bestowed on a monk.
Upon his return to Japan in 734, Gembo presented the court with a
precious gift - a complete set of the Buddhist scripture in some five
thousand fascicles. In 736 one hundred households and ten chd of
land were assigned for his support. The following year he was appointed to the office of sbjo (bishop), the highest ecclesiastical rank,
and at the same time was designated a court chaplain (naigubu s5).
Immediately putting his faith-healing skills to work, he succeeded in
curing the empress dowager of her chronic depression, which greatly
enhanced his standing with the imperial family. Gembo used his
9
10
11
12
13
14
Nihon shoki, kan 30, Jito 10 (6g6)/i2/i, NKBT, vol. 68, p. 532.
Saicho, Naisho Buppo sojo kechimyaku fu (819), DDZ, vol. 1, p. 213.
Takeuchi Rizo ed., Nara ibun, 3 vols., rev. ed. (Tokyo: Tokyo-do, 1962), vol. 2, p. 522.
Takeuchi, Nara-cho jidai ni okerujiin keizai no kenkyu, p. 26.
Horyuji garan engi narabi ni ruki shizai chd (747), BZ, vol. 97, p. 3b.
Daianji garan engi narabi ni ruki shizai chd (747), BZ, vol. 28, p. 117b.
Cambridge Histories Online © Cambridge University Press, 2008
PRELUDE TO HEIAN BUDDHISM
453
newly acquired influence to promote the fortunes of the scholar Kibi
no Makibi, who had been his fellow student in China, and collaborated with Tachibana no Moroe, who was attempting to loosen the
Fujiwara stranglehold on the court. So strong was Fujiwara resentment against the meddling of Gembo that a leading member of the
Fujiwara family, Hirotsugu, began an insurrection in the hope of
driving him from the court. Although the insurrection failed and Hirotsugu was put to death, the Fujiwara managed to regain their ascendancy and arranged, in 745, to have Gembo sent to Kyushu to
oversee the construction of the Kanzeonji. Once he was out of the
capital, an order was issued stripping him of his property. Gembo
died in Kyushu the following year, possibly at the hands of Hirotsugu's partisans.
An even greater threat to the government was posed by the monk
Dokyo, whose apparently intimate relationship with Empress Koken
became one of the most scandalous episodes in Japanese history.
Trained in the mountains where he had mastered various esoteric
rites, Dokyo had already acquired a reputation as thaumaturge when
he was summoned to the court by Empress Koken in 752. He grew
particularly close to Koken after her abdication in 758 and was credited, in 761, with having brought about her recovery from a serious
illness. After the eminent monk Jikun was summarily dismissed from
his position as shdsozu in the Buddhist hierarchy in 763 to create a
vacancy the Dokyo might fill, resentment against Dokyo began to
grow. The following year the Chancellor, Fujiwara no Nakamaro,
who eighteen years earlier had forced the banishment of Gembo, attempted to remove Dokyo by a coup, which soon collapsed. Nakamaro was captured and ignominiously executed along with his wife
and children.
Two days after the death of Nakamaro, Ddkyo was appointed his
de facto successor and given the newly created title Daijin Zenji (The
Meditation-Master Who Ranks as Minister of State). The incumbent
emperor Junnin, who had been a protege of Nakamaro, was forced
to abdicate in favor of Koken, who reassumed the throne under the
new name of Shotoku. Following a visit by the empress in 765 to
Yuge to worship at Dokyo's clan temple, Dokyo was honored with
yet another newly coined title, Daijo Daijin Zenji (The MeditationMaster Who Ranks as Chancellor), indicating that he now occupied
the highest office in the government. A year later Dokyo was granted
the title Hod (King of the Buddhist Faith) to signify his supremacy
in the religious world as well. Honors, unprecedented in the case of
Cambridge Histories Online © Cambridge University Press, 2008
454
ARISTOCRATIC BUDDHISM
a layman, were heaped on him: ministers of state were required to
do obeisance; the imperial carriage was made available for his personal use; all ordinations had to have his approval, indicated by the
imprint of his personal seal. Not content with having concentrated
both political and ecclesiastical power in his hands, Dokyo contrived, in 769, to have an oracle delivered to the court from the Hachiman Shrine in Usa that promised peace for the nation if D6ky5
became emperor. Owing to the opposition of loyalists, particularly
the highly respected Wake no Kiyomaro, the oracle was rejected in
favor of a new oracle that affirmed that only members of the imperial family might occupy the throne. Dokyo's secular and ecclesiastical powers were ultimately based solely on his close personal relationship with the empress. For the aristocratic families and Buddhist
clergy, he was an object of contempt. Less than a month after the
death of Empress Shotoku in 770, Fujiwara no Nagate, who held the
office of Minister of the Left and who had long chafed under
Dokyo's arrogant rule, succeeded in having him banished to Shimotsuke, where he lived in obscurity as superintendent (betio) of the
Yakushiji, a local temple, until his death there two years later.
THE ASSERTION OF GOVERNMENT CONTROL OVER THE
BUDDHIST CHURCH
The Gembo and Dokyo affairs dramatically illustrated the dangers
that an unchecked clergy posed to the powerful aristocratic clans
and even to the hitherto sacrosanct imperial family. Within the short
span of twenty years two monks, Gembo and Dokyo, had managed
to exercise control over the government by using their religious
charisma to win favor with ex-empresses. Their extraordinary hold
on political power is evident from the crushing military defeats suffered by the two Fujiwara chieftains who sought to expel the clerics
from court. It is important to remember, however, that neither
Gembo nor Dokyo achieved power through the church itself. While
Gembo at least might have commanded respect within the clerical
community on the basis of his scholarly credentials, Dokyo could be
viewed only as an unscrupulous cleric who rode roughshod over the
established ecclesiastical hierarchy (sogo), countenancing the dismissal of some of its most distinguished members simply to fulfill his
own personal ambitions.
Although the church as a whole may not have been responsible for
the success of such political monks as Gembo and Dokyo, the imCambridge Histories Online © Cambridge University Press, 2008
CONTROL OVER THE BUDDHIST CHURCH
455
perial family as well as the Fujiwara could not have failed to see the
perils of having the affairs of state conducted in a city like Nara,
where the offices of government were overshadowed by imposing
monasteries accommodating thousands of monks. The desire to escape interference by the church no doubt was a major factor in the
decision by Emperor Kammu, who ascended the throne in the
fourth month of 781, to abandon Nara in favor of a new administrative center. Yet it should be remembered that aside from the fear of
clerical domination there were other reasons for relocating the capital. For one thing, the location of Nara, surrounded by hilly terrain
to the north, east, and west, did not allow easy access to the port of
Naniwa, which had assumed increasing importance with the emergence of a centralized state in the eighth century. Another consideration may have been the lack of an adequate supply of water to sustain Nara's burgeoning population.15
Pressure to establish a new capital from which the country could
be more easily governed intensified after 780 when news of the rout
of government forces by the Emishi aborigines in Mutsu created
great unrest in Nara. The following year the aging Emperor Konin,
who had succeeded Empress Shotoku in 770, relinquished the
throne to his energetic eldest son, Prince Yamabe, then forty-four
years of age. The new emperor, known as Kammu (r. 781-806), immediately undertook a series of economic reforms designed to
strengthen the central government. In the fifth month of 784, Fujiwara no Oguromaro, who had three years earlier won an impressive
victory over the Ezo, was ordered to inspect the Nagaoka region in
Yamashiro Province and report on its suitability as the site for a new
seat of government. Construction of the new capital at Nagaoka,
which bordered on the Katsura River, a tributary of the Yodo River
that passed through Naniwa, was begun the following month. By the
end of the year the emperor formally took up residence in a temporary palace at Nagaoka. Despite an enormous expenditure of tax
revenue and labor, construction of the Nagaoka capital did not progress satisfactorily, partly owing to the belief that the area was
haunted by the vengeful spirits of Fujiwara noTanetsugu and Prince
Sawara, who lost their lives in 785 because of intrigues relating to the
establishment of the new capital. In 793 the government finally
abandoned Nagaoka and announced its intention of relocating at
15 Kuroita Katsumi, Kotei kokushi no kenkyu (Tokyo: Iwanami shoten, 1936), vol. 2, pp.
197-98.
Cambridge Histories Online © Cambridge University Press, 2008
456
ARISTOCRATIC BUDDHISM
Uta to the northeast of Nagaoka. Kammu, who moved here during
the following year, renamed the new city Heian-kyo (The Capital of
Peace and Tranquillity), which later came to be called simply Kyoto
(The Capital).
As descendants of Emperor Tenji (r. 661-72), Konin and his son
Kammu did not feel themselves bound to continue the unrestrained
pro-Buddhist policies pursued by the Nara rulers Gensho, Shomu,
and Koken, who all belonged to the rival lineage of Emperor Temmu
(r. 673-86). Determined to bring the northern aborigines under imperial control and to alleviate the hardships suffered by the peasantry, Konin, unlike his predecessors, concerned himself primarily
with military and economic problems. It was only in the last two
years of his reign that Konin began to consider the consequences of
a church functioning free from government supervision.
In 779 the Civil Affairs Ministry (Jibusho) issued two directives designed to curb the illegal activities of monks. One stipulated that a nationwide census of the clergy be taken to identify "official monks"
(kanso).16 By compiling such a register, the authorities hoped to weed
out those monks who had been ordained without government sanction. A second directive sought to send back to their home provinces
all monks affiliated with the kokubunji (the government-supported
monasteries established in each province after 741) who were illegally
residing in the capital.17 Emperor Konin's critical attitude toward the
Buddhist clergy was revealed in a remarkable edict written on the
twentieth day of the first month of 780, just six days after the pagodas of the Yakushiji and Katsuragidera had been struck by lightning
and destroyed in the ensuing fire.18 Citing the widely held Chinese
belief that unusual natural phenomena reflected the judgment of
Heaven, Konin declared that the destruction of the pagodas signified
such divine censure. While acknowledging in stereotyped language
his own lack of virtue, he asked how the clergy could not but be
ashamed of its conduct, which was no different from that of laymen,
and accused the church of violating both the teachings of the Buddha and the laws of the land. It can be regarded only as an irony that
the Buddhist clergy, which since Asuka times had been thought to
provide protection for the state, was seen by the end of the Nara period as the very cause of the natural disasters that it was supposed to
prevent.
16 Shoku Nihongi, kan 35, Hoki 10 (779)78/19, KT, vol. 2, p. 451.
17 Shoku Nihongi, kan 35, Hoki 10/8/26, KT, vol. 2, p. 451.
18 Shoku Nihongi, kan 36, Hoki 11 (7805/1/20, KT, vol. 2, p. 456.
Cambridge Histories Online © Cambridge University Press, 2008
CONTROL OVER THE BUDDHIST CHURCH
457
Sharing his father's fear of an unrestrained church, Emperor
Kammu, or his Council of State, in the course of his twenty-five-year
reign issued more than thirty directives that sought to correct abuses
by the clergy and reduce the threat that temples and monasteries
posed to the national economy. The various strictures applied by
Kammu were not intended to humiliate the clergy or denigrate Buddhist doctrine, for he always saw himself as a devout follower of Buddhism who founded monasteries, sponsored religious services at
court, and periodically ordered moratoriums on the slaughter of animals, albeit to a lesser extent than his predecessors in the Temmu
lineage. Kammu's political goal was to revitalize the ritsuryd state
and create a strong and fiscally sound government. It is not surprising, therefore, that many of the measures that he adopted reflected
prohibitions already found in the Soniryo, which, like the laws in the
ritsuryd itself, had been increasingly ignored during the second half
of the Nara period.
The underlying theme of Kammu's reign was sounded in a rescript
issued in 782, just one year after he ascended the throne, in which he
declared that only by bringing an end to the various construction projects, by developing agriculture, and by the practice of frugality on the
part of government, could the nation become prosperous.19 Specifically linking the decline in the price of silver to the decision to put a
stop to temple construction, Kammu ordered the disbanding of the
Office for the Construction of the Hokkeji, which was supervising
work on the still uncompleted Hokkeji, a major convent in Nara
founded by Empress Komyo, who was the consort of Emperor Shomu.
Another of Kammu's major concerns was the widespread practice
of acquiring tax-exempt land through the creation of privately controlled temples. In the years following the move of the capital to
Nara, the government permitted a limited number of powerful families to found temples that would essentially be under their own control. Since such temples were expected to offer prayers for the wellbeing of the country and to conduct services for the benefit of their
lay sponsors, they received annual payments of rice from the government as an indication of their quasi-official status. The lay sponsors {dan'otsu) and their heirs held in perpetuity the right to appoint
the clerical officers (sango) who administered these temples, which
were officially designated jogakuji (temples within the set limit). According to a decree issued in 749, jogakuji were prohibited from own19 Shoku Nihongi, kan 37, Enryaku 1 (782)/4/n, KT, vol. 2, pp. 483-84.
Cambridge Histories Online © Cambridge University Press, 2008
458
ARISTOCRATIC BUDDHISM
ing more than 100 chb of cultivated land, which was a mere 10 percent of the amount allowed for the "great monasteries" (daiji) such
as Daianji, Yakushiji, Kofukuji, Hokkeji, and the provincial kokubunji.20 Nevertheless jogakuji continued to expand their landholdings, thereby enriching their sponsors who exercised de facto control
over the temples through their handpicked clerical officers.21
To cope with this problem Kammu issued an edict in 783 chastising officials for their laxity in enforcing earlier prohibitions against
unsanctioned private temples, which, he said, if left unchecked
would in time acquire every square inch of land in the country.22 Not
only were both private temples that had not been accorded status as
jogakuji and "chapels" (dojo) to be banned, but even unauthorized
donations of land or other property to established monasteries were
forbidden. Violators were subject to severe punishment: immediate
dismissal for officials, eighty lashes of the whip for others. Kammu's
predecessor, Konin, likewise troubled by donations to monasteries,
had declared three years earlier, in 780, on the occasion of his gift "in
perpetuity" of one hundred households to the Akishinodera in Nara,
that the term "in perpetuity" must henceforth be interpreted in all
cases to mean "for one generation only."23 In 795, one year after the
capital was moved to Kyoto, the injunction against donations of land
or property to monasteries was repeated, with a new stipulation requiring registration with the central government of all land previously donated.24 Monasteries were also warned, as were families of
ranking ministers, in 784, immediately after the capital was moved to
Nagaoka, and again in 798, after the move to Kyoto, to cease acquiring plots of land in a pattern that effectively blocked access of
the common people to mountains, rivers, and marshes, which were
all in the public domain.25 By monopolizing such areas, which had
previously been open for hunting, fishing, and certain types of specialized agriculture, the powerful families and monasteries increased
their wealth but, as the edict of 784 pointed out, only at the cost of
great hardship to the people.
Even the legitimate property of recognized monasteries was
ao Shoku Nihongi, kan i7,Tempyo shoho 1 (749)/7/i3, KT, vol. 2, p. 204.
21 See, for example, the directive of the Council of State (Daijo kampu) Enryaku 24 (8os)/i/3
in Ruiju sandai kyaku (mid-eleventh century): zempen, kan 3, KT, vol. 25, p. 116.
22 Shoku Nihongi, kan 37, Enryaku 2 (783)/6/io, KT, vol. 2, pp. 493-9423 Shoku Nihongi, kan 36, Hoki 11 (78o)/i 1/6/5, KT> v o ' - 2> P- 46124 Ruiju kokushi (892): zempen, kan 182, Enryaku 14 (7955/4/20, KT, vol. 6, p. 279.
25 Shoku Nihongi, kan 38, Enryaku 3 (784)/i2/i3, KT, vol. 2, p. 503; Ruiju sandai kyaku: kohen,
kan 16, directive of the Council of State dated Enryaku 17 (JQS)/I2JS, KT, vol. 25, p. 497.
Cambridge Histories Online © Cambridge University Press, 2008
CONTROL OVER THE BUDDHIST CHURCH
459
henceforth to be subjected to careful scrutiny. A directive issued in
796 empowered the provincial governor, along with the temple officers (sango), to audit temple property, which was now subject to
confiscation if some illegality was discovered in the course of the investigation.26 While this directive affirmed the right of the lay patron, in consultation with the monks of the temple, to appoint the
temple officers, such appointments also required the approval of the
provincial governor, who could subsequently press charges against
the lay patron in the event that the temple officers were guilty of any
illegal acts.
Each recognized temple was required to submit annually to the
central government an inventory of its property (shizai chd). Such
records, however, could be easily falsified since it was often not possible to check their accuracy if the temple was far from the capital.
The Council of State therefore decided in 798, at the behest of
Kammu, to replace these annual reports with an on-the-spot verification by each newly appointed provincial governor.27
Since a major reason for the decision to remove the capital from
Nara was the venality of many clerics in the great monasteries, it is
not surprising that Kammu should have singled out the monasteries
of Nara and their clergy for particular condemnation. As early as 783
he issued a sharply worded edict denouncing specifically the Nara
monasteries for charging usurious interest rates on loans (suiko) that
inevitably led to the forfeiture of mortgaged lands and property, depriving farmers of their livelihood.28 Such practices, it was pointed
out, not only contravened the laws of the state but also signified the
secularization of the church. While the edict did not prohibit monasteries from making loans, it limited the interest they could charge,
regardless of the length of the loan, to 10 percent. This was well
below the current rate for private loans, which fluctuated between 50
percent and 100 percent per annum.
In 795, one year after the court had moved to Kyoto, Kammu was
informed by his ministers that the seven great monasteries of Nara
were still making loans at exorbitant interest rates, which had so impoverished the peasantry that they could no longer pay their taxes or
sustain their families.29 Although the ministers' proposal to restrict
26 Ruiju sandai kyaku: zempen, kan 3, directive dated Enryaku 15 (796)73/25, KT, vol. 25, pp.
11516.
27 Directive dated Enryaku 17 (7g8)/i/2o, in KT, vol. 25, p. 116.
28 Shoku Nihongi, kan 37, Enryaku 2 (783)712/2, KT, vol. 2, p. 496.
29 Ruiju kokushi: kdhen, kan 182, memorial dated Enryaku 14 (795)711/22, KT, vol. 6, p. 279.
Cambridge Histories Online © Cambridge University Press, 2008
460
ARISTOCRATIC BUDDHISM
the frequency of temple loans was readily accepted by the emperor,
it proved difficult to enforce. Kammu made no secret of his dislike
for the Nara clergy. No sooner had he moved to Kyoto than he dispatched a commissioner to Nara to investigate the clergy of the
seven great monasteries.30 In an edict issued in 798, Kammu again
lashed out at the Nara clergy, accusing them of licentious behavior,
and appointed the governor of Yamato Province, Fujiwara no Sonohito, to carry out an investigation.31
By Heian times clerical misconduct was no longer a phenomenon
limited to the Nara monasteries. Although the Soniryo, which in theory defined the standards of behavior of the clergy, remained in force
after Nara was abandoned as the capital, it became increasingly apparent that many monks and nuns were openly flouting its provisions. To bring the wayward clergy under greater control of the authorities, Kammu issued within the short span of fifteen years an
unprecedented number of directives aimed at correcting specific
abuses. A list of the principal decrees follows:
784: Edict deploring corrupt provincial bishops (kokushi) who expect to be
received with ceremonies appropriate for lay officials. Such persons, who
cause much hardship to the people, must be replaced.32
785 (and again in 798): Edict deploring the failure of the authorities to arrest corrupt monks who choose their own lay patrons or who tour villages
claiming to be able to work miracles and thus delude the ignorant masses.
Such monks should be exiled to a distant province and required to stay in
a recognized temple.33
785: Edict ordering that a roster of "virtuous monks" be compiled and sent
to the central government so that such monks might be commended as
models for the clergy.34
785 (and again in 799): Decrees prohibiting monks, nuns, or lay persons
from practicing black magic in the mountains with the aim of harming their
enemies.35
797: Edict instructing the provincial bishops Qtbji) to scrutinize internal temple affairs and correct any abuses by the clergy within their jurisdiction.36
30 Ruiju kokushi: kohen, kan 180, Enryaku 14 (795)/7/i8, KT, vol. 6, p. 257.
31 Ruiju kokushi: kohen, kan 186, edict dated Enryaku 17 (7985/7/28, KT, vol. 6, p. 300. Kammu
decried the immorality of the clergy in at least two other edicts issued in the course of this
year. See entries for Enryaku 17/4/15 and 10/17 m Ruiju kokushi: kohen, KT, vol. 6, p. 300.
32 Shoku Nihongi, kan 38, Enryaku 3 (7845/5/1, KT, vol. 2, p. 499.
33 Shoku Nihongi, edict dated Enryaku 4 (7855/5/25, KT, vol. 2, p. 508; Ruiju kokushi: kohen,
kan 186, edict dated Enryaku 17 (7985/4/15, KT, vol. 6, p. 300.
34 Shoku Nihongi, kan 38, edict dated Enryaku 4 (7855/7/11, KT, vol. 2, p. 511.
35 Ruiju sandai kyaku, kan 2, directive of the Council of State dated Shotai (goi)/2/i4, citing
an earlier directive dated Enryaku 4 (785)710/5, KT, vol. 25, p. 74; Nihon koki (840), kan 8,
Enryaku 18 (7995/6/12, KT, vol. 3, p. 22.
36 Ruiju kokushi, kan 186, edict dated Enryaku 16 (7975/8/11, KT, vol. 6, p. 300.
Cambridge Histories Online © Cambridge University Press, 2008
CONTROL OVER THE BUDDHIST CHURCH
461
798, 4th month: Edict deploring the lack of proper training of the annual
ordinands (nembun dosha), who are described as resembling monks only in
appearance but whose behavior is like that of laymen. Henceforth only people of good character and learning, over thirty-five years of age, will be eligible for ordination. The edict further prohibits those monks who engage in
commerce and "tour the villages like ordinary people" from residing in
monasteries.37
798,9th month: Edict ordering all monks who have fathered offspring to be
returned to lay life.38
799: Directive ordering provincial governors and bishops to purge the
kokubunji of corrupt monks.39
An attempt was also made to raise the intellectual level of the
clergy. According to a directive issued by the Council of State in
734, the minimum requirements that a novice had to fulfill in order
to be ordained were (1) that he be able to recite from memory passages from either the Hokekyo (Lotus Sutra) or the Konkomyokyo
(Sutra of Golden Light) - the two scriptures chanted regularly at
the kokubunji for the protection of the state; (2) that he know the
rules for proper worship; and (3) that he have completed three
years of service as a novice.40 While such training might suffice for
a monk whose primary duty was to conduct routine services at a
kokubunji, it did not guarantee that a monk had any real understanding of Buddhist doctrine. Consequently, Kammu ordered in
798 that a candidate for admission to the clergy must take a qualifying examination (kanshi) in which he successfully answers five
out of ten questions on doctrine as expounded in the scriptures before he can be granted the status of a full-fledged novice.41 On the
day of the ordination proper, the candidate was required to pass a
second, more detailed examination (shinshi) in which he was expected to answer eight of ten questions. It was further stipulated
that only those who had learned the kan'on pronunciation of Chinese, which had recently been introduced from Ch'ang-an, would
be eligible for ordination. The requirement for the second examination, administered on the day of ordination, was dropped in 801,
apparently because the scope of the first qualifying examination
37 Ruiju kokushi, kan 187, edict dated Enryaku 17 (798)74/15, KT, vol. 6, p. 313. See also the
edict dated Enryaku 20 (801)74/15 in KT, vol. 6, p. 314, which lowers the minimum age for
ordination to twenty.
38 Ruiju sandai kyaku: kohen, kan 19, Enryaku 17 (798X9/17, KT, vol. 25, p. 621.
39 Nihon koki, kan 8, directive dated Enryaku 18 (799X5/19, KT, vol. 3, p. 21.
40 Shoku Nihongi, kan 11, directive datedTempyo 6 (734)/n/2i, KT, vol. 2, p. 135.
41 Ruiju kokushi, kan 187, edict dated Enryaku 17 (798X4/15, KT, vol. 6, pp. 313-14.
Cambridge Histories Online © Cambridge University Press, 2008
462
ARISTOCRATIC BUDDHISM
was expanded to include questions on the complex doctrines of the
Sanron and Hoss5 schools.4*
SAICHO
The Heian period was dominated by two schools, Tendai and Shingon, the former established by Saicho and the latter by Kukai, two
of the most important figures in the history of Japanese Buddhism.
Although both Tendai and Shingon were originally introduced from
China, they were transformed by their Japanese protagonists into
uniquely Japanese schools of Buddhism.
Saicho was born in 767 in Furuchi-g5 (part of the present-day
Otsu city) in Omi Province.43 At the age of eleven he "left his family" (shukke) to enter the Omi Kokubunji, where he came under the
tutelage of Gyohyo, the provincial bishop of Omi. Two years later
Saicho received his tokudo (initiation as a shami, "novice"). His full
ordination (Jukai, "accepting the [250] precepts") raising him to the
status of monk (so) took place at Todaiji in 785. Immediately thereafter Saicho moved to Mount Hiei, where he devoted himself to
meditation, worship, and especially the study of scripture. His
abrupt move to Hiei was in keeping with the practice of many monks
of the time, who sought to purify themselves and perhaps even acquire supernatural powers by undergoing austerities in the mountains. In Saicho's case, as the five vows he made at the time suggest,
the move to Hiei reflected disenchantment with the corruption that
was infecting the great monasteries of Nara.
Saicho probably first heard of Tendai from his master Gyohyo,
who had been a disciple of Tao-hsuan (Japanese, D5sen), a learned
Chinese monk who was said to have been versed in the doctrines of
42 Ruiju kokushi, kan 187, edict dated Enryaku 20 (8oi)/4/i5, KT, vol. 6, p. 314.
43 The earliest and most reliable source for the biography of Saicho is the Eizan Daishi den
compiled by his disciple Ninchu and included in DDZ, vol. 5, furoku, pp. 1-48. Another
important primary source providing valuable information regarding Siacho's date and
place of birth, family background, service as a novice, and various ordinations is the collection of ordination certificates included in the same volume on pp. 101-5. For critical
modern biographies of the life of Saicho, see Paul Groner, Saicho: The Establishment of the
Japanese Tendai School (Berkeley, Calif.: Berkeley Buddhist Studies Series, 1984); Katsuno
Ryushin, Hieizan to Kbyasan (Tokyo: Shibundo, 1959); Kiuchi Hiroshi, Dengyo Daishi no
shogai to shiso, Regurusu bunko, vol. 56 (Tokyo: Daisan bummeisha, 1976); Nakao Shumpaku, Dengyo Daishi Saicho no kenkyu (Kyoto: Nagata bunshodo, 1987); Saeki Arikiyo,
Dengyo Daishi den no kenkyu (Tokyo:Yoshikawa kobunkan, 1992); Shioiri Ryodo and Kiuchi
Gyoo eds., Saicho, vol. 2 of Nihon meiso ronshu (Tokyo:Yoshikawa kobunkan, ig82);Tamura
Koyu, Saicho, Jimbutsu sosho, vol. 193 (Tokyo: Yoshikawa kobunkan, 1988).
Cambridge Histories Online © Cambridge University Press, 2008
SAICHO
463
the Kegon, Ritsu, and Zen schools, in addition to those of Tendai.44
It was only after Saicho had settled on Hiei, however, that he was
able to acquire a set of the major Tendai treatises, which he then
studied with great enthusiasm. Gradually a small group of followers,
which included monks such as Gishin and Encho who were destined
to become major disciples, gathered around him. Official recognition of his learning came twelve years after his move to Hiei, when,
in 797, he was named one of the ten court chaplains (naigubu ju
zenji) whose responsibility was to pray for the well-being of the emperor. This appointment entitled his small temple on Hiei to receive
a subsidy paid from the Omi tax revenues.
The following year Saicho invited ten monks from Nara to hear a
series of lectures on the Hokekyo and two related sutras, which together constitute the basic scriptures of Tendai. These lectures, designed by Saicho to commemorate the anniversary of the death of
Chih-i, the Chinese systematizer of Tendai, provided him with an
opportunity to expound to the scholar-monks of Nara Tendai doctrine, which had been only briefly introduced by the Vinaya master
Chien-chen (Japanese, Ganjin), who had arrived in Japan in 754 carrying the major treatises of this school.45 Known as the Hokke jikko
(The Ten Lectures on the Lotus), these memorial lectures held annually on Hiei in the eleventh month subsequently became a major
event on the Tendai calendar.
In 802, Emperor Kammu, who was troubled by the frequent
wrangling between the Nara schools, particularly between Hosso
and Sanron, ordered Wake no Hiroyo, the head of the state Academy
(Daigaku) and eldest son of the loyalist Kiyomaro, and Hiroyo's
brother, the renowned scholar Matsuna, to arrange for lectures on
Tendai at the Wake clan temple Takaosanji (the predecessor to present-day Jingoji in Kyoto). Kammu promoted Tendai, apparently
hoping to provide some common ground for a resolution of the disputes between the Sanron and Hosso schools, since Tendai teachings
included the concept of progressive revelation, according to which
each of the major groups of scriptures had its own place in a grand
design devised by Sakyamuni Buddha to lead his followers to accept
44 Gyonen (1240-1321), Sangoku Buppo denzu engi, kan 2, BZ, vol. 101, p. 115a. In the same
work Gyonen quotes the now lost Tendai fuho engi, which is attributed to Saicho, as saying
that the Chinese monks Tao-hsiian, Chien-chen, and Fa-chin all disseminated the Tendai
teachings in Japan (pp. 1260-273).
45 Genkai, To Daiwajb tdsei den (779), BZ, vol. 113, p. 120a.
Cambridge Histories Online © Cambridge University Press, 2008
464
ARISTOCRATIC BUDDHISM
the Hokekyo as his final and highest teaching.46 Kammu's plan to use
Tendai as a unifying ideology for Japanese Buddhism was not unreasonable since Chi-tsang and Tz'u-en, the two most prominent
scholar-monks of the Chinese Sanron and Hosso schools respectively, had written major commentaries on the Hokekyo.
In the course of theTakaosanji lecture, in which Saicho played the
leading role, Kammu expressed his desire to see Tendai established
as a full-fledged Buddhist school in Japan. Saicho immediately responded, in a message relayed to the emperor by Hiroyo, that this
could be accomplished only if a mission was sent to China to create
a formal link with the Chinese patriarchate in the T'ien-t'ai mountains, where the school originated and maintained its head monastery. Without such a formal transmission of doctrine, Saicho insisted, Tendai would carry little authority in Japan. He also made
clear to Kammu his own conviction that Tendai was inherently superior to both Sanron and Hosso because the latter two schools were
based on treatises written by Indian scholiasts, whereas Tendai was
rooted in the Hokekyo, a scripture preached by Sakyamuni Buddha
himself. Saicho's proposal to undertake a mission to China was
promptly accepted, and he was granted permission to make a short
visit to the T'ien-t'ai mountains accompanied by his disciple Gishin,
who was to serve as his interpreter.
Sailing on one of the four ships that transported the Japanese embassy to the T'ang court, Saicho arrived at Ming-chou (the presentday Ning-p'o) in the ninth month of 804. En route to the T'ien-t'ai
mountains he stopped briefly at T'ai-chou (present-day Lin-hai),
where he met Tao-sui, the then patriarch of the Chinese Tendai
school. By the tenth month Saicho had reached T'ien-t'ai, where he
visited the holy sites and had a chance to study Tendai doctrine at its
source. A totally unexpected reward from his visit to T'ien-t'ai was
an encounter with a monk named Hsiao-jan, who initiated him into
the Gozu (Chinese, Niu-t'ou, "Ox Head") lineage of Zen (Chinese,
Ch'an). 47 The following month Saicho returned to T'ai-chou for further instruction in Tendai doctrine from Tao-sui, and, in the third
month of 805, on the eve of his return to the embarkation point of
46 For a brief account of Chih-i's classification of the major groups of sutras, see my "Imperial Patronage in T'ang Buddhism," in Perspectives on the T'ang, ed. Arthur F. Wright and
DenisTwitchett (New Haven, Conn.:Yale University Press, 1973), pp. 284-87.
47 In addition to the Gozu lineage of Zen, Saicho claimed to have received the transmission
of the Northern School (Hokushu) through both Gyohyo and Tao-hsiian. The latter had
been a pupil of P'u-chi, the Dharma-heir to the famous (Northern) Sixth Patriarch, Shenhsiu. See Saicho, Naisho Buppo sojo kechimyaku fu, DDZ, vol. 1, pp. 210-15.
Cambridge Histories Online © Cambridge University Press, 2008
SAICHO
465
Ming-chou, received from Tao-sui the endonkai (perfect and immediate precepts), which was a Tendai ordination based on the fiftyeight bodhisattva precepts (bosatsukai) taught in the Bomrriokyo.
Upon learning after his arrival in Ming-chou that the embassy's
departure for Japan was to be delayed, Saicho decided to use the
extra time allowed him in China to visit Yiieh-chou (present-day
Shao-hsing), where he hoped to find additional Tendai manuscripts
and, perhaps, also to acquire texts belonging to the Mikkyo (Esoteric
Buddhist) tradition. During his stay in Yiieh-chou, which occupied
most of the fourth month of 805, Saicho managed to receive an esoteric initiation (kanjo, "sprinkling of consecrated water on the
head") from one Shun-hsiao, who also provided him with many esoteric texts and several implements for use in esoteric rituals. In all,
Saicho collected 120 manuscripts in T'ai-chou and 102 manuscripts
in Yiieh-chou, most of the latter being Mikkyo works.48
Saicho sailed from Ming-chou on an embassy ship in the fifth
month and reached Kyushu in the middle of the sixth month. He was
immediately summoned to the court, where he personally presented
Kammu with the manuscripts and ritual implements that he had acquired in China. Not surprisingly, the emperor ordered that copies of
the Tendai texts be distributed to each of the seven great monasteries
of Nara. But what impressed Kammu, whose health was now failing,
even more than the precious Tendai manuscripts was Saicho's newly
acquired status as a practitioner of Mikkyo, in which interest had
been steadily growing since Nara times because of its practical value
for curing illnesses, preventing misfortunes, and producing various
benefits. It might well have struck Saicho as ironic that the first service that Kammu ordered him to perform after his return to Japan
was not related to Tendai, which he had gone to China specifically to
study, but rather to the Mikkyo, which was at best only of secondary
interest to him. By imperial decree a platform-altar for esoteric initiations {kanjodari) was constructed at Takaosanji, where, in the ninth
month of 805, Saicho performed for eight monks from Nara the first
esoteric initiation rites ever held in Japan. Later the same month
Saicho was summoned to the palace to conduct an esoteric ritual that
would bring about the recovery of the ailing emperor.49
48 For a list of the titles of the manuscripts acquired in T'ai-chou and Yiieh-chou, see Dengyo
Daishi shorai Taishu roku and Dengyo Daishi shorai Esshu roku, both compiled by Saicho (T,
v
°'- 55J PP- 10553-583 and pp. io58b-6ob).
49 Eizan Daishi den, DDZ, vol. ^furoku, pp. 21-24. Kammu's faith in the esoteric Buddhism
transmitted by Saicho is attested in a proclamation issued by Kammu included in Saicho's
Kenkairon engi (821), DDZ, vol. 1, pp. 283-84.
Cambridge Histories Online © Cambridge University Press, 2008
466
ARISTOCRATIC BUDDHISM
In the first month of 806, Saicho sent a petition to Kammu requesting thatTendai be formally accorded status as one of the recognized Buddhist schools.50 To accomplish this, Saicho proposed that
the traditional system of ordaining ten monks at the beginning of the
new year to pray for the well-being of the nation be restructured and
expanded to include representatives of each of the officially recognized schools.51 The Kegon, Tendai, and Ritsu schools were to be allocated two novices each; the Sanron, to which the Hinayanist Jojitsu
school was attached, and the Hoss5, to which the Hinayanist Kusha
school was appended, were to be assigned three novices each, for a
total of twelve annual ordinands (nembun dosha). The government
promptly accepted Saicho's proposal, which had won immediate
backing from the hierarchy (sogo), but stipulated that of the two ordinands allotted annually to the Tendai school only one should devote
himself solely to the study of classical Tendai doctrine. The other candidate was to study and become a specialist in Mikkyo. Thus from its
inception the Japanese Tendai school, unlike its Chinese parent, had
Mikkyo as one of its major components. The reconciliation of Mikkyo
with classical Tendai thought was to become one of the principal
tasks for future generations of Tendai scholars.
With the death of Kammu in the third month of 806 Saicho lost a
strong supporter. The new emperor, Heizei (reigned 806-9), seeking
to reduce government expenditures, placed restrictions on the construction of new temples and the use of state revenues for religious
purposes. Because of this new policy no Tendai monks were ordained
until 810, when Heizei's successor, Saga, allowed eight Tendai
novices to be tonsured at the court, thus compensating for Heizei's
failure to honor Kammu's promise to Saicho. Saga's decision to
carry out the ordinations created practical difficulties for Saicho,
since one of the two annual Tendai ordinands had to be trained in
Mikkyo, which Saicho himself had not had the opportunity to study
properly in China. Another problem was that Saicho's library on Hiei
was lacking many important Mikkyo texts. To acquire copies of these
texts and also to supplement his obviously deficient knowledge of
Mikkyo, Saicho turned to Kukai, a monk seven years his junior,
whose understanding of Mikkyo was without equal in Japan.
50 For Saicho's proposal, the statement by the hierarchs, and the official response by the government, see Kenkairon engi, DDZ, vol. 1, pp. 292-96.
51 The practice of ordaining ten monks at the court on the last day of the year or at the beginning of the new year was begun in 696. See Nihon shoki, kan 30, Jito 10 (6g6)/i2/i,
NKBT, vol. 68, p. 532, and Saicho's Kenkairon (819), kan 3, DDZ, vol. 1, p. 150.
Cambridge Histories Online © Cambridge University Press, 2008
SAICHO
467
Unlike Saicho, who went to China to study Tendai but by chance
happened to encounter Mikkyo adepts who conferred on him lowlevel initiations, Rukai visited China with the specific objective of
mastering the doctrines and rituals of Mikkyo and received its highest initiations. Although both men sailed in the same flotilla, they
traveled on different ships and probably became acquainted with
each other only after their return to Japan. Saichd's ship, as we have
noted, landed in Ming-chou, whereas the ship carrying Kukai entered the port of Fu-chou, whence Kukai proceeded in the entourage
of the Japanese ambassador directly to Ch'ang-an, where he intensively studied Mikkyo for more than a year. When Kukai returned to
Kyushu in the tenth month of 806, he had in his possession a priceless collection of esoteric texts, ritual implements, paintings, and
mandalas (graphic representations of various divinities, often portrayed through mystical symbols and arranged according to a pattern
that emanates outward from a central point).52
In the eighth month of 809, Saicho sent a disciple to Kukai, who
had taken up residence at Takaosanji a month earlier, bearing a letter requesting the loan of twelve esoteric texts. Over the next six
years Saicho wrote almost thirty such letters, often signing them
"your disciple Saicho," even though Kukai was seven years his junior.53 Although Saicho himself had received several esoteric initiations while in China and had performed such an initiation at
Takaosanji in 805, he openly acknowledged Kukai's superior understanding of Mikkyo. Toward the end of 812 Saicho visited Kukai at
Takaosanji to request the initiation based on the kongokai (diamond
realm) and taizokai (embryo realm) mandates, which are the two
principal mandalas of the line of Mikkyo transmitted by Kukai that
subsequently came to be known in Japan as Shingon Mikkyo. Kukai
readily assented, but conferred on Saicho only a kechien kanjo (an
initiation establishing a link), which is the most elementary of the
various levels of initiation.54 Despite Saicho's eminence as a Tendai
monk, his previous esoteric initiations in China, and his subsequent
52 Kukai's catalogue, the Go-shorai mokuroku (T, vol. 55, pp. io6oa-66a) lists the titles of 216
works that he brought back to Japan.
53 Saicho's letters to Kukai are included in the Rankeiyuionshlt, Mikkyo bunka kenkyujo, ed.,
Kobo Daishi zenshu, 3rd ed., revised and enlarged (zoho), 8 vols. (Koyasan: Mikkyo bunka
kenkyujo, 1965-68), vol. 5, pp. 353-86.
54 Although virtually all Shingon scholars hold that Saicho received only the introductory initiation, there is a tradition within the Tendai school, based on a letter Saicho's disciple
Encho wrote to Kukai in 831, that Kukai conferred an intermediate level ordination on
Saicho. See my "Beginnings of Esoteric Buddhism in Japan: The Neglected Tendai Tradition," Journal ofAsian Studies 34, 1 (1974): 188.
Cambridge Histories Online © Cambridge University Press, 2008
468
ARISTOCRATIC BUDDHISM
self-study of esoteric texts, he was in Kukai's eyes still an amateur in
Mikkyo, a point that was driven home when Kukai granted Saicho
only the lowest level of initiation at a ceremony in which laymen also
participated.
Saich5 and Kukai each viewed Mikkyo differently. For Saicho,
Mikkyo and classical Tendai formed the two wings of the newly established Tendai school, a unique amalgam not found in China.55 To
Kukai, however, Mikkyo was the ultimate teaching of Buddhism and
fully constituted a school in its own right. It is not surprising, therefore, that the two men would inevitably part company. The first indication of serious difficulty was Kukai's refusal, in 814, to lend
Saicho an esoteric manuscript that he had requested. Kukai sharply
rebuked Saicho for trying to understand Mikkyo through texts
alone, which, Kukai asserted in a letter to Saichd, were no more than
the "dregs of Buddhism." Truth, in other words, Mikkyo, could be
transmitted only "from mind to mind." To teach Mikkyo without
having received a proper transmission, Kukai warned, was tantamount to "stealing the doctrine."56 The relationship between Saicho
and Kukai ended on a bitter note in 816, when Saicho's disciple, Taihan, who at Saicho's urging had gone to study Mikkyo with Kukai
four years earlier, refused Saicho's request that he return to Hiei.
The break with Kukai marked the end of Saicho's period of docility, as was indicated by his decision to circulate publicly his Ehyo
Tendai shu, a polemical work written in 813 that sought to document
the superiority of Tendai over all other schools. Once the rupture became final in 816, Saicho embarked on a tour of the Kanto region,
where he laid the basis for a future Tendai stronghold by lecturing
on the Hokekyo, establishing pagodas enshrining this sutra and proclaiming before large groups of rural people the Tznd&ilHokekyo
doctrine of One Vehicle (ichijo), namely, that the three traditional divisions of Buddhism known as the Three Vehicles (sanjo) were no
more than an expedient device created by Sakyamuni, the historical
Buddha, to lead people of different intellectual and spiritual capacities to the One Vehicle that will ultimately carry each and every sentient being to Buddhahood.
By publicly proclaiming the Tendai doctrine of universal enlightenment, Saicho openly challenged the influential and aristocratic Hosso
55 See, for example, Saicho's letter toTaihan (in DDZ, vol. 5, p. 469), written in 816, in which
Saicho declares that the teachings of the Hokekyo and those of Shingon are equally true.
56 The letter is included in the anthology of Kukai's writings entitled Henjo hakki Seirei shu,
in Sango shiiki, Seirei shii, vol. 71 of NKBT, pp. 442—50. The passage cited occurs on p. 447.
Cambridge Histories Online © Cambridge University Press, 2008
SAICHO
469
school, which took the opposite view - namely, that the doctrine of
One Vehicle taught in the Hokekyo was merely an expedient teaching
intended to encourage simpleminded people to put their faith in Buddhism and uphold its basic moral code. For the Hosso school, the ultimate teaching of Buddhism was to be found in the Gejimmikkyo
(Sutra Explaining the Profound Doctrine), its principal scripture,
which not only accepted the concept of three real, distinct vehicles,
but also held that sentient beings were inherently divided into five
groups (gosho), the lowest consisting of the luckless musho (those lacking the Buddha-nature), who, strive as they might, were destined to
wander eternally through the cycle of birth and death. Saicho's popular gospel of universal salvation was immediately denounced by the
well-known Hosso scholar Tokuitsu, who likewise was active in the
Kanto area. Over the next five years the two men produced a total of
eight works in an effort to refute each other's positions.
Having severed relations with Kukai and having become involved
in a protracted doctrinal dispute with Tokuitsu, Saicho was now
ready to dissociate himself completely from the traditional Buddhism that centered around the six Nara schools. It was Saicho's
view, but not that of Chinese Tendai, that the Shibunritsu - the disciplinary code used in both China and Japan to ordain monks and
nuns - was essentially a Hinayanist work and hence not suitable for
Mahayanist ordinations. In the third month of 818 Saicho took the
unprecedented step of formally renouncing the 250 precepts of the
Shibunritsu that he had taken at the time of his ordination atTodaiji.
Two months later Saicho requested approval from the throne for a
set of six regulations that he had formulated and wished to make
binding on all future Tendai ordinands.
Formally known as the Tendai Hokke-shu nembun gakusho shiki
(Bylaws for the Annual Ordinands of the Tendai Hokke School), the
new regulations had far-reaching implications for the future course
of Japanese Buddhism.57 Particularly significant were the following
proposals: (i)The names of candidates for ordination should not be
removed from family registers, as was the custom, but retained with
the added notation "son of the Buddha" (Busshi). Under the prevailing law, when someone was accepted as a novice his name was
deleted from the family register, which was under the control of the
secular authorities, and entered into a clerical register (sdseki), which
placed him directly under the supervision of the Nara hierarchs
57 The text is included in Saicho's Sange gakusho shiki, T, vol. 74, pp. 6z3C-24b.
Cambridge Histories Online © Cambridge University Press, 2008
47O
ARISTOCRATIC BUDDHISM
(sogo). (2) The tokudo (initiation as a novice) and jukai (full ordination as a monk) should take place in the same year. It had been the
practice to receive the tokudo from a monk at one's "home temple"
and the full ordination several years later from preceptors belonging
to the Ritsu school at one of the three monasteries authorized to
have ordination platforms (kaidari) :T6daiji in Nara,Yakushiji in Shimotsuke, and Kanzeonji in Chikuzen. By linking the jukai with the
tokudo, Saicho hoped to keep Tendai novices out of the hands of the
Ritsu preceptors. (3) Ordinations should be based on the Busshi kai
(precepts for sons of the Buddha), an ambiguous term coined by
Saicho suggestive of the bosatsukai (bodhisattva precepts) in the
Bomniokyo. (4) Newly ordained monks should be required to reside
on Hiei for an uninterrupted period of twelve years. As Saicho was
to point out later, only ten of the twenty-four annual ordinands selected between 807 and 818 remained on Hiei, the others having
been "stolen" (his word) by the Hosso (six monks) and Shingon
(one monk) schools or else having left for reasons of their own.58 (5)
Tendai monks, regardless of whether they specialized in esoteric rituals (shanago) or traditional Chinese Tendai meditation (shikango),
should view the protection of the state (gokoku) as their primary
concern. (6) Those monks who exhibit special talents after completing their twelve-year training period should be appointed to serve as
proselytizers or provincial bishops. In addition to their religious
tasks, these monks should also actively promote the public welfare
by sponsoring the construction of irrigation ditches, the reclamation
of farmland, the building of bridges, and other such projects.
The court forwarded Saicho's proposals to the Office of Hierarchs
(Sogo-sho), which did not comment on them, perhaps because
Saicho's occasionally vague language left the hierarchs uncertain
about how far he was prepared to go in establishing Tendai as a
school completely independent of the established church. In the
eighth month of 818, Saicho submitted to the court another document containing eight proposed bylaws regarding the administration
of Hiei and the training of its monks.59 Again, the Office of Hierarchs, to which the document was referred, remained silent.
Any doubts regarding SaichS's ultimate intentions were dispelled
when, in the third month of 819, he presented a third set of bylaws
58 For a list of the students, with notations indicating their reasons for leaving Hieizan and
the names of the schools to which they defected, see Saicho, Tendai Hokke-shu nembun
tokudo gakusho myocho, DDZ, vol. 1, pp. 250-53.
59 T h e document, entitled Kansho Tendai-shu nembun gakusho shiki, is included in Sange
gakusho shiki, p. 624b—c.
Cambridge Histories Online © Cambridge University Press, 2008
SAICHO
471
to the court for consideration. Reflecting his increasing impatience
with the court's failure to act on his earlier proposals for a truly independent Tendai school that he believed had been sanctioned by his
imperial patron Kammu, Saicho provocatively entitled his new set of
regulations Tendai Hokke-shu nembun dosha esho kbdai shiki (Bylaws
for the Conversion of the Annual Ordinands of the Tendai Hokke
School from Hinayana to Mahayana).60 The three chief points made
in this final set of new regulations were: (1) There are three categories of monasteries: (a) those exclusively Mahayanist, (b) those exclusively HTnayanist, and (c) those in which Mahayana and Hinayana
coexist. Tendai ordinands and "those who converted to Mahayana,"
that is, monks originally belonging to one of the Nara schools who
had subsequently joined Tendai, should be required to spend twelve
years on Hiei, which in Saicho's view would become the only truly
Mahayanist monastery in Japan. (2) There are two types of precepts:
(a) the fifty-eight Mahayanist ones of the Bommokyo and (b) the 250
Hlnayanist ones of the Shibunritsu. (3) There are two types of ordinations: (a) the Mahayanist one based on the Kanfugengyo, in which
the Buddha and two bodhisattvas act as preceptors and (b) the
Hlnayanist one based on the Shibunritsu, in which three senior
monks serve as the preceptors. Tendai novices should, of course, receive the Mahayanist type of ordination using Mahayanist precepts.
This last set of proposals signaled the beginning of a reform unprecedented in the history of East Asian Buddhism. First, Saicho's insistence that Tendai monks live in an "exclusively Mahayanist monastery" marked the emergence in Japan of the sectarian monastery,
which subsequently became one of the hallmarks of Japanese Buddhism. Horyuji, Daianji, Gangoji, and other large Nara monasteries
each accommodated groups of monks belonging to different schools.
It was only after the time of Saicho that the idea of an exclusive sectarian monastery or temple took root and became the norm.
Second, Saicho broke completely with previous East Asian Buddhist practice when he replaced the Shibunritsu precepts with those
of the Bommokyo. This latter set of precepts, which lays particular
stress on the social responsibility of the individual, had been traditionally viewed as precepts intended primarily for the bodhisattva
(bosatsukai), that is, precepts for laymen, which monks might also
voluntarily choose to accept, as Saicho himself had done in China.
Monks had always been minutely regulated in their monastic life by
60 Sange gakusho shiki, T, vol. 74, pp. 624c—25b.
Cambridge Histories Online © Cambridge University Press, 2008
472
ARISTOCRATIC BUDDHISM
the 250 precepts of the Shibunritsu, which were thought to transcend
such relativistic categories as Hlnayana or Mahayana. By labeling as
Hinayanist the Shibunritsu precepts, which had hitherto formed the
basis of all ordinations, Saicho was implying that there were no true
Mahayana monks in Japan.
Third, Saicho repudiated the ordination system prevailing in East
Asia when he categorized as Hinayanist the traditional ordination
ceremony in which three monks administer the precepts to the ordinands, as prescribed in the Shibunritsu. By proposing to substitute
the Kanfugengyo, a sutra linked to the Hokekyd and particularly esteemed inTendai, for the Shibunritsu, Saicho was creating an entirely
new ordination system peculiar to Tendai, the practical effect of
which was to make Tendai a completely independent school, no
longer dependent on the Ritsu monasteries for the ordination of its
clergy. To accomplish this, Saicho sought permission to establish his
own ordination platform on Hiei, beyond the jurisdiction of the
Nara hierarchs, where he could perform his own ordinations.61
Outraged by Saicho's third set of proposed bylaws for Tendai ordinands, the Office of Hierarchs sent a sharply worded memorial to
Saga in the fifth month of 819 denouncing Saicho's views on precepts, ordination, and the training of monks and urged the government to reject Saicho's proposals. Saicho responded some ten
months later with his famous Kenkairon (Treatise on the Precepts),
which refuted in fifty-eight articles the arguments against him put
forward by the hierarchs. Along with the Kenkairon he submitted another important work, the Naishd Buppo sojd kechimyaku fu, in which
he sought to demonstrate that he - and hence the Japanese Tendai
school - was the legitimate heir to four distinct traditions: (1) Tendai
proper through his study in China under Tao-sui and Hsing-man;
(2) Zen through his master Gy5hyo and the transmission that he received later in China from Hsiao-jan; (3) the Bodhisattva (that is,
Mahayanist) Precepts through the ordination he received at T'ient'ai shan based on the Bommokyo; and (4) Mikkyo through the initiations by Shun-hsiao and Wei-hsiang.
The Kenkairon was forwarded by the court to the hierarchs, who
61 The full text of Saicho's petition to the throne for permission to conduct ordinations is
given in Eizan Daishi den, DDZ, vol. 5,furoku, pp. 33-34. Although ninth-century records
do not specifically report that he sought the approval of the emperor to build a kaidan, it
is likely that he did so, since it had been the custom for ordinations to be performed on a
kaidan. Saicho's biography in Kokan Shiren, Genko Shakusho (1322), states unambiguously
that in the third month of 819, Saicho requested permission for the construction of a
kaidan (BZ, vol. 101, p. 149b).
Cambridge Histories Online © Cambridge University Press, 2008
KUKAI
473
declined to respond. In 821 Saicho presented the court with yet another work, the Kenkairon engi, defending the principle of independent Tendai ordinations, but again was met with silence from the hierarchs and consequently from the throne. Frustrated by his failure
to get permission from the government to conduct his own ordinations, Saich5 spent the final year of his life quietly managing the affairs of Hiei. In response to a personal appeal from one of Saicho's
disciples, Emperor Saga on his own authority promoted Saich5 to
the highest ecclesiastical rank, dai koshii, an honor that had already
been granted to Saicho's junior, Kukai, two years earlier. Less than
four months later, on the fourth day of the sixth month of 822,
Saicho died on Hiei, without having gained the permission he so fervently sought to construct a Tendai ordination hall. Seven days after
Saicho's death, Saga, without consulting the hierarchs whose opposition was well known, agreed to a petition signed by four of Saicho's
prominent lay supporters, including Fujiwara no Fuyutsugu, who
was the Minister of the Right, and Yoshimine no Yasuyo, a son of
Kammu and half brother of Saga, to allow ordinations on Hiei. The
following year an imperial decree was issued granting the name Enryakuji to Saich5's monastery on Hiei in memory of Kammu whose
reign was known as Enryaku.Two months later the first Tendai ordinands, fourteen in all, received the Bommdkyo precepts from Gishin,
who had succeeded Saicho as abbot of Hiei. The long-awaited ordination hall (kaidan'in) was completed in 827 with a grant of 90,000
sheaves of rice to defray construction costs.62
KUKAI
Kukai, the founder of the Japanese Shingon school, was born in 774
in Sanuki Province (present-day Kagawa Prefecture in Shikoku),
where his family, surnamed Saeki, exercised considerable influence.63 At the age of fourteen he was brought to the capital by his
maternal uncle, Ato no Otari, who was the Confucian tutor to Prince
Iyo, the third son of Emperor Kammu. After three years of intensive
study of the Chinese classics under the tutelage of his uncle, Kukai
entered the state Academy with the intention of eventually establishing himself as a scholar of Chinese. Shortly thereafter, however,
62 Denjutsu isshinkai mon, kan 2, DDZ, vol. 1, pp. 588-90.
63 For critical modern biographies, see Katsuno, Hieizan to Koyasan; Kushida Ryoko, Kukai
no kenkyu (Tokyo: Sankibd Busshorin, 1981), and Watanabe Shoko and Miyasaka Yusho,
Shamon Kukai, Chikuma sosho, vol. 84 (Tokyo: Chikuma Shobo, 1967).
Cambridge Histories Online © Cambridge University Press, 2008
474
ARISTOCRATIC BUDDHISM
he encountered a monk who taught him the esoteric ritual known as
Kokiizd gumonjihd, the purpose of which is to increase the powers of
memory through incessant repetition of a mystical incantation.64
After beginning the gumonjihd, Kukai had some sort of deep religious
experience, which led him to withdraw from the Academy and retreat to the mountains where he undertook austerities.
According to the traditional accounts dating from late Heian
times, Kukai entered Makiosanji in Izumi at the age of nineteen to
start his training as a novice under Gonzo, an influential Sanron
monk, and was ordained two years later atTodaiji.65 However, in his
first book, the Sangd shiiki, written in 797 when he was twenty-three,
Kukai gives no indication of being an ordained monk. On the contrary, it is apparent from the Sangd shiiki, a semiautobiographical
work that seeks to demonstrate the superiority of Buddhism over
Taoism and Confucianism, that Kukai had an aversion to the formalistic and often corrupt Buddhism of the great monasteries. His
sympathies clearly lay with the itinerant holy man (hijiri) who, although often lacking a proper ordination, spends his life searching
for truth while bringing the word of the Buddha to the common
people. In all likelihood such was the life led by Kukai between 791
when he retreated to the mountains and 804, when, according to the
Shoku Nihon koki (compiled in 869), a more reliable source than the
late Heian biographies, he first became a novice, presumably in
order to be eligible for study in China.66 His full ordination took
place at Todaiji in the fourth month of the same year.67
Virtually nothing definite is known about Kukai's study of Mikkyo
before his visit to China. The traditional biographies claim that
Kukai first learned of the Dainichikyo, one of the basic Mikkyo scriptures, in a dream.68 After locating a copy of this text beneath a
64 Most biographies of Kukai, dating from the end of the Heian period or later, identify the
unnamed monk as Gonzo, a respected cleric who stood in the Sanron lineage. This view
has been challenged by some contemporary scholars. For a summary of their arguments,
see Shimode Sekiyo, "Kukai to Shingon-shu," in Nihon hen, vol. 2 of Nakamura Hajime,
Kasahara Kazuo, and Kanaoka Shuyu, eds., Ajia Bukkyd shi (Tokyo: Kosei shuppansha,
1974), pp. 134-38, andWatanabe and Miyasaka, Shamon Kukai, pp. 34-39.
65 See, for example, Kyohan, Daishi on-gyojo shuki (1089), in Zoku gunsho ruijii, vol. 8 (Tokyo:
Keizai zasshisha, 1904), pp. 495-96, and Ken'i, Kobb Daishi go-den (first half of the 12th
century) in Zoku gunsho ruijlt, vol. 8, p. 526.
66 Shoku Nihon koki (869), kan 4, Jowa 2 (835)73/25, KT, vol. 3, p. 38.
67 Zo Daisojo Kukai Wajo denki (895), Kobo Daishi zenshu, shukan, p. 9.
68 The earliest reference to this well-known legend occurs in the Go-yuigo (T, vol. 77, p. 408c),
which purports to have been written by Kukai on his deathbed. Although it is no longer
generally recognized as an authentic work of Kukai's, it was accepted as such since Heian
times and hence served as a source for many of the biographies of Kukai.
Cambridge Histories Online © Cambridge University Press, 2008
KUKAI
475
pagoda at Kumedera in Yamato, these accounts relate, Kukai resolved to travel to China, where he could receive proper instruction
from Mikkyo masters regarding the meaning of this difficult scripture. Although the traditional biographies make Kukai's encounter
with the Dainichikyo appear miraculous, the simple fact is that esoteric texts were in use in Japan long before Kukai's birth. By the end
of the Nara period more than 130 such texts, including the Dainichikyo and the Kongochokyo, the other principal scripture of Shingon
Mikkyo, had been brought to Japan.69 More than one quarter of the
150 surviving images from the eighth century are representations of
Mikkyo divinities, which is another indication of the inroads that
Mikkyo had already made in Japan.70 Although not yet recognized as
an independent school in Nara times, iconographic and textual evidence shows that Mikkyo was widely known and practiced even before Kukai undertook his journey to China.
We have no reliable information about how Kukai managed to get
permission from the Japanese government to study in China. Traveling in the company of the ambassador, Fujiwara no Kadonomaro,
on whose behalf he drafted letters to the Chinese authorities, Rukai
reached Ch'ang-an at the end of 804. In the sixth month of 805 he
was accepted as a disciple by Hui-kuo, who was recognized as the
foremost master of esoteric Buddhism in China. According to Kukai's account, Hui-kuo, who was then ailing, declared at their first
encounter that he had been long awaiting the arrival of Kukai, whom
he formally designated as his successor.71 After receiving the initiations based on the kongokai and taizokai mandalas, Kukai was
granted the dembo kanjo (the initiation for transmitting the Dharma),
which is the highest of the three levels of esoteric initiations and signifies that its recipient has become a dembo ajari, that is, an esoteric
master {ajari) who is himself empowered to transmit the teachings.
Before Hui-kuo's death in the twelfth month of 805 he passed on to
Kukai a reputed relic of the Buddha, various paintings, images, ritual implements, texts, and mandalas. Although Kukai had originally
planned to spend twenty years in China, he decided to return to
Japan immediately after Hui-kuo's death so that he could disseminate the "orthodox" Mikkyo that he had learned from Hui-kuo.
Kukai traveled back to Japan with the embassy of Takashina no
69 Ishida Mosaku, Shakyoyori mitaru Nara-cho Bukkyo no kenkyu (Tokyo: Toyo bunko, 1930),
p. 146.
70 Katsumata Shunkyo, Mikkyo no Nihon-teki tenkai (Tokyo: Shunjusha, 1970), p. 10.
71 Go-shdrai mokuroku, T, vol. 55, p. 1065b—c.
Cambridge Histories Online © Cambridge University Press, 2008
476
ARISTOCRATIC BUDDHISM
Tonari, reaching Kyushu in the tenth month of 806. Through the
ambassador he submitted to the court a catalogue of the precious esoteric manuscripts and religious objects that he had acquired in
China, hoping to be summoned to the capital to transmit the
Mikkyd teachings. Although later biographies claim that he moved
to Makiosanji the following year, it is clear from earlier records that
despite his unique training in China, no warm welcome awaited
Kukai in Kyoto.72 His exclusion from the capital was almost certainly due to political happenstance. In 807, one year after ascending
the throne, Heizei sent his half-brother, Prince Iyo, into exile on suspicion of plotting mutiny, ultimately forcing him to commit suicide.
Kukai, through his uncle Otari, who had been Prince Iyo's tutor, had
had a long-standing friendship with the Prince, which, no doubt
tainted Kukai in the eyes of Heizei.
Kukai's fortunes abruptly changed when Saga became emperor in
the fourth month of 809. The new emperor had a deep interest in
Chinese culture, particularly literature, poetry, and calligraphyareas in which Kukai excelled. In the seventh month of 809, Saga invited Kukai to take up residence in Takaosanji, where four years earlier Saicho had erected an altar for esoteric initiations (kanjddan) at
the behest of Saga's father, Kammu. A close relationship immediately developed between Saga and Kukai through frequent exchanges of poetry and calligraphy. Saga, who was not especially concerned with religion, esteemed Kukai primarily for his profound
knowledge of Chinese culture. For Kukai the strong personal bond
with Saga provided entree to the court and aristocracy and gave him
the opportunity to lay a solid foundation for Shingon Mikkyo.
Kukai lost no time in trying to convince the court of the practical
value of Shingon. In the wake of an unsuccessful coup to restore Retired Emperor Heizei to the throne in 810, Kukai successfully petitioned Saga for permission to hold an esoteric rite at Takaosanji to
ensure the tranquillity of the country, using the occasion to remind
Saga of the prevailing Chinese custom of maintaining a permanent
palace chapel (naidojo) staffed by monks who were expert practitioners of esoteric ritual.73 Kukai's preeminence as a Mikkyd master
72 The three earliest biographies of Kukai - the Kukai Sozu den attributed to his disciple
Shinzei, the Daisozu Kukai den compiled by Fujiwara no Yoshifusa (804—72) et al., and the
Zo Daisqjo KukaiWajo denki—make no mention of Kukai's whereabouts between his arrival
in Tsukushi in Kyushu in 806 and his move to Kyoto in 809.
73 Kukai's petition, entitled Kokke no on-tame ni shuho sen to kou hyd (A Memorial Requesting Buddhist Services for the Benefit of the Nation), is contained in the Seirei shu, kan 4,
NKBT, vol. 71, pp. 228-30.
Cambridge Histories Online © Cambridge University Press, 2008
KUKAI
477
was acknowledged by everyone, including his future rival Saicho.
When Kukai acceded to Saicho's request in 812 to conduct kongokai
and taizokai initiations, close to two hundred people flocked to
Takaosanji to take part.74 Among those receiving initiations from
Kukai were not only Saicho and his disciples, but also twenty-two
monks from the major Nara monasteries, thirty-eight novices, and
more than forty laymen, including the influential Wake brothers,
Matsuna and Nakayo, who effectively controlled Takaosanji. In
recognition of Kukai's status as a Mikkyo master, Saicho instructed
five of his leading disciples to remain with Kukai until the following
year so that they might receive the highest level of initiation, the
dembo kanjb, from him. After receiving their esoteric initiations from
Kukai in the twelfth month of 812, the Wake brothers entrusted the
administration of Takaosanji to Kukai, allowing him to appoint his
own disciples as the ranking temple officers (sango). In 829 the Wake
brothers formally vested Kukai and his successors with irrevocable
authority to manage Takaosanji, which had been made an officially
sanctioned Shingon temple five years earlier.
At the very time that Saicho had embarked on a course of confrontation with the older schools, Kukai, following a conciliatory policy, was winning acceptance for Shingon at the court and among the
Nara hierarchs. In 816 Saga approved a request from Kukai that he be
given exclusive proprietary rights to Mount Koya, on which he sought
to establish a retreat for meditation and a monastery for the training
of Shingon monks. Two years later Kukai moved to Koya, where two
of his disciples were already laying foundations for the future Kongobuji monastic complex. Despite his strong desire to remain on Koya,
Kukai was summoned back to Kyoto in 819, his presence in the capital being deemed indispensable. Kukai strengthened his links with
Nara by establishing in 822 a hall for esoteric initiations {kanjo ddjo)
within the precincts of Todaiji. This hall, later known as Shingon'in,
which remained under the control of Shingon monks, played a key
role in the dissemination of Mikkyo among the Nara clergy.
In the first month of 823, three months before his abdication, Saga
ordered that Kukai be put in charge ofToji, the chief state-supported
temple in Kyoto. Saga's successor, Junna, stipulated later in the same
year that henceforth only Shingon monks would be permitted to reside at Toji, fixing their number at fifty. In 824, Junna further decreed that control of Toji would be permanently vested in Kukai and
monks of his lineage. The conversion of Toji into an exclusive Shin74 For a list of participants, see Kukai's Takao kanjo ki, Kobd Daishi zenshu, vol. 3, pp. 620-29.
Cambridge Histories Online © Cambridge University Press, 2008
478
ARISTOCRATIC BUDDHISM
gon temple was a clear indication that Rukai's Shingon Mikkyo had
become the officially sanctioned religion of the court.
Unlike Saicho, who after 816 was continually at odds with the
older schools, Kukai always remained on good terms with the Nara
clergy in the hope of persuading it to accept Shrigon as a supersectarian system of esoteric ritual that would bring an infinite variety of
benefits to both the individual and the state. Thus, while Saicho encountered stiff opposition in his efforts to build an independent ordination platform on Hiei, Kukai succeeded in establishing a hall for
Shingon services in the very heart of Nara. Even the Hosso scholar
Gomyo, who as the ranking hierarch had denounced Saicho's proposal for an independent ordination platform, felt no contradiction
in accepting an administrative position at Toji after it was designated
a Shingon temple. Nor was it unusual for Shingon monks to be
named to the superintendency (betid) of Todaiji, which was nominally a Kegon monastery. So completely had Kukai succeeded in
winning acceptance from the hierarchs that when he submitted to
the throne in 830 his Jujushin ron, a treatise in which he proclaimed
the superiority of Shingon over all other schools of Buddhism as well
as over the secular philosophies, not a word of protest was heard
from Nara.
Kukai finally received permission to return to K5ya in 832. Two
years later the court granted his petition to establish within the
palace precincts a Shingon'in (Shingon Chapel), where an annual
esoteric rite known as go-shichinichi mi-shiho was to be performed
from the eighth to the fourteenth day of the first month by Shingon
monks from T5ji. The purpose of the rite was to ensure the wellbeing of the emperor and the prosperity of the nation. In the first
month of 835, the Shingon school was officially admitted to the annual ordinand {nembun dosha) system and was allotted a yearly quota
of three state-supported novices, which exceeded by one the number allowed the Tendai school. The following month Kong5buji was
accorded the status of a recognized temple (jogakuji). Thus by the
time of Kukai's death in the third month of 835, Shingon was firmly
established within the imperial court, a position it was to hold until
the first years of Meiji.
THE TENDAI SCHOOL AFTER SAICHO
Although the Tendai school gained independence from the Nara hierarchy when it was granted the right to conduct its own ordinations
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
479
in 822, it nevertheless fared poorly in the first decades after Saicho's
death. By the 820s the Japanese aristocracy had become intoxicated
with Mikkyo rituals, which were performed to achieve specific material or spiritual ends. Saicho's disciples and followers on Hiei had
been completely overshadowed by Kukai, who had established himself in Toji, where he performed various esoteric rites commissioned
by the government. The plight of the monks on Hiei after the death
of Saicho was graphically illustrated in a letter sent in 825 by its lay
superintendent (zoku betto), Tomo no Kunimichi, to an official who
was an influential patron of Horyuji, requesting the latter to use his
good offices to find a place for the "foodless monks of Hiei" at
Horyuji and Shitennoji so that they could continue to transmit the
teachings of Tendai.75
It was clear to the Hiei community that in order to enhance its
standing with the aristocracy, which was more interested in the material rewards accruing from esoteric rites than in the lofty but impractical philosophy of classical Tendai, it would have to prove its
competence in Mikkyo, which, as we have seen, had been viewed by
Saich5 as one of the two wings of the Japanese Tendai school. This
need to acquire additional training in Mikkyo forced Encho, one of
Saicho's disciples and a future abbot of Enryakuji, to swallow his
pride and appeal, in 831, to Kukai for further instruction in the doctrines of Mikkyo, for which, he frankly admitted, he had as yet not
been able to find a suitable teacher.76 Obviously, such appeals were
humiliating and, worse still, tended to confirm the view that Rukai's
Shingon school was the ultimate authority in matters of Mikkyo.
Ennin, Enchin, and Annen
In 835, Ennin, a relatively obscure monk who had become a disciple
of Saicho in 808 at the age of fourteen, was granted permission to
travel to China with Fujiwara no Tsunetsugu, the newly appointed
envoy to the T'ang court. Although the purpose of Ennin's mission
was ostensibly to visit T'ien-t'ai shan in order to receive further
training in Tendai doctrine, it is apparent that Ennin planned from
the outset to take advantage of his stay in China to learn as much as
possible about Mikkyo.
After several false starts, the ship carrying Ennin arrived atYang75 Denjutsu isshinkai mon, kan 2, DDZ, vol. 1, pp. 592-93.
76 Encho's letter, which was also signed by nine other monks from Hiei, is included in Chbya
gunsai, kan 16, KT, vol. 29A, pp. 397-98.
Cambridge Histories Online © Cambridge University Press, 2008
480
ARISTOCRATIC BUDDHISM
chou in the seventh month of 838.77 While awaiting permission from
the Chinese authorities to travel toT'ien-t'ai shan, Ennin sought out
two Mikkyo adepts inYang-chou, who gave him instruction in shittan, the peculiar Indie script used to write esoteric formulas, and allowed him to make copies of manuals explaining esoteric rituals.
Ennin lost no time in securing an initiation {kanjo) from one of these
adepts, Ch'iian-ya, who was a third-generation disciple of Kukai's
master Hui-kuo.
In the second month of 839 Ennin was told by local officials that
there was not sufficient time for him to visit T'ien-t'ai shan since he
was, strictly speaking, not a student-monk (rugakuso) but a scholar
attached to the embassy and hence had to leave China with the embassy, which was due to sail later that month. By the time Ennin's
ship reached Wen-teng at the tip of the Shantung peninsula, the last
anchorage before crossing the Yellow Sea, Ennin had resolved to remain behind in China even though he lacked permission from the authorities. After spending the winter at a monastery in Shantung run
by Korean monks, Ennin set out in the second month of 840 for Wut'ai shan, a major place of pilgrimage in Shansi, in order to consult
with Chih-yuan and several other learned Tendai monks who resided
in one of the many temples dotting the mountain. While there is no
reason to doubt that the opportunity to meet Tendai scholars was a
principal factor in Ennin's decision to travel to Wu-t'ai, it should also
be noted that Wu-t'ai was renowned as a center for popular Mikkyo
cults, one of its most prominent temples, Chin-ko ssu, having been
founded by the great esoteric master Pu-k'ung (Amoghavajra), who
was the teacher of Kukai's mentor, Hui-kuo. During his fifty-day stay
on Wu-t'ai, Ennin spent about two weeks at the Chu-lin ssu, where
he learned the goe nembutsu, a rhythmical chanting of the name of
Amida Buddha using five different intonations. As we shall see, this
nembutsu practice was later to have a profound influence on the development of Pure Land Buddhism in Japan.
In the eighth month of 840, Ennin arrived in Ch'ang-an, where he
was to spend almost five years engaged in the study of Mikkyo while
systematically acquiring esoteric texts, formulas written in the shittan
script, and copies of mandalas. He received instruction and esoteric
initiations from three Mikkyd masters, each a second-generation dis77 The primary source for Ennin's stay in China is his very rich diary entitled Nittd guhojunrai koki. For a critical edition with full annotation, see Ono Katsutoshi, Nittd guho junrai
koki no kenkyu, 4 vols. (Tokyo: Suzuki gakujutsu zaidan, 1964-69). The diary has been
translated into English by Edwin O. Reischauer under the title Ennin's Diary: The Record of
a Pilgrimage to China in Search of the Law (New York: Ronald Press, 1955).
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
481
ciple of the illustrious Hui-kuo, who had administered the taizokai
and kongokai initiations to Kukai. Not only did Ennin likewise receive
both these initiations, but he was also granted a third type of initiation known as the soshitsuji kanjo (initiation according to the Soshitsuji
Sutra), which symbolically integrated the taizokai and kongokai initiations. Ennin took particular pride in having been admitted to this
third level of initiation, since it had not been offered to Kukai even
though it was evidently transmitted within Hui-kuo's lineage.
Life in Ch'ang-an became increasingly difficult for Ennin after the
fanatically anti-Buddhist emperor Wu-tsung began the persecution
of the clergy in 842. In the fifth month of 845, Ennin was defrocked
and ordered to return to Japan. He immediately left Ch'ang-an for
the port of Teng-chou in Shantung, which he reached three months
later. After an agonizing delay of two years, he finally found passage
on a Korean merchant ship bound for Japan. Ennin landed in
Kyushu in the ninth month of 847 carrying 584 texts, mostly esoteric, in 802 fascicles, 21 ritual implements, and a collection of religious paintings, and mandalas.78 Still more important, he had spent
five years in Ch'ang-an studying with the leading esoteric masters of
the day, whereas Kukai had been under the tutelage of Hui-kuo for
only half a year before the latter's death. Having received a total of
thirteen different initiations in China,79 Ennin was now in a position
to challenge the Shingon monopoly of Mikky5.
Ennin's extraordinary command of Mikkyo, which was promptly
recognized by the court, was instrumental in reversing the decline of
Hiei. In the sixth month of 848, a mere three months after his arrival
in Kyoto, the court accepted Ennin's proposal that it sponsor annual
esoteric initiation rites "for the enhancement of the imperial cause
and the protection of the state."80 These rites, popularly known as
Hiei kanjo, were first performed at Enryakuji in 849, the government
providing support for more than a thousand monks who participated in the ceremony.81 Until Ennin's return from China Shingon
78 Ennin, Nitto shingu shogyo mokuroku, T, vol. 55, p. 1078c.
79 This figure appears in the Jusan-ju kanjo hiroku, a record claiming to have been compiled
by Ennin and transmitted to his disciple Shoun. See Tendai kahyo, kan 5, part 1, BZ, vol.
126, pp. 526b-27b.
80 The text is included in Ruiju sandai kyaku: zempen, kan 2, Kasho 1 (848)/6/i5, KT, vol. 25,
pp. 69-71.
81 Nihon sandai jitsuroku: zempen (901), kan 8, Jogan 6 (864)71/14, KT, vol. 4, p. 126. For an
early description of the Hiei kanjo, see Minamoto no Tamenori, Sambo ekotoba, written in
984, kan 3, BZ, vol. 111, pp. 4698-708; English translation by Edward Kamens, The Three
Jewels: A Study and Translation of Minamoto Tamenori's Samboe (Ann Arbor, Mich.: Center
for Japanese Studies, 1988), pp. 349-52.
Cambridge Histories Online © Cambridge University Press, 2008
482
ARISTOCRATIC BUDDHISM
monks had complete responsibility for the various esoteric rites that
were performed at the court. This exclusive control of Mikkyo ritual
by Kukai's disciples was brought to an end by Ennin, who easily
forged close links with the imperial family and aristocracy thanks to
the exceptional prestige that accrued to him as a result of his lengthy
studies and religious training in Ch'ang-an. From 850 on Ennin was
repeatedly invited to the imperial palace to conduct esoteric rites or
administer lay precepts to members of the imperial family. His influence with the aristocracy reached its pinnacle in 856 when he performed the twofold initiation (ryobu kanjo) for Emperor Montoku,
the Crown Prince, Minister of the Right Fujiwara no Yoshifusa, the
latter's adopted son Mototsune, and Major Counselor Fujiwara no
Yoshisuke.
Ennin's standing at the court brought immediate gains to Hiei. In
850 Emperor Montoku, at the urging of Ennin, ordered the construction of Sojiin, an imperially endowed cloister on Hiei accommodating fourteen monks, which was to serve as a hommyo dojo, a
chapel in which prayers for the tranquillity of the nation were offered
to the star governing the year of the emperor's birth.82 During the
same year the Tendai school was authorized two additional annual
ordinands, one to specialize in the Kongochokyo and the other in the
Soshitsujikyo, thereby giving formal recognition to the particular version of Mikkyo transmitted by Ennin from China. In 854 the court
appointed Ennin abbot of Enryakuji, formally designating his office
zasu (head of the community), a title subsequently held by all succeeding abbots, indicating that the incumbent was not under the jurisdiction of the Nara hierarchy.83 Ennin's success in establishing
Tendai Mikkyo as the predominant form of Buddhism at the court
was clearly demonstrated in 866, two years after his death, when the
government conferred the honorary posthumous names Dengyd
Daishi and Jikaku Daishi on Saicho and Ennin respectively.84 By
contrast, Kukai was not granted the posthumous name Kobo Daishi
until half a century later, in 921.85
It was the good fortune of the Tendai school that Ennin was followed on Hiei by a line of eminent monks who could further consolidate the close relationship he had established with the imperial
82 Nihon sandai jitsuroku: zempen, kan 8, Jogan 6/1/14, KT, vol. 4, p. 126.
83 The two abbots preceding Ennin, namely, Gishin and Encho, were officially designated
Demboshi, "Master Who Transmit the Dharma." See Shibuya Ryotai ed., Kotei zoho Tendai
zasu ki (Tokyo: Daiichi shobo, 1973), kan i, p. 8.
84 Choya gunsai, kan 17, KT, vol. 29A, pp. 424—25.
85 Nihon kiryaku (ca. 12th century): kohen, kan 1, Engi 21 (921)710/18, KT, vol. 11, p. 24.
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
483
family and Fujiwara. Four years after Ennin's death, the scholarmonk Enchin was installed as abbot of Enryakuji. Like Ennin,
Enchin had spent five years in China studying Mikkyo in Ch'ang-an
and Tendai doctrine at T'ien-t'ai shan, which Ennin had not been
able to visit. In 859, the year after his return to Japan, Enchin accepted an invitation to become abbot {chdri, "head officer") of
Onjoji (also called Miidera), in Otsu, a temple belonging to the
Otomo clan. Enchin, who had not been a disciple of Ennin's, seemed
to have thought it prudent to keep a respectable distance from Hiei
during the incumbency of Ennin in order not to detract from this senior monk by appearing as a competitor.
Immediately after Ennin's death, Enchin was summoned to the
imperial palace to administer esoteric initiations to Emperor Seiwa,
Fujiwara no Yoshifusa, and other eminent figures. In 868, Enchin
was named abbot of Enryakuji, succeeding Ennin's disciple Anne,
who had died that year. In time of drought or illness, or when a new
emperor ascended the throne, the court invariably turned to Enchin
for the appropriate prayers and rituals. By the time of his death in
891, Enchin is said to have conferred the exalted esoteric rank of
ajari on over 100 individuals, personally tonsured more than 500
monks and administered the precepts to some 3,000 persons.86
Moreover, he had secured a decree guaranteeing that future abbots
of Onjoji would be chosen exclusively from monks belonging to his
own lineage.
Owing to the efforts of Ennin and Enchin, Japanese Tendai was
transformed into a thoroughly esoteric school, with scant attention
being given to the teachings of traditional Chinese Tendai doctrine.
In addition to the frequent esoteric rituals and initiations they performed in the palace or on Hiei to enhance the standing of the
Tendai school, Ennin and Enchin also sought to provide a solid theoretical foundation for Tendai Mikky5 (also known as Taimitsu, as
opposed to Kukai's Shingon Mikkyo, also called Tdmitsu) by writing a number of doctrinal works defining the relationship of the
Hokekyo to the Mikkyo scriptures. Saicho, preoccupied with his
struggle to win approval for an independent ordination platform
and his polemics against his Hosso rival, Tokuitsu, never adequately
addressed this question. His view, as indicated in a letter to his exdisciple Taihan, was that the revealed teachings of the Hokekyo and
the secret teachings of the Mikkyo scriptures were equally true since
86 Sontsu, Chishd Daishi nempu (1467), BZ, vol. 28, p. 1395b.
Cambridge Histories Online © Cambridge University Press, 2008
484
ARISTOCRATIC BUDDHISM
both these scriptures took the position that there was ultimately
only the One Vehicle (ichijo) that would lead all beings to Buddhahood. Saicho believed that both the Hokekyo and the esoteric scriptures rejected the notion that the HInayana constituted a separate
vehicle for enlightenment and hence stood in opposition to
Mahayana. Rather, HTnayana was seen as an expedient device established by Sakyamuni to lead the ignorant gradually to an appreciation of Mahayana and through that to the attainment of Buddhahood itself.
Contrary to the position of Saicho and Chinese Tendai, Ennin asserted that the Hokekyo was in fact an esoteric scripture.87 He gave
two reasons for this view. First, the Hokekyo teaches the doctrine of
One Vehicle, which, according to Ennin's unique interpretation, was
the principal criterion for defining Mikkyo. Second, the earthly Buddha Sakyamuni who preached the Hokekyo revealed himself in the
second half of the sutra to be an eternal Buddha and therefore must
be identified with Dainichi Nyorai, the Buddha who expounded the
esoteric scriptures. Thus, with respect to doctrine, Ennin viewed the
Hokekyo as being on par with the "pure" esoteric scriptures such as
the Dainichikyo and Kongochokyo. But as Ennin admitted, Mikkyo
also stresses various secret practices - incantations, hand signs, mandala initiations, and so on - which are not mentioned in the Hokekyo
but figure prominently in esoteric scriptures. Ennin concluded,
therefore, that although the Hokekyo could be characterized as "esoteric in doctrine" (rimitsu), it ranked below the "pure" Mikkyo scriptures, since the latter were esoteric in both the doctrines and rituals
that they taught (jiri gumitsu).
Enchin saw an even wider gap between the Hokekyo and the
"pure" Mikkyo scriptures. While recognizing in principle that the
Hokekyo could be deemed an esoteric scripture, Enchin nevertheless included it among the exoteric scriptures on the grounds that
it had been openly preached by the Buddha. Since the Hokekyo did
not teach the secret practices of Mikky5, Enchin held that it was inferior to the "pure" esoteric scriptures, which he collectively designated "the king of Mahayana, the most secret of the secret" (Daijdchii no o, hi-chu no saihi).88 This tendency to upgrade steadily the
status of Mikkyo within the Tendai school culminated in the extra87 For a succinct account of the teachings of Ennin, Enchin, and Annen, see Katsumata,
Mikkyo no Nihon-teki tenkai, pp. 275-91.
88 Daibirushanakyo shiki, T, vol. 58, p. 19b.
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
485
ordinary ranking of the various Buddhist doctrines by the great
scholar Annen (841-ca. 895), who devised an elaborate systematization of Tendai Mikkyo. According to Annen, who was a junior
contemporary of Enchin, Mikkyo, as revealed in the "pure" esoteric
scripture, represented the ultimate teachings of Buddhism. Despite
his own Tendai affiliation - he actually belonged to the lineages of
both Ennin and Enchin - Annen held that classical Tendai ranked
below both Mikkyo and Zen, the latter, in his view, being a
nonverbal - hence "secret" - transmission of the Buddha's enlightenment. Taking the view that Mikky5 was the highest teaching
within Buddhism to its logical conclusion, Annen referred to his
school not as Tendai but as Eizan Shingonshvi (the Shingon School
of Hiei), as opposed to the Shingon school of Kukai.
Hiei prospered as a result of the close contacts that Ennin,
Enchin, and subsequent abbots of Enryakuji had established with
the imperial family. Successive emperors founded cloisters on Hiei:
Montoku built the Sojiin and Shioin; Suzaku, the Emmyoin; Kazan,
the Joryoin; and Go-Reizei, the Jissoin. Gifts of land by ex-emperors
and princes for the general upkeep of Hiei or for the endowment of
specific annual rituals became commonplace after the 860s. The first
such major grant was made in 863 by Emperor Nimmyo's sons,
Saneyasu and Tsuneyasu, whose households together contributed
224 cho of tax-exempt land.89 In addition, substantial amounts of
tax-rice were allocated for the support of specific cloisters such as
Shakado, Joshin'in, Sojiin, and Shi5d5. By 972, Hiei controlled estates inOmi,Wakasa, andYamashiro in central Honshu and Izumi in
the west.90 In the hope of expanding its influence in the countryside,
Hiei persuaded the government to designate many of the recognized
temples (jogakuji) that were located in the provinces "Tendai betsuin" (Tendai branch temple), which signified that these temples
would henceforth be administered only by monks ordained at Enryakuji. By the end of the ninth century Tendai betsuin were found in
such diverse regions as Mutsu, Kozuke, Harima, Shinano, and
Kaga.91
89 Nihon sandai jitsuroku: zempen, kan 7, Jogan 5 (863)/4/io, KT, vol. 4, p. 110.
90 For a list of the donations made by the imperial family and the aristocracy to Hiei in the
ninth and tenth centuries, see Murayama Shuichi, "Heian Bukkyo no tenkai," in Ienaga
Saburo, ed., Kodai hen, vol. 1 of Nihon Bukkyo-shi (Kyoto: Hozokan, 1967), pp. 244-45.
91 For a list of Tendai betsuin, see Sonoda Koyii, "Saicho to Tendai-shu," in Ajia Bukkyo-shi:
Nihon hen, vol. 2, p. 131.
Cambridge Histories Online © Cambridge University Press, 2008
486
ARISTOCRATIC BUDDHISM
Ryogen
The appointment of Ryogen as abbot of Enryakuji ushered in a period of increased grandeur and power for Hiei.92 In 937, at the relatively young age of twenty-five, Ryogen won acclaim for his skill in
outwitting the learned monks of the Nara monasteries in the annual
debate on the Yuimagyo (the so-called Yuima-e) held at K5fukuji.
Word of this extraordinary achievement soon reached the Chancellor, Fujiwara no Tadahira, who, together with his influential son,
Morosuke, and grandsons, Koretada and Kaneie, subsequently became devoted patrons of Ryogen. In 964 Ryogen was named a court
chaplain (naigubu) and the following year, at the age of fifty-three,
became the youngest monk to be admitted to the hierarchy (sogo)
since its creation in the seventh century.
Three months after Ryogen's appointment as abbot of Enryakuji
in 966, Hiei was ravaged by a fire that destroyed thirty buildings, including such important structures as the Sojiin, Lecture Hall
(Kodo), Shioin, Emmyoin, and Jogyodo.93 The devastation of Hiei
was now almost complete, since Hiei had not yet fully recovered
from an earlier fire that occurred in 935 and claimed many buildings,
including Kompon Chudo, the main hall of Enryakuji.94 Undaunted
by the magnitude of the task, Ryogen vowed to rebuild Hiei during
his own lifetime. Besides the support that he could count on from
the imperial family, whose endowed cloisters such as Sojiin, Shioin,
and Emmyoin had been lost in the conflagration, Ryogen could also
draw upon the income from eleven estates that Morosuke had bequeathed to his son, Jinzen,95 who at Morosuke's urging had become
Ryogen's disciple in 958, two years before his father's death.
Ryogen mobilized all the resources at his command to carry out
the reconstruction of Hiei. In 967 the Hokkedo and Jogyodo were
completed; in 972 ceremonies were held marking the reconstruction
of the Lecture Hall and four other buildings; in 979 the Shakado, the
Sutra Repository for the Hodoin, and the Jeweled Pagoda (Hoto)
were finished, and in 984 the ornate Hodoin was completed, the gold
for the temple fittings having been supplied by Fujiwara no Tame92 For critical studies of the life of Ryogen and his contributions to the Tendai school, see
Hirabayashi Moritoku, Ryogen, Jimbutsu sosho, vol. 173 (Tokyo: Yoshikawa kobunkan,
1976), and Eizan Gakuin, ed., Ganzanjie Daishi no kenkyu (Kyoto: Dobosha, 1984).
93 Nihon kiryaku:kohen, kan 4, Koho 3 (966)/io/28, KT, vol. 11, p. 99.
94 Ryogen, Tengen sannen Chudo kuyo gammon (980), reprinted in Koji ruien, 51 vols. (Tokyo:
Yoshikawa kobunkan, 1967-71), vol. 12 (shiikyd bu 4), p. 563.
95 Hirabayashi, Ryogen, p. 58.
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
487
naga, governor of Mutsu. But the high point in the restoration of Hiei
was reached in 980, when Kompon Chudo was formally consecrated
in a service conducted by Ryogen in which Emperor En'yii, Regent
Fujiwara no Yoritada, Minister of the Left Minamoto no Masanobu,
and Minister of the Right Fujiwara no Kaneie participated.
During the rebuilding of Enryakuji, Ryogen paid special attention
to Yokawa, which together with Toto (Eastern Pagoda) and Saito
(Western Pagoda) comprise the Santo (Three Pagodas), that is, the
geographical areas into which Hiei has been traditionally divided.
The Toto area, in which Kompon Chudo, the Lecture Hall, and the
Ordination Hall are situated, was originally developed by Saicho.
Saito, which is to the west of Toto, centers around Shakado built by
SaichS's disciple Encho. Yokawa, which lies to the north of Toto and
Saito, was first settled by Ennin, who withdrew to this desolate part
of Hiei in 831 in order to undertake spiritual exercises in preparation
for his death, which he believed to be imminent. After his recovery
Ennin established Shuryogon'in, which became the central cloister
of Yokawa.
Following Ennin's appointment as abbot of Enryakuji, he moved
to Zentoin in Toto, entrusting the management of Yokawa to his disciples. Although Yokawa continued to be administered by monks belonging to Ennin's lineage, its fortunes declined rapidly after Enchin
assumed the abbotship of Enryakuji. When Ryogen's patron, Tadahira, died in 949, Ryogen, who stood in Ennin's lineage, retreated to
Yokawa where he offered prayers for the repose of the deceased
Tadahira. At Morosuke's request, he also beseeched the divinities to
grant Morosuke's daughter, who was the consort of Emperor Murakami, a male child, which would put Morosuke in the position of
being the grandfather of a potential successor to the throne. The
birth of a boy, Prince Norihira (the future Emperor Reizei), in 950
convinced Morosuke of Ryogen's supernatural powers. He immediately had Ryogen appointed "protector-priest" (gojiso) for his grandson and announced his intention of restoring Yokawa, where the
prayers had been offered. In 954 Morosuke, accompanied by his eldest son Koretada, visited Yokawa to dedicate the nearly completed
Hokke zammai'in, which was to function primarily as a prayer cloister for Morosuke's branch of the Fujiwara clan. Morosuke felt a particular attachment to the Yokawa community, which he expressed in
a vow promising support to the lineage of Ennin. Kaneie, the third
son of Morosuke, continued to promote the expansion of Yokawa,
building Yakushid5 and Eshin'in in 983. The number of monks resiCambridge Histories Online © Cambridge University Press, 2008
488
ARISTOCRATIC BUDDHISM
dent in Yokawa increased from the two or three that Ryogen found
there in 949 to more than two hundred by the time of his death.96
Although Ryogen, like his predecessors, regularly performed esoteric rituals for his imperial and aristocratic patrons and wrote several treatises on Mikkyo, he deplored the decline in classical Tendai
learning on Hiei, which became particularly acute after the death of
Enchin. As a way to stimulate Tendai scholarship, Ryogen proposed,
upon his appointment as abbot, that the annual services commemorating SaichS's death include a formal five-day debate on the
Tendai interpretation of the Hokekyo. The format of the debate
called for an examinee (rissha) to lecture on ten doctrinal problems
{dai) put to him by five questioners {monja), who would then challenge his explanation. If the monk who acted as a judge (tandai) determined that the examinee had successfully withstood the challenges, the examinee would be accorded the much coveted academic
rank of tokugd (scholar). These debates, formally known as kogaku
ryiigi (erudite disputations) or, more popularly, yama no rongi (the
[Hiei] mountain debates), subsequently played an important role in
reviving Tendai scholarship.
In addition to encouraging formal Tendai learning, Ryogen sought
to tighten discipline on Hiei. Four years after becoming abbot, he issued a code in twenty-six articles, called Nijurokka jo kisho, which
was intended to curb widespread abuses by the Hiei clergy.97 Several
articles prohibit offerings of food or gifts to priests officiating at services; monks are told that they should rehearse and be thoroughly
familiar with the various rituals they are expected to perform, that
they should attend all services, that they should not leave the mountain during their twelve-year period of training, that they must not
wear elegant garb, and that they must not keep horses on Hiei. While
some of these regulations restate proscriptions issued by Saicho in
818 and Ennin in 866, other articles in Ryogen's code clearly bespeak
the secularization of the clergy that was beginning to occur in the
powerful monasteries. In this category are the regulations calling for
the expulsion of monks who conceal their faces (kato), carry
weapons, inflict corporal punishment, or violently disrupt religious
services - behavior suggestive of the akuso (wicked monks), later
96 Shimaji Daito, Tendai kyogaku shi (Tokyo: Meiji shoin, 1929), p. 353.
97 The text of the Nijurokka jo kisho has been published inTakeuchi Rizo, ed., Heian ibun, 15
vols. (Tokyo: Tokyo-do, 1947-80), vol. 2, pp. 4318-4403. For a critical study of this important document, see Hori Daiji, "Ryogen no Nijuwkkajo kisho seitei no igi," Shiso, 25
(March 1967): 12—39.
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
489
popularly known as sohei (monk warriors). During his nineteen-year
incumbency Ryogen ordered that the names of seven hundred wayward monks who did not participate in services be expunged from
the Hiei registers.98
Ryogen was honored for the successful prayers that he offered for
his imperial and aristocratic patrons and was eventually deified for his
restoration of Enryakuji. In recognition of his having brought about
the recovery of Emperor En'yu in 981, he was granted the privilege of
being allowed to enter the palace grounds in a palanquin and was
awarded the supreme ecclesiastical rank of daisdjo (archbishop), the
first time that this rank had been conferred since the eighth century.
In 987, two years after Ryogen's death, the honorary posthumous
name Jie Daishi was conferred on him." By the end of the Heian period, Ryogen, popularly known as Ganzan Daishi (The Great Master
Who Passed Away on the Third Day of the New Year), was variously
viewed as a reincarnation of Saicho, Ennin, or even Kannon. It was
widely believed that Ryogen, instead of departing for Pure Land on
his death, remained on Hiei as its protector. Rubbings containing his
likeness were widely used in Kamakura times as amulets to ward off
malevolent spirits, cure illness, and avert natural disasters.100
The Schism within the Tendai school
The origin of the conflict that ultimately split the Tendai school into
opposing Sammon (Hiei) and Jimon (Onjoji) factions in the late
tenth century can be traced back to successional disputes among
Saicho's disciples. When Saicho fell ill in 812, he designated Encho,
his disciple of longest standing, to be his successor. On his deathbed
ten years later, however, Saich5 indicated that the leadership of the
Hiei community should be entrusted to Gishin, who had traveled
with Saicho to China, where he received ordinations identical to
those of Saicho's. After their return from China in 805, Gishin did
not remain on Hiei, but went back to his native Kanto to proselytize.
It was only in 813, a year after Saicho's illness, that Gishin took up
permanent residence on Hiei.
98 Sonoda Koyu, "Heian Bukkyo," in Kodai, vol. 4 of Ienaga Saburo, Ishimoda Sho, Inoue
Kiyoshi et al., eds., Iwanami koza:Nihon rekishi (23 vols.) (Tokyo: Iwanami shoten, 1962),
P- 19499 Nihon kiryaku: kohen, kan 9, Eien 1 (987)72/16, KT, vol. n , p. 161.
100 Hazama Jiko, Tendai shushi gaisetsu (Tokyo: Daizo shuppan, 1969; rev. by Okubo Ryojun),
p. 124. For an up-to-date list of temples containing halls or chapels dedicated to the worship of Ryogen, see Ganzan Jie Daishi no kenkyu, pp. 299-316.
Cambridge Histories Online © Cambridge University Press, 2008
49°
ARISTOCRATIC BUDDHISM
Although regarded as something of an outsider by Saicho's immediate disciples, Gishin was nevertheless accepted by them as the
first abbot of Enryakuji in accordance with Saicho's last request.
However, when Gishin, on his deathbed, ordered in 833 that the abbotship be passed on to his own disciple Enshu, who had only recently moved to Hiei, Saicho's leading disciples protested to the lay
superintendent of Hiei that Gishin was attempting to replace
Saicho's lineage with his own. After eight months of wrangling the
authorities finally agreed to dismiss Enshu from the office of abbot
and appoint Encho in his place.101
Gishin's lineage on Hiei, however, did not come to an end with
Enshu's departure. On the contrary, the appointment of Enchin, a
disciple of Gishin, as abbot of Enryakuji in 868 assured it a prominent position within the Tendai school. During Enchin's incumbency, which spanned twenty-three years, his followers came to
overshadow those of the equally prestigious Ennin, who had been a
direct disciple of Saicho. While Enchin himself had no quarrel with
his predecessor Ennin - indeed, he held him in high regard - he was
aware of the possibility of future trouble between their respective
groups of followers and hence before his death specifically urged his
own disciples to honor the memory of Ennin.102
Monks of the two lineages - Ennin's and Enchin's - managed to
avoid an open breach until 980 when Rydgen, who belonged to the
Ennin lineage, failed to invite a representative of the Enchin lineage to
participate in the lavish ceremonies dedicating the newly rebuilt Kompon Chudo.1O3 The Enchin faction, taking this as a deliberate affront
since representatives of other schools had been asked to take part in
the services, appealed to the court, which ordered Ryogen to include
Yokei, who, as abbot of Onjoji, was the head of the Enchin faction.
The first violent confrontation between the two factions occurred
the following year, when Yokei was named tenth abbot of Hosshoji,
a temple founded in 925 by Ryogen's future benefactor, Tadahira.
The nomination of Yokei infuriated the Ennin faction, which petitioned the court to remove Yokei's name on the grounds that all previous abbots had been selected from their own faction. When the
101 The relevant documents are included in Denjutsu isshinkai mon, DDZ, vol. 1, pp. 640-45.
See also Kotei zoho Tendai zasu ki, p. 8.
102 Enchin, Yuisei, BZ, vol. 28, pp. 1349b—50b, and also his Seikai mon, cited inTsuji Zennosuke, Nihon Bukkyb-shi: josei hen (Tokyo: Iwanami shoten, 1944), p. 827.
103 For a discussion of the major events leading to the split between the Ennin and Enchin
lineages, see Takagi Yutaka, "Tendai-shu no tenkai," in Ajia Bukkyo-shi: Nihon hen, vol. 2,
pp. 1820-858.
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
491
court refused to do so, asserting that spiritual attainment and high
intellect were the only proper criteria for nominating an abbot, more
than two hundred monks belonging to the Ennin lineage on Hiei, including several holding high office, carried out a violent demonstration at the residence of Fujiwara noYoritada, the Chancellor and lay
superintendent of Enryakuji.The willingness of the Ennin faction to
resort to force was sufficient to intimidate Yokei and several hundred
other monks of the Enchin faction, forcing them to quit Hiei and
seek refuge in nearby temples. Meanwhile, a stalwart band of three
hundred monks of the Enchin faction held fast at Senjuin atop Hiei,
which had been Enchin's residence and was the repository for his
manuscripts. It was so widely rumored that Ryogen was planning to
order an assault on all temples sympathetic to the Enchin faction, including Senjuin, that he was compelled to issue a denial.104 Calm returned to Hiei only when Yokei yielded to his opponents and resigned the abbotship of Hosshoji.
After the Hosshoji incident the Ennin faction was even more determined to maintain its preeminent position on Hiei, which had
been firmly established during the nineteen-year incumbency of
Ryogen. When word reached Hiei in 989 that Yokei had been nominated by the court to be the new abbot of Enryakuji, replacing
Ryogen's disciple Jinzen, who had just retired, the Ennin faction
protested vigorously, declaring that they would seal the halls of Enryakuji rather than admit an abbot from the Enchin lineage. The imperial envoy carrying the official notification found his path blocked
by irate clerics. When a second envoy was dispatched several days
later accompanied by the police officials, angry monks snatched the
court order from his hand as he attempted to read it. After an interval of several weeks the court made a third attempt, this time sending a ranking official, Fujiwara no Arikuni, backed by a large detachment of police, directly to Zentoin, where Ennin was enshrined,
in the hope that the monks would refrain from violence at so holy a
place. The court order was read aloud before Ennin's spirit tablet,
along with an indictment of the Hiei monks for their unruly behavior. But all to no avail, for when Yokei arrived on Hiei, the monks refused to perform any services under his direction. Faced with such
intransigence, Yokei had no choice but to resign as abbot, a mere
three months after his appointment.105
104 Ftiso ryakki (ca. 1150), kan 27, KT, vol. 12, pp. 250-51.
105 Nihon kiryaku: kohen, kan 9, Eiso 1 (9895/9/29 and Eiso 1/12/27, KT>
vo
Cambridge Histories Online © Cambridge University Press, 2008
'-
lx
> PP- 167-68.
4Q2
ARISTOCRATIC BUDDHISM
In 993 Yokei's disciple Josan, smarting from the humiliation suffered by his mentor, who had died two years earlier, assembled a
group of monks at Kannon'in in Kita Iwakura (in present-day
Kyoto), a stronghold of the Enchin faction, from which he carried
out a raid on Sekisan Zen'in, the cloister on Hiei where Ennin had
lived, reportedly damaging many artifacts that had belonged to
Ennin.106 The Ennin faction retaliated by destroying more than forty
buildings on Hiei affiliated with the Enchin faction and putting to
flight more than a thousand of its monks, most of whom subsequently settled at Onjoji. This temple thus became the headquarters
of the Enchin lineage, which subsequently was known as the Jimonha (Temple Branch) of the Tendai school, as opposed to the
Sammonha (Mountain Branch), in other words, the temples of Hiei
that were under the permanent control of monks in the Ennin lineage. Although the monks who had fled to Onjoji made no attempt
to return en masse to Hiei - to do so would clearly have invited more
violence - they asserted that members of their own faction should be
considered eligible for appointment to the abbotship of Enryakuji
since, in their view, the abbot of Enryakuji was the head of the entire Tendai school and not merely the chief administrator of the temples on Hiei.
This principle was first put to the test in 1038, when My5son, the
learned abbot of Onjoji, was nominated by his lay patron, Regent
Fujiwara noYorimichi, to fill the vacant abbotship of Enryakuji.107
Hiei responded with an enormous show of force: three thousand
monks descended on Yorimichi's residence demanding that he
choose a monk from Hiei. When he refused, citing Myoson's great
attainments, the outraged mob began dismantling the locked gates
of Yorimichi's residence. Troops were summoned, and a bloody
melee ensued. Several weeks later Yorimichi agreed to appoint
Kyoen, the candidate of the Hiei monks. When Kyoen died in 1047,
Yorimichi, ignoring the outcry from Hiei, managed to force through
the appointment of Myoson as Kyoen's successor, but owing to the
fierce opposition on Hiei the seventy-seven-year-old Myoson was
compelled to resign after only three days in office. By the end of the
twelfth century, seven other monks from Onjoji had succeeded in securing appointments to the abbotship of Enryakuji, but none was
able to hold office for more than a few days.
106 Fuso ryakki, kan 27, pp. 260-61.
107 For a detailed account of this episode, see Shikd, Jitnon denki horoku (1397), kan 19, BZ,
vol. 127, pp. 426a—27b.
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
493
The militarization of the clergy
As their secular power and wealth grew, the larger monasteries - particularly Kofukuji, Todaiji, Enryakuji, and Onjoji - increasingly resorted to force when they felt their interests threatened. Although
some of their fighting men fell into the category of servitors who
were recruited from monastic estates, the majority came from the
large body of minimally educated monks and temple hangers-on,
who seem to have been more at home with weapons than with Buddhist scripture. The employment of monks in a military capacity was
striking proof of the debased character of large segments of the
clergy, since the Mahayanist disciplinary code used in China and
Japan specifically forbade monks to carry weapons or engage in any
form of belligerent activity.108
The first major incident involving violence by clerics occurred in
949, when fifty-six monks from Todaiji gathered at the residence of
an official in Kyoto to protest an appointment that displeased
them.109 A brawl ensued, claiming the lives of several of the participants. In 969 a dispute over conflicting claims to temple land resulted in the death of several Kofukuji monks at the hands of monks
from Todaiji.
While such violent clashes became commonplace in the eleventh
and twelfth centuries, monasteries generally tried to achieve their
objectives through intimidation rather than brute force. Typically,
thousands of protesting monks would surround an important government building or the residence of a high official and refuse to disperse until their demands had been met. Because of their clerical
status and the sanctity of the religious implements they carried,
these monks were not often subjected to rough treatment by the secular authorities. After 968, Kofukuji clerics repeatedly terrorized the
Fujiwara by storming into Kyoto carrying branches of the sacred
sakaki tree especially consecrated for the occasion by Kasuga
Shrine, which was under the control of Kofukuji. Since Kasuga was
the clan shrine of the Fujiwara, the latter were compelled to seclude
themselves in their residences {rokyo) whenever such a demonstration took place. Another device used by Kofukuji in dealing with ob108 See Bommokyo, kan 2, T, vol. 24, p. 1005c, which forbids the possession of "swords, clubs,
bows, arrows, halberds, axes, and implements of combat."
109 For a detailed chronology of the major events involving warrior-monks, see Katsuno
Ryushin, Sdhei (Tokyo: Shibundo, 1955), pp. 156-99, which provides a wealth of information on the use of warrior-monks in the conflicts between monasteries.
Cambridge Histories Online © Cambridge University Press, 2008
494
ARISTOCRATIC BUDDHISM
stinate Fujiwara was to threaten expulsion from the clan (hoshi),
which was done by denouncing the offender at the Kasuga Shrine.
The first instance of this occurred in 1163, when ShijoTakasue was
ousted for criticizing Kofukuji.
By the end of the eleventh century Hiei monks had developed their
own unique method of intimidating recalcitrant officials. Protesting
monks would descend upon Kyoto carrying sacred palanquins
(mikoshi) from the seven major shrines on Hiei that were believed to
protect the capital from natural disasters and epidemics. The simple
act of transporting the sacred palanquins through the city streets was
usually sufficient to strike terror in the hearts of the people, who
feared that they might incur the wrath of the deities if the Hiei monks
were mistreated. A favorite tactic employed by Hiei monks when
their protests were ignored was to abandon the palanquin at a palace
building, a residence of a high official, or the Kangakuin (the office
overseeing Fujiwara affairs), which would create turmoil since laymen did not know how to dispose of so sacred an object.
The clerical violence that frequently erupted during the second
half of the Heian period was generally provoked by conflicting
claims to temple lands, attempts by a particular monastery to establish jurisdiction over lesser temples and shrines, real or imagined
slights of one monastery by another, and dissatisfaction with ecclesiastical appointments. Doctrinal disputes between different schools,
on the other hand, rarely led to armed confrontations. Thus Todaiji
and Kofukuji most often wrangled over land. Hiei and K5fukuji
clashed repeatedly because of Hiei's success in extending its influence and control over such previously Kofukuji-dominated religious
establishments as Gion and Tonomine. Acts of great violence were
often precipitated by trivial incidents. Hiei monks burned Kiyomizudera to the ground in 1165 because monks from Kofukuji, with
which Kiyomizudera was allied, had damaged a plaque bearing the
name of Enryakuji during a memorial service for Emperor Rokujo at
Ninnaji. The following day, Kofukuji monks, bent on revenge, rampaged through Kyoto in an attempt to burn to the ground all temples and shrines connected with Hiei.
Onjoji's request for permission to establish its own
ordination platform
An extraordinary degree of enmity existed between Enryakuji and
Onjoji, the two major Tendai monasteries after the expulsion from
Cambridge Histories Online © Cambridge University Press, 2008
THE TENDAI SCHOOL AFTER SAICHO
495
Hiei of monks of the Enchin lineage in 993. Enryakuji was angered
by Onjoji's persistent demand that it be accorded equal treatment
with Enryakuji itself and not be viewed merely as a branch temple
of Enryakuji. It suspected, not without reason, that Onjoji was
scheming to reestablish the dominance of the Enchin lineage on
Hiei by insisting that its leading monks be appointed to the abbotship of Enryakuji.
Hiei's distrust of Onjoji was further intensified in 1039 when
Myoson, after having been denied appointment as abbot of Enryakuji because of the intense opposition of the Hiei monks, sought
permission from the court for Onjoji to construct its own ordination platform.110 This was an understandable request since novices
from Onjoji found it increasingly difficult to receive proper ordinations owing to the hostility between their temple and Hiei, which
controlled the only Tendai ordination hall in Japan. Onj5ji openly
challenged the validity of the Hiei ordinations by asserting that the
fifty-eight bodhisattva precepts constituting the core of a Tendai ordination had been transmitted properly only within its own lineage,
which traced its origin back to Gishin, who had received these precepts directly from Tao-sui, the seventh Chinese patriarch of
Tendai. Onjoji recognized, of course, that Saicho had also received
these precepts along with Gishin in the same ceremony, but maintained with some justification that Saicho had never transmitted
them to Ennin because Saicho had not been granted permission by
the court to conduct ordinations. Since authorization to build an
ordination hall on Hiei was given only after Saicho's death, the first
true Tendai ordinations, Onjoji argued, were performed by Gishin,
when he administered the precepts to his disciple Enchin.
Enryakuji vigorously opposed Myoson's petition, arguing that the
existence of two separate ordination platforms within a single school
would only create dissension. Enryakuji's real fear was that it would
not be able to maintain its primacy within theTendai school if Onjoji
could carry out its own ordinations. Onjoji, not surprisingly, was
supported in this dispute by the Nara temples, especially Kofukuji,
which saw this as a chance to weaken Enryakuji. In the face of Enryakuji's unwavering opposition, the court, after much deliberation,
110 Jimon denki horoku, kan 19, BZ, vol. 127, pp. 426a—27b; Shunki (the diary of Fujiwara no
Sukefusa, covering the years 1038-52, with lacunae), Tankaku sosho (Tokyo: Kokusho
kankokai, 1912), vol. 1, pp. 150-51; Gempei seisuiki, kan 10, vol. 1, pp. 325-27, in
TsukamotoTetsuzo, ed., Yiihodo bunko (121 vols.) (Tokyo: Yuhodo shoten, 1914-18). For
the dispute regarding the establishment of an ordination platform at Onjoji, see Tsuji,
Nihon Bukkyo-shi: josei hen, pp. 835-38, and Katsuno, Sohei, pp. 92-95.
Cambridge Histories Online © Cambridge University Press, 2008
496
ARISTOCRATIC BUDDHISM
rejected Myoson's petition, but sought to mollify Onjoji by appointing him abbot of Enryakuji, a position, as we have already noted,
that he was able to retain only for three days. The question of
whether Onjoji should be allowed to have its own ordination platform was discussed again by the court in 1070, but still no decision
was reached. The issue was raised once more in 1074 when Raigo, a
monk of Onj5ji who was asked by Emperor Shirakawa to offer
prayers for the birth of a crown prince, declared that if the court
wished to reward him for his services, it should grant Onjoji's longstanding petition for an ordination platform. This provoked a violent
demonstration by Hiei partisans, who set fire to Onjoji.
Relations between the twoTendai monasteries had so deteriorated
by 1081 that a minor altercation at a shrine festival between underlings from Onjoji and Hiei sparked a series of increasingly serious incidents, which finally culminated in an attack on Onjoji by a force of
several thousand armed monks and laymen assembled by Hiei from
its subordinate temples and shrines.111 Although Onjoji, which at the
time consisted of fifteen imperially sponsored temples, seventy-nine
halls, fifteen sutra repositories, and dozens of other structures, was
totally devastated by fire, its abbot, as an indication of the importance Onjoji attached to the right to conduct its own ordinations, did
not seek any restitution from Hiei, but instead requested permission
to establish an ordination hall once Onjoji was rebuilt. But fear of
the yama-koshi (monks from the mountain), as the Hiei clergy were
known, was sufficient to deter the court from granting even this
seemingly modest request.
Enryakuji continued to show profound hostility toward Onjoji
after its reconstruction in 1092. On the flimsiest of pretexts monks
from Enryakuji, in 1121 and again in 1140, put the torch to Onjoji,
which consequently prohibited its novices from receiving ordinations on Hiei, thus further strengthening Onjoji's case for an independent ordination hall. In 1163, Enryakuji obtained an order from
the court requiring that Onjoji novices henceforth be ordained on
Hiei and not in the ordination hall of Todaiji, which, although denounced as HInayanist by Saicho, was the only ordination procedure
available to them if they declined to go to Hiei.112 Kofujuki joined
Onjoji in protesting the court ban on Nara ordinations for Tendai
monks and, with support from Onjoji, proposed that Enryakuji be
111 Fuso ryakki, kan 30, pp. 322-23; Tsuji, Nihon Bukkyo-shi, pp. 840-42.
112 Jinten ainosho (1532), kan 18, BZ, vol. 150, p. 443a. See also Katsuno, Sohei, pp. 99-103,
and Tsuji, Nihon Bukkyo-shi, pp. 888-91.
Cambridge Histories Online © Cambridge University Press, 2008
THE SHINGON SCHOOL AFTER KUKAI
497
made a subsidiary temple of Kofukuji on the grounds that Saicho,
Gishin, Ennin, and Enchin had been ordained in Nara. Outraged at
the audacity of Kofukuji, Hiei vented its anger by burning down
Onjoji for a fourth time in less than a century. By the end of the
Heian period neither side could claim a victory: Onjoji failed to receive approval for its own ordination hall, but Hiei was unable to enforce the decree requiring Onjoji novices to be ordained on Hiei.
THE SHINGON SCHOOL AFTER KUKAI
Unlike Tendai, which was temporarily eclipsed when Saicho passed
from the scene, Shingon continued to enjoy unwavering support
from the imperial family and aristocracy even after Kukai's death.113
Shingon monks were regularly summoned to the court to pray for
rain or to perform rites of exorcism. Shingon services likewise became an integral part of the annual imperial calendar, the two most
important rituals being the previously mentioned go-shichinichi mishiho performed in the Shingon'in (Shingon Chapel) within the
palace precincts, and the taigenho held at the same time as the mishiho in the Joneiden (Jonei Hall), which was part of the imperial residence. Introduced from China by Kukai's disciple Jogyo, the
taigenho, which became an annual observance in 851, was a complex
ritual in which various weapons were laid out at an elaborately appointed, multitiered altar in the expectation that these implements
would vanquish all enemies of the emperor and protect his person.
Kukai's disciples
The major institutional problem confronting Kukai's successors in
the ninth century was the lack of a universally recognized Shingon
center comparable to Hiei. Before his death Kvikai entrusted the administration of each of the four major Shingon establishments to a
different disciple: Jitsue, the most senior disciple, was appointed
abbot of Toji; Shinzei was given responsibility forTakaosanji; Shinga,
Kukai's younger brother and confidant, was put in charge of the
Shingon'in atTodaiji, and Shinzen, Kukai's nephew, was made abbot
of Kongobuji on Koya, where Kukai had chosen to be interred.
113 Three sources that were particularly helpful to me when I was writing this section are Shimode Sekiyo, "Shingon-shu no tenkai," in Ajia Bukkyo-shi: Nikon hen, vol. 2, pp. 205-22;
Sonoda Koyu, "Heian Bukkyo no seiritsu," in Nikon Bukkyd-shi, vol. 1, pp. 226-33; Katsuno, Hieizan to Koyasan, pp. 197-242.
Cambridge Histories Online © Cambridge University Press, 2008
498
ARISTOCRATIC BUDDHISM
Although there is no evidence to suggest that the disciples bore any
personal animosity toward each other, this arrangement could not
but create serious difficulties given their different personalities, interests, and family backgrounds.
Eager to secure the patronage of the powerful families, Rukai's
disciples sometimes found themselves on opposing sides in political
conflicts. In 850, Shinzei, who was then abbot of bothTakaosanji and
Toji, agreed to a request by Emperor Montoku to pray that the latter's eldest son, six-year-old Prince Koretaka, be chosen heir to the
throne. Koretaka's mother was a member of the Ki clan, as was
Shinzei.114 This attempt to secure the right of succession for a member of the Ki clan was strongly opposed by the Minister of the Right,
Fujiwara no Yoshifusa, whose daughter, Meishi, also a consort of
Montoku's, had likewise given birth to a son, Prince Korehito (the
future Emperor Seiwa).To counter Shinzei's efforts on behalf of Koretaka, Yoshifusa turned to Shinga, whose prayers for Korehito's succession eventually proved successful. Shinga was rewarded with the
abbotship of Toji in 860, even though this necessitated bypassing the
dembo ajari Shinsho, who was his senior with respect to ordination.
So great was the esteem for Shinga that he became the first monk
permitted to enter the palace grounds in a carriage.
Rivalry among the major Shingon temples
Even more divisive than the personal alliances between Shingon
monks and opposing factions of the aristocracy was the rivalry between Toji, which performed a vast array of esoteric rites for the
court, and Koyasan, where Shingon novices underwent rigorous
training. In 853, Shinzei, then abbot of Toji, persuaded the court to
decree that in view of Koya's remoteness the annual examination of
candidates for ordination as Shingon monks, whose number had
been increased from three to six that year, should no longer be conducted at Koya, which had been the rule since 835, but at Toji.115
Only three of the successful candidates would be permitted to proceed to Koya for the actual ordination; the remaining three were to
be ordained atTakaosanji. Several years later, however, it became the
practice for all ordinations to be done at Toji. Resentful of its loss of
114 Oe no Masafusa, Gddanshd (ca. 1107), kan 2, Gunsho ruiju (Tokyo: Keizai zasshisha, 1905),
vol. 17, p. 568.
115 For the dispute regarding the proper site for Shingon ordinations, Katsuno, Hieizan to
Koyasan, pp. 209-12.
Cambridge Histories Online © Cambridge University Press, 2008
THE SHINGON SCHOOL AFTER KUKAI
499
control over ordinations, Koya was able to have the offending decree
rescinded in 885, after its abbot, Shinzen, who a year earlier had
been asked to serve concurrently as abbot of Toji, protested vigorously. Shinzen's successor at Toji, Yakushin, who belonged to the
Toji lineage proper emanating from Jitsue and who also enjoyed the
strong support of Emperor Uda, succeeded in convincing the court
in 897 to return to the arrangement of 853, arguing that Koya, "a
branch temple located in the mountains," had usurped the prerogatives ofT6ji. ll6 The wrangling over the right to ordain continued for
another ten years when the government forced a compromise. The
number of annual ordinands was increased from six to ten, four to
be examined at Toji, three at Koya, and three atTakaosanji.
The cause of Koya was championed primarily by Shinzen, who as
a nephew of Kiikai was determined not to permit Koya, Kiikai's retreat and place of burial, to be overshadowed by Toji, which was benefiting enormously from its proximity to the center of secular power.
While Kiikai's other disciples were busily engaged in performing rituals for the aristocracy in the capital, Shinzen devoted himself single-mindedly to expanding Kongobuji and codifying its basic rituals.
When he moved to Kyoto to become abbot of Toji in 884, Shinzen
left the administration of Kongobuji to his disciple Jucho, who had
no ties to the rival Toji. In 889, at the urging of Shinzen, the court
awarded Jucho the prestigious title of zasu,117 to be passed on in perpetuity to succeeding abbots of Kongobuji, who would now enjoy
equal status with the abbots of Enryakuji. Shinzen entrusted to
Jucho, as a token of the authority of the office of the Kongobuji zasu,
a highly treasured collection of manuscripts known as the Sanjujo
sakushi, mostly in Kukai's hand, which Shinzen had borrowed from
Toji in 881, while Shinga, his master and uncle, was still its abbot.
Shinzen's efforts notwithstanding, Koya was hard-pressed to
maintain its independence from Toji, especially after 901 when Retired Emperor Uda received the dembo kanjo from the T5ji abbot
Yakushin. The relationship between the two Shingon centers became
increasingly strained because of Koya's persistent refusal to return
the Sanjujo sakushi to Toji." 8 In 915, Kangen, the powerful abbot of
Toji, obtained from Uda a decree instructing Muku, the then abbot
116 Koho, Tbbbki (1352), kan 8, Zokuzoku gunsho ruiju (Tokyo: Kokusho kankokai, 1907), vol.
12, p. 156.
117 Kaiei, Koya shunju hennen shuroku (1719), kan 3, BZ, vol. 131, p. 36.
118 For the dispute regarding the Sanjujo sakushi, see Tsuji, Nihon Bukkyo-shi: josei hen, pp.
40910.
Cambridge Histories Online © Cambridge University Press, 2008
5OO
ARISTOCRATIC
BUDDHISM
of Kongobuji, to surrender the manuscripts. Refusing to do so,
Muku resigned his office and accompanied by a few disciples fled
with the Sanjujo sakushi to Iga, where he died three years later. Only
after being presented with yet another decree from Uda did Muku's
disciples agree to hand over the manuscripts to Kangen, who, after
a triumphant display at the imperial palace, deposited them in the
Toji treasure-house. In 919, the year after the return of the Sanjujo
sakushi, the court formally recognized the primacy of Toji by decreeing that its abbot (traditionally called choja, "elder") would
henceforth also serve concurrently as overseer {kengyo) of KongSbuji.119 By combining the abbotships of the two major Shingon centers, the government was able to avoid the sort of schism that
plagued the Tendai school and resulted in the almost continuous
warfare between its Enryakuji and Onjoji factions. The Shingon
school achieved full parity with Tendai in 921, when the court, acting on a petition by Kangen, gave Kukai the posthumous name Kobo
Daishi, thereby according him the same honorary status that it had
granted Saicho and Ennin in 866.
Ninnaji and Daigoji
Beginning with the reign of Uda (887-97) th e imperial family developed a particularly warm relationship with the Shingon school,
probably as a reaction to the domination of Enryakuji by tonsured
members of the Fujiwara clan. Ninnaji and Daigoji were the two
Shingon temples most directly linked with the imperial family. Situated to the west of the palace, Ninnaji was originally established in
888 as a Tendai temple by Emperor Uda in memory of his father,
Koko, whose reign had been known as Ninna. In 899, two years after
abdicating in favor of his son Daigo, Uda was tonsured at Ninnaji by
Yakushin, the learned abbot of Toji. The following year Uda appointed Kangen superintendent {betto) of Ninnaji, thereby transforming it into a Shingon temple. In 904, Uda moved to the newly
constructed Omuro (Imperial Chamber) adjacent to Ninnaji, to become its first monzeki (abbot of imperial or aristocratic lineage). Uda
occupies a unique position in the history of Japanese Buddhism as
the only emperor to have carried out, as a dembd ajari, the formal
transmission of the Dharma to his successors. His major disciple
and spiritual heir, Kangu, became abbot of Toji in 949, and his
119 Toji choja bunin (ca. 1634)5 kan 1, Zokuzoku gunsho ruiju, vol. 2, p. 491a.
Cambridge Histories Online © Cambridge University Press, 2008
THE SHINGON SCHOOL AFTER KUKAI
501
grandson, Kancho, who had studied under both Uda and Kangu,
was named abbot of Toji in 981.
Throughout its history Ninnaji retained extremely close ties to the
imperial family, no less than twenty-two of its abbots having been
hoshinno (ordained imperial princes of the first generation), eight of
them in the Heian period alone. Ninnaji reached the pinnacle of its
influence during the incumbency of its sixth monzeki, Shukaku, who
was the second son of Emperor Go-Shirakawa. A learned monk with
a reputation as a skilled practitioner of esoteric rites, Shukaku eventually became abbot of nine temples in the Kyoto area, bringing
them under the sway of Ninnaji. So great had the prestige of
Shukaku become that when he requested Toji, in 1186, to "lend"
Ninnaji the precious Sanjujo sakushi and two mandalas attributed to
Kukai, Shunsho, the abbot of Toji, was unable to refuse. The Sanjujo
sakushi has remained "on loan" at Ninnaji ever since.
The other Shingon temple to receive comparable support from
the imperial family was Daigoji, located in the southeastern suburbs
of Kyoto. Originally built in 874 as a hermitage for Shobo, an eminent Shingon monk of imperial descent, Daigoji was richly supported by Shobo's devoted patron, Emperor Daigo, who sponsored
the construction of the Shakado,Yakushido, and Godaido, which are
three major halls of Daigoji, between 904 and 907. Daigo decreed in
919, on the tenth anniversary of the death of Shobo, that henceforth
only monks belonging to Shobo's lineage could serve as abbot of
Daigoji and that they would have the rank of zasu, the same title
used for the abbots of Enryakuji and Kongobuji. Kangen, who was
Shobo's foremost disciple and successor, was named the first abbot
of Daigoji. Since Kangen already held the abbotship of Toji and
Kongobuji and was also superintendent of Ninnaji, the practical effect of Daigo's decree of 919 was to bring all the principal Shingon
centers in Japan under the nominal control of a single monk.
Daigo's sons, the emperors Suzaku and Murakami, whose births
were attributed to prayers offered at Daigoji by Kangen, further enlarged the temple by constructing new halls and making periodic donations of land and tax revenue. In 952 a five-storied pagoda, which
still stands, was consecrated in memory of Daigo, whose grave was
placed adjacent to the temple. Suzaku and Murakami were likewise
buried at Daigoji, as were the consorts of emperors Shirakawa and
Go-Toba, as well as numerous other women of the imperial family.
Daigoji became a frequent place of pilgrimage for members of the
imperial family, particularly if they were seeking offspring, since it
Cambridge Histories Online © Cambridge University Press, 2008
5O2
ARISTOCRATIC BUDDHISM
was widely believed that the bodhisattva Jundei Kannon, who was
enshrined in Daigoji, would intercede in such matters.
Because of its extraordinarily close ties with the imperial family,
Daigoji, like Ninnaji, was able to become a virtually autonomous
Shingon temple. In 1018 it was decreed that henceforth only monks
of Shobo's lineage who had actually been trained at Daigoji could
occupy the office of abbot. The specific intent of the decree was to
exclude the appointment of monks from Todaiji, where Sh5bo, who
was also a Sanron scholar of note, had established in 904 Tonan'in,
a major subtemple for Sanron and Shingon studies. Since the monks
at Tonan'in constituted one branch of Shobo's lineage, they had, on
occasion, been selected by the court to serve as abbots at Toji and
Daigoji. Ever anxious to expand its influence,T6daiji used this as a
pretext for asserting its supremacy over the Shingon temples in
Kyoto by arguing that since abbots of both Toji and Daigoji had
come from Todaiji, Toji should be regarded as a branch temple (betsuiri) ofT5daiji and Daigoji as a subordinate temple (jnatsuji). The
decree of 1018 thus guaranteed the independence of Daigoji from
Tddaiji.
After the declaration that monks belonging to Shobo's Todaiji
lineage would no longer be eligible to occupy the highest office of
Daigoji, its abbotship passed firmly into the hands of descendants
of Emperor Daigo: the twelfth zasu, Kakugen, was the son of Emperor Kazan (Daigo's great-grandson); the thirteenth zasu, Joken,
was the son of Minamoto no Takakuni (another great-grandson of
Daigo); the fourteenth zasu, Shokaku, was the son of Minamoto no
Toshifusa (a fourth-generation descendant of Daigo); the fifteenth
zasu, Jokai, was the son of Minamoto no Akifusa (another fourthgeneration descendant of Daigo), and so on.
A new round of temple construction around Daigoji was begun in
1115, when Shokaku built Samboin, which was soon recognized as a
major center of Shingon learning. Samboin was the first of five subtemples erected in the vicinity of Daigoji during the twelfth and early
thirteenth centuries to be headed by abbots of imperial or aristocratic ancestry (monzeki). Toward the end of the Heian period it had
become customary to choose the abbot of Daigoji from one of these
five subtemples, collectively referred to as the Daigoji gomonzeki
(The Five Aristocratic Temples of Daigoji). By 1155 the Daigoji
complex consisted of 42 main halls (do), 4 pagodas, 3 imperial villas
(gosho), 4 sutra repositories, and 183 dormitories to accommodate
monks.
Cambridge Histories Online © Cambridge University Press, 2008
THE SHINGON SCHOOL AFTER KUKAI
503
The Ono and Hirosawa branches ofShingon
The growing complexity of Shingon rites led during the twelfth century to the emergence of twelve subschools (ryu), each based in a
different temple. These subschools, which differed from one another
not so much in matters of doctrine as in the minutiae of ritual, were
divided into two groups, the Ono ryu and the Hirosawa ryu, each in
turn consisting of six subschools.
The Ono ryu, which traced its lineage back to Shobo, took its
name from the Ono district of Kyoto, where the great Shingon
scholar Ningai, who was a fourth-generation disciple of Shobo and
the de facto founder of the Ono ryu, established Zuishin'in in 991.
The Ono ryu subsequently split into the two groups: (1) the Three
Daigoji Subschools {Daigo sanryu) based in the Samboin, Rishoin,
and Kongooin, which were part of the Daigoji complex; and (2) the
Three Kanjuji Subschools (Kanjuji sanryu) located in Kanjuji,
AnjSji, and Zuishin'in inYamashina to the north of Daigoji.
The rival Hirosawa ryu traced its origins back to Yakushin, the
seventh abbot of Toji, through his disciple, the tonsured retired emperor Uda. Named after Hirosawa Pond in northwest Kyoto, where
Kancho, Yakushin's disciple in the third generation, founded Henjoji
in 989, the Hirosawa ryu subdivided around the beginning of the
twelfth century into six subschools, four of which were based in subtemples within the Ninnaji complex. The two remaining Hirosawa
subschools, the Ninnikusen ryu and the very important Daidemboin ryu, evolved in temples inYamato and Kii provinces respectively.
The restoration of Koya
After it was decided in 919 to have the abbot of Toji serve concurrently as abbot of Kongobuji, it soon became apparent that the interests of the Kongobuji could not be adequately represented in such
an arrangement, since the abbot of Toji was required to reside permanently in Kyoto. In 928, Saiko, the twelfth abbot of Toji, sent an
appeal to the court in which he deplored the impoverished condition
of Kongobuji and requested funds for its renovation.120 In response
to Saiko's petition the court agreed to the appointment of a resident
"executor" (shigyd) on Koya whose primary function would be to
promote the restoration of Kongobuji. The executor, while techni120 For the rebuilding of Koyasan, see Katsuno, Hieizan to Koyasan, p. 216.
Cambridge Histories Online © Cambridge University Press, 2008
504
ARISTOCRATIC BUDDHISM
cally subordinate to the zasu, who resided in Kyoto, was the highestranking monk in Koya and hence became its de facto abbot. In 983,
with the renovation and expansion of KongSbuji almost completed,
the designation "executor" was changed to "overseer" (kengyo)> a
title that remained in use until 1875, when the original title zasu was
reinstituted to designate the head monk of Koya.
The Kongobuji was almost completely destroyed in 994 by a fire
that started when lightning struck its main pagoda. Although
Kancho, the influential nineteenth abbot of Toji, was immediately
able to secure a decree ordering the reconstruction of Kongdbuji,
the project made little headway because of a lack of funds. The situation was further aggravated when the governor of Kii, who had
been charged with the responsibility for rebuilding the Kondo
(Golden Hall), seized the K5ya estates after Kancho3s death in 998,
allegedly to meet construction costs. Deprived of their sole source of
income, the few remaining monks on Koya were compelled to quit
the mountain completely.
The reconstruction of Kong5buji and its subsequent popularity as
a place of pilgrimage were ultimately brought about by the unflagging efforts of two monks, Kishin (also known as Joyo) and Ningai.
Kishin moved to Koya in 1016 at the relatively advanced age of fiftyeight in response to a command received in a dream that he had
after imploring the bodhisattva Kannon to inform him of the whereabouts of his deceased parents.121 When he reached K6ya, it is said,
his spiritual eye was suddenly opened, enabling him to see that
Kongobuji was none other than the Inner Palace of Tosotsu (Tusita)
Heaven, the celestial realm in which Miroku (Maitreya), the Future
Buddha, is believed to reside until the time is ripe for his descent
into this world. Here Kishin discovered his parents, transformed into
bodhisattvas, seated on lotus thrones. After making his way to the
tomb of Kukai, which was almost inaccessible because of the dense
overgrowth after years of neglect, Kishin made a solemn vow to devote the remainder of his life to the restoration of Koya. Kishin's efforts bore fruit, and by the time of his death in 1047 Kong5buji enjoyed a greater degree of prosperity than ever before.
The perception of Koya as a sacred mountain was also widely promoted by the prestigious monk Ningai, whose power to summon
rain in times of drought assured him a large following at the court.
Known popularly as the "Rain Bishop" (Ame Sojo), this exception121 This story is recounted inTsuji, Nihon Bukkyo-shi, pp. 413-14.
Cambridge Histories Online © Cambridge University Press, 2008
THE SHINGON SCHOOL AFTER KUKAI
505
ally learned monk, who, as already noted, is regarded as the founder
of the Ono branch of Shingon, felt a particular attachment to Koya,
where, in 957, at the age of six he had been tonsured by Gashin, who
later became the first monk to hold the office of overseer of Kongobuji.
In 1023, Ningai persuaded the former regent, Fujiwara no Michinaga, then fifty-seven years of age, to undertake the arduous journey
to Koya.122 Ningai convinced Michinaga that by making a pilgrimage
to Koya, he would not only be sure to escape an unpleasant rebirth
but would also be present when Miroku makes his descent into this
world from Tosotsu Heaven. Accompanied by an entourage of sixteen high-ranking monks and laymen, Michinaga spent some five
days on the road before reaching the Inner Sanctum (Okunoin) of
Koya, where Kukai was entombed. In an emotional declaration of his
faith, he pledged the rebuilding of Kukai's mausoleum and the donation of land for the upkeep of Koya.
It subsequently became a regular practice for Fujiwara regents
and retired emperors to undertake such a pilgrimage to Koya.123 In
1048 Yorimichi visited Koya, followed by Morozane in 1081, Moromichi in no8,Tadazane in 1144, andYorinaga in 1148. The retired
emperors Shirakawa and Toba made three visits each to Koya (Shirakawa in 1088, 1091, and 1127;Toba in 1124, 1127, and 1132).124 As
might be expected, the pilgrimages by the retired emperors, like
those by the Fujiwara regents, were lavish affairs, the entourage usually including ministers of state, leading clerics from the main Shingon monasteries in Kyoto, and a military escort, which accompanied
the procession only as far as the Administrative Office (mandokoro)
of Kongobuji. These aristocratic and imperial pilgrims contributed
generously to the reconstruction and expansion of Kongobuji. On
the occasion of his first visit, Shirakawa formally ordered the rebuilding of the Great Pagoda (Daito), donating an estate in Bingo
for its permanent upkeep. On his second visit, he presented Koya
with an estate located in Aki. In 1124, Toba sponsored the construction of the Western Pagoda, which he subsequently endowed. After
the construction of Daidemboin in 1132, Toba donated seven estates
specifically for the training of its monks.
122 Fuso ryakki, kan 28, KT, vol. 12, pp. 276-77; Koya shunju hennen shuroku, kan 4, BZ, vol.
131. P- 66.
123 On the pilgrimages to Koya by the Fujiwara, see Katsuno, Hieizan to Koyasan, pp. 218-20.
124 The imperial visits to Koyasan at this time are chronicled in Kongobuji, ed., Koyasan sen
hyakunen shi (Koyasan: Kongobuji, 1942), pp. 76-83.
Cambridge Histories Online © Cambridge University Press, 2008
506
ARISTOCRATIC BUDDHISM
The enormous popularity of Koya as a pilgrimage site after the
eleventh century indicates the profound esteem in which Kukai was
held.125 By the end of the tenth century it was believed that Kukai
had not actually died at Koya in 835 but rather had entered into a
deep trance {nyujo), from which he would emerge upon Miroku's
descent from Tosotsu Heaven. According to a widely held belief,
when Kangen opened Kukai's tomb in 921 to read the court decree
granting Kukai the posthumous name Kobo Daishi, he discovered
that Kukai's body was intact: his facial color was normal, his skin
was warm, and his hair had grown longer - all indications that he
was indeed still alive. Since Kukai had been a devotee of Miroku, he
soon was identified with him. If Kukai was the earthly incarnation
(keshiri) of Miroku, it was only natural to view Koya, Kukai's abode,
as a terrestrial manifestation of Tosotsu Heaven.
By the late eleventh century the deification of Kukai had progressed to the point that he was being equated with the great cosmic
Buddha Dainichi, the supreme deity of Shingon. When Shirakawa,
in 1088, expressed a wish to visit Vultures' Peak in India where Sakyamuni was believed to have preached many of the Mahayana sutras,
he was urged by the court scholar, Oe no Masafusa, to make a pilgrimage to Koya instead, since K5ya, Masafusa declared, was none
other than the eternal Pure Land of Dainichi Nyorai (Mitsugon
Jodo) and hence vastly superior to Vultures' Peak in India.126 So
highly was Kukai revered that when a lock of hair belonging to the
deceased emperor Horikawa was discovered in 1108, the court decided to inter it alongside Kukai's tomb in the hope of establishing a
spiritual link (kechieri) between the deceased emperor and the deified Kukai. In 1165 a lock of hair belonging to the deceased emperor
Nijo was sent to Koya for the same purpose.
The first instance of the burial of a lay person on Koya, which was
to become a common practice - Koya now has more than 100,000
graves - occurred in 1160, when Mifukumon-in, the consort of Emperor Toba, was interred there. A devout believer in Kukai, Mifukumon-in had not been allowed to make the pilgrimage to K5ya in her
lifetime because of the ban against visits to the mountain by women.
In the final years of her life, she erected a hermitage at Arakawa,
about twenty kilometers west of Koya, from which she worshiped
daily in the direction of Kukai's tomb. In accordance with her will,
125 The following account is largely based on Shimode, "Shingon-shu no tenkai," in Ajia
Bukkyo-shi: Nihon hen, vol. 2, pp. 2100-148.
126 Washio Junkyo, Koshitsu to Bukkyo (Tokyo: Daito shuppansha, 1939), pp. 265-66.
Cambridge Histories Online © Cambridge University Press, 2008
PURE LAND BUDDHISM
507
her cremated remains were brought to Kdya for burial near the main
hall of Kongobuji, where her tomb still remains.
THE GROWTH OF PURE LAND BUDDHISM
Amida (Sanskrit, Amitabha or Amitayus) Buddha, the central figure
of the Pure Land faith, is one of the most popular divinities in the
Buddhist pantheon, his name or that of his Pure Land appearing in
more than 270 scriptures, roughly one out of every three works in the
Mahayanist canon.127 At a very early stage in the development of Indian Mahayana there emerged a cult centering on Amida, that
viewed him as a sort of savior who had, through his own boundless
merit, created a Pure Land, or haven, offering shelter to all beings
who demonstrated their faith in him by certain devotional acts. In
China the Amida cult grew steadily after the year 402, when the
monk Hui-yuan founded the first Pure Land association of lay and
clerical devotees. Faith in Amida and his vow to deliver all beings to
Pure Land became one of the dominant themes in Chinese Buddhism and was recognized as an ancillary teaching by the various
Buddhist schools, whose masters often produced commentaries on
the Pure Land scriptures and advocated the Pure Land faith, but always in such a way that it was subordinate to the principal tenets of
their own school. In addition to these sectarian interpretations of
Pure Land, however, an independent cult, which viewed Pure Land
Buddhism as the only valid type of religious practice for the present
age, began to take shape in the sixth century under the guiding hand
of T'an-luan. This cult, known later in China and Japan as the Pure
Land school, culminated in the work of Shan-tao, who, while recognizing the necessity of such traditional practices as sutra chanting,
meditations, image worship, and the presentation of offerings, asserted that the vocal recitation of Amida's name was the primary devotional act leading to rebirth in Pure Land.
Although Pure Land scriptures had already been copied and studied in the Nara period and images of Amida could be found in Nara
temples, it was only in Heian times that the Pure Land faith emerged
as a major movement within Japanese Buddhism. In seventh- and
eighth-century Japan, Amida was viewed primarily as a Buddha who
could deliver the souls of the dead to his Pure Land, commonly
127 Yabuki Keiki, Amida Butsu no kenkyu, rev. and enlarged ed. (Tokyo: Meiji shoin, 1937),
P-449Cambridge Histories Online © Cambridge University Press, 2008
508
ARISTOCRATIC BUDDHISM
called Paradise (Gokuraku).128 As the inscriptions on his images indicate, Amida was worshiped most frequently by laymen seeking to
ensure the rebirth of a deceased relative in Pure Land. Relatively few
instances are recorded of individuals commissioning statues of
Amida in the hope that they themselves might be able to enter his
Pure Land after death.129
Pure Land observances on Hiei
The systematic practice of chanting the name of Amida in the hope
of attaining rebirth in his Pure Land was introduced into Japan by
Ennin, who, as we have already noted, first became acquainted with
the goe nembutsu service during his brief stay at the Chu-lin ssu in
Wu-t'ai in 840. The founder of this monastery was the Pure Land
devotee Fa-chao (died c. 820), who had himself devised the goe nembutsu, which combined the traditional meditations on Amida with
the fervent invocation of his name as taught by Shan-tao. Ennin encountered the goe nembutsu again at the Tzu-sheng ssu, the monastery in Ch'ang-an at which he resided for almost five years. The
melodious chanting of Amida's name constituted an important part
of the liturgy used at the Tzu-sheng ssu.
In 848, the very year that Ennin returned from China, he built a
Jogyo zammai-do (frequently abridged to Jogyodo) in the Toto
(Eastern Pagoda) temple complex on Hiei. The Jogyodo, as its name
indicates, was a hall exclusively devoted to the practice of the jogyo
zammai, a meditation lasting ninety days in which one concentrated
one's thoughts on Amida while invoking his name and circumambulating his image.130 The jogyo zammai, which was one of the four
basic types of meditation taught by the Tendai school, was practiced
not primarily to bring about rebirth in Pure Land but, rather, to enhance the powers of concentration of the devotee by enabling him to
focus his mental and physical activities on Amida. But for Ennin,
who had been greatly influenced during his stay in China by the intense piety of Fa-chao, the jogyo zammai was not simply a meditative
exercise but an act of devotion to Amida that facilitated entry into
his Pure Land. So strongly did Ennin feel about Amida worship that,
128 Although Buddhism does not technically recognize the existence of a soul (Sanskrit,
atman; Japanese, ga), the notion of a soul, commonly referred to as rei or ryo, is widespread in popular Japanese Buddhism.
129 Inoue Mitsusada, Nikon Jodo-kyd seiritsu-shi no kenkyu (Tokyo: Yamakawa shuppansha,
1956), PP- 15-26.
130 Eigakuyoki (ca. 1267), kan 1, Gunsho ruiju, vol. 15, p. 542a.
Cambridge Histories Online © Cambridge University Press, 2008
PURE LAND BUDDHISM
509
in his will, he instructed his disciples to begin the practice of fudan
nembutsu, "uninterrupted contemplation of [Amida] Buddha."131
First performed in 865, the year after Ennin's death, the fudan
nembutsu played a crucial role in the dissemination of the Pure Land
faith among the Hiei clergy and their aristocratic lay supporters. The
fudan nembutsu was held annually at the Jogyodo from the morning
of the eleventh day of the eighth month through the night of the seventeenth day. The goal of the participants was to achieve rebirth in
Pure Land through the practice of the fudan nembutsu, which it was
believed had the power to destroy the effects of accumulated evil
karma. The fudan nembutsu had three components, each signifying a
different realm of human activity: (1) the continuous invocation of
Amida's name, occasionally interrupted by the chanting of passages
from the Amidakyo extolling the merit embodied in the name; (2)
the circumambulation of an Amida's image; and (3) the concentration of one's thoughts on Amida. These three components represented respectively the verbal, physical, and mental activities of the
person, in short, the totality of human actions. By dedicating his verbal utterances, physical movement, and mental processes to Amida
in this fashion, the devotee immersed himself completely in Amida,
which guaranteed his rebirth in Pure Land.132
The popularity of the fudan nembutsu grew steadily: in 893 a second Jogyodo was built with imperial sponsorship in the Saito (Western Pagoda) complex on Hiei and in 968 a third Jogyodo was established by Rydgen atYokawa, the third major temple complex, thus
enabling the two thousand or so monks on Hiei to take part in the
annual fudan nembutsu. It was not long before provincial clerics who
came to Hiei for training also became devotees of the yama no nembutsu (Amida-contemplations of [Hiei] mountain), as the fudan nembutsu came to be popularly called, and erected Jogyodo in their home
regions. By the end of the tenth century Jogyodo could be found in
such diverse places as Kyoto, Otsu (as an adjunct of Onj6ji),T6nomine, and as far east as the Izu peninsula. The predominance of
Amida pietism on Hiei is attested in a document dated 970, which
wryly observes that although novices had been expected to spend
twelve years learning all four types of Tendai meditations, they now
limit themselves to the practice of jogyo zammai alone.133
131 Sammon dosha ki (early 14th century), Gunsho ruiju, vol. 15, p. 488a.
132 For a description of ihe fudan nembutsu, see Sambo ekotoba, BZ, vol. 111, p. 467; English
translation, Kamens, pp. 342-44.
133 Nijurokka jo kisho, Heian ibun: komonjo hen, vol. 2, p. 435a.
Cambridge Histories Online © Cambridge University Press, 2008
51O
ARISTOCRATIC BUDDHISM
The Ojo yoshii and the dissemination of the Pure Land faith
among the aristocracy
T h e appearance of Genshin's Ojo yoshii {Anthology on Rebirth in Pure
Land) in 985 marked the beginning of a new phase in the development of Japanese Pure Land Buddhism. Its author, Genshin (9421017), who had been a disciple of Ryogen, won recognition early in his
career for his great learning. Instead of pursuing high ecclesiastical office, as was the custom of the Hiei elite, Genshin chose to go into
seclusion at Eshin'in inYokawa, where he devoted himself to meditation and scholarship. With more than eighty extant works attributed
to him, Genshin was one of the most prolific monks of the Heian period, his writing covering such diverse topics as Tendai doctrine,
Hlnayana philosophy, logic, esoteric ritual, and Pure Land teachings.
Although Genshin always remained within the Tendai tradition, his
interpretation of Pure Land doctrine, particularly as he presented it in
his Ojo yoshii, provided the impetus for the Pure Land faith that swept
Japan in the eleventh and twelfth centuries resulting in the rise of the
Jodo, Jodo Shin, and Ji schools of the Kamakura period.
What distinguished Genshin's approach to Pure Land doctrine
from those of his Tendai predecessors was his belief that the world
was on the verge of entering the mappo age, that is, a period of irreversible spiritual decline that called for less demanding religious exercises. Whereas in earlier periods it was possible to attain enlightenment through meditation, adherence to precepts, and cultivation
of wisdom, in the mappo age, which Genshin believed was about to
commence, these practices were exceedingly difficult, if not impossible, for most people.134 The most appropriate teaching for the
mappo age, in Genshin's view, was that of Pure Land, which Amida
Buddha had created as a refuge for all beings, both good and evil,
who sought rebirth there. Although Ennin established Pure Land as
a devotional practice on Hiei, he did not attempt to provide a doctrinal justification for it in any of his numerous writings. That task
fell to Genshin, who, in order to lay a firm theoretical foundation for
the practice of Pure Land within the Tendai school, undertook an
extensive study of the Buddhist scripture as well as of the various
treatises on Pure Land by Chinese scholars, as is attested in his Ojo
yoshu, which quotes from more than 160 different works.135 It is par134 See the opening lines of his Ojo yoshu, kan 1, T, vol. 84, p. 33a.
135 Hanayama Shinsho, trans., Ojo yoshu, Iwanami bunko, vol. 2992-96 (Tokyo: Iwanami shoten, 1942), p. 3.
Cambridge Histories Online © Cambridge University Press, 2008
PURE LAND BUDDHISM
511
ticularly significant that among the Pure Land treatises most frequently cited are those of Shan-tao and his master, Tao-ch'o, who
laid the foundation for an independent Pure Land faith in China.
Running through the Ojo yoshu was a simple theme: our present
world is one of defilement and pain in contrast to the Pure Land of
Amida, which is a blissful realm free from taint or suffering. By using
copious quotations from the scripture that provided graphic accounts of the suffering of beings trapped in the cycle of transmigration contrasted with the ecstatic existence of the inhabitants of Pure
Land, Genshin sought to warn his contemporaries about attachment
to the transitory pleasures of this life and awaken within them a
yearning for Amida's Pure Land. While Genshin in theory accepted
the traditional Tendai emphasis on the importance of meditation on
Amida as a means to achieve rebirth in Pure Land, he also recognized that in the mappo age such practices were difficult for most
people, who could not sustain for long periods of time the demanding meditations in which one contemplated each of the thirty-two
physical signs of Amida's Buddhahood, not to mention the more abstract ones in which Amida's radiant form is viewed as a manifestation of absolute truth.
For such spiritually weak people - and Genshin included himself
in this group - Genshin advocated, as a last resort, isshin shdnen,
"wholehearted invocation of Amida's name (sho) combined with
contemplation (neri)" the latter referring specifically to contemplating the act of submitting oneself to Amida or to visualizing the
deathbed scene in which Amida, accompanied by a host of bodhisattvas, descends from the sky to lead the dying person to Pure
Land. The single most important religious practice for the Pure
Land devotee was nembutsu, which signified not simply the repetition of the name of Amida as it did in Kamakura times and later, but
also included the notion of meditating on Amida as well. To Genshin, the highest expression of Pure Land faith was jinjo nembutsu, in
other words, the continuous practice of nembutsu in one's daily life.
Since Genshin believed, however, that the world was about to enter
the mappo age, he acknowledged that jinjo nembutsu remained more
of an ideal than a practical course of action. He therefore attached
particular importance to rinju gyogi, "deathbed rites," in which the
dying person, surrounded by friends and relatives urging him to
think of Amida's imminent arrival, repeatedly invokes Amida's name
while holding a cord attached to the hand of an image of Amida.
It seems likely that Genshin intended the Ojo yoshu to be a practiCambridge Histories Online © Cambridge University Press, 2008
512
ARISTOCRATIC BUDDHISM
cal guide for the Nijugo zammai-e, a spiritual mutual-help society
formed in 986 by twenty-five monks from the Shuryogon'in in
Yokawa who vowed to assist one another in attaining rebirth in Pure
Land.136 The monks agreed to meet on the fifteenth day of each
month for a service that opened in the early afternoon with a lecture
on the Hokekyo, after which the group began meditations on Amida,
while invoking his name and chanting the Amidakyo. If any member
of the society became ill, he would be cared for by the others, and if
it seemed that he was dying, he would be brought to the Ojoin
(Chapel for Rebirth), which was a hall enshrining an image of
Amida, where his colleagues would encourage him to chant Amida's
name while envisioning the arrival of Amida with his retinue of bodhisattvas. In return, the dying person promised that after reaching
Pure Land he would reappear in the dreams of his colleagues to describe Pure Land and strengthen their determination to reach it.
The rules of the Nijugo zammai-e were drafted byYoshishige no
Yasutane, a devout Pure Land believer who had taken the tonsure
just before the Nijugo zammai-e was established. Earlier, in 964, Yasutane had founded the Kangaku-e (Society for the Promotion of
Learning)137 which was a Pure Land society consisting of twenty
monks from Hiei and twenty laymen, most of whom had aristocratic
or literary backgrounds. The Kangaku-e, which met twice yearly at
temples around Hiei, contributed to the spread of the Pure Land
faith among the upper classes and helped pave the way for the allclerical Nijugo zammai-e, whose membership soon expanded to include Genshin himself and Emperor Kazan, who had renounced his
throne and taken holy orders a month after the founding of the
Nijugo zammai-e.
The Ojo yoshu was well received by the aristocracy, who were
deeply moved by its profound religious message and enthralled by
its detailed description of the palaces, lakes, and gardens of Pure
Land. The most powerful aristocrat of the day, Fujiwara no Michinaga, was an avid reader of the Ojo yoshu and an admirer of Genshin,
to whom he twice sent emissaries after becoming ill in 1004.138 A de136 For the Nijugo zammai-e, see Inoue Mitsusada, Nihon kodai no kokka to Bukkyo (Tokyo:
Iwanami shoten, 1971), pp. 162-64, a n d m e same author's Nihon Jodo-kyd seiritsu-shi no
kenkyu, pp. 148-49.
137 T h e origins and practices of the Kangaku-e are discussed in Inoue Mitsusada, Nihon
Jodo-kyd seiritsu-shi no kenkyu, pp. 150-52. For an early account see Sambo ekotoba, BZ,
vol. 111, pp. 4500-510; English translation, pp. 295-98; see also the valuable notes inYamadaYoshio, Samboe ryakuchu (Tokyo: Hobunkan, 1951), pp. 288-92.
138 For Michinaga's involvement with Pure Land Buddhism, see Inoue, Nihon kodai no kokka
to Bukkyo, pp. 165—70, andTsuji, Nihon Bukkyo-shi: josei hen, pp. 572—73.
Cambridge Histories Online © Cambridge University Press, 2008
PURE LAND BUDDHISM
513
vout Pure Land believer, Michinaga took the tonsure in 1019 and the
following year built an Amida Hall (also known as Mido or
Muryojuin), in which he installed nine 16-feet-tall gilded images of
Amida, each representing this Buddha in one of the nine traditional
divisions of Pure Land. Michinaga died at his Amida Hall, in the
manner of Genshin, clutching a silk cord that linked him to the nine
images of Amida. Michinaga's son and successor, Yorimichi, likewise built an Amida Hall, the famous Hoodo at Byodoin in Uji,
which when viewed with its surrounding landscape was intended to
be a re-creation of Pure Land in this world. The great popularity that
the Pure Land faith enjoyed among the aristocracy and the powerful clans is attested by the large number of Amida Halls built in the
eleventh and twelfth centuries. Documents of the period indicate
that at least ninety-five Amida Halls were established between 1020,
when Michinaga dedicated his Mido, and 1192, when Minamoto no
Yoritomo built Nikaido with its nine images of Amida at Eifukuji in
Kamakura.^9
The popularization of the Pure Land faith
The Amida Halls with their resplendent images and elegant gardens
suggestive of the topography of Pure Land, the meditations on
Amida, and the elaborate ceremonies called mukae-kd (welcoming
services) in which young monks wearing jeweled crowns, golden
masks and the raiments of bodhisattvas pretend to welcome their
prestigious patrons in to Amida's Paradise had little to do with the
Pure Land faith of the common man, who could neither read the
Chinese in which the Ojo yoshu was written nor participate in the rituals of the aristocratic Amida Halls.140 His faith in Pure Land was inspired not so much by the monks of the great monasteries as by the
hijiri, "holy men," who were typically clerics who had chosen to pursue the religious calling away from the main monasteries.141 Some of
the hijiri were eccentrics who shunned clerical garb in favor of deerskins; others spent long periods in remote mountain retreats practicing austerities while chanting the Hokekyo; still others lived in her139 Inoue, in his Nihon kodai no kokka to Bukkyb, pp. 171-78, provides a detailed list of the
Amida Halls of this period giving the date of construction, the name of the hall, its sponsor, and the primary sources in which the hall is mentioned.
140 On mukae-kd, see Ishida Mizumaro, Jodo-kyo no tendai (Tokyo: Shunjusha, 1967), pp.
13032.
141 For a comprehensive study of the hijiri, see Hori Ichiro, "On the Concept of Hijiri (HolyMan)," Numen 5, 2 (April 1958): 128-60; 5, 3 (September 1958): 199-232.
Cambridge Histories Online © Cambridge University Press, 2008
514
ARISTOCRATIC BUDDHISM
mitages known as bessho, "places apart," which were located on the
fringes of the precincts of larger monasteries; some others such as the
yugyo hijiri, "wandering hijiri," roamed the countryside preaching to
the common people and soliciting small contributions from them for
religious projects - a practice known as kanjin - which appealed
greatly to members of the lower classes for it enabled them to "establish affinity" (kechieri) with a hijiri and partake of his great merit.
Although hijiri appear in Buddhist sources as early as the Nara period, it was only after the middle of the tenth century that they became a prominent feature of Japanese Buddhism. Their religious
faith, for all its intensity, was usually an amalgam of esoteric rites
that were simple enough to be performed by the individual, devotion
to the Hokekyo, and belief in Amida as a savior, coupled with a
yearning for rebirth in his Pure Land. The first hijiri to play a major
role in the dissemination of the Pure Land faith was Kuya (903-72),
also known as Koya, who had spent his early years undertaking spiritual exercises in the mountains.142 After tonsuring himself in his
twenties at the Owari Kokubunji, Kuya became an itinerant proselytizer, carrying scriptures and holy images in his backpack. Like the
famous Gyoki of the Nara period, he is said to have built roads,
erected bridges, dug wells, and collected abandoned corpses for cremation, all activities that brought him close to the common people.
Known as Amida hijiri because of his continuous chanting of the
name of Amida - even the wells he dug came to be called "Amida
wells" - Kuya settled in Kyoto in 938, spreading the Amida faith
among its inhabitants and earning the appellation ichi hijiri, "the hijiri of the marketplace." Kuya's alienation from the established
church, which was common to many hijiri, is evidenced by his failure to seek formal ordination on Hiei until he had reached the relatively advanced age of forty-five. Commenting on his achievements,
the Nihon ojd gokuraku ki, which is a late-tenth-century collection of
biographies of people who have attained rebirth in Pure Land, observes that before the Tengyo era (938-47) few of the simple folk
practiced Amida devotions, but thanks to the efforts of Kuya, "who
invoked the name himself and made others invoke the name," the
whole of Japan came to devote itself to nembutsu.143
142 The most reliable treatment of Kuya's life is to be found in Hori Ichiro, Kuya, Jimbutsu
sosho, vol. 106 (Tokyo: Yoshikawa kobunkan, 1963).
143 Yoshishige no Yasutane, Nihon bjo gokuraku ki (985), in Inoue Mitsusada and Osone
Shosuke, eds., Ojd den, Hokke genki, vol. 7 of Nihon shiso taikei (Tokyo: Iwanami shoten,
1974). P- 9Cambridge Histories Online © Cambridge University Press, 2008
PURE LAND BUDDHISM
515
By the eleventh century, warriors, farmers, fishermen, and others
whose occupations brought them into conflict with the Buddhist
precepts increasingly viewed the practice of nembutsu as the principal method for achieving rebirth in Pure Land. The five ojo den
(biographies of people who have attained rebirth in Pure Land),
written in the twelfth century, contain numerous stories of such
individuals - not to mention rapacious officials, monks who had fathered children, and other assorted miscreants - who were saved
through the power of the nembutsu. Typical is the biography of
Kiyohara no Masakuni in the Shui ojo den, which claims that although "there was not an evil act that he did not commit," he nevertheless attained rebirth in Pure Land after taking up the practice
of nembutsu at age sixty, reciting the name 100,000 times daily.144
Similarly the learned monk Jungen, "who turned his daughter into
his wife," began the practice of nembutsu only on his deathbed and
yet was received into Pure Land.145
Although in most cases the nembutsu was practiced in conjunction
with other devotional acts not intrinsically connected with the Pure
Land faith, such as chanting the Hokekyd or performing simple esoteric rites, the ojo den indicate that some laymen and monks performed nembutsu alone, to the exclusion of all other types of Buddhist devotional practice. The Go-shui ojo den reports, for example,
that a Hiei monk named Ryusen (1047-1116) renounced the study of
Tendai and Mikkyo early in his career to devote himself exclusively
to nembutsu., uttering the name of Amida 120,000 times daily for
thirty years.146 The same work tells of an impoverished farmer from
Omi who was too poor to make offerings, but instead continuously
recited the name of Amida both when doing his daily devotions as
well as when working in the fields.147 As the Konjaku monogatari attests by its frequent use of such phrases as nembutsu wo tonau, "to recite the nembutsu," the term nembutsu by the twelfth century had
come to signify for many laymen and nonelite monks simply the
recitation of Amida's name without any reference to meditation.148
The concept that even evil men could attain rebirth in Pure Land,
the belief that practice of nembutsu alone was sufficient to bring
about rebirth there, and the emergence of the view that the nembutsu
144
145
146
147
148
Miyoshi Tameyasu, Shui ojo den (ca. 1111), kan 2, Ojo den, Hokke genki, p. 339.
Shui ojo den, kan 3, pp. 363-64.
Miyoshi Tameyasu, Go-shuojo den (1139), kan 1, Ojo-den, Hokke genki, p. 649.
Go-shuibjo den, kan 1, p. 651.
Konjaku monogatari shit (ca. 1120), kan 15, NKBT, vol. 24, p. 382.
Cambridge Histories Online © Cambridge University Press, 2008
516
ARISTOCRATIC BUDDHISM
consisted primarily of reciting the name of Amida laid the foundation for the Pure Land schools that were established in the Kamakura period and that came to figure so prominently throughout
Japanese religious history. Shinran (1173-1262), the founder of the
Jodo Shin school, openly acknowledged his debt to the lay tradition
of Pure Land Buddhism that emerged in the Heian period by declaring that he modeled himself after Kyoshin (d. 866), a tonsured
but married layman (shami) who expressed his deep faith in Pure
Land simply by reciting Amida's name as often as possible in his
everyday life.149
149 Kakunyo, Gaijasho (1335)) Shinshu shogyo zensho, vol. 3 (Kyoto: Oyagi kobundo, 1964), pp.
67-68. For examples of Kyoshin's piety, seeYokan's Ojofiin (1103), Jbdo-shu zensho, vol.
15 (Tokyo: Sankibo Busshorin, 1971), pp. 375b-77a.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 8
RELIGIOUS PRACTICES
Upon arising, first of all repeat seven times in a low voice the name
of the star of the year. Take up a mirror and look at your face, to
scrutinize changes in your appearance. Then look at the calendar
and see whether the day is one of good or evil omen. Next use your
toothbrush and then, facing West, wash your hands. Chant the
name of the Buddha and invoke those gods and divinities whom we
ought always to revere and worship. Next make a record of the
events of the previous day. Now break your fast with rice gruel.
Comb your hair once every three days, not every day. Cut your fingernails on a day of the Ox, your toenails on a day of the Tiger. If
the day is auspicious, now bathe, but only once every fifth day.1
As these Testamentary Admonitions of Fujiwara no Morosuke
(908-60) indicate, the life of aristocrats during the Heian period was
punctuated by a plethora of practices and precautions of a ritual
character. What appear on the surface to be common practices, such
as washing one's hands or brushing one's teeth, were directly related
to the daily performance of specific rites dedicated to various kami
and buddhas and bodhisattvas. These rites were related to Taoist
practices, to various ritual elements of Buddhism, or to beliefs and
practices subsumed under the category of yin and yang. Yet other ritual practices were of a political character, or appear to be outside
any immediately identifiable framework. This complexity of the ritual and liturgical world was part and parcel of life among aristocrats
during the Heian period.
The short quotation from Morosuke's admonitions informs us
that the human body was treated with great ritual care. This notion
might be related to the highest levels of the doctrine of Esoteric
Buddhism which stated that Buddhahood could be achieved in this
lifetime and in this very body, while it might also be related to more
ancient, autochthonous views concerning purification. No matter
1 Kujo-den no Goyuikai, translated in George Sansom, A History of Japan, vol. 1 (Stanford,
Calif.: Stanford University Press, 1958), p. 180.
517
Cambridge Histories Online © Cambridge University Press, 2008
518
RELIGIOUS PRACTICES
what those preceding views and practices might have been, there is
little doubt that Esoteric Buddhism redefined the ontological position of the human body in radical terms, and may have been an important factor in the centrality of ritual concerns with the body
throughout the Heian period.2 Furthermore, Morosuke's admonitions tell us that attention was given to the cosmic context of space
and time, thus suggesting a particular conceptualization of the body
within larger parameters. The same admonitions also indicate that
daily rituals were dedicated to autochthonous kami as well as to buddhas or bodhisattvas, thereby leaving the impression that such acts
of reverence were of an inclusive character in which both indigenous
and imported cultural and ritual practices and notions were in contact, if not in interaction. As will be seen presently, such was indeed
the case: the majority of ritual and liturgical practices of the Heian
period were cast within a vast congruity consisting of specific notions of reality laid out in a number of documents, of particular economic behavior, and of elaborate concepts of time and of space, all
of which were in turn directly related to (Shinto) shrines and to
(Buddhist) temples set in a distinctive sociopolitical framework.
Generally speaking, in their study of the overall establishment of
religion in Japan, scholars have laid emphasis on the development of
Buddhism through the support of the imperial house, and have analyzed the doctrinal aspects of Buddhism through the elaboration of
sects or schools of thought; in contradistinction, little has been said
about the presence or concurrent establishment of shrines in the immediate vicinity of temples, or about the growth and status of ritual
practices in both shrines and temples. However, this concurrence
was structurally responsible for specific practices and beliefs such as
those mentioned by Fujiwara no Morosuke.
Government as it had been set up under the ritsuryo (statutory
system) at the end of the seventh century and at the beginning of the
eighth century had established a Department of Shrines (Jingikari),
one of the duties of which was to formalize rituals performed by various sacerdotal lineages in the name of private families or of the
state, and to regulate the "worship of the kami of Heaven and Earth"
(Jingi siihai) in shrines. The Bureau continued to exist during the
Heian period and exercised much influence on the beliefs, rites, and
practices of sacerdotal lineages, of the aristocracy at large and, be2 The text positing the centrality of the body is Kukai's Sokushin jobutsu-gi; see Yoshito
Hakeda, Kukai: MajorWorks (New York: Columbia University Press, 1972), pp. 225-34.
Cambridge Histories Online © Cambridge University Press, 2008
RELIGIOUS PRACTICES
519
yond those privileged groups, of the populace in the geographical
areas under the direct control of the government.
One of the more significant developments of the one hundred and
fifty years before the beginning of the Heian period was the creation
of capital cities designed on a Chinese model.3 That development
was significant for two major reasons. First, it suggests that the state
was conceived of in a radically different way than it had been in the
past; second, it brings to light a new type of relationship between
the center and the periphery, and between the imperial lineage and
its surrounding houses and clans. In the realm of ritual and liturgical organization this meant, consequently, that centralized ritual
practices would play a strategic role in the formulation of imperial
legitimacy.
With regard to the new conception of that state, the compilation
of the Kojiki and Nihon shoki fixed the myths and symbols that became central to the definition and to the legitimation of the state.
With regard to the new relationship between center and periphery,
the creation of temples and the erection of shrines in the capital as
well as in the provinces show a new type of institutional organization
and cultic emphasis. Their establishment reveals, for example, the
delineation of geopolitical spheres of influence and an attempt to
unify ritual and liturgical notions and practices. The appearance of
those texts and institutions expresses forcefully that the cults dedicated to the kami of Heaven and Earth were being unified at the
level of the state. The need for such unification might be explained
according to the following three hypotheses. First, in maintaining a
single realm of symbols, ritual, and protocol, the state was hoping to
institutionalize its control over houses and lineages surrounding the
imperial family. Second, the control over shrines was accompanied
by an equal increase of control over Buddhist temples (although
such control was evident by the mid-Nara period, as is manifest in
the establishment of the provincial monasteries and nunneries
[kokubunji, kokubunnijt] in 741, it had become necessary by the end
of that period). And third, there arose a need to ground the definition and legitimacy of the centralized state in a symbolic and ritual
realm pertaining to both shrines and temples.
Many of the ritual practices of the Heian period can be explored
and explained in terms of these three hypotheses.
3 See PaulWheatley and Thomas See, From Court to Capital (Chicago: University of Chicago
Press, 1978).
Cambridge Histories Online © Cambridge University Press, 2008
52O
RELIGIOUS PRACTICES
THE ASSOCIATION OF SHINTO SHRINES WITH
BUDDHIST TEMPLES
The need for unification of ritual in shrines
The existence of specialists in liturgical and ritual matters is attested
in documents from the fourth century C.E. on.4 When those specialists came to form social units in which liturgical functions were
transmitted hereditarily is not altogether clear, but it seems that
some of these units have a long history. Japanese scholars refer to ritual specialists of pre-Nara periods by the word shukyo-teki bemin, literally, religious professional corps. The term be denotes, in pre-Taika
(645) society, a number of professional groups controlled or owned
by the emerging court or by influential houses. Even though the
Taika Reforms engineered by Prince Naka no Oe and his adviser
Nakatomi no Kamatari eliminated the term, the social reality remained in one form or another, especially in the case of sacerdotal
houses, whose name often contains the term be, such as Imbe, or
Urabe, and others. The term shake (literally, "shrine-house"), used
today to refer to social units charged with liturgical matters, appears
in texts only from the medieval period onward. These specialists in
ritual and liturgical matters carried separate traditions of practice
and interpretation, which the government tried to unify during the
eighth century, as is evidenced by the Kojiki, the Nihon shoki, and
the Fudoki, in the compilation of which some of the sacerdotal lineages closely related to the dominant houses cooperated. For example, one notices that the Nakatomi sacerdotal lineage (out of which
the Fujiwara house emerged at the end of the seventh century) took
some part in compiling those documents. That sacerdotal houses did
not agree on all points of liturgy, ritual, and mythology is illustrated
by the compilation of the Kogoshui in 807, in which the Imbe sacerdotal house expressed its own views. It was certainly not by chance
that the sacerdotal lineage governing shrines symbolizing the imperial state (such as Ise or Kasuga) was the Nakatomi lineage related
to the Fujiwara house.
The Taika Reforms conceived of the structure of government in
terms of a separation (and, at the same time, of an interdependence)
of cultic and policy-making duties; the Department of Shrines (Jingikan) was set above the Council of State (Daijokari), even though the
4 See InoueTatsuo, Kodaioken to shukyo-teki bemin (Tokyo: Kashiwa shobo, 1980).
Cambridge Histories Online © Cambridge University Press, 2008
SHINTO SHRINES AND BUDDHIST TEMPLES
521
rank of its members was inferior to that of the members of the
Daijokan. The Jingikan determined ritual practice pertaining to the
conduct of political affairs and of social matters within the court, but
it also functioned as an apparatus of control of the main sacerdotal
lineages of the time. Of great importance is that the Jingikan was
thought to be above the purely policy-making organs of government;
this does not mean that ritual matters took precedence over policy
making, but implies that ritual matters were central to the definition
of the state and of its social construct, and to the legitimation process
of some decisions made by the government. A corollary of this implication is that the social construct, the imperial state, and the policies made by the government were not only legitimized, but colored
by, and grounded in, ritual considerations. Since this principle determines in part our understanding of ritual practices and retained some
of its validity throughout the Heian period and beyond, it needs some
elaboration. One objective of the compilation of the Kojiki and Nihon
shoki had been to establish the supremacy of certain clans (uji) by
postulating their direct descent from specific divinities (kami) which
played an important role in the myths described in those texts. This
postulation not only created a definite prestige for the houses, it also
determined their position vis-a-vis the imperial lineage by virtue of
what were considered to be historical events. Furthermore, this postulation suggested that the structure of society was a mirror reflection
of the structure of the kami pantheon, by virtue of which rituals that
were dedicated to the kami had sociogenic characteristics. These
rituals dedicated to the kami of Heaven and Earth served to reinforce
the existing political power structure, especially since the kami of
Heaven were related to the sociocosmic construct of the state and to
its main houses, while the kami of Earth were related to other matters: as a general rule, few of those are the ancestral kami of the main
aristocratic houses. In other words, the world of shrines dedicated to
those kami which played an important role in the imperial mythology
was intensely sociopolitical; this is what accounts, in part, for the
dominantly protocolar character of ritual in shrines. When Buddhism
appeared, it was sponsored by precisely those houses and lineages immediately surrounding the imperial lineage; this means that it was expected to play, at the level of aristocratic houses and of the state, the
role played by the main shrines, and that the Buddhist clergy would
play a role similar to that of sacerdotal lineages. That is one of the reasons why shrines and temples came to be associated with each other
early in history, and why those associations retained strong sociopoCambridge Histories Online © Cambridge University Press, 2008
522
RELIGIOUS PRACTICES
litical overtones through much of Japanese history. If we keep in
mind the sociopolitical roles of the shrines and temples, we understand much better the types of practices of a ritual or liturgical character that flourished during the Heian period, and achieve a better insight into the combinatory nature of those practices and of their
related creeds.
Many of the most important temples of the Nara period were dedicated to the welfare of the spirits of departed leaders of the main
lineages, while a number of shrines were dedicated to the ancestral
and tutelary kami of the very same lineages. Concerned perhaps that
the private character of these institutions might lead to an excessive
fragmentation of society, the government transformed many such
temples into state-sponsored operations, and gave a new and unified
structure to the cult in the main shrines to ensure that it was directed toward the state rather than toward private interest. The government did not completely succeed in this attempt.
The Heian period is distinguished in part by the introduction
and remarkably swift acceptance of Esoteric Buddhism, symbolized
by the creation of the Enryakuji on Mount Hiei by SaichS (Dengyo
Daishi, 767-822) to house the Tendai school of Buddhism,5 and by
the creation of theToji (Kyoo-gokokuji) in Kyoto and of the Kongobuji on Mount Koya by Kukai (Kobo Daishi, 774-835) to house the
Shingon school.6 Even though the Tendai school was based on the
Lotus Sutra, esoteric practices of Shingon character and of Taoist
origins crept in during Saicho's lifetime and became the main
praxis of the school under the direction of Ennin (794-864) and
Enchin (814-91), to the point that one refers to Shingon esotericism as Tomitsu (i.e., esotericism [mitsu] of the Toji [td] Temple),
and to Tendai esotericism as Taimitsu (i.e., esotericism of Tendai,
here referred to as Tai). These were separated on the institutional
level and by different emphases on certain rituals which are discussed later in this chapter. These forms of esotericism tended to be
highly inclusivistic and eclectic, and became the main vehicle for
the assimilation by Buddhism of all kinds of ritual practices whose
origins were not Buddhist or which belonged either to Chinese and
Korean traditions or to Japanese indigenous practices. Thus, the
5 See Paul Groner, Saicho (Berkeley, Calif.: Asian Humanities Press, Buddhist Studies Series,
1985)6 See Hakeda, Kukai: Major Works. On the relationship between Kiikai and Saicho, see Allan
Grapard, "Patriarchs of Heian Buddhism Kukai and Saicho," in Great Historical Figures of
Japan (Tokyo: Japan Culture Institute, 1978), pp. 39-48.
Cambridge Histories Online © Cambridge University Press, 2008
SHINTO SHRINES AND BUDDHIST TEMPLES
523
main aspect of Buddhist practices throughout the Heian period
was esoteric and combinatory. This point cannot be emphasized
enough, for much of the ritual world of Heian evolved as the result
of systematic combinations between Esoteric Buddhism and the
ritual world of the shrines that, for convenience's sake, even though
inadequately, has been named Shinto.7 However, as we shall see,
the ritual world of the shrines was heavily laden with Taoist notions
and practices, as was, in fact, the Buddhist world. The cultic realm
of all classes of Heian society was combinatory, by which is meant
that it consisted of intermeshed forms of Esoteric Buddhism, Exoteric Buddhism, Taoism (though not of the institutionalized form
such as was then found in China), and various practices taking
place in shrines. I will approach this combinatory world by separating its main components, first at the institutional level, and then
at the level of practices.
Institutional considerations
Ritual and liturgical practices of the Heian period are marked by the
creation or enlargement of shrines (jinja) and temples (tera) supported by the government, by the codification of rituals performed
therein, by the emergence of new schools of Buddhism, by the appearance of a preeminence of practice over theory, and by the association of kami with buddhas and bodhisattvas worshiped in adjacent Buddhist temples. These associations resulted in the emergence
of specific cults of great importance, the analysis of which yields an
adequate understanding of varied but distinctly related dimensions
of Heian cultic life.
The various processes outlined in the paragraphs above continued to evolve after the move of the capital to Kyoto, and came to
fruition toward the middle of the Heian period. Furthermore, the
gradual association of some temples with shrines seems to have
been related to the formulation of the protection of the state {chingo
kokka), to the creation of the ideology of mutual support between
the state and Buddhism {obo buppo), and to the evolution of Heian
culture in general. This is most conspicuous in the phenomena
known as the "shrine-temples" (jinguji), and the "Twenty-two
Shrines" (nijunisha).
7 On this topic of importance, see Kuroda Toshio, "Shinto in the History of Japanese Religions," Journal ofJapanese Studies 7 (1981): 1-21.
Cambridge Histories Online © Cambridge University Press, 2008
524
RELIGIOUS PRACTICES
Shrine-temples: the jinguji
Japanese Buddhist monks backed by private support began to build
temples on the grounds of shrines by the middle of the Nara period;
such temples are known as jinguji. Even though it is most probable
that this phenomenon originated in Kyushu around the Usa Shrines
dedicated to Hachiman (where the Mirokuji was begun in 739 and
completed in 779),8 the earliest documented case occurred in
Kashima in Hitachi Province (present Ibaraki Prefecture), where a
certain Mangan erected a temple next to the shrine in 749. This is an
interesting matter, for the shrine in question was an important one
located in the easternmost part of the country controlled at the time
by theYamato government; in point of fact, it was so important that
the kami enshrined there, Takemikatsuchi no mikoto, was taken
around that time by the Fujiwara clan as its tutelary kami (ujigami)
and was subsequently enshrined in the Kasuga Shrine in Nara. The
other main tutelary kami of the Fujiwara house is Futsunushi no
mikoto, which was enshrined in Katori, located a few miles west of
Kashima. A jinguji was erected in Katori also, but the date is unclear.9 Other jinguji were built during the Nara period in shrines that
were distant from the center of government, by monks who appear
to have been associated with the Kofukuji, the main clan-temple
(ujidera) of the Fujiwara. One thus notes the creation of the jinguji of
Tado in 763 (Ise Province, present Mie Prefecture), and of those of
Kehi (Echizen Province, present Fukui Prefecture), Taga (Omi
Province, present Shiga Prefecture), Okutsushima (Omi Province),
Usa (Buzen Province, present Oita Prefecture), and Utagahama
(Shimotsuke Province, present Tochigi Prefecture) in the following
decades. This last one was created by Shodo (735-817) in 784; a detailed document showing the processes which led to its creation was
written in 814 by Kukai.loThis document indicates that Shodo made
the ascent of Mount Futara (today called Nantaizan, in Nikko), in
order to worship local kami and to achieve enlightenment, and that
this ascent was a religious experience in which he visualized the
8 See Nakano Hatayoshi, Hachiman shinko-shi no kenkyu (Tokyo:Yoshikawafcobunkan,1975).
9 On the topic of the Kashima and Katori Shrines, see the primary sources contained in
Shinto taikei hensankai, ed., Shinto taikei, vol. 22 (Tokyo: Seikosha, 1984). For a study of
the shrines, see MiyaiYoshio, Fujiwara-shi no ujigami-ujidera shinko to sobyo saishi (Tokyo:
Seiko shobo, 1978).
10 See the translation of this text by Allan Grapard, "Kukai: Stone Inscription for the sramana
Shodo, who Crossed Mountains and Streams in His Search for Awakening," in Michael
Tobias and Harold Drasdo, eds., The Mountain Spirit (New York: Overlook Press, 1978),
PP- 5O-59Cambridge Histories Online © Cambridge University Press, 2008
SHINTO SHRINES AND BUDDHIST TEMPLES
525
Bodhisattva of Compassion (Kannon) and saw Mount Futara as its
transcendental abode on earth. He then erected on the shores of the
lake (called today Chiizenji-ko) & jinguji. Besides being the first complete text documenting the ascent of mountains for religious purposes in Japan, this text reveals that, at least in Kukai's eyes, the
combined worship of kami and buddhas and bodhisattvas posed no
problem.
The jinguji, then, were institutions symbolizing early trends of
nonexclusive attitude toward autochthonous and imported creeds
and practices, while they were at the same time institutions expressing the power of houses concerned with controlling cultic matters,
such as the Fujiwara house.The Buddhist clergy administering those
shrine-temples performed Buddhist rites in front of the kami, chanting Buddhist scriptures with the avowed goal of releasing the kami
from their unenlightened state of being and guiding them toward the
Buddhist goal of awakening. This was probably one of the strategies
used by Buddhism to gain converts in remote areas, but it was also
a strategy aimed at installing, in distant places, cult representatives
of the government. Indeed, government provisions concerning the
number of monks that could be ordained by temples under government supervision stipulated that a certain number would specialize
in shrines; they were called shimbun dosha, "monks ordained for the
worship of the kami." The Heian period saw the creation of temples
on the grounds of shrines on a massive scale, but the term jinguji
tended to be abandoned. This suggests that the function of temples
vis-a-vis shrines evolved toward increasing control (if not domination), as is shown by the following analysis of the system of statesponsored shrines and temples called traditionally, though improperly, nijunisha, the "Twenty-two Shrines."
The Twenty-two Shrines system
The term nijunisha indicates twenty-two shrines that became the object of imperial support during the first part of the Heian period and
thus became major institutions symbolizing what might be called the
cultic system of the imperial lineage. By the middle of the ninth century this system showed a trend toward the unification of ritual
modes and the stabilization of a government ideology concerning
the relations between shrines and temples. At the time of ritual performance, which followed the seasonal cycle overlaid with calendrical principles originating in India and China, the emperor or some
Cambridge Histories Online © Cambridge University Press, 2008
526
RELIGIOUS PRACTICES
of his representatives would participate in rites according to strict
standards laid down by the government, make offerings the value of
which was set forth in decrees, and provide economic support of the
sites of cult. The appearance of that system heralded a significant
step in the evolution of cultic and policy-making institutions and
practices, as well as a major increase in state control. This evolution
led to the systematic institutionalization of the dominant Heian government ideology, namely, the "protection of the imperial state (by
shrine-temple multiplexes)" (chingo kokka, a term that will be discussed later), and the mutual support between the state and the
combined exoteric-esoteric system (kemmitsu taisei) of shrines and
temples, also discussed later in this chapter.
The shrines under consideration were located, for the most part,
in the Kinai area (the five provinces immediately surrounding the
capital), and most were included in the list of major shrines established by the Engi shiki in 927 because they were related to the central lineages supporting the imperial house, or to some lineages of
great historical importance.11 Those shrines included in the list of
twenty-two, but not in the Engi shiki, are sites of cults that gained
popular significance thereafter, or that were created at the time to
accommodate the court. The origin of the majority of these shrines
goes back to antiquity, when they already represented specific connections between cultic and sociopolitical concerns. It is those concerns that came to be expressed and reinforced by the creation of the
system and its accompanying ideology.
The process of quasi unification of those shrines may have begun
at the onset of the Heian period for, when the list appears for the
first time in documents in 966, sixteen shrines were included that
had long been the object of imperial offerings. These were separated
into an Upper Group of seven shrines, a Middle Group of seven, and
a Lower Group of two to which were added three shrines in 991, one
in 994, and two in 1039.
A study of these shrines brings to light two major characteristics:
a very specific relationship of those institutions to the Fujiwara
house and the presence of Buddhist institutions at most of the
twenty-two shrines.
Indeed, the classification of the twenty-two shrines into three
groups exhibits two rationales: one geopolitical, and one political. In
the first case, five of the first seven shrines are located in the province
11 See Felicia Bock, trans., Engi-shiki: The Procedures of the Engi Era, i vols. (Tokyo: Sophia
University Press, 1970-72).
Cambridge Histories Online © Cambridge University Press, 2008
SHINTO SHRINES AND BUDDHIST TEMPLES
527
TABLE 8 . 1
The Twenty-two Shrine-temples sponsored by the
imperial government
Upper seven
Ise
Iwashimizu
Kamo
Matsunoo
Hirano
Inari
Kasuga
Middle seven
Oharano
Omiwa
Isonokami
Yamato
Hirose
Tatsuta
Sumiyoshi
Lower eight
Hie
Umenomiya
Yoshida
Hirota
Gion
Kitano
Nibimokawakami
Kibune
of Yamashiro, where the capital Heian-kyo is located; the two other
shrines are Ise (in Ise Province) and Kasuga (in Yamato Province),
dedicated respectively to the ancestral and tutelary kami of the imperial lineage and of the Fujiwara house. The second seven shrines
are located in the Yamato region, with the exception of the Oharano
and Sumiyoshi Shrines; the Oharano Shrine (inYamashiro Province)
is the Kasuga Shrine that had been duplicated in the Nagaoka capital, while the Sumiyoshi Shrine (in Settsu Province) was directly related to imperial mythology and imperial enthronement ceremonies,
particularly during the Heian period. The last group of eight shrines
consists of shrines and temples located in the Kinai provinces and
specializing in placating droughts or various nocive forces that made
themselves evident during the period. In other words, the first group
is close to the capital, the second group represents the past history
of the lineages in Yamato, and the third group represents a larger
concern with the periphery.
In the case of political rationale, it must be pointed out that the
Fujiwara house made sure that its major shrines would be represented in the three classes of shrines: Kasuga in the first, Oharano in
the second, and Yoshida (the Kasuga Shrine of Kyoto), in the third.
This indicates that the Fujiwara house emphasized its presence at
the three levels of organization. Furthermore, one might emphasize
that shrines related to Kyushu are found in the three classes:
Iwashimizu (Hachiman) in the first, Sumiyoshi in the second, and
Hirota in the third. This suggests that certain aspects of mythology
remained important during the Heian period.
Equally important is the often ignored fact that the Twenty-two
Shrines were associated with Buddhist temples located on or near
Cambridge Histories Online © Cambridge University Press, 2008
528
RELIGIOUS PRACTICES
their precincts. By the middle of the Heian period the Twenty-two
"Shrines" had become major shrine-temple multiplexes (jisha) consisting of complex combinations and interactions between autochthonous (or foreign) kami and imported buddhas and bodhisattvas,
between sacerdotal lineages in the shrines and Buddhist lineages in
the temples, between administrative and economic structures of the
two types of institutions, and between modes of cult, all under the
aegis of the state symbolized by the overwhelming presence of
members of the Fujiwara house as the leading prelates of those institutions. Many of the shrines and temples had been created by the
Fujiwara house, or were restored by members of the house during
the Heian period. This indicates that the Fujiwara house consolidated its grip over the country in part through the formation and
administration of shrine-temple multiplexes to which the imperial
house made offerings and from which it requested the performance
of sumptuous rites for the protection of the realm under its governance. The soon dominant modality of relationship between cultic
and policy issues, which took the name of mutual support between
these institutions and the state (obo-buppd), was thus institutionalized in the system (taisei) of shrines and temples of exoteric (ken)
and esoteric (mitsu) Buddhism (that is, kemmitsu taisei).12
Of particular fame and importance were the shrine-temple multiplexes of Hie-Enryakuji (Mount Hiei),13 of Kasuga-Kofukuji,14 of
Iwashimizu Hachiman-Gokokuji,15 of Sumiyoshi-Shiragidera,16 of
Gion-Kankeiji (Kanshin'in), and of Kitano-Kannonji. All these
played a central role in the evolution of Heian cultic life and its political components, and continued to be major sites of cult throughout Japanese history.
Chingo kokka theories and practices
In its narrow meaning the term kokka refers to the imperial lineage
and supporting houses (ka) which govern the realm of the country
12 These issues have been studied in detail in KurodaToshio, Jisha seiryoku (Tokyo: Iwanami,
1980), andObo to buppo (Kyoto: Kozokan, 1983).
13 See Murayama Shuichi and Kageyama Haruki, Hiei-zan (Tokyo: NHK Bukkusu, 1970).
14 See Allan Grapard, The Protocol of the Gods (Berkeley: University of California Press, 1992);
and Royall Tyler, The Miracles of the Kasuga Deity (Princeton, N.J.: Princeton University
Press, 1990).
15 See Nakano Hatayoshi, Usa-gu (Tokyo: Yoshikawa kobunkan, 1985); Hachiman shinko-shi
no kenkytt, 2 vols. (Tokyo: Yoshikawa kobunkan, 1975); Allan Grapard, The Religion of Space
and the Limits of Religion (forthcoming).
16 See NishimotoYutaka, Sumiyoshi taisha (Tokyo: Gakuseisha, 1977); see also Ueda Masaaki,
ed., Sumiyoshi to Munakata no kami (Tokyo: Chikuma shobo, 1988).
Cambridge Histories Online © Cambridge University Press, 2008
SHINTO SHRINES AND BUDDHIST TEMPLES
529
(koku); in its larger meaning, however, it denotes the state at large,
with an emphasis on its geopolitical domain or territory. As such, the
term chingo kokka refers to a government ideology legitimated by rituals performed by both sacerdotal lineages in specific shrines and
Buddhist lineages in specific temples, in the name of the imperial
lineage and its satellite households, particularly the Fujiwara. Although the term is of Sino-Japanese origin, the concept of protection
of a state was also present in Buddhism in India. It was developed to
a great extent in China (one of the scriptures on which Buddhism
relied for those rituals, the Ninno-kyo, is actually an apocryph conceived in China),17 and made its way into Japanese theory and practice as early as the Nara period, to blossom during the Heian period
in the Twenty-two Shrine-temple multiplexes as well as in other sites
of cult located in various parts of the country. Thus, what is covered
by the term chingo kokka is in reality a complex assortment of Indian
and Chinese concepts and practices adapted to Japanese needs and
within the parameter of the political submission of various Buddhist
communities to the state.
It had been noted above that the term kokka may have meant originally the imperial lineage and the realm under its rule. One of the
basic notions related to that meaning was that the realm in question
was seen as literally embodied in the person of the emperor, who was
considered to be descended from kami such as Izanami, Izanagi, and
Amaterasu. These central figures of the pantheon exhibited in their
symbolic body the very structures and functions of the ideal classical society; therefore, the emperor's body symbolized the same
structures and functions, and that body came to be seen as the
"body of the imperial state" (kokutai). A primary consequence of
this view was that some of the major rites of protection of the state
were in fact rites of protection of the emperor's body. Thus, when the
emperor fell ill, rituals were requested from shrines and temples in
the name of the protection of the state; and a majority of rites for the
protection of the state tended to be related, in one way or another,
to the body of the emperor. The two were inseparable. A rite such as
that dedicated to the peacock {Kujaku-ko; Sanskrit, Mahamayuri) in
Esoteric Buddhism came to be used, in Japan, for the protection of
the state: that ritual was originally related to snake bites, poison, and
17 Ninno-kyo, properly, Ninno haramitsu-kyo, in Takakusu Junjird, Watanabe Kaigyoku, Ono
Gemmyo, eds., Taisho shinshu daizbkyo (hereafter abbreviated as 7*), 100 vols. (Tokyo:
Taisho issaikyo kankokai, 1924-34), vol. 8, no. 245 (in KumarajTva's translation), and no.
246 (in Amoghavajra's translation).
Cambridge Histories Online © Cambridge University Press, 2008
53°
RELIGIOUS PRACTICES
disease, and was not, at the time of its inception in India, connected
to the notion of realm or to the political power of a ruler.
The tradition conceiving of the body of the emperor as a symbol
of the state can be explained from several perspectives in the Indian
and Sino-Japanese contexts. In the case of India, it is well known
that enthronement rites were related to speculations concerning the
microcosmic nature of the human body in relation to the macrocosm. In the case of China, Taoism conceived of the body of the
ruler as a cosmic body as when, for example, the Yellow Emperor
was thought to be a manifestation on earth of Lao-Tzu, himself a
cosmic body.18 In Japan, the Kojiki and the Nihon shoki offer several
indications to the effect that threats to the imperial body were
threats to the structure of society, and that death of the imperial person needed to be treated by rituals protecting the realm at large.
That is how one might look at the tama-shizume {chinkonsai) rituals
aiming at restoring longevity to the imperial figure, and it is not by
chance that the term shizume ("to pacify") is found in the term
chingo kokka (in which it is pronounced chin). Therefore, when we
find in the language of rituals dedicated to the protection of the
state, even within a Buddhist context, that the object of those rituals
is the body of the emperor, we must understand that what is meant
is both the physical and symbolic body of the emperor/realm. Since
both the realm and the imperial body were conceived of as symbolic
entities, ritual was the appropriate solution to any type of "disease"
affecting them. And in that ritual cure, shrines and temples played a
structurally similar role in which they were partly competing, and
partly cooperating. Indeed, the emperor came to be represented, at
the ideal level, as a bodhisattva (as opposed to India, where the bodhisattva was described as a ruler).
Since the Twenty-two Shrine-temple multiplexes were those from
which rituals for the protection of the realm were requested, and
since the appearance of the system seemed to herald a unification
process, one might expect that the unification would have occurred,
and that a single administrative body governing the system might
have emerged. This was not the case, however, and the reason might
18 On the topic of the Yellow Emperor, see Anna Seidel, "La Divinisation de Lao-tseu dans
le Taoisme des Han," Publications de l'Ecole Francaise d'Extreme-Orient, vol. 71 (Paris:
Ecole Francaise d'Extreme-Orient, 1969). On the subject of the body in Chinese religious
thought in general, see Christofer Schipper, "The Taoist Body," History of Religions 17
(1978): 325-51. And for views of the body in India, see Frits Staal, "Indian Concepts of the
Body," Somatics (Autumn/Winter 1983—84): 31—41.
Cambridge Histories Online © Cambridge University Press, 2008
SHINTO SHRINES AND BUDDHIST TEMPLES
531
be that the unification process was true only at the level of the ritual
and legitimacy needs of the imperial house and its satellite lineages.
The system of multiplexes never did develop into a single unified
body of practice or theory, or into anything close to what might be
termed religious orthodoxy. If there was any orthodoxy in Heian
Japan, it was of a sociopolitical nature, and was embedded in rituals
performed in discrete and semiautonomous sites of cult governed by
distinct sacerdotal lineages under the direction of powerful houses,
and by ecclesiastic elites issued from aristocratic lineages. For example, starting in the late tenth century, shrine-temple multiplexes
came to be governed by aristocracy-born ecclesiasts (the monzeki
governed by kishu, "noble seeds," from main lineages, and by ryoke,
"members of good households," from lesser aristocratic birth).19
This means that multiplexes represented house-oriented private interests, even when their public rhetoric was geared to the state. A
second reason is connected with the decline of the statutory (ritsuryo) system during the Heian period. Shrine-temple multiplexes garnered vast land estates free from taxation, and thus developed into
major economic and ideological units that, by the end of the Heian
period, came into various conflicts either with the imperial government or with warrior houses, or with each other. Indeed, during the
period a few emperors attempted to curb the growing political, economic, and military power of those sites of cult, but the private interests of specific sites of cult overrode any other interest, so that violent conflict erupted between multiplexes that should have been
unified at least at the doctrinal level.20 However, these shrine-temple
multiplexes represented the cultic orthodoxy of the imperial state,
and most of them were closed, during the Heian period, to people
who did not belong to the main houses that formed the court. The
theory and practice of the protection of the state that they evolved
during the course of the Heian period were thus restricted to the imperial house and its satellite lineages. Such theory and practice continued to survive centuries after the end of the Heian period and
were revived at the time of foreign invasions or at the time of imperial succession crises.
19 Not surprisingly, the first monzeki was the Ichijo-in (created between 978 and 983) of the
Kofukuji, whose abbot was josho, grandson of Fujiwara no Moroyasu, then Minister of the
Left.
20 See Gaston Renondeau, Les Moines Guerrien du Japon (Paris: Imprimerie Nationale,
1965). See also Kuroda, Jisha seiryoku.
Cambridge Histories Online © Cambridge University Press, 2008
532
RELIGIOUS PRACTICES
RITUALIZED AND RITUALIZING ACTIVITIES
Rituals performed in shrine-temple multiplexes
During the Nara period, rituals of protection of the state had been
performed in shrines and temples in order to placate various forces
responsible for earthquakes, droughts, epidemics, and disease striking the imperial person or leading aristocrats, or the country at large.
Buddhist statues were commissioned and copying of scriptures
(shakyo) took place as preventive measures. The scriptures in question belonged to the literature of the Perfection of Wisdom {daihannya; Sanskrit, maha-prqjna paramita). Furthermore, the erection by
the government of monasteries and nunneries in the various
provinces (the kokubunji system created in 741) had been undertaken with the avowed purpose of protecting the realm at two levels:
those of ritual and education. Specific members of the pantheon enshrined in shrines and temples were regarded as active protectors of
the state in this context, and were therefore the object of ritual and
devotion; preeminent among those in the Buddhist context were
Kannon (Avalokitesvara), whose sculptures were made at the time of
major political or military unrest threatening the texture of society
and, among many more, the Four Heavenly Kings (Shitenno),
guardians of the four directions. These rituals belonged to the unorganized tradition of "mixed esotericism" (zomitsu), which was spread
widely in the western part of Japan and did not have allegiance to
particular schools of Buddhism.
However, Kukai and Saich5 came back from China at the beginning of the ninth century and created new schools of Buddhism that
were to have a tremendous impact on the ritual and philosophical
world of Heian. Both introduced to Japan new rituals of protection
of the state.
Saicho had learned during his stay in China a number of rites to
this effect: the Birushana-ho (ritual ofVairocana), which he eventually performed in the imperial palace; the Gundari-dampo (ritual of
Kundali), which is essentially martial; the Fugen-e dampo (ritual of
Samantabhadra), the Nyoirin-dampo (ritual of Cintamani), and the
Gobutcho-ho (ritual of the Five Buddhas emanated from Sakyamuni's cranial protuberance). But the full-fledged development
and performance of rituals by theTendai lineage did not occur until
Saicho's disciples Ennin and Enchin returned from China with the
latest developments in the field, and thoroughly transformed
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
533
Tendai into an esoteric ritual tradition geared to the protection of
the state.21
When Kukai returned from China in 806, his expertise was immediately sought after by the government. He set to the task of reproducing mandalas and altars for these protective rituals and performed initiatory unctions (kanjo), allowing monks and aristocrats
alike to be participants in the esoteric doctrine and the rituals connected to it. The government requested in 816 that he build an Initiation Hall within the compounds of theTodaiji in Nara for the protection of the state. In 834, one year before his death, Kukai saw to
it that a special building designed for the performance of those rites
was built within the imperial palace. That is the Shingon'in, in which
Shingon's major rite for the protection for the imperial house, the
go-shichinichi mi-shiho, was performed during the second week of the
first month.
The Shingon tradition and rituals for the state
The first ritual Kukai performed for the protection of the state was
in 810, the year after the rebellion of Fujiwara no Kusuko which signaled the decline of the Shiki branch and the surge to power of the
Northern branch of the Fujiwara house. Kukai obviously sided with
the Northern branch, since it is reported that he performed in 814
the rituals for the opening of the Nan'en-d5 at the Kofukuji in Nara,
which remained for centuries one of the salient symbols of dominance of the Northern branch of the Fujiwara house. As Kukai wrote
in a document asking for permission to perform this ritual, he
needed to teach correct procedure to his disciples; he admitted that
even though he had been initiated by his master in China and was
thus qualified to perform the rite, he still needed to train. He also
stated that past rituals for the protection of the state had emphasized
the chanting of scriptures, but he emphasized that rituals would be
far more effective if they were based on esoteric doctrine and on the
adequate formulation of potent formulas and charms (darani),
which he equated with the proper intake of medicine." The aristocrats of the time welcomed Kukai's stance because it fit perfectly
21 The best treatment of the variegated aspects of Tendai esoteric practices and ideas is in
Misaki Ryoshu, Taimitsu no kenkyu (Tokyo: Sobunsha, 1988); Mikkyo tojingi shiso (Tokyo:
Sobunsha, 1992); Taimitsu no riron tojissen (Tokyo: Sobunsha, 1994); and Nihon. Chugoku
Bukkyo shiso to sono tenkai (Tokyo: Sankibd, 1993).
22 Hayami Tasuku, Heian jidai kizoku-shakai to bukkyo (Tokyo: Yoshikawa kobunkan, 1975).
Cambridge Histories Online © Cambridge University Press, 2008
534
RELIGIOUS PRACTICES
their concern with magic, manipulation of symbols, and medicine.
Indeed, Kukai insisted in many of his writings on the therapeutic
quality of the teachings of the Buddha and, more specifically, of the
rituals of esotericism.23 To cure the body and the mind stood, in the
early Heian period, as the prime metaphor for the protection of
the state. If the cosmos was seen as a body, and the human body as
a microcosm thereof, then cosmic disasters were to be treated like a
disease in the body. That is why Kukai and his contemporaries had
a strong personal devotion to the Buddha of Medicine, and why that
Buddha became a central figure in rituals of protection of the state.
The scriptures used for those rituals were already used during the
Nara period: they were the Konkomydkyd,24 the Daihannya-kyo,25 and
the Kongdhannyakyd2% on which Kukai wrote several interpretive
documents (kaidai).27 In the time of Kukai, aristocrats requested the
performance of those rituals for themselves or members of their
family; several such requests are mentioned in Kukai's works, accompanied by the petitions he wrote for those rites.28 Another trend,
known as "day chanting of scriptures, and evening rites of penitence," surfaces as early as 833 in court orders to perform such rites
against epidemics, and in 834 against threats to crops; in both cases
the court ordered that various temples dedicate three days to readings of the Kongohannyakyo, and three evenings to rites of penitence.
Penitence was the most fundamental aspect of Mahayana ritual; it
consisted mainly in reciting lists of sins in order to purify the organs
of perception. Once accomplished, this allowed a direct and undefiled perception of the universe, which took the form of visions to
which a therapeutic character was assigned. It is clear that Kukai,
and others with him, considered natural events such as epidemics
and droughts to be caused by unethical behavior grounded in misperceptions of reality, and therefore countered them by acts of penitence leading to a more adequate perception of the natural world
and of one's place in it. The rites of penitence in question were those
associated with the scriptures dedicated to the Buddha of Medicine,
23 See, for instance, the introduction to his Himitsu mandara jujiishinron in Katsumata
Shunkyo, ed., Koto Daishi zenshu, vol. 1 (Tokyo: Sankibo, 1968), p. 210.
24 Konkomyo-kyb, T, vol. 16, no. 663.
25 Daihannya-kyo (properly, Maka hannya haramitsu-kyo), T, vol. 8, no. 223.
26 Kongohannya-kyo (properly, Kongo hannya haramitsu-kyo), T, vol. 8, no. 235.
27 See, for instance, his Ninno-kyo kaidai, in Kobo Daishi zenshu, vol. 1, p. 329, and his Kongo
hannya-kyo kaidai, in Kobo Daishi zenshu, vol. 1, p. 489.
28 See examples of these petitions in Watanabe Shoko and MiyasakaYusho, eds., Sango-shiiki,
Seireishu (Tokyo: Iwanami, 1965): Seireishu, kan 4, no. 20, for a petition for a ritual for the
imperium; for rituals for aristocrats, see kan 6, no. 46, no. 50, no. 52, no. 55, and others.
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
535
and were known as Yakushi keka. Whereas in Chinese practice rites
of penitence tended to be individual matters (even though that was
not the case in the early Taoist context), the Japanese required group
penitence for the sake of the state, aiming at the removal of baleful
omens concerning whatever might threaten the human representatives of that state. By the second half of the ninth century, however,
the evening sessions tended to be devoted to the recitation of charms
and spells dedicated to the Buddha of Medicine or to the Bodhisattva of Compassion, and were followed by the recitation of
mantras known as komyo shingon and Shaka butsugen shingon.
Beyond those rites, Esoteric Buddhism initiated the performance
of specific rites dedicated to individual members of the pantheon.
Among these were two major rites for the protection of the imperial
state, performed in competition by Shingon and Tendai. The major
Shingon rite, still performed in the modern period, is known as
Daigen(sui)-ho. The main Tendai ritual performed over the centuries
for the protection of the state is the Shijoko-ho. Various Shingon lineages also specialized in the performance of rainmaking rituals,
which were important matters of state.
The Ritual o/Daigen(sui)
The first of these rites concerned the pantheon figure Daigensui,
one of the important "Kings of Science" (myoo) of Japanese esotericism.29 In India Daigensui's name was Atavaka, an ogre who converted to Buddhism and became one of the eight attendants of
Vaisravana (Japanese, Bishamon), the guardian of the northern direction (itself the object of important rites of protection of the realm
in China as well as in Japan). In China, however, the term Daigensui designated the supreme commander of military forces; later it
came to designate in both China and Japan the emperor as commander in chief of all forces (in Japan, from the Meiji period). Because Daigensui was the commander in chief of all spirits and
demons protecting the land, his cult developed with a focus on the
protection of the state; it penetrated the court in China during the
T'ang dynasty, and was introduced to Japan by Kukai and his disciple Jogyo (d. 866), where it became one of the most important national rituals kept in the highest degree of secrecy.
29 See Robert Duquenne, "Daigensui," Hobogirin, 6 (1983): 610-40. (Usually, within the
Shingon tradition, the graphs dai, gen, and sui are written, but sui is not pronounced.)
Cambridge Histories Online © Cambridge University Press, 2008
536
RELIGIOUS PRACTICES
In Japan, various altars and purified areas were constructed for the
performance of the rituals dedicated to Daigensui; the rites involved
preliminary purification, the chanting of incantations, various offerings, and the layout of military attributes symbolizing the deity, such
as swords, bows, and arrows. The ritual was performed for the first
time in 840 in the Joneiden within the imperial palace. In 850 Jogyo
requested that the rite be performed by fifteen monks, as in the case
of the other main Shingon ritual performed in the Shingon chapel of
the imperial palace. The Daigensui ritual was performed in the
imperial palace every year, with even more regularity than the goshichinichi mi-shiho, but it was abandoned in early Meiji at the time
of anti-Buddhist movements.30
Rainmaking rituals
Rituals performed to cause rain to fall were, in Japan, important
matters of state directly related to the emperor, and were requested
from both shrines and temples. The Engi shiki, for instance, mentions the existence of as many as forty-nine shrines (of xhe yamaguchi
and mikumari types) specializing in such rituals in the Kinai area
alone. On the Buddhist side, Shingon specialized in the performance
of such rituals, but there are few records showing thatTendai monks
did them. Two major types of "Buddhist" rain rituals existed during
the Heian period: the Kujaku-ho (Rite of the Peacock),31 which became the specialty of the Hirosawa lineage of Shingon, and the
Shou-kyo-ho (based on the scripture of the same name),32 in which
the Ono lineage of Shingon specialized. Such rites were already performed during the Nara period, but in the Heian eminent thaumaturgists were called upon. A famous case is that of the Ono lineage thaumaturgist Ninkai (b. 951), who came to be known as ame
sop, literally, the "Monacal Rector of Rain." He performed the ritual
successfully nine times between 1028 and 1044.
Tradition claimed that Kukai performed the first ritual for rain at
the Shinsen'en garden in the capital in 824, though there is no proof.
But the text he is said to have authored in the wake of his success is
30 The ritual was subsequently reinstated with the provision that it be performed only once
during the reign of an emperor. It was last performed, officially, in 1928; however, extraordinary performances took place during World War II.
31 The peacock ritual is in the Dai kujaku nryoo gazb danjo giki, T, vol. 19, no. 983 [a].
32 Shou-kyd (properly, Daiunrin shou-kyo), T, vol. 19, no. 989. The ritual is Daiun-gyo kiu
dampo, T, vol. 19, no. 990.
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
537
so filled with information related to the preceding discussion of the
"body of the state" that it deserves full presentation.
How pitiful we are! Poor people living in the age of decadence of the
dharma, deaf and blind who fail to pay heed to the words of the Buddha.
Forever submerged in the drowsiness caused by the alcohol of ignorance,
unaware of the source of Innate Awakening. Eternally lost in slumber while
dreaming up the Triple World, and creating a foolish attachment to this
body composed of four elements! We perpetrate the ten evils with body,
speech, and mental activities. Ignoring filial piety or loyalty, we let the
thicket of sins grow wild and, discarding the law of cause to effect, we fail
to realize the distinction between well-being and misery. Lost in the dark of
illusion we follow wit and greed and thus strain the function of mouth and
belly. Laughing while alive, and crying upon death. Losing control we find
ourselves struck now to the east, now to the north. But the retribution of
evil acts is heavy, while that of good acts is light. That is why one day we
will see fire instead of water in rivers. We find hell in what is in fact our Buddha-body, and fail to see the value of the seven jewels with which we are
adorned. And thus even though it is time for rain to fall, the four horizons
are blazing with heat: the sun burns everything, and rice and millet ears are
all dry. The entire natural world dries up and hardens, and animals of fur
and scale perish. With nothing to see but aridity in the land, court and
peasants alike pour tears ceaselessly.
That is the moment our emperor chose to express a vow and, for the sake
of the people, to leave his palace. Filled now with wisdom, now with humaneness, he surveyed the Eight Islands. Holding in mind the Three Teachings and the Nine Sciences, he remained set in his practice of the Four
Incommensurable Concentrations and of the Six Perfections. Taking responsibility for the drought upon his own person he left his palace and, for the
benefit of all living things, has reduced his food and sits day and night in
pain. At this juncture he has requested monks in all temples to perform rituals and has sent emissaries to all mountains to ask thaumaturgists to enter
ascetic concentrations.
As the venerable monks chant the sacred scriptures, light clouds appear
in the sky, and as the practitioners of meditation hold firm to their concentration, rain clouds gather on the horizon. Then, sweet nectar of rain,
that sublime ghee obscures the skies and comes to wash mountains and valleys. Waterfalls gush forth from high peaks and soak wild animals, while
rain fills the fields enough to drown water buffaloes. Green trees are as if
decorated with gems at the tips of their leaves, and the wide dams now
overflow with water green as lapis lazuli.
Peasants, do not lament anymore! Go look and see whether early and
mature rice are dead or alive. The southern fields abound with seedlings
turning green. The eastern fields echo with the beat of the drums; songs
and rhythms fill the air all around. See the storehouses, where grain piles
up like islands, like mountains. How marvelous it is that the efficiency of
the dharma need not be explained. How felicitous it is that the power of
the emperor cannot be measured. A bodhisattva's single spit can extin-
Cambridge Histories Online © Cambridge University Press, 2008
538
RELIGIOUS PRACTICES
guish the fires of a hundred realms, and in a moment end the laments of
all subjects.
I, Kukai, herewith give teachings to those who peregrinate through the
six destinations; using the words of the Buddha, I let you know of the Buddha within you. Men and women! While holdingfirmto the single letter a,
envision the inner palace of your own heart-mind! It is the very ground of
the Triple Body of the Buddha, adornedfromthe origins widi the Five Wisdoms. Should you wish to discover this Original Source, undergo the rite
of Initiation. By doing so and holding firmly to it, you will immediately
reach the level of the Awakened One. As rain falls from Heaven, so will food
and clothing. And since you will not interfere with nature and will33be without cause to worry, you will forget even the virtue of the emperor.
Whether written by Kiikai or not, this document does indicate
that the emperor was seen as symbolizing in his body the entirety of
the realm; it was as carrier of everyone else's deeds that he performed the rituals. And it was through ascetic practices in his body
that he caused the rain to fall. Naturally, Kukai was the performer of
the austerities and rites, and much more than "in the name" of the
emperor: he was the emperor's ritual body. Another interesting aspect of this text is that it indicates that the rites of rainmaking were
somehow grounded in penitence, which implies a relatedness between ignorance and natural calamities. This relatedness is inscribed
within the larger framework of a dialectic between nature and culture, one of the results of which suggests that the natural world is
sensitive to unethical human behavior, in which case it goes awry,
and that it is also sensitive to ethical action, in which case it goes
back to its ideal workings. That is the reason why meditation and
ethics were a central part of rainmaking rituals.
The first credible record of the performance of this rite at the
Shinsen'en appears in 875 with the name of Shobo (832-909), the
Shingon organizer of mountain asceticism (Shugendo). After 915,
rain rituals were usually accompanied by the rite of the Five Dragons (Goryusai), of Chinese origin.
The Tendai tradition and state rituals
As discussed earlier, Saicho did bring back to Japan a number of
rites and made sure that they were performed in the name of the imperial state, because this meant sponsorship by the government and
the leading aristocratic lineages. However, it became quickly evident
33 Kukaij Shoryoshu, kan 1, no. 4, in Sango shiiki: Shoryoshu, p. 168. There is yet another petition for rain written by Kukai in 827, in Shoryoshu, kan 6, no. 47, p. 292.
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
539
that Shingon esotericism was superior because it provided a complex ritual framework backed by a doctrine that appealed to the aristocracy, and Tendai henceforth competed with Shingon by sending
its monks to China to bring back even more rituals. Thus, Ennin,
who went to China in 838 and stayed there for nine years, brought
back a number of rituals of importance, among which were rites for
the protection of the state: the Emmyo-ho (to prolong life); the Shichibutsu Yakushi-ho (a grand rite in front of seven sculpted representations of the Buddha of Medicine); and the Monju hachiji-ho, dedicated to the Bodhisattva of Wisdom and based on the Monjushiri
bosatsu hachiji sammai-ho translated into Chinese by Bodhiruci.34 But
the most important ritual introduced by Ennin was the Shijoko-ho,
which became the central Tendai rite for the protection of the state.
The ritual of Shijoko
The word Shijoko, which means "radiance of a vivid fire," is the
name of one of several personified emanations that sprang from the
cranial protuberance of the Buddha, five of which gradually
achieved a single status in doctrine and ritual, while the others remained in groups of five, eight, or ten.35 Even though the ritual of
Shijoko had been brought back from China by Kukai, it was never
emphasized within the Shingon tradition; it became instead the central ritual for the protection of the state in the Tendai lineage within
the Sammon (Enryakuji) tradition, in which it served as a balance
to the ritual of Daigensui of Shingon.
Ennin's first performance of the ritual took place either in 849 or
850 on Mount Hiei. The principal object of the ritual was the annihilation of natural catastrophes, but it came to be associated with the
emperor in a direct way, and was requested by the government
whenever military danger faced the court. Originally performed by
fourteen monks in the Sojiin of Mount Hiei, which had been created
by Emperor Montoku for that purpose, it came to be performed
from the tenth century on more often and for different reasons. In
34 The original tide of this scripture is Monjushiri hobozo damni-kyo, T, vol. 20, no. 1185; the
current title was dictated by common usage over time.
35 The Japanese generic term used to refer to those emanations in Butcho (Sanskrit, Buddhosnisa, or UsnTsaraja).The Shijoko emanation under discussion here is the object of the
Shijoko daiitoku shosai kichijo darani-kyb,T, vol. ig, no. 963; and of the Daiitoku kinrin butcho
shijoko nyorai shojo issai sainan darani-kyo, more commonly called Shojosainan-gyo, T, vol.
19, no. 964. The oldest extant ritual text, Shijoko dojo nenju-gi, T, vol. 46, no. 1951, was compiled by Fa-baoTa-shi (946-1032).
Cambridge Histories Online © Cambridge University Press, 2008
540
RELIGIOUS PRACTICES
905 the number of monks was changed from fourteen to forty, and
in 911 the ritual was performed within the imperial palace after there
had been an ominous occurrence involving a heron. A new accompanying scripture was given by aT'ang merchant to no less than Fujiwara noTokihira (871-909). Because that ritual was shrouded in secrecy, it did not involve the establishment of a specific mandate;
however, paintings representing the Buddha Shij6k5 do exist.36
How did a ritual aimed at preventing natural calamities come to
be associated with military threats to the court? People of the time
believed that disturbances in the world of nature were advance
warnings of impending military rebellions. This was not a specifically Buddhist phenomenon, for we read in the Fudoki that the rumbling and shaking of a mountain in Kyushu was thought to mean
that an army was on the march.37 In a similar vein, the cult dedicated
to Nakatomi no Kamatari (the ancestor of the Fujiwara house) on
Mount Tonomine in the province ofYamato was partly based on ritual counteractions to tremors that shook the mountain or cracked
Kamatari's statue.38 In that case it was believed that the direction of
the epicenter of the quake indicated specific realms of governance by
the Fujiwara house. Thus, any sudden and frightening occurrence in
the realm of nature was thought to be an omen concerning political
affairs and the security of the imperial lineage. This was also part of
the nature-culture dialectic that so profoundly impacted Heian culture and that was central to cults such as those which evolved in the
Gion and Kitano multiplexes. This, then, was the main reason why
the ritual of Shijoko was performed for the protection of the state, in
contradistinction to the ritual of Daigensui, which was geared toward military affairs from the outset.
Doctrinally, Ennin held the view that the esoteric tradition of
Shingon was equal to the exoteric tradition of Tendai, and that theory and praxis were also equal. But his successor Enchin caused an
important shift in the Tendai tradition by insisting that the esoteric
tradition was superior and that praxis was far more important than
theory. Thus, from Enchin's time on, esoteric practices and rituals
became the chief concern of the Tendai tradition. This developed to
such an extent, in fact, that Annen (b. 841?) claimed that Tendai es36 See, for instance, Hamada Takashi, Zuzb (Tokyo: Shibundo, 1969), p. 39,fig.54.
37 Hizen no kuni itsubun, in Akimoto Kichiro, ed., Fudoki (Tokyo: Iwanami shoten, 1958), p.
5*538 See Allan Grapard, "Japan's Ignored Cultural Revolution: the Separation of ShintoBuddhist Divinities and a Case-study: Tonomine," History of Religions 23, 3 (1984): 240-65.
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
541
otericism should be called Shingon. From that point on the court
looked to the shrine-temple multiplexes it sponsored for Taimitsu
and Tomitsu rituals; by the tenth century it was also asking for help
from thaumaturgists who had trained in mountains in order to gain
supernatural forces and who were reputed for the efficacy of their
rites and incantations. One began to see an increasing number of
mountain ascetics (yamabushi) at court.
The privatization of ritual activity
The position of the emperor declined while the power of the Fujiwara house grew; by the tenth century, as government came under
the control of Fujiwara regents, rituals that had been performed in
the name of the emperor came to be requested by individual aristocrats for their own welfare. The government forbade, after the exile
of Sugawara no Michizane in 901, the performance of such rituals for
private purposes. But the practice resurfaced with increased vigor, in
connection, perhaps, with the disintegration of the social system
based on the statutes, and the emergence of private concern and individual competition within given lineages. This trend toward individualization of cultic and political concerns accelerated throughout
the rest of the Heian era and was partly responsible for the acceptance of Pure Land beliefs. Indeed, most rites performed on a private
basis by aristocrats in the second half of the Heian period concerned
their private fate; eschatological concerns increased among the elite,
and private temples that would take care of individuals after their
death (bodaiji) were built at a fast pace. Other aristocrats, alienated
by the growing power of the Fujiwara regency, performed rituals to
solve public problems through private observances, while the leaders
of the Fujiwara house sponsored extravagant renditions of the Pure
Land in their private temples. This background, while providing a
fertile soil for the development of Pure Land beliefs, fostered increased interest in magical and ritual practices that would protect the
individual.39
Indeed, it came to be widely accepted that specific forces determined the fate of an individual, both from within and without as, for
example, in astrology. Therefore, aristocrats paid great attention to
the decimal and duodecimal signs corresponding to the date of their
39 See a wide-ranging discussion of these issues in Miyai Yoshio, Jodai no shimbutsu shugo to
Jbdo-hyo (Tokyo: Seiko shobo, 1980).
Cambridge Histories Online © Cambridge University Press, 2008
542
RELIGIOUS PRACTICES
birth, and believed that by performing rites of devotion to planets
and stars related to those signs, they might attract benefits of a thisworldly character (genze riyaku) and thus engineer a better existential
situation. The study of those rites impresses upon the observer the
centrality of systematic combinations between "Shinto," Buddhism,
yin-yang philosophy and practices, and Taoism. All these aspects
merged in a rather rigorous manner and in a vast coherence in which
one can read the main concerns of the people of the time, but they
also resulted in an impressive network of limitations and censorship
of the body and mind, which came to be increasingly ritualized.
The most widely performed rituals at an individual level for the
aristocrats were the Fugen emmyo-kd, dedicated to the bodhisattva
Samantabhadra to request an enhanced longevity,40 and the Fudohd, ritual of Acala, performed in a wide variety of contexts.41 Tendai
esotericism performed the Samantabhadra ritual in order to remove
the fear of death; this ritual was originally reserved for the emperor
(as a kind of chinkonsai) but, from 950 on, it was performed for aristocrats who were considered to represent the imperial state. Indeed,
the first aristocrat for whom this ritual was performed was no less
than Fujiwara no Morosuke, then chieftain of the Fujiwara house,
whose Testamentary Admonitions were quoted at the beginning of
this chapter.
The Fudo-ho, less complicated in scale, was even more widely appropriated by the nobility since it required less lengthy preparations.
It seems that Fujiwara no Tokihira, Mototsune's son, was the first to
request its performance to cure a personal disease, and then to ensure the safe birth of one of his children in 903.
Yet another rite of importance throughout the Heian period at
the state level as well as at the personal level was the Daiitoku-ho,
dedicated to the powerful and frighteningYamantaka.42 This important divinity of the esoteric pantheon, one of the "Five Powerful
Ones" (godairiki) in a wrathful appearance (funnu), was a manifestation of the Bodhisattva of Wisdom, Manjusrl. Symbolized by the
number six (it has six arms and six feet), it was conceived to be
more powerful than Yama (Emma, in Japanese), the King of Hells.
40 This ritual was based on the Kongo jumyo darani-kyo, T, vol. 20, no. 1134 [b], and is described in the Kongo jumyo damni-kyoho, T, vol. 20, no. 1134 [a].
41 There were several forms of Fudo rituals, variously emphasized in Shingon and Tendai lineages. The main texts on the topic are contained in T, vol. 21.
42 Daiitoku-ho. There are several texts describing varied rites dedicated to Daiitoku; of central
importance are those contained in T, vol. 21, no. 1214 and no. 1216. See Robert Duquenne,
"Daiitoku my56j" Hbbdgirin 6 (1983): 652-70.
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
543
In the esoteric tradition each of the Five Powerful Ones was
thought to protect one of the five directions; Daiitoku protected the
west, long considered to be a dangerous direction (the whole complex of Mount Atago, just west of the capital, had been erected to
ward off evil from that direction). Hence, he came to be seen not
only as the manifestation of ManjusrI but also of Amitabha/Amitayus. Like all major deities of the esoteric pantheon, Daiitoku was
represented in a mandala for ritual purposes, either in painting or
sculpture, in which each iconographic detail was related to a doctrinal aspect and was therefore subjected to great attention. In
terms of esoteric doctrine, the Buddha possessed five wisdoms conceived of under three aspects: Self-nature, symbolized by a buddha;
Correct Law, symbolized by a bodhisattva; and Injunction, symbolized by a wrathful manifestation of either, and whose function was
to forcefully inculcate the teachings to reticent beings, and to transform relative cognitions into a corresponding wisdom. Thus, the
five wisdoms of the Buddha were symbolized by five buddhas, five
bodhisattvas, and five wrathful manifestations, each corresponding
to the other and represented in spatially oriented mandalas. This
was propagated in scholastic milieus of Heian esoteric Buddhism by
a new version of the apocryphal Ninno-kyo, and by the ideology sustaining the protection of the state through rituals involving spatially
oriented mandalas that came to represent the land under its control. The most common rite seems to have been that of fire oblation
{goma) accomplished over a triangular altar (abisharoka), the summit of which was occupied by Daiitoku, and faced north. The officiant, dressed in black and positioned to the south, drew an image
or inscribed on the ash at the bottom of the hearth the name of the
agent to be vanquished by the ritual.
The most important texts describing those rituals date back to the
end of the Heian period or the beginning of the Kamakura period.
Documents of the time show that these rites were used for different
purposes: to separate a wife from her husband so that she became
available for the amorous ardor of a prince or emperor, or to cause
two persons to fall in love. More generally, however, the Daiitoku
rites were used to bring success in military endeavors.
The ritual world of shrines
A salient feature of the world of shrines during the Heian period was
the reformulation and codification of ritual practices under the aegis
Cambridge Histories Online © Cambridge University Press, 2008
544
RELIGIOUS PRACTICES
of the government and the dominant sacerdotal lineages. This reformulation appears in the Procedures of the Jogan Era {Jogan gishiki)
promulgated in 869-71, and in the Procedures of the Engi Era {Engi
shiki), completed in 927 but promulgated in 967, which unified the
preceding procedures. These documents indicate that the "worship
of the kami of Heaven and Earth" was redefined during the Heian
period according to specific notions held by the government. Indeed, although little information is available on ritual conduct before the Heian period, it is probable, and in some cases evident, that
most of the rituals classified and determined in great detail in those
Procedures were substantially different in the Nara period.
It must be noted that most shrines that came to be supported by
the imperial state in Heian had been either maintained or built privately by the major social lineages and clans (uji), even though some
(if not most) were subtle indicators of those lineages' claim to a privileged relationship to the imperial lineage and thereby to the state.
What changed during the Heian period was that the government,
even while recognizing the private character of those shrines, began
to acknowledge diem as its own symbols, and therefore made offerings of food and various products and in some cases set up government representatives to oversee the architectural development of
shrines with funds given by the state. In the case of the Twenty-two
Shrines, such as the Kasuga Shrine dedicated to the ancestral and
tutelary kami of the Fujiwara house, the imperial lineage would send
messengers at the time of major rites; these messengers were
charged with reading a supplique in the name of the emperor to the
kami and making predetermined offerings. The promulgation of
these procedures coincided in time with the rapprochement of Buddhist temples to the newly enlarged shrines. Several hypotheses can
be offered to interpret the meaning of those developments.
First, it is evident that those rules of the Engi shiki which concerned ritual and liturgy were applicable not only to the Department
of Shrines but also to the conduct of ritual at court, and that they
had a triple interrelated character: they attempted to determine expenditures (of money and offerings) in relation to certain practices
inscribed in space (the protocol of ritual and alliances with the periphery), and in time (the determination of ritual performance according to seasonal rhythm and within a complex framework of control of time). Money, time, and space are therefore three of the
fundamental categories from which ritual practices in shrines, temples, and at court, might be approached.
Cambridge Histories Online © Cambridge University Press, 2008
RITUALIZED ACTIVITIES
545
Second, the formulation of ritual conduct in the Engi shiki was
parallel to the establishment of the system of Twenty-two Shrines
described earlier in this chapter, and therefore represents a concerted effort to unify practices in shrines, but within the context of
shrines and temples. While this formulation of ritual represents concerns of the government, it perhaps also expresses concern on the
part of sacerdotal lineages worried by the wide acceptance of Buddhist ritual. One notes in those codes, for example, that Buddhist
prelates were no longer allowed at enthronement rituals.
Third, the developments just outlined determined aspects of the
relationships between the worlds of shrines and temples: clear demarcations prevented chaotic syncretism from occurring, and called
instead for combinations whose elements retained some of their integrity. Had it been otherwise, it is probable that one discourse,
grounded either in shrines or in temples, would have become dominant and would have reduced the other to a meaningless amorphousness. The discourse issued from temples attempted to be dominant, but it never succeeded in replacing the realm of the shrines,
and even took second place in later periods. The world of combinations exhibited by shrine-temple multiplexes during the Heian period became a central part of the cultic world during the medieval
period.
Yet another element must be mentioned at this point: the fact that
the temples related to the Twenty-two Shrines were closely connected to the leading houses of the Heian period means that Buddhism was restrained and confined within those limits of sociopolitical order. In any other situation, freethinkers and heterodox figures
would have taken Buddhism to other parts of Japanese society, with
radically different messages - which is precisely what happened toward the end of the Heian period.
Expenditures, time, and space in ritual
Even a cursory reading of the Engi shiki impresses upon the reader
the fundamental features of ritual in shrines during the Heian period: the performance of codified behavior on the part of a strictly
defined social body within the confines of particular spatial arrangements at strictly defined times of the year, month, and day, and focusing on a specific relationship to a divine entity, expressed by purification and by various offerings among which the offering and
sharing of food took precedence.
Cambridge Histories Online © Cambridge University Press, 2008
546
RELIGIOUS PRACTICES
The literature of the Heian period is striking in the total absence
of the word "Shinto." This does not mean that cultic aspects of what
came later to be called Shinto were not a significant part of the cultic and political world of the times; it simply means that the cultic
world of Heian was not conceived of according to the tentative distinctions made today on the basis of more recent history. What we
find instead is a marked emphasis on sites of cult which were, for the
duration of the period and beyond, a conglomerate of shrines and
temples, and we also find an emphasis on the observance of protocol both at court and in those multiplexes. Thus, the first striking aspect of cultic life in Heian is its relation to place within the context
of combinations between imported and autochthonous cultic systems, all subsumed under a sociopolitical ideology.
A second striking characteristic concerns the strict regulation of
expenditures by the government at the time of ritual observances;
this means that private houses were probably competing in prestige
through outlays of wealth ranging from amounts of rice offerings to
ceremonial clothing and horses. This competition, which had been
brought under control by the stipulations of the Engi shiki, returned
to the fore when the government began to cede land to centers of
cult as a reward for success in rituals. Furthermore, since many of
these land holdings were partly free from taxation, a number of people began to grant parts of their land to centers of cult for "spiritual
protection," in order to avoid taxation. That land was granted to specific shrines and/or temples, not to "Shinto" or to "Buddhism."
Hence, ritual came to be ever more connected to economics.
A third notable characteristic of ritual in Heian was the emphasis
on hierarchy and code and on the observance of protocol. This emphasis can be understood from two related perspectives: a sociological concern and a concern with purity and pollution. One way to
focus on the relation between these two concerns is to observe that
many of the multiplexes of the Heian period were dedicated to ancestor worship and to the cult of ancestral kami (associated yet different phenomena). In the temples, ancestors were worshiped in part
through the performance of rites of penitence which, as seen earlier,
were for the most part rites of purification of the sense organs; and
in the shrines, ancestral kami were worshiped in the context of a
triple type of ritual purification (pharae, ethical purification; omisogi,
physical purification; and the use of special individuals related to the
observance of taboo, mono-imi), all in the context of ritually prepared food offerings involving a high level of precautionary measures
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
547
against any type of pollution. At the sociological level, it must be emphasized that many of those rituals were used to reinforce social
bonds within clear contexts of time and space, and within frameworks of exclusion and inclusion: included were the social group
whose personal kami were worshiped, as well as the social group
charged with the governance of shrines and ritual observances before, during, and after the main rites. Excluded was any social group
that, by definition, did not worship ancestors or ancestral kami of
others. That is one of the reasons why cultic practices in Heian times
were greatly hierarchized and related to specific places. In the case of
the neighboring temples, although Buddhism was not geared toward
the exclusion of any member of the social body (with the significant
exception of the almost systematic exclusion of women), in fact the
higher echelons of the priestly hierarchy came to be reserved for
leading ecclesiasts born from those noble lineages which had dedicated or created the temples in the first place. Thus, principles of exclusion tended to remain important aspects of the social construct of
"monastic" communities, ideology notwithstanding.
A fourth and final characteristic of cultic life in Heian was the emphasis on the performance of dances, songs, and music at the time
of grand rites in shrines, and later in temples as well. Much of the
culture of the Heian period was transmitted to the present day
through the medium of those rites, which show that cult and culture
were intrinsically related phenomena, and that both were related to
performance and grounded in aesthetic concerns.43
DEALING WITH THE FORCES OF NATURE
The Bureau of Yin and Yang COmmyoryo)
The Bureau ofYin and Yang was reorganized by promulgation of the
Y6ro code in 718 by Fujiwara no Fuhito and others.44 Although it
was conceived on the model of its counterpart established by the
T'ang dynasty, it was different in scope and partly different in structure. The Bureau was governed by a director under whom four ad43 For discussions of these issues, seeYamagami Izumo, Nihon geino no kigen (Tokyo: Yamato
shobo, 1977).
44 The following discussion is based in part on Murayama Shuichi, Nihon ommyodo-shi sosetsu
(Tokyo: Hanawa shobo, 1981). The reader may also refer to Sait5 Tsutomu, Ocho-jidai no
ommyodo (Tokyo: Sogensha, 1947), and Felicia Bock, Classical Learning and Taoist Practices
in Early Japan (Center for Asian Studies Occasional Paper, Tucson: University of Arizona
Press, 1985).
Cambridge Histories Online © Cambridge University Press, 2008
548
RELIGIOUS PRACTICES
ministrators oversaw the activities of specialists in four major fields.
The director himself was responsible for astronomy and the calendar, while he also observed and interpreted climatic phenomena. In
the advent of any abnormalities (z) in the natural cycles, he reported
secretly to the emperor. Under the four administrative aides to the
director were six Masters ofYin and Yang (ommydji) who specialized
in divination (uranai) and in geomancy (fusui). The four major
fields of investigation were, in descending order of importance: Yin
and Yang, Calendar, Astronomy, and Time. Each of these fields was
under the supervision of a scholar {hakase), who taught his skills to
ten students in each of the first three fields, while two scholars
taught twenty students in the field of time measurement. These two
scholars were helped by twenty-three assistants who, in all probability, worked in shifts. Murayama Shuichi notes that by the year 720
the Bureau of Yin and Yang employed a total of eighty-nine officers
instructed by four scholars who were all either of Chinese or Korean
origin, and who were all in their fifties. By comparison, the same bureau in China employed a grand total of 1,413 people whose functions were roughly divided between astronomy and the calendar on
the one hand, and divination on the other. The preeminence of these
fields was reversed in Japan. A final distinction between the Chinese
and Japanese systems must be added: divination by turtle shell was
not assigned to the Bureau of Yin and Yang in Japan, but to the Department of Shrines, where it was performed by the Nakatomi and
Urabe sacerdotal lineages.
Many of the practices related to yin and yang were of Taoist origin
and were associated with medicine. The Chinese organs of government included an office of medicine, part of which specialized in the
"magical" cure of disease; the same was true of Japan in the Nara period, for there was a Bureau of Medicine, parts of which were staffed
by one hakase ("doctor"), two masters, and six disciples, all specializing in exorcism. The Bureau continued to exist during the early Heian
period, but did not develop because of the combined influences of
Esoteric Buddhism (within which there evolved specific practices of
medicine related to magic and exorcism), and of the "Way of Lodgings and Planetoids" (Sukuyddo) and mountain asceticism (Shugendo), in which various strains of yin and yang practices and notions
(ommyodo) evolved. Furthermore, the exorcists of the Bureau of Medicine were originally almost all of foreign descent (kikajin): Hirotari,
the infamous disciple of En no Gyoja (the putative founder of
Shugendo), was of Korean origin and a member of that Bureau.
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
549
The Nara aristocracy and common people were a fertile ground
for the growth of magical practices; the government issued several
decrees prohibiting them during the course of the eighth century. It
is possible that the use of magic was thought by the authorities to be
potentially nefarious to the ideals of the state if it was engaged in by
enemies of the state; for example, the case of the Buddhist prelate
D5ky6, whose claim to the throne during the Nara period was based
on his access to oracles from Hachiman, suggests that some ritual
activities were viewed as potent forces acting on the symbols upon
which the state rested. Another reason to prohibit such practices was
medicine: the government claimed that magic "treatment" was ineffective and led people astray with false hopes. By and large, however,
the treatment of disease remained the favorite domain of mountain
anchorites (yamabushi), and other thaumaturgists.
The case might be offered that Buddhism presented itself as a social body mainly interested in doctrinal exchange within the institutional context of a school of thought. But reality was otherwise, for
there was a need for dialectical exchange between society and the
monastic community; in this exchange, lay society was concerned
more with magic and cure of disease than it was with metaphysical
speculation, and the monastic community did need some popular
support in order to attract the masses to its teachings. And so it was
that Buddhist monks, especially those who had been to China, used
magic and charms for a variety of purposes ranging from curing disease to fighting off zealous government officials checking on their
credentials. But, all in all, the main interest of the people and of the
court was the cure of disease; this remained a central element of
many cults throughout history.
In yet another realm of practical concern, Buddhists used geomantic knowledge acquired in China to decide on sites for the construction of temples, or used their knowledge of astronomy and astrology to determine the proper placement of the body of ecclesiasts
and aristocrats in space and time. Many practices related to those issues were not looked upon favorably by the authorities, perhaps because some of them were too closely related to Taoist notions the
government did not condone. This was true also of rites related to
the Polar Star, which focused on Ursa Major; these rites in China
had long involved sexual promiscuity. The government prohibited
sacrifices of bulls in rites dedicated to fertility. But some prohibitions
were ineffective, and many rites were soon coopted by Esoteric Buddhism, which developed combinations between the Polar Star cult
Cambridge Histories Online © Cambridge University Press, 2008
550
RELIGIOUS PRACTICES
and a cult to the bodhisattva called in Japanese Myoken, or Sonsho6. The main scripture on which this ritual was based was the Hokuto
shichisho emmyo-kyo, which promised accrued longevity to those who
requested its performance.45 Yet another example ofTaoist practices
which developed in Esoteric Buddhism is the Koshin belief, also related to longevity, which involved spending a sleepless night every
sixty days on the day of the Monkey in order to prevent demonic inhabitants of the body from leaving it to report to the heavenly bureaucracy on one's evil deeds. The name of the divinity in Esoteric
Buddhism was Skomen kongo. The cult developed in Tendai esoteric
circles and quickly spread to the populace. (It is still extant today.)46
Practices based on yin and yang notions infiltrated the ritual world
of shrines as well, particularly in the evolution of practices of purification. The Nakatomi, Urabe, Abe, and Kamo sacerdotal lineages
were responsible for these rituals intended to protect the body, the
imperial city, and the country against epidemics. An excellent example is provided by the michi-ae rites, of which we have some record
in the Engi shiki.47 These rites were performed at crossroads and at
the various entrances to the capital at the time of epidemics, but in
particular also at the time of the approach of foreign missions. This
can be explained by the belief that disease came from afar along various paths, and that visitors from foreign countries were apt to bring
it, as history amply demonstrated. These michi-ae rites consisted of
offering food and of formulating incantations, probably of Chinese
origin. In this manner a large number of notions, concepts, and
practices belonging to the world of yin and yang in China found their
way into Buddhism and into the ritual world of shrines.
For reasons that are yet to be determined, the early Heian government did not support the activities of the Bureau of Yin and
Yang; it was disbanded in 820. Instead, the government relied on
both Shingon and Tendai traditions of esotericism to perform a
number of rites that in earlier years would have been the domain of
the ommyoryo. At the time of the Jowa Incident of 842, in which the
Fujiwara house began to assert its centrality in decisions concerning
the choice of emperors, the government requested the performance
of rituals by one hundred Buddhist monks within the imperial
45 Hokuto shichisho emmyo-kyo, T, vol. 21, no. 1307.
46 On this topic, see Hirano Minoru, Koshin shinko (Tokyo: Kadokawa shoten, 1969). See also
Tsubo Noritada: Koshin shinko (Tokyo: Yamakawa, 1956), and Koshin shinko no kenkyu
(Tokyo: Nihon Gakujutsu shinkokai, 1961).
47 See Felicia Bock, Engi-shiki, vol. 1, p. 86, and vol. 2, pp. 90-92.
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
551
palace, and asked them to read the Sutra of the Buddha of Medicine, to perform rites dedicated to that Buddha, and further requested readings of the Daihannya-kyo within the imperial palace.
The Jowa Incident was the first in a series of coups engineered over
a period of one hundred years by the Fujiwara house to ensure that
emperors were chosen exclusively among children born of a union
between a reigning emperor and a woman of Fujiwara birth; this enabled the Fujiwara ministers to govern in the name of their grandchildren and thus control the imperial lineage. Cultic activity was an
essential feature of their claims to legitimacy in that context.
The frequency with which both ominous and auspicious natural
events occurred during the Heian period indicates most clearly the
mood of the times. From Emperor Heizei (r. 806-9) t o Emperor GoSanjo (r. 1068-73), a s many as 1,686 natural occurrences of an ominous nature were officially recorded: 653 earthquakes, 134 fires, 89
instances of damage to the crops, 91 outbreaks of epidemics, 356
calamitous occurrences of a heavenly nature (volcanic eruptions,
comets, eclipses, thunder in clear skies), and 367 ghostly events. In
contradistinction, the same documents record a grand total of only
185 auspicious occurrences during the same period. The reign of
Emperor Yozei (876-84) seems to have been plagued by the largest
number of calamities (as many as 133 in the course of seven years),
and to have been visited by the largest number of auspicious signs as
well: as many as 31, the largest number for any single reign. This period corresponds exactly to the rise to ascendancy of the Fujiwara
house and to the regency of Fujiwara no Mototsune, a fact which reveals that political worries in Heian Japan were as if symbolically
manifested in an increased consciousness of the course of nature in
relation to human affairs, well within the framework of the nature-culture dialectic mentioned earlier in this discussion. The notable increase in records of natural occurrences interpreted as heavenly warnings or blessings is related to the evolution of the goryd
belief system (discussed later), and might be viewed as a manifestation of popular criticism of governmental policies and of the internecine struggles that took place within the aristocracy. It was just
a few years after the reign of Emperor Yozei that the exile of Sugawara no Michizane caused the people to interpret fires in the imperial palace as heavenly reprimands by the angry spirit of the
wronged statesman.
In other words, when the government began to applaud the
miraculous appearance of a white turtle in 715, 724, and 770, and
Cambridge Histories Online © Cambridge University Press, 2008
552
RELIGIOUS PRACTICES
interpreted it to be a manifestation of heavenly pleasure with its
rule, little did it know that a few years later the reasoning would be
reversed, and that people would pay great attention to ominous occurrences in order to criticize the government.48 In both cases, the
court turned to liturgical and ritual behavior: when an auspicious
sign appeared, emissaries were sent to make announcements to the
major shrine-temple multiplexes of the realm and to as many as
twelve imperial mausolea; and when ominous signs appeared, rituals to placate the wrath of cosmic powers were requested from the
same multiplexes. This may have been one reason why thaumaturgists came to play a role in Fujiwara politics, and it shows that policy making was rarely if ever separated from cultic concerns: the
more unstable the political world, the more active the multiplexes.
As time passed and political concerns grew, ritual activity increased
accordingly at all levels of society, and cultic procedure became
complex and cumbersome. Such was the price to pay for grounding
a rule in symbol and rite.
The reformulation of ritual and liturgy exemplified in the Jbgan
gishiki and in the Engi shiki compilations marked a definite impact
of yin and yang practices on ritual procedures within the world of
shrines and temples. Precisely at the same time, the practices of directional taboos originated,49 concurrently with the practice of devotional respect to the four directions (shihdhai), which was performed by the emperor for the first time in 860.5°
The practices related to directional taboos {kata-imi) concerned
travel or, in other words, movements of the human body through
space. It was thought that certain celestial forces and bodies, whose
movement was carefully followed, determined spatial zones that
should not be crossed by people at certain times and in certain directions. Thus, when aristocrats planned to travel (or build a house,
or break the earth for other purposes), they consulted specialists
who determined whether the time and space correlations were favorable or ominous. Such restrictions on the movement of the body
came to be so pervasive that palliative measures were devised, to
which the name of "directional alternatives" {kata-tagae) was given.
48 Murayama Shuichi, Nihon ommyodo-shi sbsetsu, pp. 5o-6g.
49 See Bernard Frank, Kata-imi et kata-tagae: Etude sur les interdits de direction a I'epoque Heian,
Bulletin de la Maison Franco-Japonaise, Nouvelle Serie, Tome 5, 2-4 (Tokyo and Paris,
1958)50 See Fukuyama Toshio, Chusei no jinja kenchiku (Tokyo: Shibundo, 1977), pp. 95-96:
"Chiiseijin no shinko to jinja," in which the author discusses the history and conduct of the
shihdhai ritual.
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
553
This involved moving for one day in a direction other than planned,
spending the night there, and reaching one's intended destination
the following day. As time passed, aristocrats transformed these cumbersome practices into a kind of game that involved plans to spend a
night at the home of some relative or acquaintance in order to be entertained. The system gradually degenerated and was more or less
abandoned toward the end of the Heian period. However, the notion
of the body as a cosmic entity submitted to powers in space and time
remained of importance in various Japanese cultic traditions, and a
number of practices associated with the kata-imi phenomenon have
survived up to the present.
Celestial irregularities
The world of yin and yang beliefs and practices evolved as part and
parcel of Chinese studies during the Nara and Heian periods and
centered on the definition of the position of man within the cosmos.
It provided a framework of interpretation according to which the
natural world was thought to impact the destiny of man, and, conversely, according to which human activity was thought to impact
the processes of nature. In other words, man was thought to be an
agent of cosmic (but not historical) change, at the same time he was
considered to be under strong cosmic influences, as is evidenced in
astrology. A primary aspect of those beliefs and attitudes is found in
the Sukuyodo, "Way of Lodgings and Planetoids," a complex system
resulting from a crossing of astronomy with astrology that had occurred in India, but that had been enriched in China before it
reached Japan.51 Almost all Indian science reaching Japan was
brought by Buddhist monks, so that astronomy became a primarily
Buddhist matter. Monks were not only specialists of metaphysics
and philosophy, they also tended to be healers, thaumaturgists, diviners, and astrologers. During the Heian period almost all astronomy and astrology developments that had taken place in various milieus in China were transmitted to Japan by monks of Shingon and
Tendai Esoteric Buddhism, so that that form of Buddhism was heavily laden with notions and rituals that belonged originally either to
51 little has been published on the topic ofSukuyodo. There is a popular treatment of the subject by Yoshida Mitsukuni, Hoshi no shukyd (Kyoto: Tankosha, 1970). A more detailed
analysis of the topic in direct reference to esotericism is provided by Manabe Shunsho, Kobijutsu, 35 (1971): 1-48. Other indications can be found in Murayama, Nihon ommyodo-shi
sosetsu, and Hayami, Heian jidai kizoku-shakai to bukkyb.
Cambridge Histories Online © Cambridge University Press, 2008
554
RELIGIOUS PRACTICES
Indian ritual science or to Chinese Taoist and other practices. The
Japanese did not use the term "Taoism"; instead, Taoist practices became part and parcel of the praxis of Esoteric Buddhism and
Shugendo, and of the ritual world of shrines.52 This is evident in the
history of the practice of astrology and related matters such as the
system of beliefs surrounding epidemics, the realms of combinations
between shrine and temple deities, and the cults to vengeful spirits
and ghosts.
From about the middle of the ninth century documents indicate
a formidable interest in astrology on the part of the government and
aristocrats. Official documents also indicate an alarming increase of
strange movements on the part of planets and constellations. Such
occurrences were seen as signs of impending calamities at the level
of the state; rituals to counteract these events were therefore in great
demand and were proposed mostly by the Tendai and Shingon lineages. First, changes of era names were used to deflect those celestial threats.
It was thought that the name of a year era (nengo) was closely associated with the character of life during that era. If a catastrophic
event occurred, one of the ways of dealing with it was symbolic and
entailed changing the name of the era. There were two types of
changes: the first one was positive and occurred when favorable portents appeared in the world of nature; the other was negative and occurred whenever a foreboding calamity took place. The latter case of
name change was a protective or defensive measure aimed at appeasing the worries of the leading politicians, while reinforcing their
notion that the power of naming was effective. The rituals protected
the state, which means that they kept the Fujiwara regents in control; but it also meant that the thaumaturgists in some way con52 On the place of Taoism in Japan, see Murayama, Nihon ommyddd-shi sdsetsu; Saito, Ochdjidai no ommyodo; Shimode Sekiyo, Ddkyo to Nihonjin (Tokyo: Kddansha, 1975); Fukunaga
Mitsuji, Ddkyo to Nihon bunka (Kyoto: Jimbun shoin, 1982); Shimode Sekiyo, Nihon kodai
nojingi to Ddkyo (Tokyo: Yoshikawa kobunkan, 1972); Fukunaga Mitsuji et al., eds., Ddkyo
to kodai no tenndsei (Tokyo: Tokuma, 1978); and Felicia Bock, Classical Learning and Taoist
Practices in Early Japan. The reader may also refer to Osabe Kazuo, Ichigyd zenji no kenkyu
(Kobe: Kobe Shoka Daigaku, 1963).
On the topic of Shugendo, see Murayama Shuichi, Yamabushi no rekishi (Tokyo: Hanawa,
i97o);WakamoriTaro, Shugendo-shi kenkyu (Tokyo: Toyo bunko, 197a); andWakamoriTaro
et al., eds., Sangaku shiikyd-shi kenkyu sdsho, 18 vols. (Tokyo: Meicho shuppan, 1977). See
also Gaston Renondeau, Le Shugendo: Histoire, Doctrines et Rites des Anachoreus Dits Yamabushi (Paris: Imprimerie Nationale, 1965); Hartmut O. Rotermund, DieYamabushi-Aspekte
Ihres Glaubens, Lebens und Ihrer Sozialen Funktion im Japanischen Mittelalter (Hamburg:
Kommissionsverlag Cram, De Gruyter, 1968); and Byron Earhart, A Religious Study of the
Haguro Sect of Shugendo (Tokyo: Sophia University, 1970).
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
555
trolled the aristocrats. These changes in names of eras went hand in
hand with the performance of rituals such as the Daigen(sui)-ho, the
Daiitoku-hd, the Shitenrio-ho, the Shijoko-ho, the Kujaku-o-ho, and the
Shou-kyo-ho, in which Shingon andTendai thaumaturgists competed
for rewards. The competition was fierce and the claims of success
important, for success meant the granting of land, limelight for the
ritualists, and the attention of the court. The stakes grew throughout
the Heian period, especially as political and military unrest shook
the provinces and moved closer to the capital. Aristocrats responded
to an upheaval not only by sending militias to quell it, but also by
asking the multiplexes to perform rituals to ensure the victory of the
court and pray for an end to such upheavals. The Engi shiki states
that the Daigen(sui)-ho must be performed every year during the
second week of the first month, "to protect from armed upheavals."
This was the case at the time of the rebellion of Taira no Masakado
in 939.53 In 941, the Daiitoku-ho was performed by Gikai (the abbot
of Tendai) on Mount Hiei when it was learned that Korean pirates
were sighted off Kyushu; at the same time, the ajari Ensho performed the Fudd-hd at the Hosshoji, one of the great Fujiwara temples in Kyoto. Some of those rites bear traces of yin and yang practices; all involved some combined aspects of Buddhist, Taoist, or
"native" rituals.
An early response to celestial irregularities had been the reading of
scriptures, which was the main original palliative in Japan. From the
tenth century onward, thaumaturgists were most sought after. They
performed what is known as tenku (service dedicated to the devas), or
seiku (shoku, service for the stars). Such rites were not held exclusively
for the welfare of the state, since the scriptures on which they were
based (the Kujaku-kyo and Hokuto shichisho emmyo-kyo) recom-
mended that these rites be performed on a personal basis as well.
Among the many stellar rituals, by far the most popular throughout East Asia were those dedicated to the Polar Star. Called in
Japanese hokuto, Ursa Major was the focus of rites that the government prohibited as early as 796 for various reasons. But these rites
came again to the fore during the Heian period, so much so that almost all aristocrats performed them at one time or another. The first
performance of a major stellar ritual for the state on the part of Buddhist prelates was by Gikai in 945, at the time of the passage of a
comet. Such rites were performed regularly thereafter upon the pas53 Murayama Shiiichi, Nihon omnryddo-shi sosetsu, pp. 102—10.
Cambridge Histories Online © Cambridge University Press, 2008
556
RELIGIOUS PRACTICES
sage of comets or the determination of eclipses, and with increasing
frequency during the eleventh century.
Another major practice of Buddhist prelates and, thereafter, of the
nobility, was related to what is called hommydshd. The term hommyo
(Chinese: p'eng-ming, meaning personal fate) was drawn fromTaoist
terminology and was assimilated by Buddhism in China before
being introduced to Japan sometime during the early Heian period,
where it designated the notion that certain celestial bodies (sho) held
sway over the destiny of individuals. There was some question as to
exactly what was to be designated as the hommyo celestial body of a
person, and whether some days were to be considered as times when
specific rites should be performed. There was also some question as
to whether it was to be determined in relation to the decimal and
duodecimal signs of the day of birth or of the year of birth. Originally, the observance of such rites was limited to the emperor, but it
spread to all parts of society. Minamoto Counselors argued in 943
against the observance of the hommyo rite by the emperor, but were
overruled by Fujiwara noYoshifusa. Discussions concerning the day
or the year of birth as a deciding factor for the identification of the
hommydshd took place in 961: on the side of the day of birth was the
monk Hozo, and on the side of the year was the yin-yang specialist
Kamo noYasunori. The decision to follow Kamo noYasunori's theory was taken by NichizS, the famous Yoshino thaumaturgist. Which
celestial body could be chosen as hommydshd was based upon a document entitled Zokushd hihd, written by Yi-hsing, in which it was
stipulated that the decimal and duodecimal signs of the time of birth
were to be inscribed on a diagram representing the five planets:
Mars, Mercury, Jupiter, Venus, and Saturn. However, from the time
of Ninkai (953-1046) on, the planets were abandoned and replaced
by the seven stars composing Ursa Major, which were thought to be
hypostatic forms of the planets. Thus, the rites connected to one's
"planet" consisted in worshiping one of the stars of Ursa Major on
one's birthday. Such practices became quite popular and continued
in cultic circles for centuries.
Upon the occurrence of a baleful sign (yd), the Masters of Yin and
Yang (ommydji) decreed a number of taboos, restrictions, and restraints (kinki). Baleful signs came to be seen in each and every aspect of life; for instance, after a fire destroyed a large part of Kyoto,
wearing red clothes was forbidden. From 960 onward, the Masters
of Yin and Yang began to perform rites to placate calamitous forces;
these were known as kasai matsuri, "calamity rites," and were related
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
557
to the fire that had engulfed the imperial palace that year for the first
time since the creation of the capital. This fire caused great anxiety
among the aristocracy, and the government ordered that offerings be
made immediately to Ise and other shrines, and that the fire be reported to the mausolea of emperors Tenji, Kammu, and Daigo. The
following year the first kasai matsuri was held within the compounds
of the imperial palace. Although we do not have detailed information on the conduct of the rite, documents left by the Abe sacerdotal lineage indicate that the pillars of various buildings of the imperial palace were decorated with swords on which inscriptions were
engraved. These inscriptions show that Taoism was an aspect of the
protection of imperial buildings in Heian Japan. Some of the swords
had been offered by Korea: one was called "sword to protect from
enemies," and another was called "sword to protect the body." These
swords were inscribed with the symbols of the three kings and five
emperors of China, with the Polar Star of the south, emblems of
the dragon, magic charms used by the guardians of Hsi-wang Mu
(all these on the left side of the sword), while the right side of the
sword bore inscriptions of the five stars of the north pole, the seven
stars of Ursa Major, emblems of the tiger, and magic charms "of
Lao-tzu" to vanquish enemies. To these were added inscriptions asking for protection from the four directional emblems and for the realization of all wishes, protection from calamities, and achievement
of longevity. The document of Abe no Haruaki (Abe no Seimei)
written for the occasion indicates that these inscriptions had disappeared in the fire, and that new swords bearing them ought to be ordered. A further notice indicates that these swords were to be made
on the auspicious day of the Monkey, in typical Taoist fashion:
swords made on those days were more potent than others.
Thus we realize that within the imperial palace alone absolutely all
available forms of cultic practice were used: yin and yang practices,
Taoism, "Shinto" rituals colored by Korean shamanism, and the entire panoply of rituals offered by the Taimitsu and Tomitsu forms of
Esoteric Buddhism, be they performed by the main leaders of various branches of Tendai and Shingon or by powerful thaumaturgists
who had gained their supernatural powers in the wilderness of
mountains.
Some of these ritual forms were also present among the general
populace, and may even have originated there and gradually pervaded aristocratic consciousness, as is suggested by an analysis of
oracular practice.
Cambridge Histories Online © Cambridge University Press, 2008
558
RELIGIOUS PRACTICES
Oracular practice
Oracular practice was certainly one of the prevalent aspects of cultic
life during the Nara period, but it took on even more importance
during the Heian period at all levels of society. Entire cults that developed in some of the Twenty-two Shrine-temple multiplexes described earlier, such as the Hachiman, Kitano, and Gion cults,
rested upon the practice of asking specific deities for their will, by
divination, or of being granted an oracle (takusen) in which the deity
gave its indications. The Urabe sacerdotal lineage specialized in plastromancy (divination on sea turtle shells), while the Nakatomi sacerdotal lineage specialized in scapulimancy (divination on deer
shoulder blades); both types of divination were regularly performed
in specific shrines, usually by request, and the results were transmitted to the emperor.
Oracular practice during the Heian period can be classified as
three types: the jinguji phenomenon discussed earlier in this chapter,
in which oracles played a central part, the goryo phenomenon, and
the onryo phenomenon. The most important jinguji in this respect was
the Mirokuji of the Usa Hachiman Shrine in Kyushu; the Hachiman
cult originated as an oracular cult in which female officiants, under
possession, uttered sounds that were then interpreted by a male officiant called saniwa. The Hachiman cult may have been in the early
eighth century no more than a local oracular cult, but it came to the
attention of the court in 764, when the HossS monk Dokyo claimed
that an Usa Hachiman oracle stipulated that he should become hod,
"dharma-king," of Japan. Shocked by the announcement, the court
sent one of its representatives, Wake no Kiyomaro, to Usa, where he
received an oracle declaring that Dokyd was an impostor and that
only members of the imperial lineage were entitled to rule the realm.
Dokyo was subsequently exiled, and Hachiman was enshrined in the
Tamuke Hachiman Shrine as the chinju (protector) of theTodaiji imperial temple in Nara. A few years later, Hachiman came to be regarded as an incarnation of "Emperor" Ojin, and his cult was sponsored directly by the state; indeed, a major multiplex dedicated to
Hachiman was erected in 859 at Iwashimizu, south of Heian-kyo, in
honor of Emperor Seiwa's enthronement. Hachiman, however, was
conceived of as a bodhisattva, not as a kami, and this might explain
why after the year 839 oracles were revealed, not to female officiants
of Hachiman shrines as they used to be, but to male Buddhist monks
who gradually took over their oracular functions. The Hachiman orCambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
559
acles became in the fourteenth century the object of a compilation
and interpretation by Jin'un, a Buddhist prelate born in an Usa
Shrine sacerdotal house, that is, the Hachiman Usa-gu gotakusenshu,
"Compendium of Usa Hachiman Oracles," one of the most important medieval documents of the Hachiman cult and a key source for
understanding oracular practice.54 Oracular practice thus survived
predominantly among Buddhist prelates, but also among some
women who tended to be wives of yamabushi.
The goryo and onryo cultic systems
The term goryo originally denoted the spirit of a departed person,
but it came to be applied to the notion that either the spirit of departed aristocrats forced into unjust and politically motivated death,
or some cosmic power, were responsible for epidemics. The most famous case of a goryo was at the basis of a complex cult which originated under the name of Gozu tenno ("Bull-headed king of the
devas") at the Gion shrine-temple multiplex in 876. Festive rituals
aimed at pacifying those disruptive forces took place generally in
summer, a season in which epidemics were prone to occur. These
rituals were performed during daytime, in contrast to spring and autumn rituals which tended to be performed at night. They became
popular especially in the major cities, and included prominently the
performance of popular dances, the formation of impressive processions in which floats were lavishly decorated, and the performance
of popular songs and music, as early as the year 1013 in the case of
the famous Gion Festival of Kyoto.55
Another type of cult that developed at about the same time was
known as onryo ("wrathful spirit"). It concerned the spirit of an aristocrat forced to an unjust death that remained unpacified and
resided in limbo, whence it manifested its wrath (tatari) in the form
of disease, earthquakes, droughts, and the like. This belief originated
during the Nara period around several cases mentioned below, and
culminated in the cult dedicated to the wrathful spirit of Sugawara
no Michizane. The main distinction between the two cults is that an
onryo is always the vengeful spirit of a deceased person, whereas a
54 See Shigematsu Akihisa, Hachiman Usa-gu gotakusenshu (Tokyo: Gendai Shichosha,
1986); see also Nakano, Usa-gu; Kagawa Mitsuo and Fujita Seiichi, eds., Usa (Tokyo:
Mokujisha, 1976).
55 Shibata Minoru, Chusei shomin shinko no kenkyii (Tokyo: Kadokawa, 1966). On the Gion
cult, see Yoneyama Toshinao, Dokyumento Gion matsuri — toshi to matsuri to minshu to
(Tokyo: NHK bukkusu, 1986).
Cambridge Histories Online © Cambridge University Press, 2008
560
RELIGIOUS PRACTICES
goryo may be a force of another nature. The onryo phenomenon was
accompanied by the belief that spirits could be pacified even after
the death of the person: rituals of exorcism {chobuku-ho, literally,
"rite to control and vanquish") manipulated and pacified spirits that
were forced by the power of their passion to remain in limbo and
were thus prevented from what might be an almost mechanical rebirth into any of the six destinations provided for in classical Buddhist cosmology. The thaumaturgists dealing with these spirits were
regarded with awe, for they had firsthand knowledge of the realms
of death gained through their austerities in mountains and even
through their cataleptic experiences of hell described in various texts
of the Heian period.
The main onryo cults originated with the case of Imperial Prince
Nagaya in 729, and developed thereafter for Fujiwara no Hirotsugu
in 740, and Imperial Prince Sawara ("Emperor" Sudo) in 785. During the Heian period, there were the cases of Imperial Prince Iyo in
807, Tachibana no Hayanari in 842, and finally of Sugawara no
Michizane in 903.
Sugawara no Michizane: the Kitano cult
The spirit of Sugawara no Michizane became the most famous and
important in the entire tradition of wrathful spirits. It became the
object of a major cult that left a remarkable imprint on Japanese
culture.56
Sugawara no Michizane (845-903), who came to hold the position
of Minister of the Right, was regarded by Fujiwara leaders as a major
threat to their monopoly of power. To discredit him, they accused
him of plotting against the throne. He was swiftly exiled to Kyushu,
where he died two years later, in 903. One of his attendants established an altar in 905 near his gravesite, which was transformed into
the Anrakuji in 919 by Fujiwara no Tokihira's cadet, Nakahira; this
shrine-temple multiplex came to be governed by members of the
Sugawara lineage. The men who had falsely accused the statesman
of wrongdoing died in sudden or mysterious circumstances in the
years after Michizane's death: Tokihira died in 909; his nephew Ya56 On the topic of the goryo and onryo phenomena, see Murayama Shuichi, Honji suijaku
(Tokyo: Yoshikawa Kobunkan, 1974); Kuroda Toshio, "Chinkon no keifu," in Kuroda
Toshio, Nihon chusei shakai to shukyo (Tokyo: Iwanami shoten, 1990), pp. 127-56. On the
deification of Sugawara no Michizane, see Robert Borgen, Sugawara no Michizane and the
Early Heian Court (Cambridge, Mass.: Harvard University Press, 1986), pp. 307-36.
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
561
suakira, who had been made crown prince, died in his youth in 923.
At this point the court pardoned Michizane, struck from the record
the decree of exile, and reinstated him to his rank and position, hoping to quell his wrath. But matters continued to worsen; in 930 a
lightning bolt struck the imperial palace and killed the man who had
reported Michizane's "admission of guilt" and three other members
of the court. Emperor Daigo became despondent over the string of
crises and died. By that time, the vengeful spirit of Michizane had
assumed the divinized form of lightning, a fact that precipitated the
popularization of the cult because of the thunder-lightning cults that
had been a part of ancient agricultural practices. The thaumaturgist
Nichizo claimed in 941 that, after falling into catalepsy, he had visited the hells and encountered there the spirit of Michizane surrounded by thousands of irate demons claiming vengeance.
Michizane would have then told Nichizo that, if his spirit was given
a proper cult, he would protect the state. In that document Michizane was said to have called himself Nihon Dajo Itoku-ten, a name
that might indicate an association of Daiitoku (Yamantaka, discussed earlier) and daijo-daijin (Chancellor).57 This must be linked
with the fact that Daiitoku used the bull as his mount and with the
popular rites of propitiation of rain and lightning in which bulls were
sacrificed. Indeed, the spirit of Michizane came to be associated
with the bull.
Furthermore, in 942 a woman of lower extraction claimed to have
been possessed by the spirit of Michizane; in the oracle she uttered,
the vengeful spirit requested the erection of a shrine; that is the
Ayako Temmangu located in the southern part of the capital. The
woman was Tajihi no Ayako, whose forebears were known for their
worship of thunder. In 946 a certain Miwa no Taromaru, then only
seven years old and born in the sacerdotal house of the Hira Shrine,
was also possessed by Michizane's spirit and rendered an oracle that
has been recorded in full, in which Michizane requested the transfer
of the shrine. The next year a new shrine-temple was dedicated to
Michizane's spirit on a site that had been used for the propitiation
of the kami of Heaven and Earth; that is the Kitano "Shrine," located on what was then the northern edge of the capital. In that multiplex, and in the Anrakuji of Dazaifu, Michizane came to be worshiped under the name Temman Daijizai Tenjin, a compound in
which the term Daijizai must be associated with Daijizaiten, the
57 See Iyanaga Nobumi, "Daijizaiten," Hobogirin 6 (1983): 713-65Cambridge Histories Online © Cambridge University Press, 2008
562
RELIGIOUS PRACTICES
Mahesvara (Supreme Lord) which plays an important role in the
Buddhist pantheon and which is also associated with the bull.
Finally, there must be a relation between that name and the cult
that was dedicated to Shidara in the middle of the tenth century, a
cult so popular that thousands of people participated in it, carrying
portable shrines in which one of the deities was named Ayae Jizai
Tenjin.s8
By 987 the Kitano shrine-temple multiplex came under the control of the court and Mount Hiei. Leading thaumaturgists of the
time performed rituals of exorcism and attempted to bring peace to
the spirit and to the political realm. It is possible that the sudden
growth of this cult indicates moral concerns on the part of the aristocracy as well as political consciousness on the part of the populace
of Kyoto; there is little doubt that the thaumaturgists responsible for
the institutionalization of the cult were mediators between these two
classes of interest, because what needed to be pacified were not only
the spirit of Michizane but also the worries of the aristocrats and the
fears of the populace as well. The aristocratic pangs of conscience, if
there were any, might have been laid to rest by individualized ritual,
but the fears of the populace were less easy to quell because they
were multifaceted: they arose from the belief that the natural world
was ethically reactive to human behavior, as well as from what might
have been a kind of obscure suspicion that the political manipulations of the court were not in accord with its stated ideology. Because overt criticism of politics on the part of the general populace
was not possible, only cultic behavior provided an outlet for political and, therefore, ritual concerns. The Shidara cult, the early stages
of the Kitano cult, and other structurally related cults are good examples of such symbolic action whereby some sense of popular participation was achieved.
It is often said that the Shidara cult is popular in nature; but its
roots are continental and it is not clear how this type of cult came to
exercise such a fascination among the less educated. It might be appropriate to suggest that this type of cult is related to one of the
frameworks of interpretation of the existential situation that marked
the Heian period and that has been mentioned earlier, namely, the
dialectic between nature and culture. This dialectic surrounds notions according to which human ethical behavior is an agent of cosmic change. Born in China among communities that were critical of
58 In reference to the Shidara cult, see Murayama Shuichi, Honji suijaku, chap. 8.
Cambridge Histories Online © Cambridge University Press, 2008
THE FORCES OF NATURE
563
imperial rule, these notions were carried to Japan and in time interacted with notions held by the common people, according to which
natural disasters were caused by forces that came to be linked to the
vengeful spirits of wronged statesmen. These forces therefore needed
to be propitiated in such manner that the political manipulators responsible for unjust deaths be punished so that the natural world
might return to its ideal smooth rhythm.
A document from China and its interpretation by Joseph Needham offer insight into the nature of the problematic. The firstcentury skeptic philosopher Wang Ch'ung wrote diatribes against
what he called the "phenomenalists," in which he says:
Ceremonies originatefromwant of loyalty and good faith, and were the beginning of confusion. On this score people find fault with one another,
which leads to mutual reproof (of superiors and inferiors). At the time of
the Three Rulers people sat down informally (without attending to precedence) and walked about at their ease. They worked themselves instead of
using horses and oxen. Simple virtue was the order of the day, and the people were unsophisticated and ignorant (of social distinctions). Minds acquainted with "knowledge" and "cleverness" had not then developed. Originally there were no calamities or omens, or if there were, they were not
considered as reprimands (from Heaven). Why? Because at that time people were simple and unsophisticated, and did not restrain or reproach one
another. Later ages have gradually declined - superiors and inferiors contradict one another, and calamities and omens constantly occur. Hence the
hypothesis of reprimands from Heaven has been invented.
Wang Ch'ung offered us a clue when he stated, at the peak of his attacks against the phenomenalists:
Decaying generations cherish a belief in ghosts. Foolish men seek relief in
exorcism. . . . The conclusion is that man has his happiness in his own
hands, and that the spirits have
nothing to do with it. It depends on his
virtues and not on sacrifices.59
Needham comments that governmental bureaucratic circles
frowned upon the beliefs, and that "the government felt it inexpedient that the common people should engage in arguments about the
relations between the emperor and Heaven." This means that, at the
time, some people felt that natural calamities were caused by ghosts
or spirits that could be propitiated through the performance of rituals and ceremonies, and that some people also believed that the destructive activities of those ghosts and spirits had been caused by political acts of questionable character. We thus have here two related
59 Joseph Needham, Science and Civilization in China, vol. 2: History of Scientific Thought
(Cambridge: Cambridge University Press, 1962), pp. 376-78.
Cambridge Histories Online © Cambridge University Press, 2008
564
RELIGIOUS PRACTICES
issues, one concerning a dialectic between nature and culture, and
the other one concerning political relations between structure (the
establishment), and communitas (the common people).
The goryo and onryo beliefs of Nara and Heian Japan seem to be
related to those Chinese practices and beliefs which date back to the
Han dynasties, if not earlier. In Japan, they might be characterized
as being the result of structural interactions between Buddhism and
the worship of kami, between nature and culture, and between
structure and communitas. These interactions evolved at several levels of society and came to rest in the shrine-temple multiplexes,
where they were sustained by a variety of rites originating in
shamanic possession and oracular religion, as well as in the esoteric
Buddhist traditions; new Buddhist elements appearing in history
were grafted onto the cult: elements of Pure Land and Zen became
in time integral parts of the evolution of the Kitano cult.
One understands perhaps better now what caused the Kitano phenomenon and its solution. Aristocrats were confronted with the reverse side of the coin they had so successfully played earlier in claiming that their rule enjoyed the backing of "Heaven," since, after
Michizane's demise, people claimed that natural disasters impacting
their own lives were a manifestation of discontent on the part of the
very same Heaven. And between the people and the aristocrats the
thaumaturgists stood, equally pressured by both groups to pacify all
concerned. Sugawara no Michizane was eventually deified as the patron saint of poetry and learning, and his cult has continued to develop up to the present day.
THE ASSOCIATION OF KAMI WITH BUDDHAS
The associations between kami and buddhaslbodhisattvas
The creation of shrines and temples was discussed earlier in political and ritual contexts, but we must now consider a closely related
topic: the cultic associations between various kami and buddhas
and bodhisattvas. These associations are often but inadequately referred to as a syncretism between Shinto and Buddhism.60
It should be noted that the Japanese of the Heian period never
spoke of Shinto versus Buddhism, and this for several reasons. First,
60 See Murayama Shuichi, Shimbutsu shiigo shicho (Kyoto: Heirakuji, 1957), and Honji suijaku;
Oyama Kojun, Shimbutsu kosho-shi (Kyoto: Rinsen shoten, 1975); and Kubota Osamu,
Chusei Shinto no kenkyu (Kyoto: Shinto-shi gakkai, 1959).
Cambridge Histories Online © Cambridge University Press, 2008
KAMI AND BUDDHAS
565
there is a question concerning the definition of phenomena to which
the tradition itself gave the name Shinto. When the term appears
during the medieval period, it refers to specific ritual and doctrinal
systems related to particular shrine-temple multiplexes, never to an
overarching, nationally accepted body of doctrine and ritual.
"Shinto" as a system of belief totally independent from Buddhism
did not exist before the Edo period, and even then the term denoted
a system heavily impacted by Neo-Confucianism, Taoism, and yinyang notions and practices. In the Heian period the realm of shrines
exhibited indigenous and imported creeds, practices, and institutions; these were rarely at odds with the world of temples. The overall unifier of the world of shrines and temples was, as we have seen
several times in the course of this chapter, the conglomerate of governing lineages, from whose point of view the sole difference between shrines and temples was a matter of social, cosmological, and
ritual considerations. Second, the Japanese cultic traditions emphasize, more than syncretism between Shinto and Buddhism as bodies
of doctrine and practice, locale-specific associations between the
particular kami of a shrine and the particular buddhas and bodhisattvas of the temple with which that shrine was associated throughout the Heian period and, indeed, for the most part, until the nineteenth century.
The term for the phenomenon is shimbutsu shicgo, which means,
literally, combinations (shugo) between kami (shin) and buddhas and
bodhisattvas (butsu). Thus what we are confronting here is not syncretism between "Buddhism" and "Shinto," but specific relations
between shrines and temples where those divinities were enshrined.
It is, however, true that in most cases the combinations were
grounded in a framework of interpretation issued either from the
Tendai or Shingon philosophical systems; the Japanese themselves,
during the medieval period, believed that most combinatory rationales had been authored by Kukai and Saicho, the founders of Shingon and Tendai schools. This means that, during the medieval period, the most widely circulated rationales were grounded in
esotericism. There was some syncretism in Japan; but the phenomenon under consideration at the present is structurally different, since
we are looking at systems of combinations located in clear spatial
and temporal areas, and generally placed against the background of
the chingo kokka theories.61
61 On the topic of syncretism as an altogether different issue, see, for example, Birger A. Pearson, ed., Religions Syncretism in Antiquity (Missoula, Mont.: Scholars Press, 1975)-
Cambridge Histories Online © Cambridge University Press, 2008
566
RELIGIOUS PRACTICES
The jin&iji and the Twenty-two Shrines being institutions politically related to the evolution of Buddhism in Nara and Heian Japan,
it is not surprising that shrines and temples developed interactions
at the ritual and philosophical levels as well. In these interactions.
Buddhism tended to represent the dominant discourse by virtue of
the fact that temples developed an institutional supremacy over the
shrines which they came to govern, and because the definition of the
kami was apparently left to Buddhist prelates. The interactions
focused on different aspects of the symbolic realms embodied by
shrines and temples.
It is important to emphasize that during the Heian period one
rarely finds anyone of whom it can be said with absolute confidence
that he was purely Buddhist or that his life was exclusively determined by the realm of shrines. This was true for members of sacerdotal lineages and for prelates governing the temples. Just by looking
at the calendar of festive ceremonies and rituals of the Heian period
(the nenju gydji), we must conclude that cultic life was also political,
and that it involved the members of the court in shrines and temples,
sometimes separately and sometimes concurrently. People did not reflect on these matters in an abstract manner. The separation was not
doctrinal: it was mostly social and sometimes ritual.
We thus see "Buddhist" prelates going on pilgrimage to shrines in
order to ask for the protection of the kami in their endeavor to
achieve awakening. Some even spent long periods in ascetic seclusion on the grounds of shrines, where they had erected various
buildings to house scriptures and to perform rites such as the famed
Hokke-hakko.62 Conversely, the powerful Nakatomi sacerdotal lineage, for instance, wrote texts proposing Buddhist interpretations of
purification, one of the most important "Shinto" rituals. These systematic interactions profoundly marked the Heian period at all levels of life. The Taira house built the Sanjusangendo in Kyoto and the
Itsukushima shrine-temple multiplex on Miyajima. The Fujiwara
house developed a large framework of shrines and temples throughout the country, and the Minamoto house (the Seiwa Genji) developed the Futara shrine-temple multiplex, all the while taking Hachiman as tutelary deity, and later enshrining him at Tsurugaoka in
Kamakura.
It was in this context that Buddhist prelates began to explain the
62 See Willa Tanabe, "The Lotus Lectures: Hokke hakko in the Heian Period," Monumenta
Nipponica 39, 4 (1984): 393-407.
Cambridge Histories Online © Cambridge University Press, 2008
KAMI AND BUDDHAS
567
existence and the power of indigenous kami in terms of Buddhist
doctrine, with special reference to the theory known as honji suijaku
and traditionally associated with the Lotus Sutra. According to this
theory the Body of Essence of the Buddha used expedient devices
(hoben) to manifest itself in various guises in the world in order to
guide it toward salvation. One of the many guises it would take was
as a kami. It thus became customary to identify specific kami of
shrines as hypostases of the specific members of the Buddhist pantheon worshiped in their adjacent temples. Kami and buddhas were
associated to each other according to various rationalizations usually couched in an atmosphere of mystical union expressed by linguistic associations and metaphors. For example, the main kami of
the Kasuga Shrine-whose theriomorphic emblem was the deerwas said to be a hypostasis of Fukukensaku Kannon because the
iconography of that Kannon called for a deer hide on its shoulder.
People looked at the presence of the deer in the shrine, then in the
temple, saw it as a subtle and mysterious indication of the essential
identity of the kami and of the bodhisattva, and thus combined
them. This type of rationalization pervaded the world of associated
shrines and temples, and spread among the aristocracy as well as
other strata of society.
Furthermore, it might be useful to suggest that if at the individual
level, people thought that destinies were decided upon by higher
forces such as planets and the like, at the state level it was believed
that much of the destiny of the state rested in those forces and
deities, and that this was one of the reasons for the importance of ritual. These notions paved the way for the belief that history was the
result of the divine will of kami and buddhas or bodhisattvas revealed through the medium of divination and oracles, and propitiated through the medium of rites, be those preventive or curative.63
In this specific context the associations between shrines and temples
played a central role: the realm of shrines insisted upon regular ritual return to the origins and expounded on the views of the Kojiki
and Nihon shoki concerning myth and history, while the realm of
temples expressed the notion that reality was the result of complex
63 The processes leading to certain types of historical consciousness in medieval Japan have
been studied by Kuroda Toshio in "Kemmitsu bukkyo ni okeru rekishi ishiki," in Kitanishi
Hiromu sensei kanreki kinenkai, ed., Chusei shakai to ikko ikki (Tokyo: Yoshikawa
kobunkan, 1985), pp. 505-23. See also Kuroda Toshio, "Historical Consciousness and Honjaku Philosophy in the Medieval Period on Mt. Hiei," translated by Allan Grapard in
George Tanabe andWillaTanabe, eds., The Lotus Sutra in Japanese Culture (Honolulu: University of Hawaii Press, 1989), pp. 143-58.
Cambridge Histories Online © Cambridge University Press, 2008
568
RELIGIOUS PRACTICES
networks of mutual causation (engi) over which individuals had little control.
These views were then projected at the popular level by the didactic literature of setsuwa, in which people were told that their affairs were mediated partly by personal responsibility, but mostly by
the will of entities that were the object of cults in shrines and temples to which they were invited to go on pilgrimage in order to beseech them. In what other way could they have explained the sudden intrusion of buddhas and bodhisattvas into their lives, be that
intrusion the manifestation of compassion or of supernatural wrath?
Combinations between kami and buddhas or bodhisattvas were a
central part of edifying tales such as those found in the Konjaku
monogatari-shu. The practice of Buddhism to the exclusion of the
realm of shrines did not exist in Japan before at least the Edo period;
on the other hand, the elaboration of "Shinto theses" that took place
in the Kamakura and Muromachi periods in fact owed a great deal
to Buddhist interpretations. These were not dominant issues during
the Heian period.
The combinations in Kunisaki
The Kunisaki peninsula in Kyushu exemplifies the general principles advanced in the present chapter. Kunisaki was a discrete geographical area governed by mountain ascetics who, within a complex
epistemological framework and under Tendai institutional direction,
organized a cult dedicated to the Lotus Sutra and to Hachiman with
the avowed purpose of protecting the state. The religious world of
Kunisaki was a system of ritual and philosophical interactions between elements issuing from local creeds, from continental shamanistic practices, from Tendai and Shingon esoteric rites, and from
Taoist themes and rites. These interactions caused an evolution in
the concept of sacred nation (shinkoku) that was sustained through
the performance of major rituals for the protection of the state, and
a host of related creeds, practices, and rituals. This system evolved
and came to fruition during the Heian period.
The Hachiman cult is of distant and most unclear origins, and its
structure has been reorganized several times in the course of history.64 It is probable that its source is to be found in Korea, and per64 On the Kunisaki peninsula, see Allan Grapard, "Lotus in the Mountain, Mountain in the
Lotus," Monumenta Nipponica 14, 1 (1986): 21-50, and "Enmountained Text, Textualized
Mountain: The Lotus Sutra in the Kunisaki Peninsula," in George Tanabe and Willa TanaCambridge Histories Online © Cambridge University Press, 2008
KAMI AND BUDDHAS
569
haps even farther on the continent; in Japan, its original site is in the
northeastern part of Kyushu, an island that literally fell under its economic control during the Heian period. Originally related to litholatry, to solar worship, and to sword worship, this combination of local
cults attached itself to the mythology of the state when it was stated
in the early ninth century that Hachiman was the deified form of the
fourth-century king Homuda Wake ("Emperor Ojin") and of his
mother, Tarashi Hime ("Consort Jingu"), who became an important
figure in the formation of the concept of Japanese territory when it
was claimed (in the Kojiki and Nihon shoki) that she had conquered
Korea.
The main center of the Hachiman cult came to be located in Usa
for reasons that are most unclear, and the Mirokuji Buddhist temple
was built next to the shrines during the eighth century. The sacerdotal lineage that first governed the shrines was named Oga
(Omiwa), but it was replaced some time in the course of the eighth
century by the Usa sacerdotal lineage, members of which administered the shrines, while other members (of both lineages) became
Buddhist prelates and administered the temples. It appears that the
Nara prelates of Usa were thaumaturgists (fuso), that is, shaman-like
monks who were well versed in Chinese religious practices, engaged
in cults dedicated to various kami, and used a predominantly Buddhist conception of the world. The role of the oracles of Usa Hachiman became intensely political at the time of the Dokyo incident in
769, when an oracle recommending Dokyo's accession to the throne
was questioned in some governmental circles and another oracle was
requested. That second oracle stipulated that only members of the
imperial lineage could rule Japan, and ordered that Dokyo be removed from government. As was seen earlier, Hachiman was subsequently brought to Nara to become the protector (chinju) of the
Todaiji.The Hachiman cult developed an even more intense relation
to the state when Fujiwara no Yoshifusa created the Iwashimizu
shrine-temple multiplex south of Kyoto in 859 to celebrate the accession to the throne of his grandson, Emperor Seiwa, the preceding
year. In time Iwashimizu came to administer the Usa Hachiman site
of cult.
When mountain asceticism developed in the Kinai area at the beginning of the tenth century, and especially so in the Kumano rebe, eds., The Lotus Sutra in Japanese Culture, pp. 159-89. See also Nakano Hatayoshi,
Hachiman shinko-shi no kenkyii.
Cambridge Histories Online © Cambridge University Press, 2008
57O
RELIGIOUS PRACTICES
gion, it also developed in Kyushu on Mount Hiko near Usa. As a result, the evolution of mountain asceticism in Kyushu went hand in
hand with the development of the Usa Hachiman cult. It was under
the combined influences of Kyushu Tendai Buddhism, mountain asceticism, and Hachiman that the Kunisaki peninsula religious system developed during the second part of the Heian period. The various monks who developed that elaborate institutional and ritual
framework were seen as a single bodhisattva named Nimmon, who
was in turn seen as the Buddhist aspect of Hachiman in that part of
Japan.
A number of temples were built on the slopes of the volcanic
peninsula and were put under the administration of Tendaicontrolled institutions, which attempted to gain control of the
Hachiman cult in the Usa shrine-temple complex. The Tendai
school is based on the Lotus Sutra, traditionally referred to as the
"Scripture in Eight Scrolls" or "Twenty-eight Chapters," further
subdivided by scholarly commentary into three distinct parts. As a
consequence, the people administering the land domains that were
located in the Kunisaki peninsula and belonged to the Mirokuji
Temple of Usa built as many temples as there are chapters in the
Lotus Sutra, and separated them into three administrative and ritual categories that would correspond to the scholarly tripartite distinction of the Lotus Sutra. Furthermore, it was suggested (at a
time that is unclear) that there should be as many representations
of buddhas and bodhisattvas on the slopes of the volcano as there
are graphs in the Chinese version of the scripture. The result of
these decisions was that the entire peninsula was regarded as if it
were a natural and cultural embodiment of the Lotus Sutra itself.
Metaphorical associations between elements of the Lotus Sutra
and elements of the Hachiman cult were subsequently established
so that the peninsula came to be seen as the Japanese form of the
Peak of the Numinous Eagle on which the Buddha preached the
Lotus Sutra, and as the area of manifestation of the golden radiance of Hachiman in its theriomorphic appearance as an eagle or
falcon (later seen as a dove). The major rituals performed in Kunisaki were rituals for the protection of the state against foreign
invasions.
It is clear that the Kunisaki system cannot be understood from the
perspective of Buddhism only, or from the sole perspective of the
realm of shrines. Kunisaki, like most other cultic sites which developed in Heian Japan, was a combinatory phenomenon.
Cambridge Histories Online © Cambridge University Press, 2008
KAMI AND BUDDHAS
571
The combinations in Kumano
Located at the southernmost tip of the Kii peninsula directly to the
south of Kyoto, the Kumano region is one of the most important
cultic areas of Japan. It is home to three major shrines: Kumanohayatama (Hongu), Kumano-nimasu (Shingu), and Nachi, which,
over the centuries, developed complex associations with the temples
that came to be built near them. Kumano appears in documents first
in the Kojiki, in which it is purported to be the burial site of the cosmogonic deity Izanami. But it soon gained recognition on the part
of anchorites during the Nara period for it was considered to be an
ideal site for withdrawal from the world. The Nihon ryoiki, Japan's
first compilation of didactic stories written by the monk Kyokai
around 820, mentions the place in this respect.65 The exquisite
beauty of Kumano's natural environment quickly drew the attention
of the imperial court, and the shrines of Hongu and Shingu were
awarded higher ranking in 859, precisely at the time of the Jogan era
major reformulations of ritual discussed earlier; this indicates direct
Fujiwara control. In fact, members of the court were so attracted to
the region's beauty and celebrated hotsprings that Retired Emperor
Uda made a pilgrimage to Kumano as early as 907, starting what was
to become a tradition among emperors. At the end of the Heian period, the tonsured emperor Go-Shirakawa made the journey thirtyfour times, while Emperor Go-Toba went thirty times. Aristocrats
followed close behind: Jozo, the son of Miyoshi Kiyoyuki (847-918),
took the orders and spent three years in ascetic seclusion at the foot
of Nachi waterfall; Fujiwara no Tamefusa went on pilgrimage in
1081, and thereafter the desirable position of abbot of Kumano
tended to remain within the lineage of the Ichijo branch of the Fujiwara house.
Almost the entirety of the Kii peninsula then developed as a major
center of practice for the mountain ascetics of both Shingon and
Tendai lineages, and became the nexus of a particular mode of
mountain asceticism with a great formative influence on almost all
other mountain cultic centers of Japan. The cosmology of Buddhism
and the cosmography of the Pure Land, combined with the native
emphases on purification and on eschatology also had a great effect
65 On the Kumano cult, see Kumano Nachi Taisha, ed., Kumano Sanzan to sono shinko, 3 vols.
(Tokyo: Meicho shuppan, 1984). The reference to Kumano in the Nihon ryoiki is in kan 3,
no. 1. See Endo Yoshimoto and Kasuga Kazuo, eds., Nihon ryoiki vol. 70 of Nihon koten
bungaku taikei (Tokyo: Iwanami shoten, 1967), pp. 316-21.
Cambridge Histories Online © Cambridge University Press, 2008
572
RELIGIOUS PRACTICES
on the evolution of Kumano as a cultic site without equal in the rest
of the country. Kumano came to be seen as a physical gate to the
transcendental space of the Pure Land, and then as a Pure Land in
this world. It remained one of the most important foci for the development of mountain cults throughout Japanese history.
LATE HEIAN DEVELOPMENTS
Personal practices in the late Heian period
As suggested earlier, rituals tended to leave the exclusive domain of
the emperor at the same time the emperor came to be less and less
regarded as the true body of the state, or the nation as his embodiment, and as those rites were widely adopted by aristocrats on an individual basis.
These developments accompanied the appearance of individual
concerns for salvation, and it is in this context that the Japanese of
the latter half of the Heian period did not find any contradiction in
worshiping at the same time divinities related to the esoteric traditions ofTendai and Shingon, or colored by Taoism or indigenous associations and accretions, together with Pure Land beliefs and practices. An element common to all these was the reference to the
individual as the locus of cultic belief and practice within a sociocosmic framework. This is best illustrated, perhaps, by the popularization of the nembutsu, or chanted invocation to the Buddha Amitabha (Japanese, Amida), a movement that appeared against the
background of political troubles and natural calamities. When massive esoteric ritual counteraction on the part of the state proved to
be ineffective or was perceived to be geared exclusively to the protection of aristocrats in positions of power, personal salvation came
to be emphasized. Furthermore, a Buddhist doctrine of time and
history gained wide acceptance around the eleventh century: in its
framework of reference, history was not seen anymore as the symbolic manifestation of the will of the kami and their associated buddhas and bodhisattvas, but as a devolutionary process over which
human beings had little or no power. Time was conceived of as an
inexorable process leading to the final conflagration of all worlds, an
event whose date was proposed in several scriptural sources; the final
period of history (mappo) leading to the ultimate conflagration was
believed in Japan to begin in the year 1052. It was thought that one
of the marks of such a period was political and social unrest, which
Cambridge Histories Online © Cambridge University Press, 2008
LATE HEIAN DEVELOPMENTS
573
was precisely the case in Japan at the time. In that context, salvation
was proposed as residing mainly in a total surrender to the saving
grace of Amida, through the medium of chanting his name. It was
said that the Buddha would remember those who called on him, and
would therefore allow them rebirth into his Pure Land, an ideal
space of residence where the practice of Buddhism would be much
easier than in the troubled human world. An important corollary
was that it was claimed that sole reliance on the Buddha Amida ensured rebirth into the Pure Land, and that therefore the worship of
autochthonous katni was not necessary. This caused immediate divisions, whose nature was obviously political, because the rejected
kami were in fact those that represented the ideology of the state in
the shrine-temple multiplexes. Indeed, it was from such multiplexes
that massive opposition to Pure Land ideology stemmed.
The late Heian period was thus pulled between opposing tendencies of a cultic and political character, each tendency being echoed
in ritual or practice. On the one hand were rituals for the protection
of the state; and on the other hand were personal devotions within
the context of grace. In an almost direct cause-to-effect relation,
when leading Fujiwara figures lost their grip on political affairs they
"converted" to personal surrender to the grace of Amida and hoped
to be reborn into a better world.
The last part of the Heian period saw a frenzy of ritual activity.
The Insei period, which began in 1086 and during which emperors
took the tonsure in the expectation of wielding administrative power
more effectively, was marked by stately rituals that had one major
goal: the reestablishment of imperial authority through the use and
manipulation of symbols. One might say that, in the process, the emperor was drowned in a mass of ritual activity. Most of the practices
described up to the present point were aspects of the discourse of
political power in classical Japan where, as usual, the delimitation of
purely political power as different from cultic power was fundamentally ambiguous. One can propose that the more intense the political conflicts were, the more intensified the ritual activity. Legitimacy
being a symbolic phenomenon, ritual manipulation was an important tool of both politicians and thaumaturgists.
Ecstatic mass movements
If the Heian period saw the systematic development of ritual activity on a private level in relation to the world of shrine-temple multiCambridge Histories Online © Cambridge University Press, 2008
574
RELIGIOUS PRACTICES
plexes associated with the ideology of the government and its various lineages, it also saw, by contradistinction, the phenomenal rise
of antithetic, ecstatic mass movements.
We have seen that the Shidara cult and a number of gory5-re\ated
cults spread among the Kyoto populace. Other popular cults that
developed during the Heian period were related to the spread of
Pure Land doctrine and devotional practices. The precursor of such
movements of Pure Land Buddhism was Kuya (sometimes called
Koya, 903-72), a man of aristocratic descent who became an itinerant monk (yugyoso) and who, upon his return to Kyoto in 938, kept
popularizing invocation to the Buddha Amida (nembutsu) among
prisoners, common folks, and courtiers alike. He became famous for
adding musical rhythm and dance to public invocations, in which he
invited people to join him and dance; this was known asyuyaku nembutsu or, more simply, odori nembutsu, "danced invocation." Kuya's
predilection for public places earned him the title of ichi hijiri, "saint
of the marketplace."
Further popularization of the devotional practices of the Pure
Land was achieved through the work of Genshin (942-1017), more
particularly his Anthology on Rebirth in Pure Land (Ojdyoshu), which
generalized a mood of impermanence befitting the troubled times,
a strong sense of piety linked to paradise and hell imagery, and a
fervent devotion to the Buddha Amida as the sole means to salvation. However, the popularization of the nembutsu among the lower
classes was further enhanced by Ry5nin (1072-1132), who proposed a magic-oriented, mass recitation of the invocation to
Amida, known as Ylizu-nembutsu, "universally transferred merit of
the invocation."66
It might be proposed that the mass devotional movements of the
second part of the Heian period indicate the growth of eschatological concerns in a society that was witnessing the gradual erosion of
classical rule in the midst of warfare and natural calamities. Such
movements have tended to appear in Japanese history in the
decades preceding momentous sociopolitical changes, and further
investigation of symbolic behavior and popular cults within the context of power relations seems necessary before a complete assessment is offered.
66 On Kuya, see KonToko et al., eds., Rokuharamitsuji (Kyoto: Tankosha, 1969); Hori Ichiro,
Kuya (Tokyo:Yoshikawa kobunkan, 1963). On Genshin and Ryonin, see Ono Tatsunosuke,
Jodai no Jodo-kyo (Tokyo: Yoshikawa kobunkanj 1972).
Cambridge Histories Online © Cambridge University Press, 2008
LATE HEIAN DEVELOPMENTS
575
Conclusion
The Japanese of the Heian period attempted to establish structural
relations between institutions, ideologies, and the rituals that either
sustained or developed them. The creation of the Twenty-two
Shrine-temple system, coupled with the reformulation of ritual procedures, was a momentous event indicative of what came to be the
dominant ideology of the state.
It is quite clear that, overall, the realm of combined shrines and
temples was one of the structuring devices of the ritual formulation
of the state (and of the state sponsorship of ritual), and that individuals who needed personal salvation out of the framework provided
by the state could find it in those multiplexes only with great difficulty; indeed, pilgrimage to those sites of cult seems to have been engaged in by courtiers only. This explains in part the surge of heterodox movements and mass ecstatic movements that characterize the
second part of the Heian period.
The "rule of taste" of the classical age of Japan belongs to the Fujiwara dominance and to the combined world of shrines and temples
that protected it. The slow demise of the statutory system brought
the ideal state and the rule of taste to a dramatic close, as the Heian
period ended in warfare and in the emergence of other social, political, and cultic ideals.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 9
INSEI
In late Heian times, the retired sovereigns Shirakawa,Toba, and GoShirakawa dominated the court as few former sovereigns, reigning or
retired, had done. Later historians referred to this form of political
domination as inset or "cloister government," a term derived from
the fact that abdicated emperors resided in well-appointed villas or
religious cloisters (in) from which they conducted politics (sei). The
ex-sovereign's office had numerous subordinate bureaus staffed by
courtiers who supported his interests at court and supervised extensive provincial estates. Retired emperors so dominated the Heian
court that historians call the period 1086-1185, stretching from late
Heian to the beginning of Kamakura times, the Insei period.
Once a person has attained the highest office in the land, it is difficult to revert to the status of a common citizen; the aura of authority remains. In modern times, this is true of former kings, presidents, and prime ministers; however, the more hierarchic the
society, the more we should expect the residual authority to remain
viable. In Heian Japan, substantial authority remained with a former
sovereign. Although rulers everywhere abdicate, nowhere has the
practice been as common as it was in Japan. And nowhere else did
former sovereigns regularly influence national affairs so significantly.
A brief recapitulation of some features of the Japanese dynasty is
necessary background for an understanding of this phenomenon.
ABDICATION, REGENCY, AND THE JAPANESE THRONE,
645-IO68
Abdication was not peculiar to late Heian times, nor were powerful
ex-sovereigns unknown in other eras. The first recorded abdication
occurred in 645 when the Empress KSgyoku stepped down in favor
of Emperor Kotoku and simultaneously confirmed her son Prince
Naka no Oe (future emperor Tenji) as crown prince. From that time
until the abdication of Go-Sanjo in 1073-his abdication marks the
576
Cambridge Histories Online © Cambridge University Press, 2008
THE JAPANESE THRONE
577
prelude to the inset - twenty-three of the thirty-four or thirty-five sovereigns abdicated.1 In fact, abdication became the standard means of
transferring the dynastic title.
Japanese sovereigns abdicated for many reasons: ill health, belief
in spirits, the pressures of office, pressure from Fujiwara ministers,
the desire to enter the monastic life, or because they were female.2
At the basis of all acts of abdication, however, lay the critical problem of transferring royal succession. Whatever the reasons behind
the timing of an abdication, the politics of the act was crucial. The
frequency of abdication in Japan seems to be due to the fact that experience proved it the least problematic way of deciding royal succession.3
In early Japanese history, indeterminate succession practices often
provoked bitter struggles for the throne. During the Asuka and Nara
periods (the seventh and eighth centuries), the dynasty that today
still reigns in Japan made great efforts to strengthen the imperial position, including experimentation with several methods of imperial
succession to avoid armed conflict. Examination of the succession
practices of this period reveal the following:
1. The dynasty was becoming gendered: adult males were normally
considered to be proper holders of the kingship.
2. The concept of lineal succession in the senior male line was
adopted in principle and included in the official codes.
3. The crown prince was designated early in the reign and frequently enjoyed broad powers of co-rulership.
4. Female rulers were frequently enthroned, often in times of dynastic crisis; but after the late Nara period, female rule was
abandoned.
5. The practice of abdication was introduced.
These phenomena were of course interrelated. The ancient Japanese dynasty was attempting to deal with the problem of transferring
a scarce resource - the kingship - among a large dynastic group with
a minimum of conflict. But conflict was prevalent, resulting in
armed rebellion and the deaths of a number of claimants to the dy1 The number of sovereigns differs, depending upon whether or not one accepts Emperor
Kobun's accession. He is included in the "official" Japanese list, but there is great disagreement among Japanese scholars.
2 These are the kinds of reasons listed in the section of Koji ruien dealing with ex-emperors;
the noted insei scholar Kono Fusao reiterates them in Heian makki seijishi (Tokyo: Tokyodo,
1979)J PP-
7-8-
3 G. Cameron Hurst III, Insei: Abdicated Sovereigns in the Politics of Late Heian Japan,
1086-1185 (New York: Columbia University Press, 1976), pp. 47-49Cambridge Histories Online © Cambridge University Press, 2008
578
INSEI
nastic title. The first three practices, in somewhat different form, are
part of the dynastic practice in Japan today; however, the latter two
features - female rule and abdication - are especially associated with
the Asuka and Nara periods, most likely relied upon to curtail conflict. One proved successful and was retained for many years, the
other failed and was abandoned. The important point, however, is
that they seem linked, as two parts of a process to smooth dynastic
transfer.
Eight of the ten or eleven female Japanese sovereigns reigned in
the Asuka and Nara periods;4 no women were enthroned after the
death of Empress Shotoku in 770 until the Tokugawa period a millennium later. Empresses, at first principal consorts of a deceased
emperor but later other female members of the dynasty as well, were
enthroned in situations of actual or potential conflict over succession
until an appropriate male member of the dynasty could be chosen.5
But when and how to transfer the kingship proved problematic.
During Suiko's reign (592-628), her nephew Prince Shotoku enjoyed
broad powers of co-rulership as regent. She was apparently preserving the dynastic title for him, but he predeceased her, provoking yet
another dynastic dispute. This precedent may have well been in the
mind of Empress Kogyoku, who was enthroned under somewhat
similar circumstances in 645, when she attempted to abdicate early
in her reign.
All seven of the female rulers after Suiko abdicated. (Two, however, abdicated and then reascended the throne.) Since these female
rulers were essentially enthroned as caretakers - Professor Inoue
refers to them as nakatsugi ("transitional heirs") - to reign temporarily in lieu of a male sovereign (which by no means implies that
they were not fully able to rule in their own right), then abdication
4 The counting here is somewhat misleading. There were only six women enthroned (technically) in ancient Japan, but two of them - Kogyoku/Saimei and Koken/Shotoku - ascended
the throne twice. And the rule of one woman, Iitoyo, is disputed.
5 Inoue Mitsusada, Nihon kodai kokka no kenkyu (Tokyo: Iwanami shoten, 1965), pp. 179-218,
passim. Joan Piggott argues persuasively that co-rulership - "gender-complementary chieftain pairs" - were common in the Yayoi and Tomb periods and that that precedent legitimized female rulers in their "participation in the routinization of dynastic succession in the
sixth and seventh centuries." Joan Piggott, "Chieftain Pairs and Co-rulers: Female Sovereignty in Early Japan," in Hitomi Tonomura, ed., Gender in Japanese History, for Japanese
Studies, University of Michigan, forthcoming), p. 2.1 also agree with her that they were not
"sovereigns-as-usual" (p. 36) but were stand-ins, or stakeholders. Unlike Inoue, however, I
do not mean that these female rulers were figureheads. On the contrary, they could and did
rule in precisely the same way as male sovereigns. I use these terms solely in consideration
of the reasons for their selection as ruler in a dynasty that was in transition to a male-dominated one; they were enthroned in, precisely as Piggott says, "liminal moments, moments
of succession crisis" (p. 36).
Cambridge Histories Online © Cambridge University Press, 2008
THE JAPANESE THRONE
579
was the only means to transfer the kingship when conditions were
appropriate. Postponing the transfer until death would have caused
further complications. Thus, it appears that abdication was intimately linked to female rule in early Japan and developed as part of
the process of routinizing dynastic succession during a period of
great fluidity.
The first instance of male abdication occurred in the Nara period:
Shomu abdicated in 749. It was soon recognized that this was the
safest means of transmitting the kingship to the desired heir. Female
rule had not been successful in stemming conflict over the throne,
and it was also believed that the Buddhist priest Ddkyo had been able
to manipulate Empress Shotoku's emotions. In early Japanese history
crown princes or strong claimants to the throne had often been eliminated; however, during the Nara period there were no depositions of
reigning sovereigns until Junnin and perhaps only two cases of regicide. Once a claimant was enthroned, relative security could be assumed, and the safest way to effect that enthronement was to abdicate. Thus pre-mortem royal succession became standard.
During Heian and Kamakura times, dynastic succession practices
showed much greater stability than in the Asuka and Nara eras.
1. Only male members of the dynasty could hold the royal position,
but it was no longer required that they be adult.
2. lineal succession remained important in principle, but in practice
other forms, especially fraternal succession, occurred frequently.
3. The crown prince was named early in a reign but enjoyed no regency power.
4. Fujiwara clan regency became institutionalized, but it was possible for ex-emperors to wield imperial-house regency powers.
5. Female rule did not occur.
6. Abdication increased in frequency; pre-mortem succession became standard.
Politics in the Asuka and Nara periods was characterized by powerful prince regents and female sovereigns. By contrast, in the Heian
period after several strong emperors, imperial regents of the Fujiwara clan and retired emperors enjoyed maximum political power.
All these political styles, however, derived from the same source, a
dynasty that claimed legitimacy on the basis of descent from the Sun
Goddess. With the universal acceptance of this claim, the dynasty
had been placed permanently above other groups; no one could duplicate its genealogy.
Cambridge Histories Online © Cambridge University Press, 2008
580
INSEI
Kingship was an estate right of the dynasty as a group, although
title lay with the person of the emperor.6 Other dynastic members
had a stake in their common estate right, however; and shared
authority, which permitted co-rulership, was a feature of the Japanese dynasty. However, authority and power were normally separated. The emperor was the repository of dynastic authority, but
he usually remained outside the daily scramble for political and
economic power. Power could be - and usually was - delegated to
someone else: a member of the dynasty (prince regent or exemperor), another noble (the Fujiwara regent), or a military hegemon (the shogun). In modern times, power has been delegated to
the cabinet and the parliament.
The sovereign's primary function was sacerdotal rather than political, but since the imperial institution was buttressed by Chinese
Confucian political ideals, the sovereign was never totally powerless.
If other political factors permitted, it was possible for an emperor to
exercise direct political power (shinsei); however, this was unusual.
Delegating power ensured the longevity of the Japanese ruling
house: competition for power was confined to the right to control
royal authority rather than to usurp it. This provided relative security for the dynasty and the possibility for other groups to enjoy political hegemony.
Abdication was standard practice in Heian times; retired emperors were important political figures at court even before inset became
a regular feature of society. In fact, during the Heian period one can
discern an extended competition between the Fujiwara clan and exemperors, on behalf of the imperial house, for control over the right
to "possess" what the dynasty "owned."
From Kammu's time through the death of Emperor En'yu in 991,
eleven of sixteen sovereigns abdicated the throne. After abdication
an ex-emperor was given the honorific title dajo tenno ("great abdicated emperor"), but was more commonly known by a shortened
version, jokb. If he later took the tonsure, the title was dajo hob
("priestly retired sovereign") or in abbreviated form, hob. In
courtier diaries, he was often simply referred to as in ("cloister"),
hence the term inset. From early Heian times, ex-sovereigns enjoyed
some of the prerogatives of reigning monarchs, although the imperial codes did not stipulate a formal position for them. They appear
6 See Cornelius J. Kiley, "Estate and Property in Late Heian Japan," in John W. Hall and Jeffrey P. Mass, eds., Medieval Japan: Essays in Institutional History (New Haven, Conn.: Yale
University Press, 1975)) PP- 109—24, for a discussion of the estate rights of corporate groups.
Cambridge Histories Online © Cambridge University Press, 2008
THE JAPANESE THRONE
581
to have issued edicts through the same channels the emperors used.
They were provided with retirement palaces and complete incomeproducing lands, which became private, heritable assets of the imperial house.
The presence of both former and reigning sovereigns proved disruptive, however. In 809, Emperor Heizei abdicated in favor of his
brother Saga. But Heizei did not abdicate of his own volition; he was
following a plan prescribed by their late father, Emperor Kammu.
Consequently, there was great animosity between the brothers.
When some courtiers followed Heizei and settled in the old Nara
capital, there were in effect two separate and hostile courts.7 After an
abortive rebellion, Heizei remained in Nara as a monk. His son was
demoted and replaced as crown prince by one of Saga's sons. Saga
emerged all-powerful, but people at court were obviously confused
over the relative degree of authority accruing to the reigning and abdicated sovereigns.
It was Saga who tipped the balance in favor of later ex-emperors.
He abdicated in 823 but did not become active in politics until his
son Nimmyo became emperor in 833. He issued edicts and determined affairs of state on Nimmyo's behalf, his power and prestige
deriving from his position as head of the imperial kin group. This
marked a clear departure from earlier periods, when it appears that
ex-sovereigns did not exercise this private, familial role. Previously
the emperor in his public position of head of state seems to have
taken precedence.8
But in Saga's time things changed. Clearly the fact that he was
Emperor Nimmyo's father determined their respective roles. To initiate the new year the emperor paid Saga formal visits of respect,
speaking to him as a subject would address his sovereign. Saga had
emerged as the senior figure in the imperial house, and his patriarchal authority gave him the informal power to influence politics. He
was more powerful in this essentially co-regent capacity than Asuka
period princes had been.
There were retirement palaces and lands allotted specifically for
Saga's support. He appointed men who had served him closely when
he was emperor to be inshi, or ex-emperor's officials. His patriarchal
authority extended beyond the strict confines of the dynastic group
to include his offspring who had been made Minamoto nobles. The
7 Nihon koki, in Kokushi taikei, vol. 3 (Tokyo:Yoshikawa kobunkan, 1965), p. 85, Konin 1/9/10.
8 Hurst, Inset, pp. 61-62.
Cambridge Histories Online © Cambridge University Press, 2008
582
INSEI
practice of dynastic shedding helped to limit claimants to the throne
and curtail demands upon the imperial treasury. Saga had fiftyseven children; he gave many of them the surname Minamoto,
formally cutting them off from the royal house. Nevertheless, he
summoned them to participate with imperial house members in ceremonies at which he presided as patriarch. Saga can perhaps be seen
as the prototype of the abdicated emperors of some two and a half
centuries later.
After Saga, retired emperors developed separate institutions for issuing documents and administering the needs of a growing retinue
of personal retainers. The first extant ex-sovereign decree was issued
by Uda in g28. 9 The ex-sovereign's private office (in-no-cho), with a
complement of subordinate bureaus and officials who could issue
decrees in the name of the ex-emperor was in existence at least by
the time of En'yu's abdication in 984.10
Early Heian period ex-emperors, especially Saga, Uda, and En'yu,
were able to assert themselves forcefully in political affairs. As heads
of the imperial house, they exercised paternal authority over sons
and grandsons to influence decisions favorable to the dynasty. This
was especially true at times when the emerging Fujiwara clan regency was weak, or when there was no regent, as during the reign of
Daigo when Uda was ex-sovereign. But it was true even when a
strong Fujiwara regent was appointed. Kaneie, for example, lamented the fact that ex-emperor En'yu's requests were "weighty and
difficult to refuse."11
This is a telling statement, striking at the heart of the competition
between the imperial house and the Fujiwara regent's line (sekkanke). The major development of early and mid-Heian was the establishment of a permanent Fujiwara regency that dominated the
throne through marriage relationships. Disparity of status made it
impossible for the Fujiwara to aspire to the position of sovereign, but
they enjoyed an interest in the dynasty in the form of a permanent
regency supported by intermarriage, which over time reduced the
solidarity of the imperial house as a cohesive kinship group. Under
the marriage customs of the mid-Heian period,12 an imperial prince
9 Todaijiyoroku, in Zoku zoku gunsho ruiju (Tokyo: Gunsho ruiju kankokai, 1909), vol. 11, pp.
160-61, Encho 6/8/28.
10 Hurst, Insei, pp. 85-87.
11 Shoyuki, in Tokyo daigaku shiryo hensanjo, comp., Dai Nihon kokiroku (Tokyo: Iwanami
shoten, 1952), series 10, vol. 1, p. 162, Eiso 1/2/21.
12 William McCullough, "Japanese Marriage Institutions in the Heian Period," Harvard Journal ofAsiatic Studies 27 (1967): 103-67.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SANJO AND THE PRELUDE TO INSEI
583
born of a Fujiwara woman was reared in the residence of his grandfather or uncle. Strong ties of affection developed. When the child
became emperor, it was natural that he would be dominated by his
Fujiwara kin. However, Fujiwara clan heads were aware that their regency was different from a royal regency.
After En'yu's death, there were no abdicated sovereigns of any political importance until Go-Sanjo's time. This reduction of the power
of retired emperors corresponds to the apogee of Fujiwara regency
control, when Michinaga's daughters produced several sons who became emperor. Michinaga's son Yorimichi, however, was not blessed
with daughters who produced potential heirs to the throne, and thus
faced the distinct possibility of the reemergence of a strong exemperor who might exercise familial power over the reigning sovereign. During Yorimichi's long tenure as regent, he never permitted
an emperor the opportunity to evolve into a forceful retired sovereign; each abdicated just prior to death.
In effect, two forms of regency developed during early and midHeian times. The Fujiwara clan controlled the right to provide imperial consorts and thus aided by the marriage practices of the day, exercised regency power through the maternal side. But an ex-emperor,
as head of the royal house, could also exercise a kind of regency
power from the paternal side. The marriage connection made the Fujiwara regency innately fragile. Yorimichi (and probably Kaneie and
other clan chieftains as well), realizing that a royal regency exercised
by an ex-emperor was potentially more powerful than a nonroyal regency, guarded against early abdication.
GO-SANJO AND THE PRELUDE TO INSEI,
1068-1073
Emperor Go-Sanjo (1034-73) came to the throne in 1068. He was
the first emperor since Uda, eleven reigns and 170 years earlier,
whose mother was not a Fujiwara. Go-Sanjd, the second son of Emperor Go-Suzaku, was born in 1034 when his father was still crown
prince. His mother, Princess Teishi (better known asY6meimon-in),
was a daughter of the late emperor Sanjo (see Figures 9.1 and 1.1).
The possibility of Go-Sanjo's accession threatened Fujiwara domination of the imperial position; Regent Yorimichi (992-1074) tried to
prevent it. The primary reason for Yorimichi's worry was the dearth
of possible Fujiwara heirs to the throne. His sister Shoshi was
mother of both emperors Go-Ichij5 and Go-Suzaku, and another
sister, Kishi, had borne Go-Reizei. While Yorimichi enjoyed extenCambridge Histories Online © Cambridge University Press, 2008
584
INSEI
Fujiwara no Kinnari
Minamoto no Motohira
I
Moshi
(adopted by Fujiwara
noYoshinobu)
Atsuko
1
I
— GO-SANJOI'1 = = Motoko
Morofusa
Akifusa
SHIRAKAWA!2)
.
— Kenshi
(adopted by Fujiwara
HORIKAWA!3! Reishi
Fujiwara no Tokushi :
(Bifukumon-in)
: TOBAW :
Teishi
Atsufumi
(Ikuhomon—in)
- Fujiwara no Shoshi
(Taikemmon-in)
I
KONOEl6!
Fujiwara no Ishi = j = GO-SHIRAKAWA!7! = p = Taira no Shigeko
(Kenshummon-in)
SUTOKUI5!
TAKAKURA! 10 )
ROKUJOI9)
I
ANTOKUl11!
I
GO-TOBAI12!
Key: Names in small capital letters = emperors
Double line = marriage or liaison
Single line = parental relationship
Figure 9.1. Genealogy of emperors during the Inset period.
sive power as maternal uncle to several rulers, his position was much
less secure than it would have been had his own daughters produced
sons. Two of his daughters were married to Go-Reizei, but no
princes were born.
In 1044, when Go-Suzaku became seriously ill and desired to abdicate in favor of Go-Reizei, Yorimichi faced a crisis: there was no
sekkanke heir as candidate for crown prince. It was customary to
name a crown prince at the time of accession. Go-Suzaku wanted
Go-Sanj6 (then Prince Takahito) to be appointed because Go-Reizei
had no male children. Having no alternative, and because the wishes
of a dying emperor were not easily disregarded, Yorimichi grudgingly
consented to this arrangement.
Go-Sanj6 was crown prince for an unprecedented twenty-three
years, during which time he was the focus of Yorimichi's animosity.
There may even have been an attempt upon his life.13 In the past, sev13 Kawaguchi Hisao, Oe no Masafusa (Tokyo:Yoshikawa kobunkan, 1968), p. 62.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SANJ5 AND THE PRELUDE TO INSEI
585
eral crown princes had been forced into resignation, victimized by
plots that resulted in their removal, or killed. Less than a half-century
earlier, for example, Michinaga had forced Go-Sanjd's maternal
uncle Prince Atsuakira to step down. Yorimichi stalled Go-Reizei's
abdication as long as possible, hoping for the birth of a Fujiwara
prince. But when Go-Reizei died in 1068, he could not block the accession of the thirty-three-year-old Go-Sanjo. Yorimichi's frustration
is well mirrored in his response: he yielded the regency to his younger
brother Norimichi and retired permanently to his villa at Uji.
Thus, the first emperor in almost two centuries not hampered by
close Fujiwara relatives began his short, active, and exemplary rule.
In their accession edicts emperors since Emperor Mommu, who
reigned at the start of the Nara period, had emphasized that they
would rule wisely, and that subjects should serve faithfully. But the
Japanese preference for indirect rule, with a separation of formal authority and actual power, dictated that sovereigns reigned but only
infrequently ruled. The present-day Japanese emperor, like his
British counterpart, for example, performs a legitimizing and symbolic role. Unlike his British counterpart, however, the Heizei emperor's reign is in harmony with that of most of his predecessors.
And yet, as we have seen, the Japanese regarded direct imperial
rule (shinsei) without the intervention of prince regents, imperial regents, ex-emperors, or shoguns, as the ideal pattern. Therefore when
the opportunity presented itself, emperors sought to enter into the
political arena. There had been few such opportunities, especially
since the institution of the Fujiwara regency virtually relegated emperors to a permanent state of minority. But Go-Sanjo was in his
prime. He had been educated in the Chinese classical tradition and
assiduously prepared for ruling. There was no Fujiwara hindrance to
his rule. Norimichi was regent, but, lacking the strong maternal link,
he could not restrain a determined emperor. Go-Sanjo was able to
exercise a vigorous imperial rule quite unusual by Japanese standards. He was in fact much closer to the Chinese model than emperors Uda and Daigo (who reigned 887-930), who had been especially regarded by later generations as representing a "golden age" of
imperial strength.
During the nearly five years of Go-Sanjo's reign, he was actively
involved in government. This involvement extended from a deep
concern with appointments in the official bureaucracy to attempts to
reform the economic woes that plagued Heian society. The Kojidan,
an early-thirteenth-century compilation, praises his reign as a period
Cambridge Histories Online © Cambridge University Press, 2008
586
INSEI
of "just rule." Hoping to break the monopoly of the highest-ranking
nobles, especially sekkanke nobles, over government posts, Go-Sanjo
sought the participation of middle-ranking courtiers, provincial governors, and scholars. He also attempted to open the way for imperial offshoot clan members, especially the Murakami and Daigo lineages of the Minamoto, to move into the kugyo (nobles of Third
Rank and above) council so long dominated by the sekkanke. Since
Yorimichi had actively attempted to block Go-Sanjo's accession, the
thrust of Go-Sanjo's rule was directed toward reducing the power of
Yorimichi and the sekkanke.
This is obvious from an investigation of the appointment practices
during Go-Sanjo's reign/4 When Yorimichi was regent in the reigns of
Go-Ichijo and Go-Suzaku (1016-45), the highest positions in the
kugyo were dominated by the sekkanke. Fujiwara clan members held
a virtual stranglehold over kugyo membership, never dipping below
an 80 percent majority. During Go-Reizei's reign, as the succession
problem became acute, Yorimichi's power began to slip. Previously,
there had never been more than four or five kugyo of the Minamoto
clan, particularly the Murakami branch (Murakami Genji, see Figure
9.2); now there were seven or eight. In 1067, Go-Reizei's last year,
there were ten Minamoto kugyo, more than one-third of the total.
{Kugyo membership averaged about twenty-eight during the reigns of
Go-Ichijo, Go-Suzaku, Go-Reizei, and Go-Sanjo.) Moreover, dissension broke out between Yorimichi and his younger brother Yoshinobu,
the house member most closely related to Go-Sanjo, and Yorimichi's
ability to control the sekkanke began to weaken.15
The Council of State gained in importance in the mid- and late
Heian periods. No longer restricted to its original statutory role of
deliberative council, it was transformed into a decision-making
body. At the height of Fujiwara power, the decisions of the Council
had been implemented by the regent acting on the sovereign's behalf, but from Go-Sanjd's era, the regent was merely one of the participants in the Council's deliberations. The ex-emperor normally
implemented decisions. Despite the power of Fujiwara regents and
ex-emperors, however, the noble council had never become a body
composed solely of their private officials.16 It remained the major de14 Hurst, Inset, p. 108, table 2, based on Kugyo bunin.
15 Yoshinobu was Michinaga's fifth and last son; since his daughter was maried to Go-Sanjo,
he stood to gain by the accession of this emperor.
16 Hashimoto Yoshihiko, "Kizoku seiken no seiji kozo," in Iwanami kbza Nihon rekishi, Kodai
vol. 4 (Tokyo: Iwanami shoten, 1975), pp. 26-27.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SANJO AND THE PRELUDE TO
587
INSEI
Morofusa
(1008?-77)
I
I
Reishi
(wife of
Fujiwara no
Morozane,
mother of
Moromichi)
I
Moroko
(wife of
Fujiwara
no Tadazane,
mother of
Tadamichi)
I
daughter
(mother of
Bifukumon—in)
I
Kenshi
(consort of
Shirakawa)
T
Ninkan
I
Masatoshi
T
I
Akifusa
(1037-94)
(same mother
as Toshifusa)
Morotada
Shokan
T
T
Masazane
(1059-1127)
Akinaka
Shokaku
Toshifusa
(1035-1121)
(mother was
daughter
of Fujiwara no
Michinaga)
T
Moroshige
T
Morotoshi
Masasada
(1094-1162)
T
Morotoki
Moroyori
Akimichi
(1081-1122)
I
Masamichi
(1118-75)
I
Michichika
(1149-1202)
Figure 9.2. Genealogy of the Murakami Genji (Minamoto).
cision-making organ of the state in which consensus decided affairs.
Consequently Go-Sanj6 and the ex-emperors of the inset period attempted to control the noble council by cultivating its members.
The Fujiwara main line became less influential in the noble council when, during Go-Sanjo's reign, Minamoto courtiers not only
composed one-third of the kugyo membership but also held the
most important ministerial positions and the major counselor posts.
These courtiers also tended to cooperate closely with Go-Sanjo as
an "imperial faction."17 With the sekkanke fragmented and unable to
dominate the noble council as in the past, Go-Sanj6 was relatively
free to carry out policies of his own making with little opposition.
Among the courtiers closest to Go-Sanjo, three in particular are
17 Hayashiya Tatsusaburo, Zusetsu Nihon bunkashi taikei (Tokyo: Shogakukan, 1957), vol. 5,
Heianjidai 2, pp. 63-64.
Cambridge Histories Online © Cambridge University Press, 2008
588
INSEI
worthy of mention: Oe no Masafusa, Fujiwara noTamefusa, and Fujiwara no Korefusa. These men became known to later historians as
the "former three fusas" in contrast to the "latter three fusas" who
served Emperor Go-Daigo in the fourteenth century. Although only
Korefusa became a kugyd during Go-Sanjo's lifetime, these three
middle-ranking courtiers served the emperor loyally throughout his
life. Their career patterns are instructive because they represent the
emergence of a type of courtier who, in the Insei period, would become the founder of a client family of the imperial house.
All three men served in the Chamberlains' office (Kurododokoro)
and were thus intimately involved in governmental affairs. Fujiwara
no Sanemasa had been Go-Sanjo's tutor when he was crown prince
and was influential as a Confucian adviser; he was also appointed to
serve as teacher for Go-Sanjo's son, Prince Sadahito (future Emperor Shirakawa). After Go-Sanjo's abdication, these men became
superintendents (betio) in his in-no-cho. In sum, they spent their entire career in close service to the sovereign. Go-Sanjo was able to rely
on them because none was closely connected with the Fujiwara regent's house. Tamefusa was a member of the Kanjuji branch of the
Fujiwara; Korefusa, although of the Northern branch, was not from
the sekkanke; and Masafusa was from the minor Oe clan, which traditionally trained its sons as Confucian scholars. Sanemasa, too, was
of non-sekkanke Fujiwara lineage.
Thus, Go-Sanjo enjoyed an unusual situation. He was an active
emperor in the prime of life. There was neither a powerful Fujiwara
regent nor a retired sovereign to exercise any "advisory" function.
Sekkanke control of the noble council was greatly weakened, and
there was an "imperial faction" operating there. His closest advisers
were non-sekkanke courtiers who seem to have been predisposed in
favor of a strong emperor. Such a background allowed Go-Sanjo to
deal forcefully with political issues.
There is a paucity of source materials for a study of Go-Sanjo's
reign, but it appears that he focused mainly on shoring up the declining economic foundations of the imperial state system. He
adopted concrete policies of shoen regulation, standardization of the
official masu dry measures used in taxation, price control and market regulation, and quality control of silk and hemp cloth.
Enacted in 1069 (Enkyu 1), Go-Sanjo's shoen (estate) regulation is
known as the Enkyu shoen regulation ordinance. For more than a
century there had been criticism of shoen policy as well as several
Cambridge Histories Online © Cambridge University Press, 2008
GO-SANJO AND THE PRELUDE TO INSEI
589
attempts at regulation.18 Earlier attempts had not been noticeably
effective, however, as the continued issuance of legislation amply
demonstrates. The major obstacle to effective regulation of the
growth of these private, tax-exempt landholdings was the lack of an
enforcement mechanism. To be fully privatized for long periods of
time, shden required certification from the major organs of the central government. In effect, that meant cooperation or collusion with
the major nobles in the central government, the people supposedly
responsible for maintaining the viability of taxable lands. Tenure
over shden holdings could be obtained from provincial governors,
but such lands were inherently insecure because governors rotated
every four years, and new governors often disregarded their predecessors' arrangements.
Go-Sanj6's estate regulation, though from all evidence the most
successful attempt of its kind, did not challenge the existence of estates per se. On the contrary, like all other regulatory attempts, it
confirmed the right of persons or institutions to possess private
holdings, merely defining more strictly the legal ground rules by
which estates could be acquired. The relative success of the Enkyu
legislation was due to the establishment in 1069 of a special government commission to handle regulation, the Office for the Investigation of Estate Documents (Kiroku shden kenkeijo). This kirokusho, or
Records Office, as it is better known, was charged with carrying out
the ordinance that stipulated: (1) all estates established after 1045
(the date of Go-Reizei's accession) were to be declared illegal and
confiscated; and (2) older estates with improper documentation, or
estates that hampered the conduct of provincial administration,
would suffer a similar fate.
The main source of information on the Enkyu regulation is a document concerning a subtemple of the Iwashimizu Hachiman Shrine.
Of thirty-seven estates belonging to the subtemple Gokokuji for
which documents were submitted, twenty-four were recognized as
adhering to the legally prescribed form; but thirteen, some 30 percent of the temple's portfolio, were confiscated.19 A few conclusions
can be drawn from this document. First, the Records Office had established more rigorous standards for confiscation of illegal estates;
18 See Hurst, Insei, pp. 110-14, an<^ Elizabeth Sato, "Oyama Estate and Inset Land Policies,"
Monumenta Nipponica, 34, 1 (1979): 73-99.
19 Takeuchi Rizo, comp., Heian ibun, 15 vols. (Tokyo: Tokyodo, 1963-74), vol. 3, pp. 10921107, doc. 1083, Enkyu 4/9/5.
Cambridge Histories Online © Cambridge University Press, 2008
59O
INSEI
previous enforcement attempts had been entrusted to individual
governors. Second, the primary targets were estates with weak documentation. Third, also targeted were lands for which the conditions
of tax liability were unclear, that is, shden whose paddy fields were
not clearly fixed {ukimen no sho, or "floating field estates") or land on
which control of paddy fields and labor was shared or split between
estate holder and provincial government. Even estates recognized as
legal by the Records Office suffered confiscation of lands that had
been added to their original exemption since 1045.2O
The officials known to have served in the Records Office, men like
Oe no Masafusa, who is regarded as the author of the legislation, Fujiwara no Tamefusa, and Otsuki no Takanobu, were all close associates of the emperor and members of the provincial governor (zuryo)
class with no connection to the Fujiwara regency line. This suggests
it was the zuryo who urged estate control and that the shden of the
Fujiwara main house, perhaps the largest estate holder, were targets
for confiscation.
Provincial governor involvement in the estate process is, however,
somewhat puzzling. The Enkyu regulation ordinance did apply to
weakly documented shden exempted by zuryo, called kokumen no sho
(provincially exempt estates), that had been established since 1045.
On the other hand, a crucial result of the work of the Records Office
is that provincially exempt estates with clear documentation were
recognized as legal, ranking with estates that possessed central government documents. Thus we have a policy initiated by zuryo seeking to limit the growth of estates due to the increased power of the
same zuryo class, but at the same time increasing the actual power
of zuryo to exempt estates in their own provinces. This seeming contradiction derives from the dual nature of the zuryo as both public
official and private courtier: it was in the governor's interest to control estates and increase his public income while in office, and
equally in his interest to permit shden establishment at the end of his
tenure and reap private benefits. The Enkyii effort represented the
attempts of conscientious zuryo to restrain the corrupt actions of
other zuryo, but it was necessarily less than thorough.21
It is difficult to document how the Records Office dealt with estates of the Fujiwara main line. It was once believed thatYorimichi's
20 Wakita Haruko, Nikon chusei shogyo hattatsushi no kenkyu (Tokyo: Ochanomizu shobo,
1969), pp. 99-101. See alsoYasuda Motohisa, Insei to Heishi, vol. 7 of Nihon no rekishi, 33
vols. (Tokyo: Shogakukan, 1973-77), PP- 33-3521 Wakita, Nihon chltsei shogyo hattatsushi, pp. 102, 108.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SANJO AND THE PRELUDE TO INSEI
591
strong objections to the legislation forced Go-Sanjo to exempt the
sekkanke from the process, but it appears thatYorimichi actually did
submit documents to the Records Office, and that the house lost estates. At least one source refers to the confiscation of the Doi-nosho, an estate belonging to Fujiwara no Morozane.22 Of course, the
major estate owners were unhappy with Go-Sanjo's attempt to increase national finances at the expense of private holdings. The great
Todaiji temple of Nara, for example, was half a year late in submitting its paperwork to the Records Office.
Although the purpose of regulation was to confiscate lands and return them to a public, that is, tax-paying status, Go-Sanjo seems to
have had other intentions for some of these lands. Imperial edict
fields and the holdings attached to the goin, retirement palaces of abdicated sovereigns, did not provide extensive royal income. With the
decline of public revenues, all courtier families experienced a loss of
income, but the imperial house suffered most. Unlike the sekkanke
and other noble houses, it was unable to accumulate much privately
held land as a source of income. Just why this was so is not clear.
There seems to have been more of an institutional, almost psychological, restraint than a legal stricture. Perhaps it was unseemly for
the emperor, as head of a state that expressly rejected private ownership in favor of state-owned taxable lands, to establish shoen.
But Go-Sanjo took actions that paved the way for the acquisition
of private landed wealth by the imperial house. From among the
confiscated estates, Go-Sanjo established new chokushiden or "edict
fields" (the exact number is unknown) that were virtually indistinguishable from the shoen belonging to courtiers and religious institutions.23 This demonstrates the dual purpose of Go-Sanjo's estate
regulation policy: to curtail illegal estates belonging to other proprietors and to acquire private holdings for the imperial house. The effect of Go-Sanjo's regulation of estates was momentous. By establishing a bureaucracy that demanded the submission of shoen
charters and then ruled on their legitimacy, Go-Sanjo in effect legitimized those estates whose documentation was in order. As Thomas
Keirstead has phrased it, the emperor's actions represent an "emblem of the reordering of the realm that produced the shoen system."
The Records Office - "Go-Sanjo's archive" - went beyond simple
22 Go-Nijd Moromichi-ki, in Dai Nihon kokiroku (Tokyo: Iwanami shoten, 1951) vol. 2, p. 197,
Kanji 5/12/12.
23 OkunoTakahiro, "Go-Sanjo tenno goryo," in Nihon rekishi daijiten (Tokyo: Kawade shobo,
1956-61), vol. 4, p. 525.
Cambridge Histories Online © Cambridge University Press, 2008
592
INSEI
legitimation of estates and, by embedding the estate within a public
domain and private land system circumscribed by maps, land surveys, charters and the like, essentially "established a new syntax of
landholding . . . [I]n Go-Sanjo's archive we can perceive the first
condition of the possibility of the estate system as system."24
Go-Sanjo's other economic policies are less well known and certainly not as far-reaching as his shoen regulation, but they are related.
Not only had imperial control of the country's paddy fields broken
down, there was also laxity in the application of uniform standards
throughout the country. For example, the size of the masu, a measuring unit established in Nara times, was no longer standard but
varied from place to place. It sometimes also varied according to
purpose: a shoen owner might use one size masu when measuring
payments, and another, somewhat larger, when receiving rents. To
overcome the economic problems caused by such inconsistency, GoSanjo issued an edict in 1072 that determined the official size of the
masu to be used throughout the realm. The measure became known
as the "imperial edict measuring box of the Enkyvi era" (Enkyu no
senji masu). The masu was important as the basis for the system of
equal provincial levies (ikkoku heikin no kayaku) initiated in GoSanjS's era. These levies were made on all lands, public or shoen, to
pay for such national projects as the reconstruction of the palace or
the Ise Shrine.25 This unit of measure remained in use nationwide
through the Kamakura period.
In 1070 there seems to have been another attempt at standardization: an ordinance was issued controlling the quality of silk and
hemp cloth. Both materials had long been in use as media of exchange, and most likely this was another attempt to stabilize economic conditions by fixing the rate of exchange between these materials and rice, the official medium of payment. In 1072, another law
was issued covering prices and other market matters. Since the establishment of the capital in the late eighth century, commercial activity had been increasing and government ability to regulate the exchange of goods and services had been decreasing. Although the
terms of this law are not known, it was probably an attempt to bring
the activities of merchants under greater official control.
Early in 1073 Go-Sanjo abdicated in favor of his son Sadahito
(Emperor Shirakawa). On the same day, Prince Sanehito, Go24 Thomas Keirstead, The Geography of Power in Medieval Japan (Princeton, N.J.: Princeton
University Press, 1992), pp. 18-19.
25 Yasuda, Insei to Heishi, pp. 39-40.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SANJO AND THE PRELUDE TO INSEI
593
Sanjo's second son, was named crown prince. Since the time of the
Buddhist priest and historian Jien, writing in the thirteenth century,
it was generally assumed that Go-Sanjo's motive for retirement was
to set up a system of rule by abdicated emperors (inset). No evidence
for this theory exists, and recently scholars have come to reject it. A
consciously worked out, elaborate, institutional scheme probably
never occurred to Go-Sanj6. But to say that Go-Sanjo did not attempt "to establish the inset" is not to deny that he fully intended to
exercise a guiding influence in government. The history of abdicated
sovereigns prior to the heyday of Fujiwara sekkanke power must have
given him ample expectations for an active postretirement career.
The timing of Go-Sanj5's abdication may have had something to
do with the state of his health, although it is hard to judge from the
sketchy information available. In a document Go-Sanjo dedicated to
the Iwashimizu Hachiman Shrine, which he visited in the second
month of 1073, he mentions having been ill during the preceding
winter and again in the spring, and seeks the deity's aid in regaining
his health.26 In the third month an amnesty was proclaimed to benefit his health. In the fourth month he took the tonsure; one source
gives the reason as sudden illness.27 One can tentatively conclude
that while Go-Sanjo was ill near the end of his reign, it was no
chronic illness, and that for his four years on the throne he was relatively healthy. After his abdication, however, at least from the time
of the amnesty proclamation, Go-Sanjo was seriously ill. However,
illness alone probably does not explain his decision to retire.
The most plausible explanation for Go-Sanjo's abdication is his
concern for royal succession. It is instructive to compare the designation of crown princes at the accessions of Go-Sanj5 and Shirakawa. Go-Sanjo's enthronement ceremony took place on the twentyfirst day of the seventh lunar month in 1068; Shirakawa was not
named crown prince until some nine months later, even though he
was Go-Sanjo's eldest son and already fifteen years old. By contrast,
when Go-Sanjo abdicated in favor of Shirakawa, he appointed his
own one-year-old son, Sanehito, as crown prince on the same day.
The identity of the mothers of the two princes reveals the reason behind Go-Sanjo's actions (see Figure 9.3).
Shirakawa's mother was Fujiwara no Moshi, the adopted daughter of the sekkanke courtier Fujiwara noYoshinobu. Despite the an26 Heian ibun, vol. 3, pp. 1113-14, doc. 1091, Enkyu 5/2/2.
27 Fusoryakki, vol. 12 in Kokushi taikei (Tokyo: Yoshikawa kobunkan, 1965), pp. 315, Enkyu
5/3/18.
Cambridge Histories Online © Cambridge University Press, 2008
INSBI
594
Fujiwara no Kinnari
Moshi :
Minamoto no Motohira
GOSANJO =
Motoko
I
Princess Atsuko
SHIRAKAWA
Sukehito
(San no miya)
Sanehito
Key: Names in small capital letters = emperors
Double line = marriage or liaison
Single line = parental relationship
Figure 9.3. Family relations of Go-Sanjo.
imosity between Yoshinobu and Yorimichi's senior line, there remained the inherent danger that an influential Fujiwara regency
might reemerge. Sanehito's mother was Motoko, daughter of Minamoto no Motohira. Go-Sanjo also had a consort whose father was
Go-Ichij5; he decided to wait almost a year before confirming Shirakawa as crown prince because he hoped one of these non-sekkanke
consorts might present him with an heir.
When Prince Sanehito was born in 1071, he was named an imperial prince within six months. When his father abdicated in favor of
Shirakawa, Sanehito was named crown prince, although he had not
yet reached his second birthday. The difference in Go-Sanjo's handling of the succession issue leaves little doubt as to his abdication
motive. Furthermore, when Motoko gave birth to a second son,
Prince Sukehito, in 1073, Go-Sanjo made it clear that he expected
succession to follow in a fraternal line from Sanehito to Sukehito.
Thus it appears that Emperor Go-Sanjd abdicated in order to
guarantee the future direction of imperial succession and to protect
the imperial house from a possible resurgence of Fujiwara regency
power. He was able to enthrone his eldest son, who did have tenuous sekkanke connections, but he appointed a Minamoto heir as next
successor and indicated another Minamoto offspring to carry on
after that. He was laying out a long-term succession plan in which
the emperor would be unencumbered by Fujiwara maternal connections. The history of Japanese succession practices had demonstrated that abdication was the surest means to guarantee one's succession plans. As a young retired emperor (Go-Sanjo was only
thirty-eight at abdication), he could certainly have expected to exerCambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INSEI
595
cise familial supervision over his son Shirakawa in the manner of
earlier Heian ex-emperors, indeed to participate in his kingship.
While we may say the Go-Sanjo did not conjure up the idea of insei,
it is certainly possible to see in his reign and abdication the prelude
to such a development.
SHIRAKAWA AND THE NORMALIZATION OF INSEI,
IO73-H29
Go-Sanj6's son Shirakawa (1053-1129, r. 1073-87) died at age seventy-six after living one of the most active and involved lives of any
Japanese sovereign of the premodern period. It was during Shirakawa's lifetime that the inset became an almost formalized political
norm, if by insei we mean a political situation in which the retired
sovereign, operating as the head of the imperial kin group and "borrowing" the legitimacy of the titular sovereign, served as the de facto
arbiter of political and economic interests of the court. It is, however, not so easy to say that Shirakawa's abdication was motivated by
a desire to establish such a system.
Shirakawa was nineteen years of age when Go-Sanjo abdicated in
his favor early in 1073, presenting him with a succession pattern that
effectively bypassed his future progeny. But fate gave Shirakawa an
opportunity to amend his father's scheme. Shirakawa's abdication in
1087 was not intended to establish a new political system; it had the
more limited political goal of bringing about the accession of his
own line instead of perpetuating his half-brother's. And yet other
factors produced a situation in which Shirakawa emerged as the
major political force in the realm. Whether or not he had originally
conceived of an insei-like arrangement, by his death in 1129, the pattern of an actively ruling retired sovereign was firmly established.
Just what was the nature of that system, and how did it develop so
extensively during Shirakawa's reign?
Shirakawa reigned for fourteen years in much the same manner as
his father had, exercising considerable personal authority because of
the weak position of the Fujiwara regency. Morozane served as regent, and although the sekkanke remained the most powerful nonroyal house, he was unable directly to affect imperial action. Instead,
Shirakawa exercised direct personal rule, employing the same kind
of bureaucratic personnel practices as Go-Sanjo had. The "former
three fusas" and Fujiwara no Sanemasa were among his chief advisers; all of them moved into the ranks of the kugyo noble council.
Cambridge Histories Online © Cambridge University Press, 2008
596
INSEI
The Minamoto also continued to increase their influence at the
expense of Fujiwara courtiers. This did not escape the attention of
Shirakawa's contemporaries Fujiwara (Nakamikado) no Munetada,
whose diary Chiiyuki is one of the major sources of information for
the period, comments several times on the advance of the Murakami
Genji. In 1069 the ranking Murakami Genji, Morofusa, became Minister of the Right and the next year Minister of the Left, second only
to Morozane. By 1083 his sons held both the Left and Right ministerial positions, a rare occurrence indeed. Some eleven years later,
Munetada warned the sekkanke to be wary, since both ministers and
the two major captains were Minamoto, as were three of five counselors,fiveof six captains of the guard, and four of seven controllers.28
Two political matters that seem to have greatly concerned Shirakawa were the continuing problem of shden regulation and the increasing frequency of lawless behavior on the part of soldier monks
of the great capital area temples. Shirakawa seems to have inherited
his father's desire to limit the illegal acquisition of estate holdings; he
issued a regulation edict that affected at least two Todaiji estates in
Mino.29 That he was forced to do so indicates that even Go-Sanjo's
thorough legislation was insufficient to deal with the problem. Considering the number of shden that were being established, by both
legal and illegal means, it seems that Shirakawa was no more successful than his father.
On a number of occasions unruly monks, the so-called akuso from
Mount Hiei and other great Buddhist establishments, threatened the
peace of the capital. In 1079, for example, an ecclesiastical appointment precipitated a disturbance by more than a thousand armed
Mount Hiei monks and the court was forced to send warriors
against them. In 1081 an old dispute between Mount Hiei and
Onjoji broke out again, resulting in considerable fire damage to both
temples. (See Chapter 7.) There was such unrest that when Shirakawa visited the Iwashimizu Hachiman shrine south of the capital,
he required the protection of an entourage of warriors led by Minamoto noYoshiie. Akuso outrages proliferated, ultimately becoming
the nobility's chief cause for concern. Shirakawa and his successor
ex-emperors were forced to adopt new forms of military organization to cope with them.
During his reign, Shirakawa's main worry, however, was the poli28 Chuyuki, vols. 9—15 of Zoho hiryo taisei, 48 vols. (Kyoto: Rinsen shoten, 1965-89), vol. 1, p.
108, Kanji 7/12/27.
29 Heian ibun, vol. 3, pp. 1128-29, doc. 1118, Joho 2/8/23.
Cambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INS El
Fujiwara noYoshinaga
Michiko
597
Minamoto no Akifusa
=
— Kenshi
(adopted by Fujiwara
no Morozane)
SHIRAKAWA '•
I
\
\
I
Teishi
HORIKAWA
I
Reishi
Teishi
(Ikuhomon-in)
I
Atsufumi
TOBA
Key: Names in small capital letters = emperors
Double line = marriage or liaison
Single line = parental relationship
Figure 9.4. Family relations of Shirakawa.
tics of succession. Shirakawa had been presented with a predetermined succession plan, which gave precedence to his half brothers
Sanehito and Sukehito, sons of Go-Sanjo's Murakami Genji consort
Motoko. When in 1075 his own consort Kenshi gave birth to a
prince, Shirakawa could foresee the possibility of royal succession
passing to his own offspring; but the boy died a few years later in an
epidemic that swept the land. In 1079, Kenshi gave birth to Prince
Taruhito, the future Emperor Horikawa, and once again the possibility of altering his father's plan presented itself to Shirakawa (see
Figure 9.4). No contemporary records even suggest that Shirakawa
originally planned to overturn his father's succession plan. Extrapolation from later actions suggests, however, that succession was
much on Shirakawa's mind. Fate also played a hand in the matter
when in 1085, at the age of fourteen, Crown Prince Sanehito died of
smallpox.
According to Go-Sanjo's instructions, Shirakawa should have immediately invested Prince Sukehito as crown prince, but he made no
decision for almost a year. Shirakawa was certainly considering the
appointment of his own young son, but there were obstacles. His father's injunction was clear. Furthermore, Go-Sanj5's mother,Yomeimon-in, was still alive and not likely to look favorably upon Shirakawa ousting her grandson. So he procrastinated.
By the eleventh month of 1086, Shirakawa finally resolved the
dilemma. He made Taruhito crown prince then abdicated in his
favor. Shirakawa had chosen to shift succession to his own line, cutting off Prince Sukehito from the throne. He had already constructed
Cambridge Histories Online © Cambridge University Press, 2008
598
INSEI
a magnificent retirement palace south of Heian. To at least one observer, the movement of men and supplies appeared to resemble the
move of the capital. Many historians consider the building of this
Toba-dono complex tantamount to establishment of the inset.30
Given that there was no respectable palace for a retired sovereign
to use, however, it is not surprising that Shirakawa constructed one.
In the late 1070s and early 1080s there had been numerous fires at
the great mansions of the imperial family, the sekkanke, and other
kugyo. Shirakawa was constantly moving between one temporary
palace, normally the mansion of the regent or other major noble,
and another. If he had intended to dominate politics in retirement,
it would have been more logical to build a palace in the city close to
the center of activity, as, for example, Saga had. The Toba-dono
seems never to have been the center of his activities. He was not even
named posthumously for this palace, but rather for the Shirakawain, a palace on the eastern fringes of the city. Shirakawa seems to
have utilized the Oi-dono more frequently as a palace; his trips to
Toba were always termed "visits." So the establishment of a retirement palace at Toba-dono did not necessarily signal the raising of
the curtain on the insei.
In fact, for some years after his retirement Shirakawa was not very
active politically. His assumption of a major governmental role appears to have been as much the result of circumstance as of a plan.
Once he decided to devote himself to rule in abdication, Shirakawa
was very willful and powerful, but the evidence does not indicate
that he had been that way from the outset. For example, shortly after
Shirakawa's abdication, the court was forced to deal with the rising
power of Minamoto no Yoshiie (1039-1106). The ex-emperor does
not seem to have played an active role. Yoshiie had emerged from the
Earlier Nine Years'War (1051-62) as the leading warrior in the land,
chieftain of the Seiwa Genji. During Shirakawa's reign, he and his
vassals had frequently been employed to guard the emperor on outings. After the Later Three Years'War (1083-87), Yoshiie's reputation
spread, and local landholders began to commend shoen to him. At
this point the nobles began to turn against Yoshiie, and the regents
Morozane and Moromichi took the lead, not Shirakawa.
Traditionally the sekkanke had relied on the military power of the
Seiwa Genji, but now they turned on Yoshiie. The nobility had long
regarded the warrior as a samurai ("one who serves"), but Yoshiie
30 Murayama Shuichi, Heian-kyo (Tokyo: Shibundo, 1957), pp. 146-76.
Cambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INSEI
599
was threatening to compete with them on equal terms. Morozane,
Moromichi, and the nobles began to support Yoshitsuna against his
elder brother Yoshiie and, in 1091, precipitated a quarrel over leadership of the Minamoto warrior league. Afraid that Yoshiie would
dispatch troops, the court forbade his men from entering Heian and
prohibited local landholders from commending estates to him. Thus
in the 1090s the Fujiwara regents Morozane and Moromichi led the
court nobility - though Shirakawa may have been in agreement - in
taking concrete actions against a threat to their collegial leadership.
Emperor Horikawa reigned from 1087 to 1107 and, according to
contemporary documents, took his duties, especially appointments,
seriously and was judicious in political matters.31 Fujiwara no Nagako's description of Horikawa portrays him as a sensitive and tender man, a considerable flautist and poet who was the focus of a poetic circle that included a wide range of people at court, not the least
among whom were his Minamoto maternal kinsmen.32 Though Horikawa was only seven at his succession and passed away at twentyeight, Shirakawa does not seem to have exercised much, if any, political authority during Horikawa's reign.
As we have seen, the Fujiwara regents Morozane and Moromichi
appear to have played major roles in the government, without, however, the kind of dominance earlier regents had attained. Shirakawa
was strong enough to prevent that. Also, Horikawa's mother, Kenshi, was Morozane's adopted daughter, which obviated the strong
maternal tie necessary for regency domination. (Kenshi was the real
daughter of Minamoto no Akifusa, and her Murakami Genji relatives enjoyed considerable influence at this time.) Although the accident of document survival makes any conclusions extremely tenuous, it is interesting to note that there are eleven extant imperial
edicts issued by Horikawa. There is only one extant Shirakawa edict
and, more important, there are no documents issued by ex-emperor
Shirakawa or his office during Horikawa's reign. The first extant inno-cho document is dated Tennin 2 (1110), early in Toba's reign.33
Although he had dispossessed Prince Sukehito, Shirakawa continued to feel pressure from that quarter and would not rest secure
until Horikawa had produced an heir to the throne. Shirakawa was
31 Jien, in his Gukansho, in Nikon koten bungaku taikei (Tokyo: Iwanami shoten, 1967), vol. 86,
p. 204, makes this claim; other sources, including Munetada, also give high marks to
Horikawa.
32 Fujiwara Nagako, The Emperor Horikawa Diary, trans. Jennefer Brewster (Honolulu: University of Hawaii Press, 1977), passim.
33 In Heian ibun, Horikawa's documents are all collected in vol. 4, except for one in vol. 9.
Cambridge Histories Online © Cambridge University Press, 2008
6OO
INSEI
preoccupied with the problem, but he had himself to blame. In his
zeal to avoid the possibility of a reestablished Fujiwara maternal
connection, Shirakawa had married Horikawa to his aunt, Princess
Atsuko, even though she was twenty-one years his senior (see Figure
9.3). The marriage took place in 1091, when Horikawa was only
twelve; it is not surprising that, despite Shirakawa's diligence in offering prayers for the birth of an heir, Atsuko failed to conceive. In
1098, Shirakawa finally arranged a more congenial marriage for his
son with Ishi, daughter of Shirakawa's uncle and close associate, Fujiwara no Sanesue.This time he was more fortunate: in 1103 a prince
(the future emperor Toba) was born to the couple.
Shirakawa was greatly relieved. Horikawa was twenty-four and
had fallen ill; complications would bring his life to an end in four
years. Shirakawa had been very worried. Although he had abdicated
and become a monk, he had avoided taking the final vows and
adopting a Buddhist name so that he would be able to reascend the
throne if anything were to happen to Horikawa before he had a
son.34 Horikawa had been without a designated successor for eighteen years. But when Toba was five months old, he was formally
named Prince Munehito and in his eighth month was made crown
prince. There can be little doubt that Shirakawa had intended to replace Go-Sanjo's Minamoto heirs with his own line.
The birth of Munehito created a rare situation in the history of the Japanese
imperial house. Within the house, Shirakawa, Horikawa, and Munehito grandfather, father, and son - served as abdicated sovereign, emperor, and
crown prince. This was the first instance of such imperial continuity since
Uda, Daigo, and Suzaku had held similar positions in the early tenth century, and it was interpreted by at least one observer
as a great sign, a matter
for rejoicing by both the court and the people.35
Such imperial unity, unencumbered by stifling sekkanke maternal relations, had the potential for providing the basis for a powerful and
independent royal house. Although the imperial succession did not
follow the course outlined by Go-Sanj5, Shirakawa accomplished
the goal his father had envisioned.
But Shirakawa's worries were not over. In 1103, Sukehito had a
son, Arihito, a prince with a Minamoto mother. Morotada, chieftain
of the Murakami Genji, was aware that his house would prosper
even further if this boy were to become emperor. Horikawa died, not
unexpectedly, in 1107. Faced with a four-year-old emperor, Toba,
34 Taiki, vols. 23-25 aiZohb shiryo taisei, vol. 1, pp. 67-68, Koji 1/5/16.
35 Hurst, Insei, p. 137.
Cambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INSEI
6oi
and with no desirable crown prince in sight, Shirakawa must have
been concerned about the future of his line. Significantly, it was only
at this point that he became an active participant in the political
process as ex-emperor. He left the position of crown prince unfilled;
having expended so much effort, he was not about to yield to his half
brother.
Shirakawa's decision to move forcefully into politics adversely affected the fortunes of the Murakami Genji. They had been close allies of the imperial house during the reigns of both Go-Sanjo and
Shirakawa; but having reached a position of influence rivaling that of
the Fujiwara main line with two powerful candidates for crown
prince, they had become more of a threat to Shirakawa than an
asset. Ironically, the fate of the Minamoto princes Sukehito and Arihito was sealed by a precipitous act of their own clan members.36 In
1113a plot against the life of Emperor Toba was uncovered; it had
been engineered by priests of the Daigoji, Shokaku and Ninkan,
both sons of Minamoto no Toshifusa (see Figure 9.2). The noble
council, after debating their guilt, exiled several Genji courtiers.
This proved to be a damaging blow to Genji prestige, but
strangely, it resulted in Shirakawa's adopting a conciliatory attitude
toward Sukehito. The ex-emperor also adopted young Arihito. In
1119, when Toba's consort gave birth to a prince, Shirakawa gave
Arihito the Minamoto surname and treated him with extraordinary
honors. Nevertheless, Minamoto no Morotoki notes in his diary
that although there was an outside possibility of Arihito's ultimate
succession, his chances were virtually dead. He also saw that Arihito
had more in common with two earlier frustrated Minamoto
claimants to the throne, Minamoto no Takaakira and Koichijo-in,
than with Emperor Uda, the only sovereign who, being dispossessed
as a royal clansman and granted the Minamoto surname, had gone
on to become emperor.37
Thus, in the decade after 1100, a number of factors combined to
prod Shirakawa into action. One was his concern with maintaining
imperial succession in his own line and the continued threat from his
Minamoto half brother Sukehito, Sukehito's son Arihito, and family.
Another factor was misfortune within the sekkanke. Moromichi had
succeeded his father Morozane as regent for Horikawa, but died
prematurely in 1099. Morozane himself died in 1101, leaving Mo36 Fujiwara no Tadazane records the event fully in his diary Denryaku, in Dai Nihon kokiroku
(Tokyo: Iwanami shoten, 1968), vol. 4, pp. 59-62. Eikyu 1/10/5-22.
37 Choshuki, Zoho shiryo taisei, vol. 1, p. 157. Gen'ei 2/8/7.
Cambridge Histories Online © Cambridge University Press, 2008
6O2
INSEI
romichi's son Tadazane, then only twenty-two, as chieftain of the
clan. Minamoto noToshifusa was the senior court figure at the time
and outranked Tadazane. Tadazane did not become regent until 1105
but was still too young to be the major figure at court.
When the young Toba came to the throne in 1107, Shirakawa was
faced with the threat of a Minamoto crown prince at a time when he
had no candidate of his own. The influential Minamoto clan could
be expected to press the candidacy of young Arihito; and the
sekkanke was at a particularly weak point. Shirakawa was almost
forced into taking action. His pre-1100 actions do not accord with
later characterizations of him as the embodiment of the despotic
inset ruler. If we wish to see Shirakawa in that role, we must turn to
the post-1100 period.
A perusal of documents and diaries of the quarter century before
Shirakawa's death in 1129 leave no doubt that he was the political
focus of the Heian court. Most of the personal statements we have
about Shirakawa date from this period. Our image of him is thus
shaped by these views. The Heike monogatari gives perhaps the quintessential depiction of Shirakawa as despotic ruler of the realm, stating that the only things outside Shirakawa's control were the Kamo
River floods, sugoroku game dice, and the unruly Mount Hiei monks.
This statement is strikingly similar to those scattered through the
pages of Munetada's Chuyuki, which is the most complete record of
the Shirakawa era. Munetada remarks that Shirakawa, "yielding to
his desires, makes appointments unbound by law" and that "the
grandeur of the abdicated sovereign is equal to that of His Majesty,
and at the present moment this abdicated sovereign is sole political
master."38
Shirakawa appears to have been an uncomplicated man who
acted decisively on the basis of his own likes and dislikes, caring little for established custom. He lavished favors upon those he liked,
and depended heavily on them. He loved and hated in the extreme.
His great love for his consort Kenshi colored his feelings for their
son, Toba, and daughter, the lady Ikuhomon-in.The latter, if we are
to believe Munetada, had a powerful influence on political decisions.39 Also, Shirakawa's close relationship with his retainers (kinshin) gave them considerable influence over politics; sources cite
38 Chityuki, vol. 3, p. 410. Tennin 1/10/28.
39 Takeuchi Rizo, Bushi no top, vol. 6 of Nihon no rekishi, 31 vols. (Tokyo: Chuo koronsha,
1965-67), pp. 225-26, quotes Munetada to the effect that Shirakawa loved this lady above
all his other offspring, and thus power in the realm accumulated in her alone.
Cambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INSEI
603
several of them as being of crucial importance to Shirakawa in decision making.
It is, of course, difficult to depend solely upon the diary entries of
a few courtiers to reconstruct Shirakawa's personality. In the history
of the Japanese imperial house, no one had been as politically active
for such a long period; opinions of him must certainly have changed.
On certain occasions, Munetada praises Shirakawa: when Shirakawa
died, Munetada remarked on the long years of his tenure, concluding "his majesty filled the four seas and subdued the empire - since
the time of Emperor Kammu there has been no such example. He
should be called a sage ruler, a sovereign for eternity."40 Conversely,
on the death of Emperor Horikawa in 1107, Munetada noted the
sadly deteriorating conditions in the Latter Day of the Law and was
at pains to point out that the source of the problem was not Horikawa but Shirakawa - "it is because the priestly sovereign [Shirakawa] still lives, and the affairs of the land are split in two."41
Enough documents survive to confirm that Shirakawa was a dominating figure dedicated to the pursuit of the good life, that he lavished favors on the Buddhist establishment and reacted with strong
personal likes and dislikes to persons and incidents. It seems that his
temper flared frequently, and that he did not forgive easily. According to a famous story from the Kojidan, on one occasion when there
was to be a dedication ceremony for his private temple, the Hosshoji,
it rained hard and the ceremony had to be postponed. The next day
it was again postponed due to rain, then a third day, and finally a
fourth day. The irate Shirakawa became so upset, he ordered that
rain be collected in a vessel and thrown into prison.42
Perhaps a more famous - and more credible - example is the case of
Fujiwara noTadazane (1078-1162).43 Once Shirakawa had emerged
supreme at court, and the sekkanke appeared to pose no great threat
to the imperial house, Shirakawa was quite attentive to the sekkanke
nobles. In 1120 he even suggested that Tadazane's daughter Taishi be
admitted to Emperor Toba's women's quarters. Tadazane refused,
incurring the wrath of the ex-emperor. Later Tadazane changed his
mind and approached Shirakawa. This reversal seems to have enraged Shirakawa so much that he ordered Tadazane stripped of his
post as regent; Tadazane went into seclusion until Shirakawa's death
40 Chiiyuki, vol. 6, p. 65, Daiji 4/7/7. 41 Chuyuki, vol. 3, p. 231, Kao 2/7/19.
42 This story is told in detail in Yasuda, Inset to Heishi, p. 56.
43 Hurst, Insei, p. 149, quoting from the diaries of Munetada and Morotoki.
Cambridge Histories Online © Cambridge University Press, 2008
604
INSEI
twelve years later. Taishi was not only immediately prohibited from
becoming consort, but Shirakawa wrote in his testament that she
should never be made consort. He even dismissed four Hosshoji
priests who were observed close to Tadazane.
Shirakawa was the towering figure of his age, political arbiter, imperial family head, and devout Buddhist. With his emergence as the
supreme political figure at court we recognize the "establishment of
the inset." But we must be careful to define, on the basis of extant
materials, precisely what we mean by this. The medieval historian
Kitabatake Chikafusa (1293-1354), for example, states in his Jinnoskoibki that from Shirakawa's time the ex-sovereign's documents
took precedence over imperial edicts, since the ex-emperor had become the repository of sovereignty. But this does not seem to have
been the case.
Sources hint that the in-no-cho of various ex-sovereigns issued
documents, but there is only one, an inzen (decree) of the exemperor Uda, extant from before Shirakawa's time. Beginning with
Shirakawa's time, however, these documents become more numerous; we find inzen and two documents issued by the in-no-cho on
his behalf, kudashibumi (order) and cho (communique). But there
are only six extant from Shirakawa's forty-three years as retired
sovereign, three inzen, two kudashibumi, and one cho. By contrast,
we can identify at least eleven edicts of Emperor Horikawa, who
reigned for twenty of these years, and eight of Emperor Toba, who
reigned for only sixteen. Importantly, none of the extant retired
emperor documents dates from Horikawa's reign; the first is dated
the twenty-second day of the last month of Tennin 2, which in the
Western calendar would be early in m o . 4 4 Granted, Shirakawa's
activities far exceeded those of earlier ex-emperors and extant documents are more numerous, but he did not issue a large number of
documents that would become the law of the land.
Like previous Heian ex-emperors, Shirakawa opened an in-no-cho
upon his reitrement. He staffed it with the same kinds of personnel
as had his predecessors: men who had served close to him (and to
Go-Sanjo) while he was emperor. But Shirakawa's office was larger
than that of earlier ex-emperors, reflecting his great power. Also, we
can discern patterns of recruitment that tend to corroborate, with
certain qualifications, Jien's claim that the ex-sovereign's officials
had become extremely powerful.
44 These are all contained in Heian ibun. Also see note 31, above.
Cambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INSEI
Kurododokoro
(private household office)
Fudono
(secretariat)
Tsukae-dokoro
(attendants' office)
605
Mizuishin-dokoro
(office of official bodyguards)
Mushadokoro-cho
(warriors' office)
In-no-hokumen
(office of ex-sovereign's warriors
of the northern quarter)
Meshitsugi-dokoro
(servants' office)
Shimmotsu-sho
(ex-sovereign's kitchen)
Gofuku-dokoro
(office of ceremonial dress)
Saiku-sho
(ex-sovereign's workshop)
Mimaya
(ex-sovereign's stable)
Bechino-sho
(special warehouse)
Figure 9.5. Structure of the in-no-cho. Adapted from Hurst, Insei,
p. 220.
It is difficult to grasp the size of Shirakawa's official in-no-cho staff.
Organization closely resembled the house organization of the
sekkanke and other major courtier houses. This is hardly surprising,
since the original purpose of the in-no-cho was to provide a private
organization, the lack of which had put the imperial house at a disadvantage in its political and economic competition with the leading
noble families. The in-no-cho included among its offices a household
office, attendants' office, servants' office, kitchen office, workshop,
stable, and several offices for bodyguards and other warriors (see
Figure 9.5). At Shirakawa's death, Munetada noted that there were
twenty betid (superintendents), five scribes, four secretaries,fiveassistants, and in excess of eighty warriors in the in-no-hokumen, the
in's guard unit.45 This was quite an expansion from the modest beginnings in 1086, when Shirakawa had appointed to the in-no-cho
five superintendents and six scribes, making only a few more appointments, including ten warriors to a mushadokoro (warriors' office), over the next several months.46
Most of the men who became retainers of ex-emperor Shirakawa
were middle-ranking courtiers and zuryo, that is, governors or vicegovernors who served in the province of their appointment. These
45 Chiiyuki, vol. 6, p. 76, Daiji 4/7/15.
46 Yanagihara-ke kiroku, in Tokyo daigaku shiryo hensanjo, comp., DaiNihon shiryo (1901- ),
series 3, vol. 1, pp. 3-10.
Cambridge Histories Online © Cambridge University Press, 2008
606
INSEI
men provided the economic foundation for Shirakawa's regime. The
expansion of zuryo economic power was aided greatly by shoen regulation that increased the public lands, which zuryo could exploit to
enhance their income. Another result of regulation was the accumulation of a large number of imperial estates und;r the ex-emperor's
control. The role of the zuryo in the process was crucial.
The importance of the Shirakawa-srur^o relationship was evident at
the time. When Shirakawa died, Munetada noted seven things that
first came to pass during Shirakawa's era. Five of these involved
zuryo:47
1.
2.
3.
4.
Zuryo made enormous "contributions' (ko) to the government.
Boys in their teens became zuryo.
Appointments to more than thirty governorships were fixed.
Three or four members of one family served as zuryo simultaneously.
5. Zuryo failed to pay levies due temples, shrines, and courtiers.
The most obvious way in which zuryo aided Shirakawa was
through "contributions." This was a kind of institutionalized bribery,
known asjdgo, that had arisen in Heian times as public revenues decreased. Those who made private contributions toward the construction of public buildings or the expenses of court or religious
ceremonies were rewarded with appointment to rank and/or office.
Although Go-Sanjo had attempted to curtail the practice, Shirakawa actively encouraged zuryo "contributions." The reward was
reappointment or transfer to a more lucrative governorship. During
Shirakawa's insei, provincial governors contributed toward the construction of major temples and shrines, imperial residences, and especially to new chapels built for the imperial house under Shirakawa's sponsorship. During his hegemony four of six imperial clan
temples (rikushoji) located in the Shirakawa area east of the Kamo
River were built, all by zuryo "contributions."
The zuryo were in the ideal position to make such contributions
since they were little more than tax managers for the central government. As long as they forwarded the required levies to the capital, no one interfered with their attempts to make further exaction
on the citizenry. Zuryo greed was a fact of life. If several members of
a family held provincial governorships, the family could easily make
the kind of "contributions" to which Munetada referred.
47 Yasuda, Insei to Heishi, pp. 77-78, quoting Munetada's diary entry at Shirakawa's death.
Chuyuki, vol. 6, p. 65, Daiji 4/7/7.
Cambridge Histories Online © Cambridge University Press, 2008
SHIRAKAWA AND INSEI
607
Those serving most closely to Shirakawa were known as kinshin,
or retainers.48 While most of them were zuryd, they were not all of
the same social origin. Some were from traditional zuryd families;
others held a governorship in addition to their main bureaucratic
post in the capital. Some had emerged from the lowest ranks of the
nobility, while others had come from the warrior class, become zuryd
and later served in Shirakawa's in-no-hokumen. The term kinshinzuryo appears often in sources.
Shirakawa's retainers established more than a personal patronclient relationship with the ex-emperor. The tie extended to their
families over time, so that sons, grandsons, and great-grandsons
served the successive ex-sovereigns Shirakawa, Toba, and GoShirakawa. They became, in effect, client families of the imperial
house. Among Shirakawa's retainers whose descendants became
client families were Fujiwara no Tamefusa, Fujiwara no Akisue, Fujiwara no Akitaka, Takashina no Tameaki, and Taira no Masamori.
All but one of Shirakawa's in-no-cho offices had analogues in previous in-no-cho; the in-no-hokumen, or ex-sovereign's warriors of the
northern quarter, was his own innovation. The in-no-hokumen,
named for the location of their quarters in the ex-sovereign's palace,
was a personally recruited elite guard, composed of essentially two
kinds of warriors. There were small-scale warrior-landholders from
the immediate capital region, who had come to Heian to serve as
lower-level warriors, and middle-ranking provincial officials able to
mobilize bushidan organizations through their local connections.49
But the hokumen was not the only military organization upon
which Shirakawa relied for the support of his political power. While
these warriors formed a standing military guard, others of greater
stature were recruited privately to serve close to him. A good example was the Seiwa Genji chieftain Minamoto no Yoshiie, who, as we
have seen, had emerged from the wars in the northeast as the greatest warrior in Japan. Although Shirakawa and the collective noble
leadership had been sufficiently wary ofYoshiie that they prohibited
landholders from commending estates to him,50 the ex-emperor did
ultimately receive Yoshiie in audience - which was tantamount to es48 Hurst, Insei, chap. 9, pp. 237-53, and appendix 2, pp. 285-311, discusses the kinshin in some
depth.
49 Ishii Susumu, "Insei jidai," in Hoken shakai no seiritsu, vol. 2 of Nihon kenkyukai, comp.,
Koza Nihonshi, 10 vols. (Tokyo: Tokyo daigaku shuppankai, 1970-71), pp. 202-3.
50 Go-Nijd Moromichi-ki, vol. 1, p. 284, Kanji 3/10/10, contains the record of kugyo discussion
of the commendations to Yoshiie. The same diary, vol. 2, p. 133, Kanji 5/6/12, mentions the
edict.
Cambridge Histories Online © Cambridge University Press, 2008
608
INSEI
tablishing a patron-client relationship - and ordered him to serve
closely. Later the Ise Heishi rose from Masamori's service in the
hokumen to become higher-level clients of Shirakawa and his successors; their military service was also of a highly private nature.
This kind of coercive force was needed neither to maintain Shirakawa's hegemony against possible opponents among the collegial
leadership of the small circle of court nobles, nor to defend the prerogatives of the nobility as a class against revolution by the peasantry. It seems mainly to have protected the person of the ex-sovereign and the court against the armed monks of the major temples
and enforced the widespread shoen regulation attempts at the
provincial level. In 1118, for example, when the monks of Enryakuji
stormed into the city, over one thousand of Shirakawa's hokumen
and their vassals were mustered out at the Kamo riverbed. The establishment of expanded military forces, however, did nothing to
prevent such uprisings. During the Insei period hardly a year passed
without several incidents involving the rowdy monks of the great
temples either fighting among themselves or demanding some redress from the court.
With the support of kugyo in the noble council, where Fujiwara
sekkanke strength had waned, and with a large private organization
including military forces, Shirakawa was able to dominate Heian society as had few other imperial house members before him. It is perhaps his dismissal ofTadazane in 1120 that marks his acquisition of
supreme power. His emergence to such a position and the concomitant decline of the Fujiwara regent's house resulted in the formation
of the insei, which was to remain the dominant political style of
courtier rule until Go-Daigo's time in the fourteenth century.
THE HEGEMONY OF TOBA,
1129-1156
When Shirakawa died in 1129, insei rule was continued by his grandson Toba (1103—56, r. 1107-23), a man as arbitrary, strong-willed, and
domineering as Shirakawa had been. Throughout Toba's reign under
Shirakawa's insei, the relationship between grandfather and grandson
had been antagonistic. As a result, Toba tended toward direct opposition to the policies Shirakawa had adopted.
The seeds of antagonism were sown early in Toba's life. Toba was,
of course, the long-awaited heir of Shirakawa's son, Emperor
Horikawa; he was made crown prince shortly after his birth. When
Toba was four, Horikawa died, and Toba was enthroned. From this
Cambridge Histories Online © Cambridge University Press, 2008
609
THE HEGEMONY OF TOBA
SHIRAKAWA
Fujiwara no Tadazane
Fujiwara no Nagazane
Kayanoin
HORKAWA ^
Fujiwara no Kinzane
TOBA
Shoshi
(Taikemmon-in)
Tokushi
(Bifiikumon
1
KONOE
n
GO-SHIRAKAWA
SUTOKU
: : : : Assumed relationship
Assumed parentage
Figure 9.6. Family relations of Toba.
time Shirakawa emerged as the premier power at court, and during
his sixteen years on the throne Toba had little opportunity to act as
sovereign. Shirakawa chose as Toba's consort Fujiwara no Shoshi, a
lady whom Shirakawa had adopted. Shoshi, better known by her
honorary title Taikemmon-in, became consort to the fourteen-yearold Toba when she was sixteen and soon gave birth to Prince Akihito, the future emperor Sutoku (1119-64, r. 1123-42). Court gossip
suggested that Sutoku was actually Shirakawa's child and not Toba's.
Toba apparently was quite aware of the child's paternity. According
to one source, Toba referred to Akihito as his "uncle" (ojigo).51 Still,
Toba seems to have had affection for Taikemmon-in, who was
known as quite a beauty. She bore him four other sons and two
daughters (see Figure 9.6).
In 1123, Shirakawa forced Toba to abdicate in favor of Sutoku. For
more than six years, during the height of Shirakawa's rule, Toba languished as junior retired sovereign with no political influence at
court. Toba must surely have been frustrated with Shirakawa and
Sutoku for dominating the two primary positions in the imperial
house, while he had sons who were also possible candidates for
the throne. The three retired, honored members of the house Shirakawa, Toba, and Taikemmon-in (the san'in, or "three in," as
they were known) - seem to have made a number of pilgrimages together and enjoyed the wealth and prestige due their station, but
considerable enmity brewed beneath the surface.
51 Yasuda, Insei to Heishi, p. 127.
Cambridge Histories Online © Cambridge University Press, 2008
6io
INSEI
TABLE
g.l
Reigning and retired emperors in the Insei period
Reigning emperor
Go-Sanjo
Shirakawa
Horikawa
Toba
Sutoku
Sutoku
Konoe
Go-Shirakawa
Go-Shirakawa
Nijo
Rokujo
Takakura
Antoku
Go-Toba
Go-Toba
Reign dates
1068-1073
1073-1087
1087-1107
1107-1123
1123-1129
1129—1142
1142-1155
1155-1156
1156-1158
1158-1165
1165-1168
1168-1180
1180-1185
1183-1192
1192-1198
Senior retired
emperor
Junior retired
emperor
Shirakawa
Shirakawa
Shirakawa
Toba
Toba
Toba
Toba
Go-Shirakawa
Go-Shirakawa
Go-Shirakawa
Go-Shirakawa
Go-Shirakawa
-
Sutoku
Sutoku
Rokujo (to 1176)
Takakura (to 1181)
Source: Kodansha Encyclopedia of Japan (1983), adapted from George Sansom, A History of Japan to 1334 (1958).
At his grandfather's death in 1129, Toba moved quickly to inherit
the exalted position Shirakawa had carved out for the retired sovereign. He immediately seized control of the numerous wealth-filled
storehouses Shirakawa had built in the capital. Although some courtiers grumbled about the inappropriateness of this act, it was logical
and necessary for Toba to secure this economic base as he began his
tenure as abdicated sovereign. Shirakawa had been a dominant exemperor for so long (and the sekkanke had sunk so low) that no one
seemed seriously to have expected the political scene not to continue
in the insei form. From the start Toba's approach was different from
Shirakawa's, even if his manner was similar. The three main differences involved Toba's attitude towards his retainers, his feelings
about shoen, and his relationship with the sekkanke.
Toba initially dismissed a number of retainers who had been especially loyal to Shirakawa, replacing them with men loyal to himself.
This is typical of almost any authoritarian ruler who attempts to impose personal control over the political scene. In Toba's time, there
were also changes in the types of people who served the retired sovereign; the importance of zuryo, especially, seems to have decreased.
This was partially due to zuryo advancement in rank and position at
court. Furthermore, with the development of the proprietary provCambridge Histories Online © Cambridge University Press, 2008
THE HEGEMONY OF TOBA
6ll
ince (chigyokoku) system, vinder which high-ranking nobles became
provincial proprietors (chigyo kokushu) and received the tax revenues
from the province, zuryo became less important. They could no
longer monopolize tax receipts. (Some zuryo did advance to become
provincial proprietors under the ex-emperor's patronage.) But the
main reason for the diminished importance of the zuryo under Toba
was that he accumulated a large independent economic shoen base so
that zuryo economic support was no longer necessary.52
Similarly, his policy of conciliation toward the sekkanke was motivated by a desire to undo the precedents set by Shirakawa. In 1120,
Shirakawa, in perhaps his greatest exhibition of arbitrary power, had
stripped the Fujiwara chieftain Tadazane of his regency. The clash
with Tadazane had involved Taikemmon-in, then called Shoshi, who
had become consort to Toba. When Shirakawa tried to have
Tadazane's daughter Taishi afforded the same honor, the Fujiwara
leader had refused, infuriating Shirakawa. Tadazane probably opposed Shirakawa because of Shoshi's rather unsavory reputation.
Tadazane found her a "strange and unusual consort";53 she had had
numerous liaisons with other men at court, including an arranged
marriage with Tadazane's sonTadamichi, which had been postponed
several times, then canceled.
Less than a year and a half after Shirakawa's death, Toba succeeded in luring Tadazane back to court; Munetada, at least, found
this reunion of sovereign and subject praiseworthy.54 Once again, in
contrast to Shirakawa's last years the imperial house and the
sekkanke appeared to be in harmony. But the regent's house was a
pale shadow of its former self. Tadazane's return marked the beginning of a new kind of relationship, a clientization or vassalization of
the sekkanke, as its members, too, became retainers in service to the
ex-sovereign. Tadazane regained the clan headship and governmental positions (in fact, if not immediately in name) from his son
Tadamichi, widening the breach that already existed between them
and further weakening the sekkanke. The breach between Tadazane
and Tadamichi and, in the imperial house, between Toba and Su52 Ishii, "Insei jidai," p. 208.
53 Denryaku, in Dai Nihon shiryo, series 3, vol. 18, pp. 421-23, Eikyu 5/12/13, contains
Tadazane's record of the events surrounding Shoshi's entry into the women's quarters.
Three times Tadazane uses the phrase kikai fukashigi ("strange and unusual") to describe
Shoshi. Kikai actually conveys a strong sense of "outrageous" or "scandalous." He refers
to the whole procedure as the "most scandalous affair in Japan" {Nihon daiichi kikai na
koto).
54 Chuyuki, vol. 6, p. 301, Daiji 6/11/17.
Cambridge Histories Online © Cambridge University Press, 2008
6l2
1NSEI
toku, would be major factors in the Hogen Disturbance, which broke
out at Toba's death. Such were the problems brought about by the
arbitrary actions of retired emperors.
Perhaps the greatest difference between the policies of Toba and
his grandfather was in the area of shoen regulation. Whereas Shirakawa inherited from his father the desire to control the acquisition
of shoen, Toba wholeheartedly backed acquisition. There were a
number of shoen regulation attempts in Shirakawa's time, but there
are none recorded during Toba's years. While there are few documents extant from Shirakawa's era, we find an abundance of the
three types of in documents in Toba's time, all relating to the matter
of estate acquisition and management, demonstrating quite clearly
that the main business of the in-no-cho was the management of estates belonging to the ex-emperor, other imperial family members,
and clients.55 Even the number of cho (documents directed to offices
with which there was no superior-subordinate relationship) increased as communication about shoen control became more important to the imperial house.
While it is undeniable that Shirakawa was responsible for acquiring vast assets and building many great temple complexes, he did
not accumulate many shoen. He tended to raise support through
court-awarded grants, direct commendation, and contributions
from his many provincial official clients. By contrast, Toba actively
sought the establishment of new shoen. His documents largely concern guarantees of exemption from taxation for landholdings that
normally had been commended to one of the numerous imperial
Buddhist chapels or to Shinto shrines. Toba's insei was an epochal
period of shoen acquisition for the imperial house.56 The Anrakujuin Temple (a complex within the Toba-dono) is a good example: of
forty-seven holdings, four date from Shirakawa's time and fortythree from Toba's. This seems to reflect rather well their respective
attitudes toward shoen.
The Noto Province otabumi (land register) compiled in 1221 lists
public and shoen landholdings. It shows that more than 70 percent of
cultivated lands were in some form of shoen holding. Three-quarters
of that land (52 percent of the total landed area) was established in
Toba's time.57 Furthermore, Anrakuju-in documents suggest that
55 Hurst, Insei, pp. 223-36, passim. Much of the information on documents is gleaned from
the unpublished manuscript of the late Suzuki Shigeo, "Inseiki in no cho no kino ni tsuite:
in no cho Hakkyu monjo o tsujite mitaru."
56 Ishii, "Insei jidai," pp. 210-13. 57 Ishii, "Insei jidai," p. 209.
Cambridge Histories Online © Cambridge University Press, 2008
THE HEGEMONY OF TOBA
613
commendation estates were spread widely throughout Japan, beyond the Kinai region to middle-distant and outlying provinces.
Though Toba engaged in extensive shoen acquisition for the imperial house, acquisition was conducted somewhat differently than in
other great houses, such as the sekkanke. Estates were not held in the
name of the retired emperor, nor was he necessarily the highest level
shiki holder. Virtually all holdings, like those of the Anrakuju-in,
were lands for the support of temples or chapels belonging to some
member of the imperial house, which was headed by the retired sovereign. This may in part have been a necessary device, disassociating
the ex-emperor from a process that, though legal, was thought to run
contrary to the spirit of the eighth-century statutory system, to
which even the most fiercely independent ex-sovereign remained
somewhat deferential.
Toba, like Shirakawa before him, continued to build numerous
temples and sponsor construction on behalf of house members, to
whom accrued substantial landholdings in the form of shoen commendations. Toba also began the process of consolidating these scattered holdings into blocs or portfolios, which he then used as economic resources for distribution to imperial ladies. The purpose
evidently was both to make transmission of the holdings more secure
and to provide for the ladies. For example, it was Toba who created
the massive portfolio of estates known as the Hachij6-in-ry6.
Hachijo-in was the palace name (ingd) of Princess Shoshi, daughter
of Toba and Bifukumon-in. In 1140, Toba gave the princess twelve
holdings; when Bifukumon-in died in 1160, Hachijo-in inherited all
her mother's lands as well. Together they came to be known as the
Hachijo-in-ryo and were handed down in the imperial house as an
indivisible bloc. In the Kamakura period, when the house effectively
split into two branches, these holdings numbered some 220 separate
estates.
Taken together, the evidence suggests thatToba's period as retired
emperor represented a high point in the process of shoen development and that a considerable portion of these new shoen came to
endow the Buddhist chapels that formed the basis of the imperial
house shoen holdings. But this was true not only of chapels sponsored by imperial family members. It seems that while virtually all
new estates of the period were of the commendation type (kishinchikei shoen), a considerable number were commended to temples
and shrines rather than to individual courtiers. Some scholars believe that during Toba's time, and to a lesser extent during Go-ShiCambridge Histories Online © Cambridge University Press, 2008
614
INSEI
rakawa's, somewhat loosely controlled temple shoen, originally of the
zoyakumen (miscellaneous post exemption) type, became more fully
controlled, physically distinct shden.5S (Zoyakumen estates developed
from the tenth century on, when a portion of the obligations peasants
owed to the provincial government was exempted, that is, ceded to
a temple or shrine. The form was literally "miscellaneous": it might
mean rice, oil, roof tiles, straw rope, other products, or service labor.
Such estates came under the dual control of provincial government
officials and the estate holder.)
Politically, Toba was not the outstanding figure his grandfather
had been, perhaps because his life was shorter. He continued to manipulate the traditional symbols of authority. The collegial noble
council, again staffed increasingly by clients in his service, decided
matters of state. But perhaps we can see him intruding more personally into the process and arrogating to himself more of the prerogatives of imperial power; for example, he is reputed to have issued
his own inzen (not extant) to summon Fujiwara no Tadazane back to
court.
During Toba's era, there was also expansion of military organizations under the ex-emperor's control. One very important development was the rise ofTaira no Tadamori (1096-1153) as a military
hegemon and Toba's reliance upon him. The Ise branch of theTaira
had already established a close relationship with ex-emperor Shirakawa. Tadamori's father, Masamori, had commended some
twenty end of land in Iga Province to the Rokuj5-in, a chapel dedicated to the repose of the soul of Shirakawa's beloved daughter
Ikuhomon-in. For this "contribution," Masamori received Shirakawa's fervent support as well as appointments to lucrative zuryo
posts. He also continued to serve in the in-no-hokumen (see Taira
genealogy, Figure 9.7).
Tadamori inherited that military position, but he rose far beyond
his father to become one of ex-emperor Toba's major retainers. This
was in part due to the unsettled times and Toba's need for a military
force to deal with local unrest and akuso outbursts, but it was also
the result ofTadamori's skillful maneuvering. When he was governor
of Bizen in 1132, Tadamori "contributed" to the construction of the
Tokuchoju-in chapel and was admitted to imperial audience. Because of his lowly background and warrior vocation, Tadamori was
resented by the rest of the nobility. According to Heike monogatari,
58 Ishii, "Insei jidai," pp. 212-13.
Cambridge Histories Online © Cambridge University Press, 2008
THE HEGEMONY OF TOBA
615
Emperor Kammu (50)
r781-806
Prince Kazurawara
d853
I
Taira noTakamune
804-867
Kammu Heishi
Prince Takami
I
Taira noTakamochi
H889
Kammu Heishi
I
7 generations
I
Tomonobu
Koretoki
Masanori
I
I
Naokata
Masahira
(Hojo)
Masamori
I
Tadayori
Masakado
d940
Sadamori
Korehira
Ise Heishi
Yoshibumi
I
I
Koremasa
d981
I
Yoshimasa
Kunika
d935
I
Tadatsune
967-1031
I
Nobunori
1102-87
Tokinobu
d 1149
Tadamori
1096-1153
in
Tokitada
1127-86
Tokiko
dll85
r
Shigeko
1142-76
II
r
Kiyomori Tsunemori
1118-81 1125-85
I
Tokizane
1141-1213
Shigemori
1138-79
Tadamasa
d 1156
r
T
\
Norimori
1128-85
Yorimori
Tadanori
1144-84
T
I
Motomori Munemori
1139-?
1147-85
I
Tomomori Shigehira
1152-85
1156-85
I
I
Tokushi
Moriko
1155-1213 1156-79
1
Koremori
1158-84?
1
Tsunemasa Atsumori Michimori Noritsune Narimori
dll84
1168-84 d l l 8 4
1160-85
Kiyotsune
d 1183
Figure 9.7. Genealogy of the Kammu Heishi (Taira). Adapted from
Kodansha Encyclopedia of Japan.
Cambridge Histories Online © Cambridge University Press, 2008
I
616
INSEI
they openly ridiculed him, even plotting his murder. But ex-emperor
Toba favored Tadamori, and his house prospered.
In his late teens, Tadamori had subdued the bandit Natsuyaki
dayu, thus establishing his military reputation; but his greatest exploits were against pirates in the west. He was first dispatched by the
court in 1129 on an inconclusive campaign. The court sent him out
again in 1135, choosing him over another great warrior hegemon,
Minamoto no Tameyoshi. Sources reveal that he was chosen at
Toba's recommendation, because the Taira already had a secure
power base in the west.59
Tadamori soon returned in triumph with the head of the pirate
leader Hidaka Zenji and some seventy prisoners, staging an elaborate entry into the capital and dazzling the citizenry. There were
widespread rumors that these "pirates" were actually only local warriors who had refused to become Tadamori's vassals, but at least the
uprising had been quelled. Whether or not Tadamori had merely engineered a successful public relations campaign, the Taira house was
rewarded - Tadamori's son Kiyomori (1118-81) was promotedand Tadamori's reputation as a warrior was enhanced. Perhaps more
important, the expedition enabled Tadamori to solidify his power in
the area; he organized local warriors into a vassal band (Jbushidari)
with himself as warrior chieftain (toryo).®0
Toba already had the in-no-hokumen, which had been established
in Shirakawa's time. The Imperial Police (kebiishi) could also be
mustered out for protection of the imperial house and defense of the
capital. But Tadamori's band was a rather different kind of force;
Toba's use of Tadamori can in a sense be regarded as the establishment of a central government mercenary army. After the abandonment of conscript armies in early Heian times, local uprisings (such
as Taira noTadatsune's rebellion and the wars in the northeast in the
eleventh century) had been handled by recruiting local warrior
bands or troops under the command of the provincial governor.61
Local uprisings were handled by local means, whether the court appointed a central official as general or not.
But under Toba's hegemony, capital warriors like the in-no-hokumen and others in his service were mustered out like mercenaries
59 Yasuda, Insei to Heishi, pp. 147-48.
60 Yasuda, Insei to Heishi, p. 150. Bushidan is a later, scholarly term rather than a contemporary one.
61 For a discussion of how the Heian court effectively harnessed local warrior power, see Karl
F. Friday, Hired Swords: The Rise of Private Military Power in Early Japan (Stanford, Calif.:
Stanford University Press, 1992).
Cambridge Histories Online © Cambridge University Press, 2008
THE HEGEMONY OF TOBA
617
and sent to quell local uprisings. Tadamori's bushidan, incorporated
with other "public" forces, was mobilized to perform "public" functions. At the heart of the force was Tadamori's personal vassal band,
a mercenary group organized by Toba, which came to function as a
kind of central government army.62 Tadamori became one of Toba's
main sources of support; in this respect, Toba's insei had developed
beyond Shirakawa's.
Actually, things seemed on the surface to be somewhat less complex under Toba. The sekkanke was brought back into the political
process more fully, and there appeared to be little question as to
where actual power at court lay. But the changing nature of the political situation - the rise to power of the ex-sovereign and the consequent increased power and wealth of other imperial house members, the corresponding decline in the fortunes of the Fujiwara, and
the infiltration, through the dyadic contract with Shirakawa and
Toba, of middle-ranking courtiers into the previously sacrosanct circles of the kugyd - produced a number of personal and factional rivalries that were to surface when Toba was near death.
Tensions were high among members of both the imperial and Fujiwara lineages, largely because of Toba's favoritism. We have already
seen that he had called back into service the old patriarch of the
family,Tadazane, whom Shirakawa had banished in ii2O.Tadazane's
return had seriously compromised the position of his son, Tadamichi; matters became worse as Tadazane looked increasingly to his
younger son,Yorinaga, as the future sekkanke leader.
BothTadamichi andYorinaga saw their best chance for future success through the traditional technique of establishing a marital alliance with the royal house, and they engaged in fierce competition
to provide consorts to Emperor Konoe, each hoping that a son
borne by his daughter might become emperor. Both succeeded in
having their daughters appointed consort, but this only intensified
the rivalry. In 1150, Tadazane took the symbols of Fujiwara headship
from Tadamichi and had them bestowed on Yorinaga. Soon the regency, too, passed to the younger Fujiwara lord; from that time
Tadamichi was completely estranged from his father and brother.
The imperial house was not at harmony either, owing to the late
Shirakawa's disposition of matters. Toba was senior retired emperor
and the acknowledged leader of the Heian nobility, and his son
Konoe had been enthroned in 1142 at the age of two. But ex62 Yasuda, Insei to Heishi, pp. 151-53.
Cambridge Histories Online © Cambridge University Press, 2008
618
INSEI
emperor Sutoku was still alive and eager to redeem his political
position. He had been emperor when Konoe was born to Toba and
Bifukumon-in and had been forced off the throne soon thereafter by
Toba. Although the two ex-emperors, Toba and Sutoku, were supposedly father and son, Toba had no reason to desire that succession
continue in Sutoku's line. Sutoku, however, was waiting in the
wings. He had pinned his hopes on the succession of his son, Prince
Shigehito.
Besides these rifts in the two most important noble houses, there
was a serious split in one of the major warrior leagues of the day,
the Seiwa branch of the Minamoto.The Seiwa Genji, along with the
newly risen Taira, were heavily relied upon as the military arm of
the Heian court. Yoshitomo and his father Tameyoshi were, however, engaged in a dispute over the position of Minamoto warrior
hegemon. In Heian society, while certain estate rights, such as the
emperorship or the regency position, were the exclusive possession
of discrete groups, factional alignments between members of noncompetitive groups could easily band together to further their own
particular goals. This kind of factionalism was common and can be
seen in the patron-client relationships between people from various
levels of society and the shoen relationships in shiki.63
Thus members from all three groups - the imperial house, the
sekkanke, and the Minamoto warrior league - soon found common
ground to form distinct factions. When the young but sickly emperor
Konoe died in 1155, followed quickly byToba's illness and demise,
these factional divisions at court erupted into a brief but crucial
armed conflict known as the Hogen Disturbance.
GO-SHIRAKAWA AND THE TAIRA, 1 1 5 6 - 1 1 8 5
It was prophetic that Go-Shirakawa's era was ushered in by the outbreak of the Hogen Disturbance, for it was during his lifetime that
the age of the warrior came to Japan. The clash of arms was precipitated by complex rivalries in the Fujiwara and imperial families focusing on the problem of succession and political control. Although
he would succeed to the throne and eventually enjoy more than
thirty years of power as ex-emperor, Go-Shirakawa (1127-92) was
not a prime candidate for the imperial position in the mid-iisos.
Go-Shirakawa's mother, Taikemmon-in, was no longer close to
63 Kiley, "Estate and Property" (see note 6, above).
Cambridge Histories Online © Cambridge University Press, 2008
GO-SHIRAKAWA AND THE TAIRA
619
Toba; and he was some eleven years older than his half brother
Konoe. Also, it was his father Toba's opinion that Go-Shirakawa was
not qualified to become ruler.64 Yet events worked in his favor.
Tadazane and Yorinaga hoped that Konoe would produce an heir,
but the emperor was sickly. Toba's favorite, Bifukumon-in, supported
Go-Shirakawa's son, the young prince Morihito. Morihito's mother
had died and he had been adopted by Bifukumon-in, who held great
hopes for his future. Her collaborator, Tadamichi, also saw in him an
opportunity to regain power. When Konoe died in the seventh month
of 1155, Bifukumon-in, Tadamichi, and Toba discussed the possibility of enthroning Morihito, but concluded that it would be inappropriate to skip over Go-Shirakawa and enthrone his young son. Thus,
Go-Shirakawa became emperor, and Morihito was appointed crown
prince. At the age of twenty-seven, Go-Shirakawa had unexpectedly
come to the throne, the result of complex factional alignments
largely unrelated to his own personal qualities.
Go-Shirakawa's accession further intensified rivalries at court. In
the Fujiwara house, it strengthened Tadamichi's position against his
father, Tadazane, and younger brother Yorinaga, who were both at
the time estranged from Toba; they had no influence on the succession decision. To make matters worse, it was rumored that they
had cursed Konoe, causing his early death. The grief-stricken Toba
gave credence to the rumors. Yorinaga could not convince Toba of
his innocence and dispel the ex-sovereign's increasing hatred. At
the end of 1155 his younger sister and Toba's consort, Kaya-no-in,
died, cutting his only tie to Toba. Thus the situation at court appeared desperate for Yorinaga and Sutoku, and developments
forced them together.
Toba died almost exactly a year after Konoe, on the second day of
the seventh month of 1156. Within eight days of his death, the opposing forces were in place for a major confrontation.65 The Hogen
Disturbance, which broke out on the eleventh, was brief and resulted
in a victory for the forces supporting Go-Shirakawa: Yorinaga died of
an arrow wound, Sutoku was exiled to Sanuki, and Tameyoshi and
many other Minamoto warriors were executed, in the first case of
capital punishment in Kyoto in about 350 years. Go-Shirakawa,
Tadamichi, Kiyomori, andYoshitomo were the clear victors.
Another important figure in these events was the lay priest Shin64 Taiki, vol. 2, p. 104. Nimpyo 3/9/23.
65 Heihanki, 5 vols. in Zoho shiryo taisei, is the most complete primary source for the Hogen
Disturbance; see vol. 2, pp. 112-26.
Cambridge Histories Online © Cambridge University Press, 2008
62O
INSEI
zei, Fujiwara no Michinori, one of the retainers who had been closest to the late retired emperor Toba. Shinzei, whose wife Kii-no-nii
was Go-Shirakawa's wet-nurse, seems to have had a major advisory
role in the decision to enthrone Go-Shirakawa, and in the aftermath
of Hogen, he had great influence over the new emperor. He was a
leading scholar who worked with Go-Shirakawa to reinvigorate the
declining organs of imperial government;66 but he was also an upstart, a schemer, a man whose political successes were achieved at
the expense of others. He engendered considerable animosity among
certain entrenched groups at court, which ultimately led to another
outbreak of violence and Shinzei's deadi.
Go-Shirakawa reigned for only three years before abdicating in
favor of Prince Morihito. During that time, mainly on Shinzei's advice, efforts were made to reform "evil practices of the degenerate
age." For example, a seven-article addendum (shinsei) to the imperial codes, issued in late 1156, attempted to curtail some of the illegal
activities of the major Buddhist centers that continued to plague the
court. The first two articles, however, dealt with continued shden
growth, declaring confiscate all improperly documented shden established since Go-Shirakawa's accession in 1155. They also declared
that added lands (kanoderi) incorporated within originally exempted
shden boundaries were to be seized immediately. An important exclusion to this general principle, however, affords insight into the relative nature of the power and authority of ex-emperors in this inset
period.
A stipulation was made that holdings with documents of exemption from ex-sovereigns Shirakawa and Toba should not be immediately confiscated; instead, the owners should forward imperial edicts
and in-no-cho documents in their possession to the central government for perusal and await imperial decision. Following the example
of Go-Sanjo, Go-Shirakawa reestablished the Records Office to carry
out the legislation. From this case, one can see that ex-emperor documents were considerably more prestigious than others but not more
authoritative than imperial edicts. They did not guarantee clear title
to land: formal imperial approval obtained through the issuance of an
edict was ultimately most valuable.
Go-Shirakawa and Shinzei also attempted to restore the throne to
the center of the political stage, first symbolically by the reconstruc66 Michinori was born a Fujiwara but took his wife's family name, Takashina, and was sponsored by the family. The Takashina traditionally produced Confucian scholars. Later,
Michinori reverted to the Fujiwara surname, by which his sons were known.
Cambridge Histories Online © Cambridge University Press, 2008
GOSHIRAKAWA AND THE TAIRA
621
tion of the imperial palace, which had burned so many times during
the period. Special levies were made on all public and private lands,
and in the third month of 1157 the ceremony for raising the ridgepole was held. Soon the emperor was able to move into the new
palace, where court ceremonies could be conducted as in former
times. Shinzei was eager to restore all the old practices and ceremonies that had fallen into disuse.
In 1158, however, Go-Shirakawa abdicated in favor of his son
Morihito, who ruled for the next seven years as Emperor Nijo
(1143-65). Some have argued that Go-Shirakawa's early abdication
indicated his lack of interest in politics, evidenced also by his preoccupation with Buddhism and the poetic craze of imayo. Actually,
Go-Shirakawa abdicated in anticipation of playing a major political
role. He knew that his son Morihito was expected to become emperor. Having abdicated in Morihito's favor, it would be entirely in
accord with recent precedent for him to enjoy political supremacy as
father of the sovereign. This move would also be advantageous to his
advisers, Shinzei in particular. Shinzei was of low rank and had become a lay monk, two factors that hindered his operation within the
regular bureaucracy. As a kinshin of the ex-emperor, however, these
were less serious handicaps to the exercise of power.
Like his two predecessors as senior retired sovereign, Go-Shirakawa abdicated fully intending to exercise his powerful influence at
court. But for a while, at least, his ability to dominate the court was
seriously hampered, and it was some time before he approached the
level of dictatorial control that Shirakawa andToba had enjoyed. This
may have been due in part to his own unpreparedness for ruling - he
had only reigned for three years after his unexpected accession. However, a more important reason was the presence of a strong and mature emperor in Nijo, who desired to exercise the authority due him
as the formal head of state. The insei was subject to somewhat the
same kind of institutional restraint as the regency power of the Fujiwara main line. Both forms of indirect control by familial authority
worked best when the emperor was a minor. Thus, the insei tended to
perpetuate the tradition of enthroning children.
When there was a more mature emperor, an abdicated emperor
was considerably less able to dominate the court, sometimes not
making the attempt despite his superior position within the imperial
house. At the beginning of the insei period, it will be remembered,
Shirakawa did not attempt to dominate Emperor Horikawa at all; his
real role in government began in the reign of the young Emperor
Cambridge Histories Online © Cambridge University Press, 2008
622
INSEI
Toba. When Go-Shirakawa abdicated, he was faced with the
strongest emperor to rule during the tenure of an ex-sovereign, his
own son, Emperor Nijo (r. 1158-1165). Nijo had been chosen for the
imperial position before his father and thus most likely expected to
be an active ruler even though he was only fifteen. As soon as Nijo
became emperor, there arose several factional alignments, among
which Nijo's faction, supporting direct imperial rule, obtained the
upper hand. Certainly Heian politics had exhibited factional tendencies before, but it is obvious that the development of two potential foci of imperial authority in the inset period exacerbated the situation immensely.
This faction included Fujiwara no Tsunemune, whose younger
sister was Nijo's mother, and Fujiwara no Korekata, son of Nijo's
wet-nurse. While by no means the traditional sekkanke type of maternal connection, these close bonds with the sovereign provided a
basis for cooperation. Others were drawn to this group because of
their opposition to Shinzei, whose power at court had won him
many enemies. Shinzei even made enemies of others among GoShirakawa's retainers, such as Fujiwara no Nobuyori. There were
also a military faction, the major figures being Taira no Kiyomori
and Minamoto no Yoshitomo. Since their emergence in the Hogen
affair, warriors had become a political force with which to reckon.
At the outset of Nijo's reign, the imperial faction may have held the
upper hand, but relationships between the aspirants for power was
fluid: antagonisms flared up not only between factions, but within
them as well.
In the last lunar month of 1159, these animosities culminated in
the Heiji Disturbance. In this conflict the anti-Shinzei and anti-insei
faction among the nobles included the Fujiwara courtiers Tsunemune, Nobuyori, and Korekata. When they were joined by the disaffected Minamoto warrior chieftain Yoshitomo, a bitter enemy of
his military rival, Taira no Kiyomori, and of Shinzei, the faction had
military power sufficient to contemplate a coup to topple Shinzei
and Kiyomori.
The details for the Heiji Disturbance as well as for the Hogen
fighting need not concern us here; they are related in Chapter 10. Although the plotters managed to kill Shinzei, the rebellion ended with
the victory of Kiyomori's troops. The Genji forces fled in defeat,
Yoshitomo and his family and followers attempting to reach the
safety of the east, where local warriors might rally to their cause.
Many, including Yoshitomo, were killed, but at the behest of the
Cambridge Histories Online © Cambridge University Press, 2008
GO-SHIRAKAWA AND THE TAIRA
623
Taira lady Ike-no-zenni, Kiyomori spared Yoshitomo's thirteen-yearold son,Yoritomo, exiling him to Izu. It was a magnanimous gesture
Kiyomori would later regret.
The political situation immediately following the Heiji affair did
little to increase Go-Shirakawa's power at court. It was a low point
for insei power, as the courtiers who had eliminated Go-Shirakawa's
key political supporter Shinzei drew closer to Kiyomori in support
of direct imperial rule by Nijo. This group included the wily Tsunemune and Korekata, who had shifted sides several times during the
era to enhance their positions. Shinzei had once complained to Kujo
Kanezane that Go-Shirakawa was the most ignorant sovereign in
Japanese and Chinese records for failing to recognize these plotters
from the beginning.67 Go-Shirakawa was primarily an observer in
the Heiji affair and enjoyed little political support in its aftermath.
Conflict soon arose between the ex-emperor and his son. In 1158,
at the outset of Nijo's reign, both the emperor and his regent
Motozane had been only fifteen years of age. Their fathers, GoShirakawa and Tadamichi, had acted in advisory capacities for their
sons, with considerable influence on policy making.
But by the early 1160s, Nijo and Go-Shirakawa began to clash
over the degree of influence the ex-sovereign ought to enjoy. There
were several plots by Go-Shirakawa and his retainers against
courtiers serving Nijo, which suggests that the relationship was
strained. In 1161, Nijo struck back twice at supporters of Go-Shirakawa. On one occasion he exiled Taira noTokitada and dismissed
from office six courtiers, who were also Go-Shirakawa's retainers,
for plotting to make Norihito - the infant son of the ex-emperor
andTokitada's sister - crown prince.68 Within a few years the antagonism between Nijo and Go-Shirakawa was openly acknowledged
at court, but clearly Nijo was in control. The man in the middle, the
man who was using both sides as much as possible to further the interests of his own house, was Kiyomori.
The Gukansho refers to the post-Hogen-Heiji eras as the "age of
the warrior" (musha no yd), and to an extent this was true. Kiyomori's military power overshadowed all others, and his rise in court
was rapid. A year after the Heiji affair, he became a kugyo and thereafter was a full participant in important state affairs. Because of unsettled conditions in both the capital and the provinces, the court
67 Quoted inTakeuchi, Bushi no tojo, p. 375, and Yasuda, Insei to Heishi, p. 253.
68 Hurst, Insei, pp. 194-97, discusses the incidents.
Cambridge Histories Online © Cambridge University Press, 2008
624
INSEI
had to rely on Kiyomori's forces. Yet the 1160s (despite some scholars' arguments) is still too early to postulate the establishment of a
"Taira polity."
It is important to note a change in the strategic maneuverings of
Kiyomori and his kinsmen. During the Heiji rebellion, Kiyomori acquired control over a number of courtier shoen holdings and replaced Yoshitomo as warrior hegemon by incorporating a number of
local warrior-proprietors into his military league. But as soon as he
became a member of the elite courtier circle, Kiyomori gave up the
strategy of organizing local landholders against central authority. Instead, he became a conservative supporter of the political system,
adopting the age-old strategies of forging marital alliances with the
imperial family and the sekkanke and dominating the kugyo bureaucratic apparatus. He became a military-courtier and thus missed an
opportunity to unify the various frustrated local warrior elements
into a powerful revolutionary force to topple the traditional elite.
That task had to await Minamoto no Yoritomo.
The key to understanding the political situation at court during the
crucial two decades before the 1180-85 war between the Taira and
Minamoto lies in following the maneuverings of Go-Shirakawa and
Kiyomori. Each used the other to further his own goals, cooperating
for a considerable time until Kiyomori became powerful enough to
dominate his former patron. In this sense, Go-Shirakawa's insei, at
least until the outbreak of war, was less dictatorial and less powerful
than those of Shirakawa and Toba. Go-Shirakawa could not dominate the court during Nijo's reign, but even after Nijo died in 1165,
Go-Shirakawa was unable to dominate his grandson, Emperor
Rokuj5 (1164-76, r. 1165-68). Kiyomori was in the way.
An example of competition, within a larger framework of cooperation, between the ex-sovereign and the Taira leader can be seen in
their influence on the direction of succession. Nijo had no heir. As
crown prince he had been married to Bifukumon-in's daughter, but
after an extended illness, she became a nun. He then shocked the
court by taking as his empress, Tashi, consort of the deceased sovereign Konoe, who came to be known as the "empress of two reigns"
{nidai no kisaki). Nijo had no offspring from these two unions or
from his marriage to Tadamichi's daughter Takamatsu-no-in. Consequently, the birth of the Taira prince Norihito threatened the possibility of the imperial line continuing among Nijo's descendants.
In 1164 a palace lady finally bore Nijo a son, and the child was immediately given to Takamatsu-no-in to raise in preparation for sueCambridge Histories Online © Cambridge University Press, 2008
GO-SHIRAKAWA AND THE TAIRA
625
cession. Nijo's health was already poor, and he became critically ill
shortly after the child's birth. Without even formally making him
crown prince, Nijo abdicated in favor of his son. He opened the inno-cho and appointed his officials, but died a month later. Thus the
new emperor, known as Rokujo, came to the throne as an infant less
than one year old, incredibly young even for a Heian sovereign.
The death of Nijo, and shortly thereafter of his loyal regent Motozane, made Go-Shirakawa the most powerful figure at court and
within the imperial house. The exiledTaira noTokitada was recalled,
dismissed courtiers were reinstated, and Norihito was named crown
prince only six months after Nij5's death. Go-Shirakawa was in
the unusual positions of grandfather of the emperor and father of
the crown prince. He was unchallengeable within the imperial house
and master of the political situation. The inset form of government
was firmly intact, but Kiyomori, manipulating the ex-emperor while
himself being used, proved a stumbling block to Go-Shirakawa's
complete domination of the court.
The rise of the Taira was the major development of the postHogen Heian court. Much of the Taira success can be attributed to
the nature of court politics in the inset period: the possibility of a bifurcation of authority between two imperial figures could lead to
greater factional infighting than in earlier Heian times, and the use
of force increasingly became necessary to settle disputes. It is also
widely believed that Kiyomori was the illegitimate son of former emperor Shirakawa.69 But one cannot overlook Kiyomori's skill at political maneuvering. After entering the ranks of the kugyd in 1160 as
the first warrior-courtier, Kiyomori's rapid rise was due not only to
his military power or presumed royal birth but to his ability as well.
When his younger sister-in-law, Shigeko, entered Go-Shirakawa's
service and caught the attention of the ex-emperor, Kiyomori
worked skillfully to serve both factions, but necessarily drew closer
to Go-Shirakawa.
By 1162 he had risen to the second rank and in 1164 managed to
have his daughter Moriko marry Regent Motozane. The marriage
placed him in a close relationship with the Konoe branch of the
sekkanke, in effect preventing the Fujiwara from becoming an antiTaira force. When Motozane died in 1166, Kiyomori had control of
the greater portion of sekkanke property pass to Moriko, under the
pretext that Motozane's successor, Motofusa, was too young. This
69 Takeuchi, Bushi no tojd, pp. 388-92; Hurst, Inset, pp. 181-82.
Cambridge Histories Online © Cambridge University Press, 2008
626
INSEI
gave Kiyomori control over these Fujiwara assets. In 1162, Kiyomori
completed the construction of the Rengeo-in Temple complex for
Go-Shirakawa, a project that enhanced his relationship with the exemperor. His reward for this "contribution" was appointment as
governor of the rich Bizen Province.
By early 1167, Kiyomori had progressed through several major
court offices to become Chancellor {daijo daijiri) and held the First
Rank. This was all the more astonishing because he had risen this far
in only the eight years following the Heiji affair. Much was due to
his relationship with Go-Shirakawa, which became even closer with
the appointment of Shigeko's son Norihito as crown prince. The two
men seem not only to have gravitated toward each other for political
reasons; they seem to have been very alike in character as well. It was
perhaps this similarity in personality that led them into cooperation
for a time and later to confrontation.
Both Go-Shirakawa and Kiyomori seem to have been determined
to accomplish their objectives regardless of cost. Shinzei once criticized Go-Shirakawa severely for his stupidity but granted that he did
possess two virtues.70 First, he could not be restrained from accomplishing his goals. Second, he never forgot anything he heard. There
are many stories that describe Kiyomori in the same fashion, perfectly willing to break tradition to accomplish an objective. In the
Gion Shrine incident of 1146, for example, he shot arrows at the sacred palanquin, showing none of the hesitation that usually crippled
the court in dealing with obstreperous monks. Also, despite his basic
Buddhist and Shinto faith, Kiyomori put little stock in the efficacy
of prayer to cure illness or bring rain during a drought.
By the time Kiyomori had become Chancellor there were several
Taira kugyo, who provided backing on the noble council. Kiyomori
resigned the office after only three months because of illness, but
"retirement" did nothing to minimize his power. He was supported
by Go-Shirakawa and wielded power as a "former chancellor." He
and Go-Shirakawa, ex-chancellor and ex-emperor, were on similar
institutional grounds.
Early in 1168, Kiyomori fell ill and became a lay monk.71 The illness was serious enough that many courtiers, including Go-Shirakawa, were afraid he might die, endangering the security of the
realm. Go-Shirakawa ordered prayers for Kiyomori's health and de70 Kanezane records this statement about Go-Shirakawa made by his close associate Shinzei.
Yasuda, Inset to Heishi, pp. 266-67.
71 Gyokuyo (Tokyo: Meicho kankokai, 1971), vol. i, p. 40, Nin'an 3/2/9-11.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SHIRAKAWA AND THE TAIRA
627
clared an amnesty. The ex-emperor was so worried that he hastily
arranged for Rokujo to abdicate in favor of Norihito, who became
Emperor Takakura (1161-81, r. 1168-80). With Go-Shirakawa's son
on the throne, the pro-imperial rule faction would have been unable
to work through Rokujo in the event of Kiyomori's death. GoShirakawa could thus reasonably hope for the continuation of his
insei rule. So while Kiyomori was alive, the ex-emperor acted in concert with the Taira chieftain to effect this change.
Kiyomori did not die. During Takakura's reign, he and his clan
rose to even greater heights. This was partly because of Shigeko's
status as mother of the emperor. She was raised to the position of
empress dowager and given the palace name Kenshummon-in.
Takakura's accession, however, brought Kiyomori and Go-Shirakawa into competition over the emperor. In form at least, Go-Shirakawa, as father of the emperor and senior figure in the imperial
family, continued his insei rule; but Kiyomori, with close maternal
relations to the emperor and significant military and economic assets, was able to enjoy much more power than Go-Shirakawa
desired.
At least until 1175, the relationship between the two appeared
cordial, but as Taira power increased unchecked Go-Shirakawa's
frustrations grew, especially because, in the face of continued trouble with the major temples, he was dependent on Kiyomori's warriors. (Even though Go-Shirakawa had become a monk in 1169, his
ability to deal with the recalcitrant armed monks who demanded
both ecclesiastic privileges and material rewards was severely limited.) He solidified his ties to the Taira by adopting Kiyomori's
daughter Tokushi and making her consort to Takakura. The birth of
a son would put Kiyomori in the enviable position of earlier Fujiwara regents. Taira influence, seemingly with the endorsement of the
ex-emperor, continued to expand.
It would have been strange if other courtiers had not resented the
Taira, upstarts who had reached the pinnacle of success at court, had
married into the sekkanke and the imperial house, and held offices
that in the past would have been denied them. The Taira were a new
phenomenon at court, having started as provincial warriors but risen
to become kugyo courtiers in the traditional mold. Yet they were distinguishable from other courtiers. They maintained military preparedness and skills that set them apart, though they did not integrate the local warrior populace into a separate authority structure
asYoritomo later did. In Rokuhara, a walled-off Taira suburb outside
Cambridge Histories Online © Cambridge University Press, 2008
628
INSEI
the borders of the Heian capital, they created family and residence
arrangements different from those of other courtiers.
The Rokuhara area was located across the Kamo River, east of the
capital proper. There the Taira maintained a cluster of clan houses
that combined residential and military functions.72 In twelfthcentury warrior society the nuclear family became a residential unit,
with fathers and sons, brothers and sisters, and other members of
the extended family building their residences in proximity to form a
community of kinsmen. Family residences were surrounded by the
houses of retainers. This became possible through the adoption of
the yometori form of marriage, in which the wife is brought into the
husband's family. There was a basic need for self-defense in the warrior's world, and such a residential community, centered around the
soryo or clan head, reflected bushidan organization.
The warrior Taira had brought this family organization with them
to the capital in the time of Kiyomori's grandfather Masamori. As
they became more powerful and numerous, Rokuhara expanded accordingly. In giving a figure of more than 5,200 Heike-related houses
in Rokuhara, Heike monogatari may be exaggerating, but the whole
area from present-day Rokujo to Shijo along the east bank of the
Kamo River was a Taira suburb including Kiyomori's Izumi-tei,
Shigemori's Komatsu-dono, Yorimori's Ike-dono, and other famous
mansions. The Taira did not break precedents in organizing their
power structure in the 1170s, and they did their best to follow traditional patterns; but they were still a breed apart. The physical separateness of the suburb, the community of family residences, and the
military superiority of the Taira made the very name "Rokuhara"
threatening to both the nobility and the commoners of the capital.
From 1175, Kiyomori and his clansmen began to clash rather frequently with their longtime patron Go-Shirakawa, and there were
several threats to the continuation of the Taira presence at court. In
1177 the friction broke into the open when a number of courtiers
close to Go-Shirakawa were discovered plotting against Kiyomori at
Shishigatani. Plotters were executed or banished, but no direct action was taken against the retired sovereign. The warning was
thought to be sufficient, but Go-Shirakawa continued to plot against
the Taira.
In early 1179, Go-Shirakawa took actions that could only be interpreted as a challenge to Kiyomori's position. Kiyomori's daughter
72 Yasuda, Insei to Heishi, pp. 272-74.
Cambridge Histories Online © Cambridge University Press, 2008
GO-SHIRAKAWA AND THE TAIRA
629
Moriko was the widow of Regent Motozane; at his death in 1166 she
had inherited the bulk of the sekkanke holdings. When she died in the
sixth month of 1179, Go-Shirakawa broke precedent and seized those
holdings for his own. Two months later, when Kiyomori's eldest son,
Shigemori, died, Go-Shirakawa seized his proprietary province of
Echizen. He soon moved to isolate the Konoe line of the Fujiwara,
which was related to Kiyomori through Moriko's marriage.
In the eleventh month of 1179, Go-Shirakawa's actions came to
naught: Kiyomori staged a coup, demanding a rash of dismissals and
new appointments. Thirty-nine courtiers, from the Chancellor on
down, were dismissed, most of them kinshin of Go-Shirakawa.73 This
time Kiyomori moved against the ex-sovereign as well. Titles to
his landholdings were seized for examination, and Go-Shirakawa was
placed under house arrest in the Toba-dono. Early the next year, a
goin, an office that managed imperial finances in anticipation of abdication, was established forTakakura. In the Second Month Takakura
was forced to abdicate, and the infant Prince Tokihito came to the
throne as Emperor Antoku (1178-85, r. 1180-85). Antoku was the son
of Takakura and hisTaira empress Tokushi, Kiyomori's daughter. The
Taira clan had reached the height of its domination.
Anti-Taira sentiment remained strong, however, and Kiyomori
felt it necessary to move Go-Shirakawa to a Taira residence where
he could be watched more closely. Kiyomori instituted unusually
severe military rule, dispatching more than three hundred of his
spies (kamuro) throughout the capital to terrify the citizenry into
compliance. It was in this kind of military atmosphere that, in the
fourth month, Antoku's accession ceremony was held. That year,
the Taira held fifteen proprietary provinces (chigyokoku), and a Taira
kinsman or adherent served as governor in thirteen provinces.74
With the coup d'etat of 1179, the confiscation of opponents' shoen
holdings, and a move to dominate local power structures, the Taira
had created their own polity.
But it was a transitory one. The Taira had alienated too many
groups to enjoy much support. The great temples around Heian became dangerous enemies, partially owing to Go-Shirakawa's machinations. With their military might, they were the only power strong
enough to challenge Kiyomori. Ex-emperor Takakura made his first
pilgrimage, at Kiyomori's insistence, to the Taira family shrine at
Itsukushima in Aki Province, not to one of the traditional shrines
73 Gyokuyo, vol. 2, p. 310, Chisho 3/11/17.
74 Yasuda, Insei to Heishi, pp. 292-93.
Cambridge Histories Online © Cambridge University Press, 2008
630
INSEI
like Iwashimizu Hachimangu. At this the Onjoji monks rose up, calling on the Enryakuji to join them. Although they failed in their attempt to rescue the ex-sovereigns Go-Shirakawa and Takakura from
Kiyomori's hands, it was the first time these large institutions had
cooperated in a common enterprise. Kiyomori was forced to take
precautions to guard Go-Shirakawa.
By the time Takakura had returned to the capital, Go-Shirakawa's
third son, Prince Mochihito, a frustrated twenty-nine-year-old who
had several times been passed over for succession, had issued an
edict calling for the Genji and other forces to rally and oust theTaira
from their seat of power. His co-conspirator was the aged Minamoto
warrior Yorimasa, the lone Genji presence at court during the Taira
heyday. The Taira tried to arrest and exile Mochihito, but he fled to
the sympathetic Onjoji monks. Several days later, however, Mochihito, Yorimasa, and the Minamoto warriors in their service were
killed by Taira forces in a battle at Uji.
To the consternation of the citizenry, Kiyomori then moved the
capital from Heian to his family headquarters at Fukuhara (near
modern Kobe), but that did not stop the anti-Taira movement.
Mochihito's call had reached the eastern provinces, and by the ninth
month Kiyomori heard rumors that Minamoto no Yoritomo
(1147-99) n a d rebelled. In the eleventh month, the defeated Taira
general Koremori returned with tales of a huge enemy concentration
of troops. Kiyomori moved back to the capital in order to be close
to the center of activities. The Gempei War had begun.
Go-Shirakawa was kept from political activity until 1181, when
both Kiyomori and Takakura died. He reached the height of his
power only when the Taira fled the capital in the seventh month of
1183, taking Emperor Antoku with them. Go-Shirakawa enthroned
his grandson Go-Toba (1180-1239, r. 1183-98) - even though Antoku was technically still emperor and the Taira courtiers had the imperial regalia in their possession - and proceeded to dominate the
Heian political world as he had been unable to do earlier. Go-Shirakawa was the only political figure able to deal with the military factions of the Genji warriors - Yoshinaka, Yoshitsune, and Yoritomo as well as the Taira. Although later chronicles characterize him as
wily and treacherous, Go-Shirakawa was trying to defend the prerogatives of the traditional court society against the onslaught of the
warriors. To a great extent, it was because of his efforts that rapprochement was reached at the end of the war, and that the traditional state expanded to incorporate Yoritomo's newly created
Cambridge Histories Online © Cambridge University Press, 2008
GO-SHIRAKAWA AND THE TAIRA
631
bakufu as an arm of the government. One can give credit to GoShirakawa for helping to effect a courtier-warrior union that would
persist through most of the Kamakura period, until the warriors
wrested all power from the court.
Go-Shirakawa was abdicated sovereign for thirty-four years, from
his abdication in 1158 until his death in 1192, but this was by no
means a period of uninterrupted total control. He had initially been
kept from exercising dictatorial control by his son, Emperor Nijo,
and was later checked by the rising power of the Taira. It was only
after the death of Kiyomori and the flight of the Taira that Go-Shirakawa was able to direct politics at will. Because of the complex factional alignments, which were in part a result of the rise of the insei,
for most of his tenure Go-Shirakawa was unable to act in as dictatorial a fashion as Shirakawa and Toba had.
Like his predecessor Toba, Go-Shirakwa was active in the acquisition of imperial estates, and the huge block of Ch6k5d6-ryo estates,
put together in his time, became one of the two major landed assets
of the imperial house in the medieval era. Go-Shirakawa may appear
to have maintained some ambivalence about shoen acquisition, for at
least twice, as emperor and ex-sovereign, he attempted regulation.
But shoen were an economic necessity. As Fujiwara no Koremichi,
one of ex-emperor Toba's kinshin, stated: "Nobles today receive no
official allotments at all. Without shoen how could they fulfill their
public and private (obligations)?"75 Public stipends for rank and office had disappeared. Nobles, including ex-emperors, could generate
income privately through shoen, or in a quasi-public way through the
proprietary province {chigyokoku) system, which allowed private
control of provincial tax revenues. (For imperial house members,
they were called allotment provinces, or bunkoku, but the system was
similar.)
Go-Shirakawa enjoyed extensive income from the far-flung network of shoen holdings that supported his religious and secular residences and elaborate lifestyle. Since he had formally become a monk
in 1169, he often stayed in Buddhist chapels where all his needs were
met. A document survives that lists the specific requisitions made of
the various shoen for the support of the Chokodo chapel.76 For vir75 Murai Yasuhiko, Kodai kokka kaitai katei no kenkyii (Tokyo: Iwanami shoten, 1965), p. 337,
quoting Koremichi's diary Taikai hishd.
76 Takeuchi, Bushi no tdjo, pp. 282-87. For the full document, seeTakeuchi, comp., Kamakura
ibun (Tokyo: To kyodo, 1971- ), vol. 1, pp. 426-50, Kenkyu 2/10/no day recorded. The information on Fukiya-no-sho appears on p. 427.
Cambridge Histories Online © Cambridge University Press, 2008
632
INSEI
tually every day of the year, different shoen provided various goods
and services for the support of the ex-emperor and/or the chapel.
One of the Chokodo shoen was Fukiya-no-sho in Ubara district,
Settsu Province, southwest of the Heian capital. Fukiya-no-sho was
responsible for certain provisions required during the first three days
of the new year (1192): a number of sets of new bamboo blinds (SMdare), five tatami mats of differing types, cloth curtains, andfiveryo
of sand for use in Buddhist ceremonies at the temple. The shoen was
also required to provide two laborers to serve at the temple during
the sixth month and seven warrior guards to be posted during the
sixth month at the gate facing Abura-no-koji Street and during the
first ten days of the seventh month at the inner gate.
Many other levies were made on Fukiya-no-sh5 for the Chokodo,
from feed for oxen in the twelfth month to go boards. The necessities and luxuries for temples and other estate proprietors were provided by their estates; the complex requirements for each shoen were
carefully spelled out to cover the entire year's needs. Even a glance
at such documents confirms the accuracy of Fujiwara no Koremichi's statement about the need for shoen.
During the Insei era under Shirakawa, Toba, and Go-Shirakawa,
the imperial house established a vast compex of shoen holdings, later
called the "lands accumulated during the three reigns (sandai gokisho
no chi)."77 This powerful economic base helped ex-emperor GoShirakawa and his predecessors to dominate the Heian court as few
imperial figures had in earlier eras.
By the time of Go-Shirakawa's death, insei government was so ingrained that the form would be retained for much of the Kamakura
period. The warriors were so used to dealing with ex-emperor GoShirakawa that it was no problem for them to shift their allegiance
to his successor, Go-Toba. Even after the Jokyii War (1221), the H5jo
leaders did not abolish the position of ex-emperor; instead they
made Go-Takakura "ex-sovereign," although he had never been emperor. This was unprecedented but indicative of how institutionalized the insei had become.
FOREIGN RELATIONS, 1 0 7 0 - 1 1 8 5
Another source of Taira power in late Heian times was wealth obtained through control of the China trade. As seen in Chapter 1,
77 Hurst, Insei, pp. 254-74, discusses Insei-period imperial shoen.
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
633
after the Japanese court ceased formal relations with the Chinese in
838 and with the Po-hai kingdom in 920, it remained basically isolationist. However, the contact with Koryo, established in the 1019 attack of the Toi (Jurchen pirates), blossomed into a brisk trading relationship in late Heian times, especially under the sponsorship of
Kiyomori.
King Munjong (r. 1046-83) ruled during the apogee of dynastic
development, when Koryo firmly institutionalized the imported
Chinese bureaucratic system. Munjong took steps to clear up a border dispute with the Khitan in the north, as well as to reestablish
trade with the Sung after a lapse of about thirty years. Munjong also
seems to have been favorably inclined toward Japan, presenting the
opportunity for Japanese ships to reopen trade with China, using
Koryo as middleman.
The first recorded visit of Japanese trade ships to Koryo occurred
in 1073, when forty-two Japanese traders traveled to Korea and presented King Munjong with a number of Japanese specialty products,
including bows and arrows, ink stones, and a saddle inlaid with
mother-of-pearl.78 Koryo records for the reigns of Munjong and his
two successors show that Japanese ships made at least twenty-four
trips to Koryo. Since the Japanese normally bore the title of envoy
(tsukai) from Japan or Tsushima, it appears that this exchange of
goods took the form of tribute and was conducted through Dazaifu
channels.79 In other words, it was public, not private, trade and
served as an opening wedge in the reestablishment of quasi-legal
trade with the Sung.
The import of Chinese goods through Koryo whetted courtiers'
appetites, undermining the passively negative official policy against
foreign trade. Courtiers and temples with shoen interests in Kyushu
and rich merchants in the area attempted to escape Daizaifu controls and work out their own trade arrangements. During Toba's
insei, certain courtier-officials with pro-trade leanings emerged, and
a quasi-legal Sung-Japan trade was instituted.
At about the same time, however, the brisk trade with Koryo
abruptly ended, owing to the rise of a military faction in Koryo and
the ensuing internal chaos. Koryo came to regard Japanese trade
ships as pirates; ships were sometimes stopped, the goods confiscated, and the crews executed. Consequently, Japanese traders, de78 Miura Keiichi, "10 seiki-13 seiki no Higashi Ajia to Ninon," in Hoken shakai no seiritsu, vol.
2 of Koza Nihonshi, p. 258.
79 Miura, "10 seiki-13 seiki," p. 257.
Cambridge Histories Online © Cambridge University Press, 2008
634
INSEI
nied open access to Koryo, began to turn exclusively to piracy, ravaging the Koryo ports. These were the earliest of the so-called wakd,
or freebooters, who were to play a crucial role in Korean-Japanese
relations during the medieval period. There had been considerable
improvement in navigational techniques since the previous century,
so it was no longer necessary for traders to rely on Koryo as an intermediary. Sailing directly to China was easier and offered greater
profits.80
Trade with Sung China began on a private basis, but as Japanese
ships ventured into the China trade, profit-seeking courtiers inToba's
retinue began to adopt a more positive attitude toward foreign relations. As relations improved it appears that certain Japanese circles,
and especially the more intellectually oriented of the nobles, developed an interest in Sung culture. This made Chinese goods more desirable and moved Japan along a course of receptivity toward continental contact, which had not been seen for several centuries.
Shinzei is an example of the kind of person at Toba's court interested in things Chinese. Far more than just a political figure, he was
the author of the historical compilation Honcho seiki and, along with
Fujiwara noYorinaga, was considered one of the great scholars of his
day. Once during a discussion, Shinzei amazed a visiting Chinese
priest with his brilliance. He apparently understood Chinese and expected to be dispatched as an envoy to the Sung court.81 Such people were responsible for the positive attitude of the court toward foreign trade.
In late Heian times, no one was more interested in Sung trade
than the Taira. The base for Taira trade domination seems to have
been prepared by Kiyomori's father, Tadamori, during Toba's era.
Tadamori's connection with the Sung trade is undocumented, but
during the course of his several commissions to chastise pirates in
the Inland Sea area, Tadamori established impressive regional bases
along the Inland Sea, in both Shikoku and Honshu. These later developed into a network of bases, to which the Taira retreated during
the GempeiWar of the 1180s. With bases in the west, Tadamori may
have begun to participate in the Sung trade, which was then vigorously under way in northern Kyushu.
One incident, noted in Choshuki, the diary of Minamoto no Morotoki, hints at such an involvement.82 In 1132, Tadamori was serv80 Yasuda, Insei to Heishi, p. 167. 81 Yasuda, Insei to Heishi, p. 167.
82 Yasuda, Insei to Heishi, pp. 168-69, based on Choshuki.
Cambridge Histories Online © Cambridge University Press, 2008
FOREIGN RELATIONS
635
ing as governor of Bizen. He was also a retainer of retired emperor
Toba serving as azukaridokoro on the imperial estate Kanzaki-no-sho
in the province of Hizen. That year a Chinese merchant vessel
landed at Kanzaki-no-sho, and Dazaifu officials, in accordance with
precedent, carried out an exchange for the merchandise on board.
Tadamori, however, objected, calling this interference. He forged a
document, which he claimed was a decree issued by the ex-emperor,
giving him - as azukaridokoro of the imperially owned estate - authority to deal with the ship. Since the ship had landed on estate
property, he claimed that Dazaifu officials did not have jurisdiction.
The local officials complained to the court, severely criticizing
Tadamori's behavior. It is likely that Toba's support allowed Tadamori to escape censure.
It is Taira no Kiyomori who is best known for having an interest
in foreign relations. As governor of Aki, Kiyomori ordered construction of the main hall of Itsukushima Shrine as a display of reverence
for the tutelary god of navigation and to serve as a base for maritime
activities on the Inland Sea. Kiyomori's concern with the control of
the Inland Sea may have been an indication of his desire to monopolize overseas trade. A more important step came with Taira monopoly of the post oidazai no daini. Kiyomori was appointed in 1158
and named a kinsman to act for him. But whenYorimori succeeded
his brother in the post, he went to Kyushu and was influential in establishing the Taira family base on the island. With control of this influential Daizaifu post secure, the Taira were in an excellent position
to monopolize Sung trade.
Although Chinese vessels entered a number of ports along the
Japan Sea and East China Sea coasts, the official port of entry was
Hakata. For obvious reasons, foreign ships were prohibited from entering the Inland Sea; there was even a watch point at Moji. After
Kiyomori became Chancellor in 1167, however, permission was
granted for foreign ships to enter the Inland Sea, and a special port
was constructed at Owada-no-tomari (Kobe) to accommodate them.
Thus, the Kyoto nobles had easy access to Chinese goods.83 Kiyomori's interest in the trade is reflected in his constructing a villa at
nearby Fukuhara.
Fortuitously, the period of Kiyomori's political dominance coincided with the reign of the Southern Sung emperor Hsiao Tsung
(r. 1163-90), whose domestic policies and interest in stimulating for83 Yasuda, Insei to Heishi, p. 170.
Cambridge Histories Online © Cambridge University Press, 2008
636
INSEI
eign trade made him one of the leading figures in the history of the
dynasty. He lowered customs rates to encourage visits by foreign
merchants, and clarified tariffs and standardized procedures governing commerce to stimulate Chinese traders to go abroad. The purpose behind the revived trade was, of course, to obtain profits for the
imperial treasury.
In 1172, Hsiao Tsung requested that the Japanese court send an
official envoy to Sung. Naturally, some nobles held to the traditional
position that it was outrageous for the Chinese to request, in effect,
tribute and use the term Nihon kokuo- the "king of Japan" - to refer
to their sovereign. But Kiyomori, it seems, chose to meet with the
Sung envoy, draft a reply, and send gifts to the Chinese emperor,
thus initiating formal trade between the two countries. With Japanese and Chinese vessels plying the waters of the East China Sea,
Japan was drawn into the East Asian trading orbit.84
The major Japanese trade items were sulfur, mercury, mother-ofpearl, fans, lacquerwork in gold and silver, and swords. Heian
courtiers were eager for high-quality medicines, scents, porcelains,
fabrics, and books from China. However, as domestic Japanese commerce began to develop, Chinese copper coins came to be the primary object of Japan's trading interests. Coins minted in Japan (the
so-called kdcho juniseri) had gone out of circulation by the end of the
tenth century and had been replaced by rice, silk, or gold and silver.
But the expansion of shoen held by capital nobles and monasteries
led to an increase in the exchange of goods between province and
capital, and by the twelfth century there was a demand for a circulating currency. The Japanese turned to Sung coins.
These coins were accepted for the purchase of residential land and
paddy fields and the payment of shoen rents; soon they were being
used in commercial transactions, in the payment of wages for building temples, shrines, and the like. The Northern Sung had previously
forbidden the export of these coins, when the trade had become so
extensive as to present a potential problem; but the prohibition was
lifted in 1074, in the reign of Emperor Shen Tsung. Export volume
immediately increased. Sporadic attempts to limit the export of coins
proved impractical in light of the great demand. In 1171, during
Southern Sung emperor Hsiao Tsung's reign, there was a final, unsuccessful attempt at prohibition.
Coins from the mainland were most likely received in trade by
84 Yasuda, Insei to Heishi, p. 171.
Cambridge Histories Online © Cambridge University Press, 2008
THE INSEI IN RETROSPECT
637
owners of shden in the Kyushu area, who then used the money to pay
shden rents and other fees due to capital proprietors, so that the
coins ultimately came to be concentrated in Kyoto. Of course,
money also moved through the hands of Kyoto merchants to nobles
and priests and into the possession of powerful members of rural society. Contemporary documents show that these people had money
for the purchase of land. Similarly, coins filtered into the hands of
nonagricultural merchants, craftsmen, and the like. From the middle of the twelfth century on, Sung copper coins spread rapidly
throughout Japan.85
One of the main reasons for the increased demand for Sung coins
was Japan's inability to produce enough gold and silver. When coins
stopped being minted in Nara and early Heian, gold and silver came
to be employed quite widely by Heian nobles. Both were used to
purchase goods from Chinese trading vessels at Dazaifu. When gold
and silver production dropped, and these metals no longer functioned as substitutes for minted coins, the demand for Sung coins
naturally rose, not only among nobles and clerics but also among
new local landholders and members of the peasant class involved in
crafts production and commerce.86
This did not go unnoticed by the ruling noble class. The increasing development of a monetary economy threatened their control of
society. A serious attempt was made to curtail the import of Sung
coins; there was a sense that people were becoming overly concerned
with the acquisition of money. It was said that society was suffering
from a "money malady" (zeni no yamai).87 This, of course, simply reflected a considerable degree of societal change, which presaged the
rise of lower-stratum members of Japanese society, a development
that was to become more obvious in medieval times. The nobility was
aware that nonaristocrats were obtaining unprecedented economic
power and, further, that nonagricultural groups - merchants, craftsmen, and the like - were expanding. But ultimately the nobility was
too greedy for domestic and foreign produce to stop this process.
THE INSEI IN RETROSPECT
Since the work of early historians like Jien and Chikafusa, Japanese
scholars have focused on the late Heian period as a time when re85 Miura, "10 seiki-13 seiki no Higashi Ajia to Nihon," pp. 266-67.
86 Miura, "10 seiki-13 seiki," p. 267.
87 Miura, "10 seiki-13 seiki," p. 267;Yasuda, Insei to Heishi, p. 174.
Cambridge Histories Online © Cambridge University Press, 2008
638
INSEI
tired emperors dominated the political system. They have used the
term insei to characterize this political form, in which the emperor,
after abdication, established an in-no-cho, which was the repository
of state power.
There has been an attempt to link the insei with Fujiwara regency
politics (sekkan seiji), which reached a high point under the Fujiwara
ministers Michinaga andYorimichi. Not long ago, it was common to
claim that under regency politics, national administration was privatized, and the household Administrative Office (mandokoro) of the
Fujiwara regent's house in effect replaced the constituted offices of
the state, with its documents deciding important issues. Inset was
seen as a form of government in which these functions shifted from
the Fujiwara mandokoro to the in-no-cho, with ex-emperor documents replacing imperial edicts. Though not without some truth,
this was a rather mechanistic view of Heian politics according to
which consciously instituted changes in the governmental structure
compromised imperial authority. Both traditional and modern
scholars were consequently critical of sekkan seiji and insei as aberrations from the political norm.
In the last few decades, however, there has been considerable revision of this viewpoint in Japan and the West. Neither sekkan seiji
nor insei now is believed to have been a cohesive system of political
control.88 Fujiwara documents did not replace imperial edicts, nor
were decisions of state the personal prerogative of the Fujiwara regent's house. The in-no-cho, especially during the Shirakawa and
Toba eras, was not an instrument of national government but,
rather, a body for the furtherance of imperial house interests, similar to the mandokoro of the sekkanke and other noble houses. The late
Suzuki Shigeo demonstrated that through 1183, in-no-cho documents did not replace imperial edicts but dealt essentially with estate holdings of the members of the imperial kin group and their
clients. The private decrees of the ex-emperor (inzen), however, were
powerful enough to move the Council of State to action.89
In short, contemporary scholarship sees both sekkan seiji and insei
in a new light. Neither replaced the old governmental structure but
used it to further the political fortunes of the Fujiwara in one case
and the imperial house in the other. In particular, the insei resulted
88 See, for example, Hashimoto Yoshihiko's articles on sekkan seiji and insei in his Heian kizoku shakai no kenkyu (Tokyo: Yoshikawa kobunkan, 1976), pp. 85-114.
89 Suzuki's work is mentioned in the studies by Takeuchi and in Hashimoto (see note 88), but
the fullest exposition of his work remains the thesis mentioned in note 55.
Cambridge Histories Online © Cambridge University Press, 2008
THE INSEI IN RETROSPECT
639
in control over the constituted offices of the traditional government
by making in officials of courtiers holding important posts, or conversely, having in officials appointed to posts close to the emperor or
to lucrative provincial governorships. The patron-client relationships
cultivated between the ex-sovereign and his retainers were crucial in
controlling the Heian political system.
Of course, the key was the ability to appoint in officials to important posts. What was the nature of the position of ex-emperor in
Heian society - and even later - that made such appointments possible? It was not spelled out in the imperial codes, but neither was
the role of the emperor set down clearly. At least two points should
be raised here, both of which have been touched upon in the earlier
discussion.
First, the familial role of the ex-emperor developed over the early
Heian period. The ex-emperor, as father or grandfather of the emperor, normally served as paterfamilias of the imperial kin group,
and parental authority over family members was quite strong. Some
scholars have tried to equate the positions of regent and ex-emperor.
The regent, by virtue of his maternal relationship with the emperor - maternal grandfather or uncle - exercised a kind of parental
authority over the sovereign. That is why the sekkanke favored young
emperors. The ex-emperor, working from the paternal side, could
enjoy a similar familial control over an emperor, especially with the
beginning of a shift from a matrilocal to a patrilocal marriage system
in late Heian times. It was difficult for an ex-sovereign to exercise
much familial authority over a brother who was emperor, or over an
adult sovereign.
The ability to control an emperor through the exercise of natural
familial authority is only part of the explanation, however; there was
a very real difference between the respective positions of the regent
and the ex-emperor. As was pointed out earlier, one cannot easily
step down from the highest position in any complex organization, especially a dynasty, and return to a common existence. Authority
tends to remain with a person. Thus in Japan, one who had been
sovereign, not merely the recognized symbolic head of the state
structure but also the religious link in a monarchy considered to be
descended from the Sun Goddess, enjoyed an exalted position.
From the beginning, there was considerable confusion as to the respective roles of the sovereign and abdicated sovereign. We have seen
that before the late Heian period, ex-emperors were often able to
make their will known through edicts and verbal commands and to
Cambridge Histories Online © Cambridge University Press, 2008
64O
INSEI
exercise substantial power and influence. Examination of historical
materials allows us to conclude that sovereignty remained largely
with the emperor - if for no other reason than that there had to be an
emperor, while an abdicated emperor was not a political necessity.
Two well-known examples give evidence of the confusion over authority. In 1163, a dispute broke out between the local officials in the
province of Kai and Yashiro-no-sho of the Kumano Shrine.90 The
province tried to abolish the shoen on the basis of the regulation ordinance issued a few years before, during Go-Shirakawa's reign. The
shoen officials claimed exemption on the grounds of a document
from ex-emperor Toba's in-no-chd. In the capital, Doctor of Law
Nakahara no Naritomo was ordered to consider the case. In his
opinion, Naritomo wrote that there was no distinction between the
emperor and the retired emperor, and therefore retired emperor
documents were equivalent to imperial edicts.
In another famous example, Kuj5 Kanezane records in his diary a
conversation with Minamoto noYoritomo in 1190, when the latter
made his first visit to the Heian capital since he had been exiled as
a youth.91 According to Kanezane, Yoritomo recognized that exemperor Go-Shirakawa controlled politics and said that Emperor
Go-Toba was like a crown prince. This captured the reality of the
day perfectly - and Yoritomo noted that since both Go-Toba and
Kanezane were young, their best chance for political success would
come when Go-Shirakawa died and control passed to Go-Toba.
There are other documents indicating that for virtually all of the
Heian period imperial edicts were, strictly speaking, more authoritative, more orthodox than ex-sovereign documents. But obviously
there was uncertainty in many quarters, and the distinction may be
clearer to us today than it was then.
What historians call the insei was not a legally based system of political control but a more informal order of organization, which
sought to utilize the existing political system. With the development
of the insei the imperial house acquired the familial resources that
other noble houses {kemmon seika) enjoyed as private interest groups.
To be a kemmon it was necessary to have (1) a well-organized house
administrative structure; (2) shoen holdings as an economic base;
(3) private military forces; and (4) the principle of patriarchal control
by the house head.92
90 Takeuchi, Bushi no wjo, pp. 188-89. 9 1 Gyokuyo, vol. 3, p. 635, Kenkyu 1/11/9.
92 Kuroda Toshio, Shoensei shakai, vol. 2 in Taikei Nihon rekishi, 6 vols. (Tokyo: Nihon hyoronsha, 1967-68), pp. 114—15.
Cambridge Histories Online © Cambridge University Press, 2008
THE INSEI IN RETROSPECT
641
The in-no-chd provided a structure for such a private house organization for the imperial house, but during the heyday of Fujiwara
clan power, the imperial house was strongly hampered from expanding this familial organization. From Go-Sanj5's time on, however, Fujiwara control over the royal house was broken, and successive heads of the imperial house, whether emperor or ex-emperor,
exerted great efforts to increase family power and wealth. Attracting
numerous clients, including warriors, and building up considerable
estate holdings and other forms of wealth, the imperial house enjoyed tremendous power and influence in the years before military
rule came to Japan. The head of the house could rule with almost
equal political influence as emperor (shinsei) or through the emperor
as ex-emperor (inset).93
Although the in-no-chd never lost its private function, serving primarily as a private organ for the management of royal estate holdings, at some point the abdicated emperor became the acknowledged leader in national affairs; as a result, the in-no-chd came to
play a more public role. Exactly when this shift took place is extremely hard to discern, since it was an informal rather than a legal
development. Most scholars agree that the in-no-chd was not a national form of government in the time of either Shirakawa or Toba,
although the power and authority of the ex-emperor increased.94 The
shift came during Go-Shirakawa's tenure as ex-sovereign.
Some scholars will point to an 1174 in-no-chd order issued to
Todaiji's Kuroda-no-sho in Iga Province, deciding a dispute between local officials and the estate owner.95 There is also an 1176
order addressed to the local officials of all the provinces rather than
to a specific addressee in one estate or local office. This is an indication that ex-sovereign authority was extending into broader areas of
jurisdiction, transcending its earlier, narrower role.
The main national issue in late Heian and early Kamakura was
shden and provincial control of productive lands throughout Japan.
The in-no-chd seems to have exercised great authority in this area understandably, since it had almost always been connected with
shden management. Sometime in Go-Shirakawa's era, its functions
expanded to deal more broadly with the regulation and control of
93 Kuroda, Shoensei shakai, pp. n o .
94 See, for example, Ishimaru Hiroshi, "Inseiki chigyokokusei ni tsuite no ichikosatsu," in
Kurokawa Takaaki and Kitazume Masao, eds., Kamakura seiken, vol. 4 in Ronshu Nihon
rekishi (Tokyo: Yuseido, 1976), p. 17.
95 Ishimaru relies on both of these examples to conclude that the in-no-chd becomes an instrument of national policy in Go-Shirakawa's era, p. 19.
Cambridge Histories Online © Cambridge University Press, 2008
642
INSEI
land. In the opinion of some scholars, the in-no-cho may have become a body possessing final authority to decide disputes during
Go-Shirakawa's era.96 Perhaps it happened during the Gempei War,
when the Taira fled the capital with the young emperor Antoku, and
Go-Shirakawa became the supreme authority and greatest defender
of noble interests against the rising eastern warriors. An examination
of extant documents shows that from this time in-no-cho kudashibumi
and inzen are addressed to a broader audience and cover a greater
range of affairs than before.97
National government was not a necessity in the insei period. Land
and labor had fallen under the control of influential noble houses
and local proprietors. Besides defense against foreign attack, what
was really necessary at a national level was the ability to mediate disputes among the noble houses and local proprietors, and to suppress
rebellions that would upset the balance of power. "Thus national
power was a tool to support the private control of the influential
noble houses."98 Given that virtually all once-public functions and
offices had been privatized, it is not so strange that the privatized
family organization of the ruling house enjoyed the greatest degree
of power and authority among the private groups at court in late
Heian times.
The insei may be a classic case of what Kent Flannery, in his discussion of the evolution of civilizations, has called "promotion."99
Promotion is an evolutionary mechanism induced by socioenvironmental stresses. "An institution may rise from its place in the control hierarchy to assume a position in a higher level; it may in the
process go from 'special purpose' to 'general purpose.' " loo This is essentially true for the in-no-cho, which started as a special purpose institution for the organization of imperial house interests. In the
twelfth century, as the crises facing the Japanese state became
acute - there was a loss of central power, a rise of provincial warriors, and increasing lawlessness among the unruly monks of the
great temples - it was "promoted," or, rather, promoted itself, to a
general interest institution. It became less system-serving and more
self-serving.
Insei is a term that has a long history. It refers to a series of de96
97
98
99
Ishimaru, "Inseiki," p. 19.
See the appropriate periods in both Heian ibun and Kamakura ibun.
Kuroda, Shoenshi shakai, p. 117.
Kent Flannery, "The Cultural Evolution of Civilizations," Annual Review of Ecology and
Systemics 3 (1972): 399-426.
100 Flannery, "Cultural Evolution," p. 413.
Cambridge Histories Online © Cambridge University Press, 2008
THE INSEI IN RETROSPECT
643
velopments in late Heian and Kamakura Japan when aristocratic
control of the social and economic resources of the country started
seriously to weaken in the face of the rise of a new class with fundamentally different values. Out of this situation rose several exemperors whose dominant personalities allowed them to exercise
considerable influence over the affairs of the day and create political configurations that distinguish this period of Japanese history
from all others.
Cambridge Histories Online © Cambridge University Press, 2008
CHAPTER 10
THE RISE OF THE WARRIORS
ORIGINS OF THE WARRIORS
The term for the Japanese warrior, bushi ("martial servitor"), which
became common from the late Heian period on, was preceded in
earlier centuries by the words tsuwamono and musha.1 Bushi, like the
other two words, specified the professional warrior as distinguished
from peasant conscripts, court military officials, and palace guards.
Tsuwamono is a native Japanese noun of unknown morphology, and
musha (or musa) a Sino-Japanese binom meaning "martial one."
The professional warriors were private fighters who at first fought
entirely in their own interests and for their own ends. They later entered the personal retainerships of court nobles or, more commonly, military lords, the greater of whom were called toryo, another Sino-Japanese word, the constituents of 'which denoted
"rooftree and rafters," that is, the chief members of a roof and, by
extension, also "chief person." The term was commonly used to
mean the chieftain of a major clan or family. As private fighters, the
tsuwamono were fundamentally different in character from government troops, who were conscripted by authority of the state and
fought solely for it.
During the seven hundred years from the twelfth century to the
nineteenth, Japanese society was dominated by warriors, who first
appeared as a potent independent political force in the country at
the time of the revolt ofTaira no Masakado in the mid-tenth century.
Masakado and many of his supporters and opponents were profesi This chapter was translated by Patricia Sippel and adapted and expanded by Regine Johnson and the editors. It is based on Japanese primary and secondary works cited in the following footnotes. Studies in English on warriors of the Heian period include William
Wayne Farris, Heavenly Warriors: The Evolution of Japan's Military, 500—1300 (Cambridge,
Mass.: Harvard East Asian Monographs, 1992); Karl F. Friday, Hired Swords: The Rise of
Private Warrior Power in Early Japan (Stanford, Calif.: Stanford University Press, 1992);
and George B. Sansom, A History of Japan to 1334 (Stanford, Calif.: Stanford University
Press, 1958).
644
Cambridge Histories Online © Cambridge University Press, 2008
ORIGINS OF THE WARRIORS
645
sional fighting men, mounted archers descended in their skills perhaps from the prehistoric hunters found in the archaeological record
of the eastern provinces. Mounted warriors are known in Japan from
before the seventh century, but they first became professional and a
locus of autonomous political power during the tenth century, when
their arms were employed not on behalf of the government for its
purposes but in pursuit of private interests. As the organizational
skills of such warriors grew, they formed retainer-based military
bands of increasing size and power, structures that eventually enabled them to contend for control of the country. Their triumph over
a weakened imperial court in the twelfth century led to the establishment of a new, warrior-ruled system of government in the Kamakura period (1185-1333).
The popular modern image of the Japanese warrior has been
shaped largely by colorful literary and pictorial representations of fifteenth- and sixteenth-century battles in which swordplay figured
prominently, and also by the knightly figure of the samurai as found
in the status system of the Tokugawa bakufu. The Heian warrior was
neither. He was primarily a mounted archer who wielded the dagger
and sword when his supply of arrows was gone and fighting had become hand-to-hand. At this time the bow rather than the sword was
the weapon of choice. The warrior's status was not, like that of the
Edo-period samurai, clearly established near the top of a centralized
military society, nor did he see himself, his superiors, or his society
in the ideal categories employed by later Confucian pedagogues. Derived from a verb meaning to attend on a superior, samurai was used
in that sense (in its earlier pronunciation of saburai) in an earlyeighth-century chronicle to designate civilian attendants assigned to
the old or infirm or in the menial service of the sovereign, and it continued to be used in that nonmilitary sense in the Heian period for
low-ranking servants of the nobility. As a title within court society,
samurai was also applied to military attendants.
The main forces of the statutory (ritsuryo) imperial court in the
eighth century were peasant conscripts, mostly a foot soldiery organized into units of three hundred or fewer to a thousand or more
men. Military service was a heavy economic burden for the conscripts, who were required to furnish their own weapons, clothing,
and other equipment, specified by law to include a bow, bowstrings,
arrows, a quiver, two swords, leggings, boots, a reed hat, a whetstone, and containers for water, food, and salt. Each ten-man unit
was also required to supply a tent, tools, cooking equipment, and
Cambridge Histories Online © Cambridge University Press, 2008
646
THE RISE OF THE WARRIORS
moxa.2 The conscription of a single man, it was said, could bring
about the collapse of his entire family.
Modeled on the armies of T'ang China, the statutory conscript
system provided the imperial court with a well-articulated military
force, lightly armed and trained. But the operation of the system was
plagued with trouble from the beginning; individual units sometimes
degenerated into little more than work gangs in the employ of officials. In the absence of major internal enemies after the middle of the
eighth century and the fading of the perceived threat of Chinese and
Korean aggression, the system was generally abandoned in 780 and
792, retained then only in what were considered strategic "frontier"
areas, namely, the northeast (the provinces of Dewa and Mutsu) and
the Dazaifu provinces of Kyushu. It was partially replaced by small
groups of "stalwart youths" (kondei), elite mounted fighting men who
first appeared in the eighth-century east, where their martial skills
were likely learned in some part from their redoubtable Emishi enemies, especially in the matter of fighting from horseback. (For an account of the expeditions against the tribal Emishi in the northeast
around 800, see Chapter 1.)
Little is known about kondei duties and functions. In 733, three
hundred conscripts from Mutsu and Dewa skilled in mounted
archery were named kondei. It is significant that kondei were first appointed in the military districts responsible for the subjugation of the
Emishi, and that mounted archers were recognized as being more effective than foot soldiers in combat with those tribal people. The
kondei of 792 and after were members of powerful local families
igunji, "district officials") and cultivators sufficiently prosperous to
equip, mount, and support a warrior, recruited from what was likely
a substantial pool of hunters in the area.3 Provisions and two grooms
(Jbatei) aged seventeen to twenty were supplied for each kondei, one
of the grooms actually functioning perhaps as a batman.4
Although the kondei were no doubt militarily superior to the peasant conscripts of the old system, their small numbers nationwide indicate that they could not have fully replaced the conscripts. There
were fewer than four thousand scattered in units of twenty to two
hundred throughout Japan, and they could have functioned, one
2 Ritsu. Ryo no gige, (Shintei zoho) Kokushi taikei, ed. Kuroita Katsumi, vol. 22 (Tokyo: Yoshikawa kobunkan, 1939), pp. 183-85.
3 Takeuchi Rizo, "Togoku no 'tsuwamono,' " Miura kobunka, 25, 5 (1979).
4 Ruijii sandai kyaku, (Shintei zoho) Kokushi taikei, vol. 25 (1936), kan 18, p. 548, 22/1 l/Enryaku 14 (795), and pp. 553-55.3/11/Tencho 3 (826).
Cambridge Histories Online © Cambridge University Press, 2008
ORIGINS OF THE WARRIORS
647
may assume, only as a constabulary in the protection of government
facilities and operations.5 The central government had in effect
ceded responsibility for the preservation of the public peace outside
the vicinity of the capital to the landed class.
The conscript system was not disbanded in Kyushu until 928, replaced there in part by units of "select youths" (senshi), who were
similar to the kondei in origin and presumed functions but were
somewhat better attended.6 Although there were fewer horses and
archers in Kyushu than in the east, the creation of senshi units stimulated a similar, if belated, development of a warrior class there, too.
The disestablishment of the conscript armies and their replacement by small units suitable only for the security of government offices and facilities signaled the abandonment of state responsibility
for the peace of the country. Reflecting that shrinkage of government
intention, the sources for the period record no actions involving either kondei or senshi. The name kondei office (kondeidokord) survived
within provincial-government structures until the end of the Heian
period, but the offices came to have merely titular existences.
The development of a warrior class in the tenth century and after
cannot be directly linked to either the kondei or senshi, but the men
of those units were clearly products of socioeconomic circumstances
similar to those that led to the emergence of the private professional
warrior. They were mounted fighters and shared common origins
with the direct ancestor of the warrior class, the tsuwamono. Both the
kondei and the tsuwamono came largely from a prosperous landed
class, especially in the east, which was accustomed to hunting from
horseback and had a long tradition of bearing arms. Where the kondei were insufficient in number to suppress banditry, provincial landlords became more active in military preparedness. They relied increasingly on arms also to settle their disputes and to extend their
holdings when they found the opportunity.
Under the statutory laws the provincial governor was conceived of
as a civilian authority and his staff was not permitted to bear arms.
However, in the ninth century, because of lawlessness in the countryside, governors of certain provinces were granted permission to
engage armed men to protect themselves and their headquarter
compounds and to enforce their orders. By the beginning of the
tenth century and perhaps earlier, when serious disorders occurred
5 Ruiju sandai kyaku, kan 18, pp. 558-59, 14/6/Enryaku 11 (792).
6 Ruiju sandai kyaku, kan 18, pp. 553—54, 3/11/Tencho 3 (826).
Cambridge Histories Online © Cambridge University Press, 2008
648
THE RISE OF THE WARRIORS
in a province, the court issued additional police or military titles to
the governor or his deputies to empower them to raise and lead warriors in order to apprehend miscreants and restore order. These special commissions were intended to be temporary to meet emergencies, but they often became long-term and sometimes hereditary. In
the eleventh century, provincial headquarters continued to develop
as military centers, regularly garrisoned, and the governor was authorized to engage the services as needed of professional warriors
{tsuwamono) of the locality.
The word tsuwamono in its earliest known uses referred broadly to
the means of warfare - weapons, provisions, and skills - and also to
the persons who made use of those to wage war - warriors. By the
ninth century, the term was restricted to the human element of the
definition, the fighter, and in the tenth century tsuwamono came to
designate more specifically the professional fighter. It was in that
sense thatTaira no Masakado's "fame as a tsuwamono" was recorded
in Shomonki.7 Tsuwamono acquired retainers (rotd, "lads" or, more
literally, "young companions"), but initially not in the relationship of
a lord to his vassals. The term roto referred rather to the fellow tsuwamono whose interests were in common with those of their leader and
who fought under him. Since they were not strictly subordinate to
their leader, they are perhaps better described in their earliest form
as companions. Rudimentary organizations of tsuwamono and row
developed by the middle of the tenth century, thereafter growing in
size and strength, but roto tended to maintain their relationship with
their leader only as long as it served their own interests, a tendency
that remained to some extent even after the subsequent development of personal bonds of loyalty and subordination had led to the
formation of the large-scale warrior retainerships that dominated the
political history of the twelfth century.
In acquiring retainers, tsuwamono may have been following the example of provincial governors. At any rate, the first occurrence of the
word roto is the case of a former governor of Tosa, Ki no Tsurayuki,
who mentions roto among his retinue as he sets off on his return voyage from Tosa to the capital in 934. Roto are not found with tsuwamono until a few years later, at the time of Masakado's revolt. The
practice of acquiring retainers seems to have been widespread
among governors. The value of retainers to the governor came to be
more than simply personal security; they were employed for admin7 Shomonki, in Kodai seiji shaken shiso, ed. Yamagishi Tokuhei, Takeuchi Rizo, Ienaga Saburo, and
Osone Shosuke, vol. 8 of Nihon shiso taikei, 67 vols. (Tokyo: Iwanami shoten, 1979), p. 213.
Cambridge Histories Online © Cambridge University Press, 2008
ORIGINS OF THE WARRIORS
649
istrative work and were also available for police or military functions.
They appear to have been key instruments in the extraction of
wealth from a province. In one case, it is reported, the dispersal of a
governor's retainers after his death resulted in the impoverishment
of his family. But a governor was a civil official, and when his term
of office expired, his retainers might remain to serve the next governor or seek employment with the governor of another province. The
tsuwamono, on the other hand, was a professional, hereditary warrior
vested with interests in land, and his row might continue in service
with him and his family indefinitely.
The tsuwamono warrior was the product of sweeping changes in
landholding that in the eighth century and after transformed the
public, imperial lands and people of the statutory regime into a
stratified society of wealthy, powerful private landholders and a subservient peasantry.8 The new landholders were in some cases local in
genesis, their lands and labor force growing from an original family
and its state allotment of land, and in other cases stemming from external authority, descending from officials and military officers appointed in a province by the central government. As the holdings of
the new lords expanded and control of an increasingly independent
peasant workforce grew more difficult, and as commendation converted land into shoen estates with complex and often conflictive external relationships, military might became a useful and necessary
adjunct of landholding. The nucleus of the military forces created on
the great landholdings was the tsuwamono, who enforced the will of
the landholder, defended the holding against external threats, official and private, and kept the peace.
The earlier shoen were owned by higher nobles and by major temples and shrines that, as absentee landlords, sent a manager to oversee cultivation and remit revenues. However, by the tenth century,
most new shoen were established by families, local or newly arrived
from the capital, who commended land to a high noble or religious
institution or member of the imperial family (ryoke or honjo) with influence enough to protect the shoen from interference by provincial
authorities and from provincial exactions. In return the ownermanager paid his protector a percentage of the rents collected from
farmers. Both the new shoen and the older type founded by an owner
in the capital required local military force to keep order on the shoen,
to collect rents and taxes from the cultivators, to stave off the gover8 On statutory landholding institutions, see Cambridge History of Japan, vol. 1, chapter 8. On the
evolution of these institutions in the early Heian period, see Chapters 3 and 4 of this volume.
Cambridge Histories Online © Cambridge University Press, 2008
650
THE RISE OF THE WARRIORS
nor's men, to resist encroachment by nearby shden, and to expand
into kokugaryo (lands of the provincial government) or other lands
as opportunity arose. Tsuwamono were often the owners or managers
of shden. In the tenth and eleventh centuries they were also found increasingly among the administrators in the offices of the provincial
headquarters and managers of the province's lands.
As military might became a necessary element in successful landholding, two noble warrior clans established themselves as tsuwamono
both in the vicinity of the capital and in the provinces, especially and
most fatefully in the east. Their success was partly attributable to their
family lineage, the multiple lines of both clans originating from imperial offspring removed from the imperial family in the cost-cutting
of Emperor Saga's time or after and given the noble clan names of
Minamoto orTaira. (On that process, see Chapter 1 and Chapter 2.)
Many of those demoted princes remained in the capital pursuing conventional court careers, they and their descendants sometimes rising
to high office and rank, while others sought their fortunes in the
provinces, turning there to the profession of arms. But the most
prominent of these warriors maintained their connections at court
and, from time to time, held guard or police offices in the capital.
The Minamoto name was granted to one or more princes by
eleven emperors from Saga (r. 809-23) to Sanjo (r. 1011-16). The
multiple lines created in that manner are distinguished in nomenclature according to the name of the ancestral emperor: Saga Genji,
Murakami Genji, and so on, "Genji" being the Sino-Japanese reading of the characters with which "Minamoto clan" is written. The
bakufu of the Kamakura period (1185-1333) was founded by the line
of the Minamoto usually known as the Seiwa Genji, so called because they were descended from a prince named Tsunemoto (d.
961), who was supposed to be a grandson of Emperor Seiwa. But a
document discovered in the nineteenth century indicates that Tsunemoto was actually a grandson ofYozei (and thus a great-grandson
of Seiwa). Not all students of the subject accept the testimony of the
document, but if it is correct, we may suspect that the name of the
criminally inclined andfinallydeposed Yozei was avoided by family
historians in favor of a more auspicious genealogical origin for the
great line.9Tsunemoto's son, Mitsunaka (912-97), served the North9 "Iwashimizu Tanaka-ke monjo, Kawachi-no-kami Minamoto no Yorinobu komon," in
Takeuchi Rizo, ed., Heian ibun, Komonjo-hen, vol. 3 (Tokyo: Tokyo-do, 1963), p. 775; and
Hoshino Tsune, "Yo no iwayuru Seiwa Genji wa Yozei Genji naru kangae," in Shigaku sosetsu (Tokyo: Fuzambo, 1909), pp. 108-47.
Cambridge Histories Online © Cambridge University Press, 2008
ORIGINS OF THE WARRIORS
651
Emperor Seiwa
850-81
Pr Sadazumi
884-916
Emperor Yozei
869-949
I
Minamoto no Tsunemoto
d961
Mitsunaka
912-97
Yorimitsu
948-1021
Yorinobu
968-1048
Kawachi Genji
I
Yoriyoshi
988-1075
3 generations
Yoshiie
1039-1106
Yoshichika
d 1108
I
Yoshitsuna
d 1134
Yoshimitsu
1045-1127
Yoshikuni
dll55
Yoshitada
d 1109
Yoshikata
dll55
Yukiie
d 1186
Tameyoshi
1096-1156
Yorimasa
1104-80
Yoshitomo
1123-60
1
1
Yoshinaka
1154-84
Yoshihira
1141-60
Tomonaga
d 1159
Yoriie
1182-1204
Yoritomo
1147-99
Sanetomo
1192-1219
Noriyori
d1193?
Yoshitsune
1159-89
I
Ohime
Figure 10.1. Genealogy of the Seiwa Genji (Minamoto). Adapted
from Kodansha Encyclopedia of Japan.
Cambridge Histories Online © Cambridge University Press, 2008
652
THE RISE OF THE WARRIORS
em (regental) House of the Fujiwara as its military arm in overawing its rivals. For three more generations until the end of the
eleventh century the Seiwa Genji held court positions in the guards,
appointments to lead punitive expeditions, and rotations as provincial governors, and gained a reputation as the most formidable warriors in the land.
"Minamoto" was a common noun meaning "source of water,
origin." In the clan name, it was used in the sense of "source of
court officials." An alternative interpretation assumes that the
choice of the name by Saga, its first grantor, was influenced in the
court of that sinophile emperor by Chinese precedent, giving it
the sense "of the blood." The precedent occurred in China in the
reign of an emperor of the Northern Wei (386-534), who, in recognition of the common Hsien-pei origins of a Mongolian prince with
himself and the officers of his court, bestowed as a family name on
the prince the character with which Minamoto is written - Yuan (in
a modern Chinese pronunciation).
The princely recipients of the clan name of Taira were typically of
a remoter degree of imperial descent than their Minamoto counterparts: they were never sons of emperors (as Minamoto usually
were), but grandsons or genealogically even further removed from
the throne. There were several lines of the clan, springing like the
Minamoto from the progeny of different emperors (Kammu,
Koko, Nimmyo, and Montoku), but it was the first created lines, the
four Kammu Heishi lines (Heishi is the Sino-Japanese reading of
"Taira clan," descending from Emperor Kammu, that produced the
Taira who figured most prominently in the history of the period. See
Taira genealogy Figure 9.7). First bestowed in the ninth century, the
Taira name, in its orthography and Sino-Japanese reading, echoed
that of Kammu's new capital at Heian, the characters for which were
sometimes also read "Taira no Kyo" or "Taira no Miyako."
Less closely related to emperors than most of the Minamoto and
more thoroughly excluded from the higher offices of the court, some
Taira early left the capital for the provinces, where their prospects
were more encouraging. There were two chief branches of the
Kammu Heishi, one a courtier line of mostly minor and middlinglevel officials; the other, founded by Kammu's great-grandson Takamochi (fl. 889), became a warrior line that included many of the
best-known military men of the final century of the Heian period,
and it is almost always Takamochi's line that is meant when the term
"Kammu Heishi" is used. Takamochi was appointed vice-governor
Cambridge Histories Online © Cambridge University Press, 2008
REVOLTS OF MASAKADO AND SUMITOMO
653
of Kazusa around the end of the ninth century, married locally in the
province of his appointment, and expanded his landholdings in
Kazusa and Shimosa. His descendants included in the third generation Masakado, whose campaigns, from 935 to 940, may be said to
have marked the emergence of professional-warrior power in Japan.
REVOLTS OF MASAKADO AND SUMITOMO
The revolt in the east began inconspicuously in 935 as an armed
squabble among members of a Taira family who had settled in the
Kanto area perhaps forty-five years earlier. They had grown into
great landowners there, exercising control over a broad area that included the provinces of Kazusa, Hitachi, and Shimosa (the northern
part of Chiba Prefecture and the southern half of Ibaraki). As the
conflict evolved in fits and starts, one of the Taira leaders, Masakado,
emerged from his Shimosa base as the chief military power and principal arbiter of disputes throughout the southern Kanto area.10 The
disputes in which he involved himself centered on the resistance of
landowners to provincial exactions, Masakado taking the part of the
aggrieved landholders. The fighting increasingly assumed the nature
of rebellion against provincial authorities, who were mostly Masakado's kinsmen. But perhaps because the eastern provinces had a
long history of banditry, unrest, and minor revolts, the central government at first paid scant heed to Masakado and his fighting, intervening only when a suit was brought against him by one of his victims, and then simply to assess a mild punishment that was almost
immediately canceled in a general amnesty.
The military forces of Masakado and his opponents comprised
the leader's close followers (Jusha), a permanent bodyguard or army
of mounted warriors (400 in the case of Masakado), banrui (allies),
and provincial troops recruited from among the cultivators. Banrui
were minor warriors, usually mounted, who were independent collaborators and quick to retire from the battlefield when the fighting
went against them. Both Masakado and his enemies used the tactic
of burning down the houses of their opponents' banrui and peasant
foot soldiers and destroying their stored grain to weaken their re10 The principal source concerning Masakado's uprising is the Shomonki (see note 7), an account in hentai kambun ("variant Chinese"), believed by most scholars to have been completed a few months after Masakado's death in 940. See Judith N. Rabinovitch, trans.,
Shomonki: The Story of Masakado's Rebellion, Monumenta Nipponica Monograph 58
(Tokyo: Sophia University, 1986).
Cambridge Histories Online © Cambridge University Press, 2008
654
THE
RISE OF THE WARRIORS
solve, from which we can understand that they were men of meager
resources.
Although the organization of the armies of Masakado and his enemies were generally the same, they differed in one decisively important respect: the anti-Masakado forces were an alliance of warrior
leaders, at times as many as five or sixTaira and also a Minamoto and
a Fujiwara, each with his own group of close followers. These alliances were known as to (associations), largely egalitarian leagues of
warriors that joined together to further their common interests. (The
term to is used in modern times to refer to political parties.) Shortly
before Masakado's time, for instance, pack-horse haulers had formed
"hire-horse associations" (shuba no to) and were terrorizing and looting agricultural villages along the roads of the Tokaido and Tosando,
their depredations becoming so intolerable that the court established
barrier stations (sekisho) at key eastern passes to exercise some control over their movements. Nonhierarchical in membership at first,
such leagues in the twelfth century developed into small, loosely knit
warrior bands like the famous Seven Musashi Associations (Musashi
shichiio) in the Kanto or the Matsura Association of Kyushu.
Professional warriors of Masakado's time were archers on horseback. They did not by choice engage in hand-to-hand combat with
swords. Nor did the mounted men carry lances, as they did in Europe. Masakado's forces also included peasant foot soldiers armed
with spears and shields, employed in mass tactics, in the tradition of
the earlier imperial conscript armies.
The fighting between Masakado and his opponents, as it evolved
through its several stages, was destructive and murderous, but the
participants continued throughout the first years to appeal to the
court at Kyoto for justice. Imperial authority, although conspicuous
by the absence of its physical presence in the east, was still recognized
even by Masakado himself. Then, at the end of 939, Masakado became involved in a dispute over taxes between a member of the local
gentry in Hitachi and the governor of the province. Before the affair
ended, he had burned the provincial headquarters, seized the official
signet of the province, and made off with the key to the provincial
storehouse. Even the inward-looking court at Heian could scarcely
ignore so direct a challenge to its authority or income.
At the urging of a certain Prince Okiyo, a former provincial official in Musashi who had thrown in his lot with Masakado in the
course of another dispute with local officials in 939, Masakado set
out then to seize control of all the eight Kanto provinces, acting on
Cambridge Histories Online © Cambridge University Press, 2008
REVOLTS OF MASAKADO AND SUMITOMO
655
Okiyo's famous observation that the punishment for rebellion in
many provinces was no worse than for that in one.11 Claiming that
he was obeying an oracle from the Great Bodhisattva Hachiman and
justifying his action on the ground that he was a descendant of Emperor Karamu, Masakado proclaimed himself the "New Emperor"
(shinno), in contradistinction to the old tenno at Heian. Masakado
appointed new officials to the provinces and he also began to fashion a rustic version of the central statutory government centered in
Shimosa.
Masakado sent a message to the capital addressed to Fujiwara no
Tadahira, under whom in his youth he had enrolled himself as a follower. He sought the regent's understanding of his actions, and also
suggested that his ambitions did not extend beyond the Kanto, thus
proposing in effect a division of the country between the regental
Fujiwara and his own warrior family. By that time, however, the authorities at court were thoroughly alarmed, convinced that Masakado's forces would soon be descending on Kyoto. They adopted
a three-pronged policy aimed at ending the rebellion: (1) prayer;
(2) the appointment of Pursuit and Apprehension Agents (tsuibushi)
in the eastern provinces and, in the spring of 940, three or four
months after Masakado's attack on the Hitachi provincial headquarters, the dispatch of a court commander, Fujiwara no Tadafumi, to
the east; and (3) promises of reward to provincial leaders who succeeded in subduing Masakado. Significantly, it was the third prong
of the policy - reliance on provincial warrior leaders - that actually
brought Masakado down, although the court subsequently insisted
on attributing its success to the divine efficacy of the first. While the
specially deputed court commander was still en route to the east at
the beginning of 940, Masakado was surprised at his base in Shimosa
with a depleted force and was killed by the joint forces of his cousin,
Taira no Sadamori, and the Shimotsuke Suppression and Control
Agent (oryoshi) Fujiwara no Hidesato.
The failure of the revolt seems to have been chiefly a result of
organizational weakness. Masakado's lack of a retainership system
meant that his coalition force of antigovernment landholders
tended to fragment after initial successes had achieved the aims of
the landholders, and the dispersal of his large force of cultivators to
their agricultural pursuits left him exposed at a critical moment
to the overwhelming alliance of hostile forces. (The Shomonki says
11 Shomonki (text), p. 208.
Cambridge Histories Online © Cambridge University Press, 2008
656
THE RISE OF THE WARRIORS
that Masakado's force, which had once numbered as many as six
thousand, had been reduced to a thousand men.)
The imperial court's military response to the revolt may not have
been as ineffective as it seems. The punitive expedition dispatched
from the capital against Masakado arrived at the scene after his defeat, but its expected advent in the fighting presumably entered into
the calculations of the contending sides. Hidesato participated in the
fighting as an imperially appointed officer, holding one of the key titles recently created by the court as it reconstituted its military.
Following the abandonment of the conscript-army system in the
late eighth century and afterward, the statutory regime continued as
before to rely for military strengdi on the richer and more powerful
elements of rural society, asserting authority and control through a
loose, evolving structure of ad hoc and permanent titles that deputized individual warriors to exercise force in defense of the regime or
for the maintenance of public order. The social and provincial origins
of the regime's military strength remained much the same as earlier.
But the nature of the military's relationship to the court shifted from
the employed service of the Nara period, when men were recruited at
the government's pleasure and recompensed for their labor, to deputized agency in the ninth and tenth centuries, when the power of private warriors was confirmed and legitimatized by conferral of court
titles that may have been useful in the building of retainership systems. Such deputized warriors might fight the battles of the court
and enforce its laws, but first and foremost they fought in their own
interests and took the offensive when they had most to gain.
The titles awarded to warriors and the functions of the title holders as the functions evolved during the period gave the government's
military arm more than simply a paper existence. The resulting military and police system enabled the regime to survive rebellion and
rampant banditry for close to four centuries, and the new system
was in that sense a successful adaptation of statutory institutions to
a changing society. But lacking significant numbers of men and
weapons, and having no major military leaders of its own, the court
at Heian was able to do little more than survive. It did not have the
strength, certainly, to impose its unilateral will on the society.
The dispatch by the court of a force under the command of a senior noble to put down Masaskado's revolt continued the military
traditions of the statutory regime, but the ineffectiveness of that effort signaled the approaching end of the traditions. The commander
of the army subsequently led a more successful imperial army against
Cambridge Histories Online © Cambridge University Press, 2008
REVOLTS OF MASAKADO AND SUMITOMO
657
the pirate-rebel Sumitomo at the Dazaifu, but after that there were
no more such armies.
Instead, there came to be four chief, continuously filled titles
given to warriors entrusted with military and police duties outside
the capital. One, "General of the Pacification and Defense Headquarters" (chinjufu shogun), was a title of the Nara period that acquired a markedly different meaning as it came to be used during
the Heian period. Another, "Offenses Investigation Agent" (kebiishi), was originally created for police purposes in the capital city in
the ninth century, but in the tenth it expanded nationwide, such
agents then being established in each province by the local government. Two new titles formed the backbone of imperial military
framework: "Suppression and Control Agent" (prydshi) and "Pursuit
and Apprehension Agent" (tsuibushi).
The Pacification and Defense Headquarters had been established
in Mutsu Province in the first half of the eighth century to provide
security for Japanese colonizers against Emishi in the region. Its
large army of conscript soldiers from other provinces played a crucial role in the subjugation of the Emishi, continuing operations
against rebellious and bandit elements all through the ninth century.
It was staffed by four grades of military officers, headed by a general
{shogun). Beginning in the tenth century, the office of General of the
Headquarters tended to become hereditary, remaining in a single
clan or family line for as many as five successive generations. The
title fluctuated between lineages but remained always in three leading warrior families directly descended from high nobility at the
capital: sublineages of the Uona line of the Fujiwara, the Kammu
Heishi, and the Seiwa Genji.
When the Emishi submitted to the Japanese presence in the
northeast during the ninth century, the Pacification and Defense
Headquarters {chinjufu) lost its military raison d'etre, becoming
thereafter mainly a civil administrative organ, little or no different
perhaps from a normal provincial-government office. The holder of
the title of General typically held concurrent appointment to the
governorship of Mutsu, which in the eleventh century was associated with great wealth. Despite the loss of the office's primary military function, there was no other military title that brought so much
prestige to an eastern warrior, but the title seems to have been primarily a recognition of existing power and a reward for military exploits, not a grant of power or the commission of military duties.
In response to the upsurge of banditry in the countryside, the poCambridge Histories Online © Cambridge University Press, 2008
658
THE RISE OF THE WARRIORS
lice and judicial powers of the provincial governments came to be
separated and vested in special officers called, like their counterparts
in the capital. Offenses Investigation Agents. Such officers existed at
latest by the middle of ninth century, and by the tenth they are
found in nearly one-third of all provinces, from Musashi in the east
to the Dazaifu in Kyushu. Multiple kebiishi appointments might be
made for a single province, and two, presumably especially lawless
provinces, are known to have installed kebiishi also at the subprovincial, or district, level. Initially selected from rich and powerful families in a province, by the tenth century rankless peasants were holding the title, and their appointments, which had originally been
made by the central government, were in the hands of the provincial
headquarters. The appointees then tended to be functionaries of the
headquarters.
Fujiwara no Hidesato held the title of Suppression and Control
Agent when he and Taira no Sadamori defeated and killed Masakado in 940, and Minamoto noTsunemoto (d. 961), the ancestor of
the Seiwa Genji who was a member of the court's army dispatched
against Masakado, held the title of Pursuit and Apprehension
Agent. The two titles had emerged in the ninth and tenth centuries
as ad hoc court appointments of private warriors to mobilize and
lead their own and sometimes other local military forces in the suppression of revolts and banditry. Originally different in function, the
two titles came to be substantively the same in the tenth century,
when they lost their ad hoc character and were continuously filled,
each appointee having jurisdiction in a single province (Hidesato
was, for instance, the oryoshi of Shimotsuke) and in principle acting
under the direction of the provincial-government headquarters. Appointed by order of the central government, the agents were usually
chosen from among powerful local families, frequently from those
whose members were already part of the provincial headquarters
staff. The posts often became hereditary, and as officials actually
resident in a province, appointees were often able in the late Heian
period to take over the military and administrative authority of the
provincial government.
During the half year before Masakado set the fateful fire at the Hitachi provincial headquarters, the court had been rocked by news of
a major revolt by resettled Emishi in Dewa and a renewed outbreak
of piracy in the west, in the Inland Sea (Seto Naikai) area, where
piracy had been a continuing problem since the last half of the ninth
century. The piracy was soon contained, but then just two weeks after
Cambridge Histories Online © Cambridge University Press, 2008
REVOLTS OF MASAKADO AND SUMITOMO
659
word of the burning of the Hitachi headquarters reached Kyoto, the
court learned from the vice-governor of Bizen that a powerful pirate
leader named Fujiwara no Sumitomo, a great-grandson of Mototsune's biological father, Nagara, and therefore a distant cousin of the
current regent, was about to take to the sea again. A few days later
came the terrifying news that Sumitomo had caught up with the vicegovernor in Settsu as the latter was returning from Bizen to the capital, cut off his ears and slit his nose, killed his son, and taken prisoner his wife. Buffeted by alarming reports from east and west, the
court seems to have been on the verge of panic, fearing that Sumitomo and Masakado were acting in concert and that their joint forces
would soon be assaulting the virtually defenseless capital itself.
However, Sumitomo was not a tsuwamono of the Masakado type.
His rebellion followed its own course in its origins and resolution.
Sumitomo had formerly held a minor provincial post in Iyo (present
Ehime Prefecture) on the island of Shikoku, where he built a large
piratical following that was active from as early as 936. Sumitomo organized bands of pirates operating in the area, most of whom were
likely poor fishermen and petty seamen, and he began to plunder the
sea lanes and ports of the Inland Sea, the transportation routes for
the movement of government revenues and products from the western provinces to the capital.
After the renewal of his activities in 939, he attacked governmental headquarters in the provinces of Shikoku and along the Honshu
coast, burned ships, and spread havoc throughout the coastal areas.
Sumitomo's extensive control of the sea grew into a direct challenge
to governmental authority, but it was not until after the defeat of
Masakado in early 940 that the central government was able to turn
its full attention to the pirate chief. The authorities in Kyoto had earlier attempted to check his depredations by appeasement (he was
granted Fifth Rank), but now a government force made up of provincial leaders finally succeeded in attacking and, with the aid of a
turncoat, reducing Sumitomo's stronghold in Iyo. The pirate leader
fled then to Kyushu, where he occupied and sacked the lightly defended Dazaifu headquarters, continuing his predatory ways until at
Hakata he suffered another disastrous defeat in 941 at the hands of
court-dispatched land and sea forces supported by deserters from
his own ranks and he met his death.12
12 On Fujiwara no Sumitomo, see John W. Hall, Government and Local Power in Japan, $00 to
1700 (Princeton, N.J.: Princeton University Press, 1966), pp. 129-31.
Cambridge Histories Online © Cambridge University Press, 2008
660
THE RISE OF THE WARRIORS
The revolt of Masakado and the piracy of Sumitomo proved to be
not as serious threats as the court imagined; but some key elements
in the origins, prosecution, and suppression of the revolts were emblematic harbingers of the great historical changes that came during
the following two and a half centuries. The suppression of Masakado's revolt did not result in an assertion of court authority in the
east. Rather, the revolt and its suppression confirmed the private
wealth and military strength of the local leaders. Masakado and his
principal opponents, Taira no Sadamori, Fujiwara no Hidesato,
and Minamoto no Tsunemoto (the vice-governor of Musashi who
brought the initial charge of rebellion against Masakado at court),
were the ancestors of most of the important warrior leaders in the
east during the remainder of the Heian period, and it was from the
east that the decisive military power flowed in the fateful twelfth
century. All four men were alike in being descendants of the highest
strata of the court nobility (the imperial line and the Fujiwara clan).
Their ancestors had received appointments in the ninth and tenth
centuries to provincial or military posts in the east and had settled
and prospered there after the terms of their offices had expired.
In the typical case of the founder of a major warrior family in the
east, an ancestor of noble extraction first came to the area with the
court-conferred title of General of the chinjufu in Mutsu, and his descendants often continued to hold that title after him in a limited
hereditary fashion. The powers of the office of the chinjufu General
in the tenth century probably depended mostly on an appointee's
own private military resources, but it was a title of great prestige,
conferring on its holder what amounted to the titular leadership of
eastern warriors, and its official civil-military functions were still important. Although the Emishi had been decisively defeated by Japanese forces in the ninth century, large numbers in the northern and
central parts of Dewa and Mutsu continued to live in a semiindependent tribal society beyond the purview of regular provincial
and district officials. It was the responsibility of the chinjufu to manage and control them.
Bearing the old and distinguished title of chinjufu General, or the
somewhat less grand title of governor or vice-governor of a province,
and tied to the court not only by blood but also by occasional employment at the capital in court or noble-family military functions,
these men and their descendants quickly became major loci of
armed power in the eastern provinces, providing military protection
for local gentry against their enemies and contending among themCambridge Histories Online © Cambridge University Press, 2008
REVOLTS OF MASAKADO AND SUMITOMO
66l
selves for domination in an area. Sometimes called "military nobles"
(gunji kizoku) by historians, they were also, in a sense, trouble waiting to happen, ready to join in any fray that promised to be of advantage to their own interests, as Masakado did in the spring of 939
when he rushed to the relief of a district official in Musashi who had
come under attack by a provincial governor bent on the collection of
revenues and other wealth. Masakado himself was a professional
warrior, a tsuwamono in the contemporary lexicon, a predecessor of
the late-Heian bushi. "He gathered many fierce tsuwamono as his
companions and engaged in battles as his occupation," in the words
of a story about him in the twelfth-century Konjaku monogatarishu.13 He differed from the earlier warrior of the statutory state who
was in fundamental ways an employee of the state who fought for its
purposes. The professional warrior of the tenth century and after
was a private fighter who did battle as his own interests dictated, legitimizing his actions when he could as government business.
The failure of the two tenth-century revolts is attributable chiefly
in both cases to the fragmentation of the rebel forces, alliances
chiefly, in Masakado's case, of a warrior leader and rural gentry, and,
in Sumitomo's, of a renegade official and local piratical forces. After
Masakado's successes against the provincial government forces of
Hitachi and other Kanto provinces, the local gentry, their chief purpose achieved, dispersed, leaving Masakado with fewer than a thousand men in his final, losing battle with Sadamori and Hidesato. The
pattern was similar in the case of Sumitomo, who lost two of his pirate chieftains and twenty-five hundred men to the bribes of the governor of Iyo in 936 and who was later routed when his subcommander surrendered to government forces and led them to the rebel's
base in Iyo Province. There were no adequately cohesive social
bonds to hold the rebel forces together once their individual interests were served. Such bonds emerged only with the strong vassalage
system that in the late eleventh and twelfth centuries transformed
rural society.
Although the government's forces were perhaps inherently no
stronger than the rebels', the court was able, nevertheless, to field
punitive forces sufficiently powerful to suppress the revolts, thanks to
13 Yamada Yoshio, Yamada Tadao, Yamada Hideo, and Yamada Toshio, eds., Konjaku monogatari-shu, vols. 22-26 of Nihon koten bungaku taikei (Tokyo: Iwanami shoten, 1962), vol. 4,
kan 25, story 1, p. 362. Fascicle 25, devoted to tales about Heian warriors, has been translated by William Ritchie Wilson, "The Way of the Bow and Arrow: The Japanese Warrior in
the Konjaku Monogatari," Monumenta Nipponica 28 (1973): 190-233, see p. 190.
Cambridge Histories Online © Cambridge University Press, 2008
662
THE RISE OF THE WARRIORS
its own surviving military and financial resources and to its success
in allying itself with private warriors very like the rebels themselves provincial "military nobility" with strong familial and formal ties to
the court. For such figures, their services to the government in the
suppression of rebels must have seemed simply the continuation in a
different framework of the private battles that marked the expansion
of warrior strength in the provinces.
The revolts of Masakado and Sumitomo, although not revolutionary in effect, were true watersheds - in the opinion of some, the
beginning of the "middle age" (chusei). At least they were the first
major efforts by provincial military leaders to assert their autonomous strength. As the old ties of the military nobility with Kyoto
weakened and the strength and unity of the provincial military grew,
the court's position became ever more isolated and precarious,
maintained perhaps as much by ideological, religious, and sentimental values as by any important material strength.
The Masakado rebellion marked the advent of the private professional warrior in Japanese political history, and with his appearance
seems also to have emerged the first lineaments of a recognizable
warrior ideal. The loose code of conduct that in the Meiji era became generally known as bushido ("way of the warrior") is not directly traceable to the tsuwamono, but several prized attributes of the
tsuwamono may be thought of as having constituted an initial essay
at a warrior ethic. Knowledge of the early ideal is slender, since the
men left little personal record of themselves, and other sources are
problematic, existing as unique, single texts of uncertain authenticity, as in the case of Shomonki, or surviving as tales probably oral in
origin and collected over a century after Masakado's time (Konjaku
monogatari-shu). The picture found in such sources is sketchy, with
none of the detail of knightly perfection displayed for our admiration
in later sources from the thirteenth century and after, but some important elements are already in place.
Defining characteristics of the tsuwamono's conception of himself
are early seen in a letter written at the beginning of 940 by the prototypical tsuwamono Taira no Masakado. Written just at the time of
his assumption of a royal title at the head of his newly established
state in the east, and addressed in explanation of his actions to the
imperial regent at Heian, the letter as recorded in Shomonki justifies
the new state on three grounds: (1) the precedents of the past, history having recorded earlier instances of usurpation of central auCambridge Histories Online © Cambridge University Press, 2008
REVOLTS OF MASAKADO AND SUMITOMO
663
thority by arms; (2) Masakado's descent from an emperor, a lineage
that could entitle him to exercise sovereignty over the country; and
(3) his military power, with which he had been endowed by Heaven,
not by any governmental authority.14 Masakado's argument followed
a rationale suited to an age when the government had largely abdicated its defense and peacekeeping responsibilities, and the security
of the country was by default mostly in the hands of tsuwamono, or
so they claimed.
The bushi of later ages was depicted as placing supreme value on
loyalty to his lord, but the tsuwamono as represented in Masakado's
letter found his central values in historical practice, familial descent,
and military might. Those values underlay, or were expressed in, several characteristics of the ideal tsuwamono. Familial descent perhaps
more than anything else determined the warrior's status, for the profession of arms, like all other professions at the time, was hereditary
and closed at least in its upper reaches to those outside the tsuwamono families, the "houses of martial valor" (buyu no ie) or "bowand-arrow houses" (yumiya no ie), as they were sometimes called.
There was in consequence a strong emphasis on family honor and a
tendency to attribute one's accomplishments to the family line. On
being praised and rewarded by his master the crown prince for a
seemingly impossible arrow shot that killed a fox on the roof of Fujiwara no Kaneie's house one night, Minamoto no Yorimitsu
(948-1021), a court warrior and founder of a major sublineage of the
Seiwa Genji, is said to have responded, "It was not I alone who shot
that arrow. I was aided by my family's guardian deity in order not to
bring shame on my ancestors."15 Prayer to the guardian deity of a
family and to the family ancestors was essential. That centrality of
the family line led to the warrior's ritual proclamation before battle
of the names of his more notable ancestors, a practice that became
customary by the twelfth century, at least in literary accounts. That
recitation of ancestral names reflected an ancient belief in "word
spirit" (kotodama), according to which words were thought to be endowed with spirit and the intoning of a name gave the speaker the
protection and aid of the spirit of the name's referent.
Although military might may have been in practice the most significant trait of the tsuwamono, victory in battle was not necessarily
the highest goal. What seems to have been prized most in a warrior
14 Shomonki (text), pp. 210-13; Rabinovitch, trans., Shomonki pp. 113-16.
15 Konjaku monogatari-sku, vol. 4, kan 25, story 6, p. 382.
Cambridge Histories Online © Cambridge University Press, 2008
664
THE RISE OF THE WARRIORS
was his readiness to lay down his life in battle, to lead his men himself on the battlefield, and his ferocious tenacity in the face of overwhelming odds. His utter disregard for his own life in pursuit of military honor could require him to be indifferent even to the welfare of
his own children. A well-known Konjaku tale tells the story of an
eleventh-century hero named Fujiwara no Chikataka, who was reprimanded for his weakness when he reported to the renowned warrior and governor of Chikataka's province, Minamoto no Yorinobu,
that his son, a child of only five or six, had been taken hostage and
was being held under threat of death by a desperate gang of bandits
trapped at his house. Yorinobu laughed and said:
I understand your tears, but they do no good. A warrior must be resolved
to grapple with anyone, be he devil or god. Blubbering away like that is stupid. Let them kill him - he's just one child. A tsuwamono has to be prepared
for that. He's not worth much if he lets concern for wife and children give
him pause. Fearlessness
depends on thinking neither of oneself nor of one's
wife or children.16
On the other hand, other Konjaku and Skomonki stories exemplify
that the true tsuwamono spared and protected women.17
REVOLT OF TADATSUNE
Following the defeat and death of Masakado in 940, the Taira line
descending from Kammu continued to exercise a wide dominion
over the eastern provinces. Sadamori, who had overthrown his
cousin Masakado, was appointed governor of Mutsu and chinjufu
General in 974, signifying his leadership of eastern warriors. Four of
his sons received appointments as governor or vice-governor of an
eastern province. His influence was further expanded by the adoption of numerous grandsons and nephews. He and his clansmen are
said to have treated the eastern provinces as their private domain.
The death of Masakado in 940 failed to bring peace in the east.
Governmental authorities there came under repeated attack, the
sorely tried governor of Shimosa in 950 complaining, "Lawless
bands rage everywhere in the eastern provinces, looting and inflicting injury, so that there is no peace day or night."18 Accompanied to
16 Konjaku monogatari-shu, vol. 4, kan 25, story 11, p. 391.
17 Konjaku monogatari-shu, vol. 4, kan 25, story 5, pp. 374-81; Shdmonki (text), pp. 216-17;
Rabinovitch, trans., Shomonki, pp. 124-26.
18 Choya gunsai (Shintei zoho) Kokushi taikei, vol. 29, 1 (1938), kan 22, p. 511, 20/2/Tenryaku
4 (95°)Cambridge Histories Online © Cambridge University Press, 2008
REVOLT OF TADATSUNE
665
their provinces by large retinues of grasping relatives, friends, and
officials, the governors themselves sometimes contributed to the
general turmoil of the countryside by their own rapacious behavior.
One famous late-tenth-century case was so extreme that the district
officials and cultivators of a province were forced to appeal for relief
directly to the central government, citing some thirty illegal acts of
the governor on which they sought justice. But violence and fighting
remained small-scale, and no major armed defiance of governmental authority occurred anywhere in Japan throughout the heyday of
the Fujiwara regency. Then in 1028, just months after Michinaga's
death, revolt broke out again in the east. This revolt, too, was led by
a Taira, a son of a cousin of Masakado and Sadamori named Tadatsune (967-1031).19 Vice-governor at one time of both Kazusa and
Shimosa, Tadatsune seems to have become involved in the plunder
of government tax receipts in those two provinces. The same area
had been the scene of another attack on provincial authority a little
more than two decades earlier. That attack on the Shimosa provincial headquarters in 1003 by Taira no Koreyoshi (d. 1022) had
ended in Koreyoshi's flight before a court-dispatched Suppression
and Control Agent (oryoshi), but Koreyoshi was subsequently
granted court rank and given appointment as chinjufu General, apparently in response to generous gifts he made to the great Fujiwara
leader Michinaga.20
Tadatsune's plundering was not so easily managed. He had inherited a vast domain in the east, the official levies on which he had refused to pay to the provincial authorities, and his large force of warriors quickly turned back the meager troops the provincial governor
could call upon to demand payment.
In 1028 Tadatsune's rebellion spread rapidly through Shimosa,
Kazusa, and Awa (roughly Chiba Prefecture). Tadatsune occupied
the government headquarters in Kazusa, then burned the Awa headquarters and killed the governor there. In 1030 he put such fear in
the new governor of Awa that that unfortunate official fled the
province, abandoning the provincial seal, enabling Tadatsune to requisition supplies from the peasantry and obtain the tax-storehouse
keys. His rebellion became as formidable as that of Masakado.
19 Sources include Minamoto no Tsuneyori's diary, Sakeiki; Fujiwara no Sanesuke's diary,
Shoyuki (Oukt); and Konjaku monogatari-shu. For further information, see Omori Kingoro,
Buke jidai no kenkyii (Tokyo: Fuzambo, 1923), vol. 1, pp. 196-240; and Ishimoda Sho, Kodai
makki seijishi josetsu (Tokyo: Miraisha, 1964), vol. 1, pp. 182-96.
20 Takeuchi Rizo, Bushi no top, vol. 6 of Nihon no rekishi, ed. Inoue Mitsusada, 31 vols. (Tokyo:
Chuo koronsha, 1973), p. 20.
Cambridge Histories Online © Cambridge University Press, 2008
666
THE RISE OF THE WARRIORS
As usual, it was the stoppage of tax payments that moved the
court to action. Dependent for military force on tsuwamono, the
court dispatched against Tadatsune two warriors of the kebiishi office, the senior of whom was Taira no Naokata, a great-grandson of
Sadamori. (See the Taira genealogy, Figure 9.7.) Naokata had actively sought the commission, as he hoped to eliminate Tadatsune as
a rival for power among the Taira in the east, and to gain a conclusive victory in the long-standing feud between their two families. The
court, on its part, adopted the same tactic it had used in the
Masakado and Koreyoshi uprisings, sending a Taira warrior to suppress an insurgent from a rival Taira family. Notwithstanding the urgency of the situation, it was nearly two months after Naokata was
ordered to the attack before he sallied forth from the capital with a
core force of some two hundred men under his command. His departure had been delayed by the Fujiwara Minister of the Right
Sanesuke, who insisted that in a matter of such importance it was essential that the expedition wait for a calendrically auspicious day on
which to set out. The greater part of his force was recruited in eastern provinces. After two years of inconclusive fighting, the government force had still not succeeded in suppressing Tadatsune. The
two principal commanders were cashiered and Minamoto noYorinobu (968-1048), then governor of Kai, was appointed in their place.
Although Minamoto penetration of the east had begun well before
Yorinobu's day, it was from his time that the clan began to establish
its dominance in the region. Yorinobu was a "warrior of the capital,"
a son of the Mitsunaka who had earlier served Michinaga.21 He and
his father were ideally situated to initiate the creation of the powerful, enduring retainership system of landholding tsuwamono that
eventually contributed to the establishment of the Kamakura bakufu
in the time of Yorinobu's sixth-generation descendant Yoritomo (see
Figure 10.1). A chief goal of the tsuwamono was the protection of
their land from taxation and governmental interference, protection
that Mitsunaka and Yorinobu could help obtain because of their official positions and high connections in the central government.
Both held military offices, and both served in the household of the
regent. Those positions, their high birth, their appointment to lucrative gubernatorial posts in the provinces with the attendant opportunities that such appointments provided for developing a retinue of
followers, and the broad, court-conferred military authority they
21 Takeuchi, Bushi no tojo, p. 35.
Cambridge Histories Online © Cambridge University Press, 2008
REVOLT OF TADATSUNE
667
were given, all encouraged the establishment of a vassalage system
under them. The system was the ikusa, or, as it is now more frequently called, the bushidan, the "warrior band."
Like Naokata before him,Yorinobu delayed his departure for the
east several months after his appointment to command the court
forces fighting Tadatsune, perhaps in response to messengers Tadatsune sent to the capital to plead the case that he was not in revolt
against the government. The court's decision went against him, and
his messengers were arrested, putting him on notice that he could
soon expect another major assault by court forces. By the late spring
of 1031, Yorinobu had arrived in Kai with a son of Tadatsune, a
monk, in his train. He was about to launch a campaign against
Tadatsune when the rebel suddenly appeared before him. Tadatsune
had taken Buddhist vows, changed his name to Joan, and with two
sons and three roto, surrendered without a fight.
Yorinobu, in his report on the campaign addressed to the guardian
deity of his clan, boasted that he had achieved victory "without causing the people to flee or laying waste the land, without beating
drums or unfurling banners, without drawing bows or loosing arrows, neither hiding nor attacking." "Just by being there," he crowed,
"I captured the enemy."22
Although Yoshinobu claimed that his reputation as a warrior was
reason enough for Tadatsune to surrender, there seem to have been
other considerations as well. In submitting, Tadatsune may have
been honoring the obligation of a personal-service bond made with
Yoshinobu some two decades earlier (to be discussed later). However, exhaustion may have been a consideration in the surrender: the
provinces of Awa, Kazusa, and Shimosa were by that time devastated
from the protracted fighting: in Kazusa, it was reported, there was
almost no paddy left in cultivation.23 This may be why Tadatsune,
while holding off the attacks of Naokata, had tried to negotiate with
the court and then had taken religious vows as a signal that he did
not wish to continue his resistance.
En route to the capital in the custody of Yorinobu, Tadatsune became ill and died. However, his surrender to Yorinobu secured the
survival of his family: his sons became retainers of Yorinobu, and in
the twelfth century, his descendants, by then known as the Chiba,
played a major role in the rise of Yorinobu's descendant Yoritomo.
22 "Minamoto no Yorinobu komon," Heian ibun, vol. 3, p. 777.
23 Takeuchi, Bushi no tojo, pp. 32-33.
Cambridge Histories Online © Cambridge University Press, 2008
668
THE RISE OF THE WARRIORS
The explanation thatTadatsune submitted toYorinobu in 1031 because he was under a personal obligation to him derives from an incident described in the Konjaku monogatari-shu. About twenty years
earlier, whenYorinobu was vice-governor of Hitachi and attempting
to exercise his influence in the neighboring provinces of Kazusa and
Shim5sa, Tadatsune, who held lands in Shimosa, was flouting his
obligation to forward taxes to the provincial headquarters. In the
Konjaku story, he boasted: "My influence is extremely far-reaching.
In Kazusa and Shimosa, I do as I please. I do not pay public duties,
nor do I give heed to the governor of Hitachi." Yorinobu found this
behavior intolerable and resolved to chastise Tadatsune. He made
common cause with a local strongman, Taira no Koremoto, who had
old scores to settle with his rival clansman. Yorinobu with two thousand horsemen and Koremoto with three thousand (certainly inflated numbers) advanced quickly on Tadatsune's base. In the face
of this overwhelming force, Tadatsune surrendered without a fight,
proffering his name tag (myobu) and a written message of apology
for his deeds. Name tags, inscribed with a man's name, official title,
and rank, were conventionally presented by a subordinate taking service with a superior, as when someone became a functionary in a
noble household, or a disciple pledged himself to a master, or a warrior became a follower of a military lord. Men who presented name
tags were known as "house men" (keniri) of the recipients of the tags.
In the Konjaku story, the tag was presumably offered in token of
Tadatsune's submission toYorinobu.24
Although the Konjaku story may contain fictional elements, the
resolution of the conflict by the presentation of a name tag points to
a conspicuous feature of the emerging warrior society. In the Heian
period, house men were initially of diverse types. They might themselves be nobles or warriors, and they could also include kin of the
house. They were like the lads (rdto) in being personal retainers of
the lord, but superior in status because of their greater autonomy:
they did not usually live in the lord's house or vicinity, but maintained a separate existence. The scope of the term's usage narrowed
in the late Heian period, coming then to refer primarily to warrior
retainers, a nomenclatural development that was eventually incorporated in the name used of the shogunal retainers of the Kamakura
bakufu, the gokenin.
Yorinobu may have delayed his departure from the capital in 1031
24 Konjaku monogatari-shu, vol. 4, kan 25, story 9, p. 387.
Cambridge Histories Online © Cambridge University Press, 2008
REVOLT OF TADATSUNE
669
because Tadatsune was his kenin and he wished to afford him additional time to negotiate with the court. Yorinobu's personal loyalties
may have taken precedence over service to the state, an indication of
the increasing strength of the bushidan relationships, according to
some scholars. However, Yorinobu's appointment to replace Naokata as commander of the expedition against Tadatsune probably
came after it became known in Kyoto that Tadatsune was seeking a
way to surrender, and that this outcome was being resisted by
Naokata, who wanted his destruction. As the court hoped, Yorinobu's old relationship with Tadatsune enabled him to bring about
Tadatsune's surrender and an end to the destructive struggle that
had continued for almost four years.
Tadatsune's submission toYorinobu was a shock to theTaira, who
had long acted as if they owned the Kanto. From this time on they
fomented no further rebellions in the east. Instead, many became
followers of the Minamoto, with their own groups allying with other
military groups under the head of the Minamoto family. The prestige that came toYorinobu for his "victory" over Tadatsune, together
with his noble ancestry, attracted other warrior groups in the east to
ally themselves with his bushidan organization. The Minamoto became increasingly influential in the remainder of the eleventh century, particularly in the time of Yorinobu's son,Yoriyoshi (988-1075).
Although Yorinobu had effectively carried out his mission, the
court, which had little regard for warriors, useful though they were,
was slow to reward him. A year later he was appointed governor of
Mino, his second-choice position.
In 1046, as governor of Kawachi, the elderly Yorinobu began the
worship at Iwashimizu Hachiman Shrine of the martial Hachiman
as tutelary god of the Genji. Since 859, when religious practices
were imported from the Hachiman Shrine at Usa in Kyushu, there
had been some aristocratic support of the shrine. However, Yorinobu was the first to claim the deity as a patron. He explained his
decision as an act of reverence toward his ancestors and their marvelous history of defending the country, tracing his lineage back
twenty-two generations to the legendary emperor Ojin, said to have
participated in the conquest of Silla while still in the womb of Empress Jingu.Ojin was considered an avatar of the bodhisattva Hachiman, god of archery. Shrines to Hachiman soon sprang up throughout the country wherever there were Minamoto to invoke his name.
Yorinobu's son Yoriyoshi attributed his success in the Earlier Nine
Years' War to Hachiman and in 1063 secretly erected a shrine to
Cambridge Histories Online © Cambridge University Press, 2008
670
THE RISE OF THE WARRIORS
Hachiman at Yui-no-go in Sagami, moved to Kamakura by Yoritomo in 1191.
EARLIER NINE YEARS WAR
Two lengthy wars followed the Tadatsune revolt at twenty-year intervals, keeping the east - in this case the far northeast end of Honshu beyond the Kanto provinces - in a state of unrest for nearly sixty
years. The two later eleventh-century wars are called the "Earlier
Nine Years'War" (zen kunen no eki: 1051-62, or 1056-62 according to
some) and the "Later Three Years'War" {go sannen no eki: 1083-87).
There is disagreement between these traditional designations and
the dates given, even when one allows for reckoning based on a
count of calendar years.25
Since early times a major concern of the court had been the subjugation of the tribal people who occupied the far northeast. They
were called Ezo (or in earlier times usually referred to as Emishi or
Ebisu), names indicating that they were "eastern barbarians." (On
their names and characteristics, see Chapter 1.) The vast, untamed
territory that they inhabited was divided into two very large provinces: Mutsu (present Fukushima, Miyagi, Iwate, and Aomori Prefectures) on the east, and Dewa (present Yamagata and part of Akita
Prefectures) on the west. The Ezo continued a dogged resistance for
centuries, harassing the encroaching settlers and launching raids by
mounted archers on the government's palisades and administrative
bases. The innumerable expeditions of conscript troops sent against
the Ezo in the eighth century had few successes until the government recruited skilled horsemen from the Kanto area and adopted
some of the combat tactics of the enemy. In 801, with the large expeditionary force led by Sakanoue no Tamuramaro (758-811), the
Barbarian-subduing Generalissimo (seii taishoguri), conquest was almost complete. The Mutsu Ezo, against whom the court had concentrated its efforts, soon ceased resisting.
The way was then open for officials appointed by the court to exploit the territory and its people. The biographer of Fujiwara noYasunori (825-95), governor of Dewa, writing in 907, said: "The fields
(of Dewa) are fertile, and there are many rare products. There is no
limit to the land a powerful official can acquire, on which he se25 On the two wars, see Toyoda Takeshi, ed., Tdhoku no rekishi, vol. 1 (Tokyo: Yoshikawa
kobunkan, 1967), pp. 173-84; and Takahashi Tomio, Kodai Ezo: sono shakai kozo (Tokyo:
Gakuseisha, 1974).
Cambridge Histories Online © Cambridge University Press, 2008
EARLIER NINE YEARS' WAR
671
cretly raises taxes and conscripts as much labor as he desires."26The
administration of Yoshimine no Chikanari, vice-governor of the
Akita Fortress in Dewa was so oppressive that there was a revolt in
878 in a twelve-village area along the Noshiro River. The rebels were
unable to coordinate their efforts, however, and the uprising was
soon suppressed.
The cost of dispatching expeditionary forces to put down rebellions of this kind and to maintain armies stationed in the area was
prohibitive.27 A different method of maintaining order in the conquered area was essential. To this endYasunori, the Dewa governor,
introduced a compromise policy that proved effective in the long
term.28 He disbanded the army that had been sent to control the
area, keeping only a small force of about a thousand tsuwamono divided into three units strategically placed. The Ezo chief who had
surrendered was charged with the responsibility of administering his
people. "Using the barbarian to control barbarians," the land was divided into villages and districts, all with Ezo heads. The resultant
landscape of numerous independent units was similar to the depiction of Japan in the first century as described in Han shu, the official
Chinese history. Thanks to this policy, rebellions in the northeast
gradually decreased.
A similar policy was in effect in Mutsu from the time of the Dewa
uprising. The chieftain of the local Abe clan was appointed district
magistrate in Mutsu in 878, and the Abe held this position for generations. During this time the Abe built at least twelve palisades
(saku) at strategic points, and the area came to resemble a semiindependent region. Even in 939, when the Dewa Ezo revolted
again, peace was maintained in Mutsu.
The Abe lineage is unclear. According to the Mutsu waki, a chronicle of the Earlier Nine Years' War compiled shortly after the event,
the Abe were Ezo who had submitted to the court.29 It was a rare exception to the court's policy in Heian times to entrust the administration of a region to a powerful local family. Once it was entrenched
in an official position, it was difficult to check the family's appetite to
increase its influence and wealth. Before the middle of the eleventh
century, the Abe chieftain, who had been entrusted with the admin26 Fujiwara no Yasunori den, in Kodai seiji shakai shiso, p. 68.
27 Takeuchi, Bushi notojo,pp. 47-48. 28 Fujiwara noYasunori den, pp. 64-69.
2g Mutsu waki, in Kodai seiji shakai shiso, see p. 230. For an annotated translation, see Helen
Craig McCullough, "A Tale of Mutsu," Harvard Journal of Asiatic Studies 25 (1964-65):
178211.
Cambridge Histories Online © Cambridge University Press, 2008
672
THE RISE OF THE WARRIORS
cV 1 0
Semboku^—_
/ui'
5m *)
4^s
/
/
\
/ - , Dewal
J>s.
J
,_r Districts
/
)
/ i
y ^ /
*~\
"';
'-
i"~*-.s'
r -''
Echigo
\ ,-K } '
Shinano
/
Mutsu
( Kozuke ,
/
Shimotsuke
)
,'''
/
_X
J-Jr r
y" \
1
K_J—I '
Vl-
*"\ Musashi , i Hitachi /
1
,'}
\
'"~y''"\ •> " ''' ^
;Saga
! \__
C<1
--''' '''' Sii'ruea"V
X S l \ ) r Kazusa
CrLLTotomi; ^
~~-^ J
Awa
d
1
1
!Shim6sir"\\
\
y
1
2
3
4
5
6
7
8
9
10
11
12
Kamakura
Taga Fortress
Koreharu Fortress
Hiraizumi
Koromogawa Stockade
Isawa Fortress
Torinomi Stockade
Kuriyagawa Stockade
Noshiro River
Akita Fortress
Kanezawa Stockade
Numa Stockade
Map 10.1. Battle sites in the northeast.
istration of the "Back Six Districts" (Okurokuguri) north of Koromogawa Fort (near the site of the modern city of Hiraizumi in southern
Iwate Prefecture), further extended the area under his control toward the south. Provoked by the chieftain's disobedience and his refusal to pay taxes and corvee dues, the governor of Mutsu and the
vice-governor of Akita Fortress attacked him with a force of "several
thousand" men, but suffered a crushing defeat. The court then
turned to a famed and wealthy warrior, Minamoto noYorinobu's eldest son, Yoriyoshi, who was dispatched to Mutsu in 1051 as goverCambridge Histories Online © Cambridge University Press, 2008
EARLIER NINE YEARS' WAR
673
nor of that province and General of the chinjufu. He led a force made
up of eastern warriors recruited by court order, as well as warriors,
also from the eastern provinces, who were his retainers.
Twenty years earlier, Yoriyoshi had aided his father in suppressing
the rebellion of Tadatsune. He then served at court, where his exceptional skill in archery made him a favorite hunting companion.
His talents apparently caught the eye of Taira no Naokata (leader of
the first force sent against Tadatsune), who offered Yoriyoshi his
daughter's hand in marriage. It was through Naokata's gift of land
around Kamakura in Sagami - land Naokata had acquired while
leading the expeditionary force - that the Minamoto first established
themselves in Sagami, which was to become their stronghold.
When Yoriyoshi and his force reached Mutsu, the Abe chief, Yoritoki, took advantage of a general amnesty that was proclaimed to
submit to the powerful new governor without a fight.30 Peace continued until the last year of Yoriyoshi's tenure in 1056, when fighting
erupted, supposedly initiated by Yoritoki's son Sadato (1019-62) because of a personal affront, but perhaps actually provoked by
Yoriyoshi himself. The court responded by again issuing an imperial
order calling for the suppression of the Abe, raising troops for that
purpose, and reappointing Yoriyoshi governor of Mutsu.
In the course of Yoriyoshi's offensive, Yoritoki was killed in 1057 at
Torinomi Stockade by assimilated Ezo who had been recruited by
Yoriyoshi. Sadato tenaciously continued his resistance and, two
months later, defeated Yoriyoshi's undermanned force. It was not
until 1062, when Yoriyoshi succeeded in obtaining the help of the
Kiyohara, a powerful, partially assimilated Ezo family that controlled
Dewa, that he was able at last to prevail against the Abe. In the summer of that year, Kiyohara no Takenori arrived in Mutsu with an
army of "more than ten thousand men," divided into seven units,
each commanded by a family member. By fall this force and Yoriyoshi's "three thousand"-man army had taken the last Abe stronghold at Kuriyagawa (on the site of the modern city of Morioka in
Iwate Prefecture) and killed Sadato, bringing an end to the Earlier
Nine Years'War.
Yoriyoshi's term as governor of Mutsu had expired before the end
of the war and, in the view of the court, he fought the final battles as
30 An amnesty was declared because of the illness of Michinaga's daughter, J6tomon-in Fujiwara no Shoshi, 988-1074. When he submitted, the Abe chieftain, out of respect, took
the name Yoritoki in place of his former name, Yoriyoshi, pronounced the same as the governor's name.
Cambridge Histories Online © Cambridge University Press, 2008
674
T H E
RISE
O F
T H E
WARRIORS
a private warrior. Nonetheless he was promoted to the Fourth Rank
and appointed governor of Iyo. His sonYoshiie (1039-1106) became
Fifth Rank and governor of Dewa. Even when rank meant little
more than title, it served to impress local tsuwamono. The Fifth Rank
was sought after by warriors; it brought eligibility for appointment
as governor or chinjufu shogun and could, on occasion, be purchased.
The real victor in the war was Kiyohara no Takenori, whose large
force, and also whose knowledge of conditions in the northeast, were
probably crucial to Yoriyoshi's success. The court placed the Back
Six Districts of Mutsu under Takenori's control, gave him Fifth
Rank, and appointed him to the old and prestige-laden title General
of the chinjufu, the first man of local extraction to receive this distinction. Together with his original Semboku domain in Dewa (the
eastern part of Akita Prefecture) adjoining the west boundary of
Mutsu, that expansion of Kiyohara power gave the family control of
both Mutsu and Dewa.
It had been policy from Nara times to relocate Ezo who surrendered to Dazaifu. The court followed precedent, deciding not to execute the surviving Abe leaders but to remove them to Iyo on Shikoku and Dazaifu. As a result of this dispersion, Abe descendants
proliferated in other parts of the country and became powerholders
in the medieval period. Like Taira no Masakado, who came to be
worshipped as a deity after his failed insurrection, the Abe also
achieved a belated fame. Although disdained by the court, the Abe
left a legacy of two centuries of strong leadership and cultural attainment in the remote northeast. Several Abe had been "frontier
lecturers" (sakai no koshi), monks with religious duties in Ezo territory. Two fanciful anecdotes are sometimes cited as evidence of the
Abe's ability to compose poetry. At the Battle of Koromo River a poetic exchange with Sadato moved Yoshiie not to release his notched
arrow at the fleeing Sadatd, according to the Kokon chomonju (^254).
In the Heike monogatari, Sadato's brother, the captive Muneto, composed a poetic response to jeering crowds in the capital.31 The culture that took root during the long peace before the Earlier Nine
Years' War was a forerunner of the twelfth-century efflorescence of
Hiraizumi, seat of the so-called Oshu Fujiwara.
The Kiyohara are thought to have been descendants of Kiyohara
noYoshimochi, who accompanied Fujiwara noYasunori to Dewa in
879 at the time of the Emishi insurrection. That line of the Kiyohara
31 Takeuchi, Bushi no tdjo, pp. 62-64.
Cambridge Histories Online © Cambridge University Press, 2008
LATER THREE YEARS' WAR
675
traced its ancestry to Emperor Temmu's son, Prince Toneri. Although there is no reliable evidence on the point, it is unlikely that
the court would have given the title of General of the chinjufu to the
head of the Dewa Ezo in 1062 if he had not been able to trace his
lineage to the Kyoto nobility.3*
LATER THREE YEARS WAR
The peace achieved in Mutsu and Dewa in 1062 was upset two
decades later in the time of Takenori's grandson Sanehira, when his
attempt to strengthen his authority as head of die Kiyohara clan's
main line caused dissatisfaction among clan members. A doyen of
the clan, Kimiko no Hidetake, refused to accept the arrogant behavior of his junior, Sanehira, and moved into open opposition to him.
Hidetake acted not only out of an old man's sense of propriety and
amour propre, but also in resistance to a restructuring of clan power
that was replacing the more egalitarian practices of the kin alliance
or confederation of Takenori's day with a hierarchy culminating in
clan lordship by the head of the main line. Sanehira launched a
punitive attack in 1083 on his refractory senior relative in Dewa, only
to find himself assaulted in Mutsu by his half-brother Iehira and by
his stepbrother Kiyohira (1056-1128), to both of whom Hidetake
had appealed for support. Kiyohira was a son of Fujiwara no
Tsunekiyo, who had fought for Abe no Yoritoki and had been executed at die end of the earlier war. His young son Kiyohira and his
widow, a daughter of the Abe chief, Yoritoki, went as spoils of war to
Takenori's son,Takesada. He adopted Kiyohira and took the widow
as wife; she subsequently bore him Iehira. Hence, Sanehira, Kiyohira, and Iehira were all half- or stepbrothers, but none of them a full
brother to either of the others. (See Figure 10.2, the genealogies of
the Kiyohara and Oshu Fujiwara.)
Confronted by hostile forces in Dewa and on his rear in Mutsu,
Sanehira remained at his headquarters in the northern Mutsu districts until the new governor of the province, Minamoto noYoshiie,
now head of the Seiwa Genji, arrived at his post in die autumn of
io83.Yoshiie had a towering reputation as a skilled and courageous
warrior, earned in the service of his father Yoriyoshi in die earlier
war. Yoshiie was quick to intervene in the Kiyohara family feud, ostensibly as a mediator on Sanehira's behalf, but in fact he intended
32 Takeuchi, Bushi no top, pp. 65-66.
Cambridge Histories Online © Cambridge University Press, 2008
676
THE RISE OF THE WARRIORS
Abe noYoritoki
d. 1057
I
Sadato
d 1062
Kiyohara no Mitsukata
I
Takenori
I
Muneto
F. noTsunekiyo == female ==Takesada
d 1062
Kiyohira
1056-1128
Oshu Fujiwara
Iehira
d1087
Takehira
d 1087
Sanehira
d 1083
Motohira
d 1157?
Hidehira
d 1187
I
Yasuhira
dll89
I
I
Kunihira
dll89
Figure 10.2. The Kiyohara and Oshu Fujiwara.
to supplant Kiyohara power in Dewa and Mutsu.Yoshiie petitioned
the court for a commission to subdue Iehira, which would authorize
him to mobilize troops and requisition supplies in neighboring
Kanto provinces. The court refused to intervene, declaring the conflict a private Kiyohara affair. Unlike rebels such as Masakado,Tadatsune, and the Abe, the Kiyohara had obeyed the orders of the
provincial governors, paid their taxes, and performed their official
duties. The court saw no reason for punitive action. Yoshiie was not
dissuaded from intervening.
Assured of Yoshiie's backing, Sanehira at last set off to attack
Hidetake in Dewa. Taking advantage of Sanehira's absence, Kiyohira
and Iehira attacked his headquarters but were in turn assaulted and
routed by Yoshiie. Sanehira fell ill and died en route to Dewa, however, obviating further hostilities, and the two half brothers thereupon submitted to the authority of Yoshiie. He mediated the issue of
the governance of the Back Six Districts of Mutsu by assigning three
districts to each brother. This decision irked Iehira, who now considered himself head of the Kiyohara, while Kiyohira, as the only
survivor of the Abe clan remaining in the northeast, may have considered himself heir to the Six Districts. Friction between the brothCambridge Histories Online © Cambridge University Press, 2008
LATER THREE YEARS' WAR
677
ers mounted, and when Ieharu's forces raided Kiyohira's base and
slaughtered his wife and children, Kiyohira appealed toYoshiie.
In the fall of io86,Yoshiie besieged Iehira in Numa Stockade with
a force of "three thousand mounted warriors." Heavy snow, cold,
and hunger forced Yoshiie to lift the siege. His younger brother,
Yoshimitsu, left his post at the capital to come to his aid, and after a
larger force was collected, they laid siege to Iehira, now in Kanezawa
Stockade, a short distance to the north (at what is now the city of
Yokote in southeastern Akita Prefecture). After another difficult
siege the brothers prevailed at the end of 1087. Iehira was slain, and
the main Kiyohara line was at an end33 (see Map 10.1).
The court again denied Yoshiie's request for rewards for himself
and his men and official recognition of their hard-earned victory,
maintaining with much justice that the fighting was a private quarrel. Yoshiie had used his considerable resources as governor of
Mutsu, which was the principal center of gold mining, to conduct
the war and provision his troops. He was now obliged to compensate
his retainers and recruits from his own resources. Furthermore, the
court disciplined Yoshiie, ending his appointment as governor of
Mutsu the next year and requiring him to repay tax revenues from
the province that he had diverted to support the war.
Fighting had its own, more important rewards for Yoshiie and his
descendants. His victory over Kiyohara no Iehira further enhanced
the prestige of his Minamoto line in the eyes of eastern warriors. In
the two wars in the northeast, first Yoriyoshi, and then Yoshiie recruited relatives, retainers, and allies in the east to mount campaigns
against tribal chieftains. In the first war Yoriyoshi, as recipient of a
court commission, had troops requisitioned on court orders, a force
he commanded as provincial governor, and garrison guards he led as
General of the chinjufu. In the second war Yoshiie was able to raise
and later compensate an army using mostly his own resources. The
wars were long and arduous, at least in part because the Minamoto's
forces did not have numerical superiority. In both wars victory came
in the end thanks to alliances made with contending local tribal
forces. The Minamoto leaders, nonetheless, gained fame as unrivaled warriors.
The combined governorships of Mutsu and Dewa, which may
have been one of Yoshiie's goals, went instead to Kiyohira. He profited spectacularly from the war. As the sole survivor of the Abe and
33 For a fuller discussion of the wars, seeToyoda, Tohoku no rekishi, vol. 1, pp. 185-205.
Cambridge Histories Online © Cambridge University Press, 2008
678
THE RISE OF THE WARRIORS
Kiyohara families, he gained the lands once held by both families in
Mutsu and Dewa, and was appointed General of the chinjufu. He resumed the Fujiwara clan name of his father (who claimed descent
from the Hidesato who defeated Masakado). Kiyohira moved his
residence to the former base of his erstwhile Abe relatives south of
the confluence of the Kitakami and Koromo rivers, a militarily
strategic point in the southern part of modern Iwate Prefecture that
had figured in the early Heian campaigns against the Emishi and
had also more recently been the scene of fighting in which Kiyohira's
father, Tsunekiyo, had participated. Known, probably from Kiyohira's time, as Hiraizumi, the site grew into a vigorous urban cultural center distinguished by its prolific, luxurious religious architecture, its Buddhist art, and its landscape gardens, modeled in good
part on the style and standards of the imperial capital at Heian.
The remarkable flowering of Hiraizumi culture under the lordship of
the Oshu Fujiwara (the "Fujiwara of Mutsu"), as Kiyohira and his
three direct lineal descendents are called, depended on the political
and military supremacy of those lords in Dewa and Mutsu, and on
the great wealth they were able to extract from that large, rich territory, especially gold but also horses and agricultural produce. (The
warmer climate prevailing at the time probably made the region better farming land than it was subsequently.)
Kiyohira himself established or, according to temple legend,
reestablished, the Chusonji Temple at Hiraizumi, perhaps for the salvation of the war dead in the recent battles fought in the area. Under
the Oshu Fujiwara, the temple's mountain precincts came eventually
to encompass scores of buildings and hundreds of monastic cells, the
buildings including most notably the Konjikido ("Golden Hall"),
one of the best-known surviving historical structures in Japan. Constructed in 1124, the chapel-like hall provides beneath the altars of
its Amidist statuary a final resting place for the mummified remains
of the first three Oshu Fujiwara lords (Kiyohira, Motohira, and
Hidehira) and for a severed head that is tentatively identified as that
of the fourth and last lord,Yasuhira (1155-89). The hall and a scripture storehouse were the only Chusonji structures to survive a wildfire that swept the mountain in 1337, but even by itself, the hall's
artistry is more than sufficient to impress upon a modern visitor a
keen sense of the puissance, wealth, and cultural ambition of its Fujiwara patrons.
Some understanding of the organization of the private warriors,
mounted archers, recruited by the Minamoto leaders during the two
Cambridge Histories Online © Cambridge University Press, 2008
CONDITIONS IN THE CAPITAL
679
wars can be gained from two sources: the Mutsu waki on the earlier
war and the commentary on the twelfth-century hand-scroll painting concerning the Later Three Years'War.34 These works show that
the roto were organized more systematically than in Masakado's
time, and that they formed the core of the bushidan (warrior band).
Many of the roto recruited as retainers by Yoriyoshi and his son
Yoshiie were from the provinces of Sagami and Mikawa. The Hogen
monogatari (ca. 1219-22) quotes the warrior Kamata no Masakiyo as
stating that he began his service to the Minamoto when Yoriyoshi
was governor of Sagami; at this time he would have been a kenin.
Thereafter he came to address Yoshiie and then Yoshiie's son and
grandson as "master," which indicates that he had placed himself
under Yoshiie's protection and was considered a roto.
The primary change in the bushidan involved the strengthening of
the lord-vassal relationship. There are numerous tales of vassals
who, having accepted service with a lord, willingly sacrificed their
lives against hopeless odds defending him.35 Similar tales do not appear in the earlier Shomonki or Konjaku monogatari-shu. They are evidence that loyalty to one's lord, the keystone of the later warrior
ethic {bushido, was being articulated as a guiding principle, at least
in literary accounts. The commitment of vassals grew stronger when
they looked to lords to guarantee their rights to land in return for
military service. Bonds between lord and retainer were also forged
in the hardships and sacrifices they suffered together in the long and
arduous campaigns of the two wars. As the lord came to have more
control over his vassal's life, he assumed greater responsibilities toward him. In 1091 Yoshiie and his brother Yoshitsuna, who had once
been companion in arms in campaigns of the Earlier Nine Years'
War, were on the verge of combat over a land dispute between their
respective roto. This incident is often cited as evidence of the strong
commitment of mutual support between lord and vassal in the bushidan of this time.
CONDITIONS IN THE CAPITAL
Outlawry had been a daily occurrence in the life of Heian-kyo since
the ninth century, requiring an extensive system of military guards
34 Go sannen kassen ekotoba, ed. Komatsu Shigemi and Miya Tsugio, vol. 15 of Nihon emaki
taisei, 27 vols. (Tokyo: Chu5 koronsha, 1977).
35 See, for example, stories in Mutsu waki, pp. 236-37, 250; "A Tale of Mutsu," pp. 192, 203.
Cambridge Histories Online © Cambridge University Press, 2008
680
THE RISE OF THE WARRIORS
and attendants to provide security for the court and its officers.36The
statutory code provided for several bodies of guards to man the many
gates of the imperial palace and to serve as ceremonial escorts for the
emperor. Their effectiveness, however, largely evaporated in the
Heian period as their manpower dwindled, leaving even the palace
itself vulnerable to arson, robbery, and murder. The government had
sought since the Nara period to assure the personal safety and dignity of higher officials by assigning attendants {toneri) to them for
guard and escort duty in numbers varying according to their ranks,
ranging from twenty at the lower end of the scale up to four hundred
at the highest.37 Although the toneri attendants were merely provincial conscripts and had minimal military skills, their use by nobles for
private purposes came to be disruptive of public order,38 and they
were replaced in the case of the highest officials with somewhat better-born guard-officials called zuijin ("escorts"), who were organizationally attached to the Imperial Bodyguards. Fewer in number than
the toneri, the handsomely uniformed zuijin escorts were prized more
for their appearance than for their military accomplishments, and
great nobles were typically forced to augment their security with privately recruited forces of tsuwamono warriors.
Police officers known as kebiishi (Offenses Investigation Agents)
are first mentioned in 816. Their mission was the maintenance of
public order in the capital and its environs and the investigation of
crimes. By 834 a kebiishi office was established within the palace
grounds as Imperial Police. The remarkable expansion of the office's
functions and its involvement in activities in other provinces is discussed in Chapter 1. And yet it seems to have fallen short of meeting all of the needs for security in Kyoto. Various guard units continued to proliferate within the Great Imperial Palace and elsewhere
in the capital. Not even the emperor felt safe. In 890, to protect himself from the deposed and dangerously hostile emperor Yozei, Emperor Uda found it expedient to establish a special unit of Imperial
Bodyguards called the takiguchi ("water spout"), supposedly named
after the location of the headquarters for the group near the entrance of a stream or irrigation canal into the grounds of the Seiryoden, the chief ceremonial palace of the Emperor's Residential
Compound. Numbering only ten men at its founding, the takiguchi
36 Tanimori Tomoo, Kebiishi no chiishin to shitaru Heian jidai no keisatsu jotai (Tokyo: Kashiwa
shobo, 1921), pp. 56-121.
37 Ritsu. Ryd no gige, pp. 196-97.
38 Ruiju sandai kyaku, kan 20, pp. 636-37, 11/4/Engi 2 (902).
Cambridge Histories Online © Cambridge University Press, 2008
CONDITIONS IN THE CAPITAL
68l
group expanded to thirty in the reign of Go-Shirakawa. Its members
were non-nobles of the Sixth Rank attached to the Chamberlains'
Office (kurddodokoro), in the late Heian period the appointments becoming hereditary in warrior lines of theTaira and Minamoto clans.
Two similar guard groups were instituted for the palaces of the retired emperors, the first either in the time of Uda (ex. emp. 897-931)
or later by En'yu (ex. emp. 984-91), and the second under Shirakawa
(ex. emp. 1087-1129). The earlier group, called the mushadokoro
("Martialists' Office") was initially composed of ten warriors recruited mainly from the ranks of the takiguchi, its complement gradually increasing until by the time of ex-emperor Shirakawa it numbered thirty. The second, later, group was called the in-no-hokumen,
or simply hokumen ("north face")- Organized from low-ranking officers without imperial audience privileges, the hokumen became the
core of the military strength of the ex-emperors, including in its ranks
most notably Taira no Masamori, the grandfather of the first military
dictator, Kiyomori. As the Taira moved into high offices at court in
the twelfth century, acquiring audience privileges and establishing
themselves as the paramount warriors of the capital, the importance
of the guards declined, and hokumen appointments instead came to
be used as a means of bringing otherwise socially debarred favorites
and useful supporters into the court of the ex-emperor: artists and
performers, kebiishi, and provincial officials. The poet Saigyo (111890), for instance, was a hokumen. After theTaira seized control of the
government, the hokumen under ex-emperor Go-Shirakawa served as
personal attendants of the ex-emperor, becoming an important focus
of anti-Taira sentiment; they were involved, for example, in the
Shishigatani plot of 1177.
The official kebiishi and guard groups were not always able to ensure the safety of even the most exalted court figures, as ex-emperor
Kazan discovered in 996 when he is reported to have come under
arrow attack at the direction of a disgruntled Fujiwara lover. Lesser
leaders of the society were even more vulnerable to the outlawry of
the times. Princes, regents, chief ministers, and other prominent nobles often found it necessary to establish private guard units within
their own palaces and mansions in quarters called the saburaidokoro
("Attendants' Office"). They recruited tsuwamono who had proved
their prowess in fighting. Tsuwamono who attended (saburau) court
nobles at their beck and call came to be called saburai, which later
became samurai. This name for the hired guards at first lacked the
honor associated with the word tsuwamono. But warriors who previCambridge Histories Online © Cambridge University Press, 2008
682
THE RISE OF THE WARRIORS
ously may have used and sharpened their military skills in banditry
or revolt could thus sometimes find themselves transformed into defenders of their former enemies and victims. Enrollment in the service of a great noble at the capital brought high political connections
and prestige that could be used in attracting retainers into a warrior's own service, and the chieftains (toryo) of the warrior clans
were thus willing to accept service with regents and ministers. But as
the might of the warriors came to be increasingly decisive in the affairs of the capital, and as the warriors themselves rose in government rank and office, they posed an ever growing threat to the political authority of their noble lords.
The court was frequently under threat from the armed monks
(sdhei) of the great temples near the capital. Although Buddhist
teachings forbade monks to carry arms, that proscription had early
been ignored in China, and clerical warriors were known also in
Japan by the eighth century. But it was not until the tenth century,
following a rapid growth in temple populations, that larger, wealthier temples organized armies to protect and assert their own interests and to control and manage a not always peaceable brotherhood.
The armies were recruited from the more unruly elements within
the temples themselves and also from the temples' shoen.
In 1006, Kofukuji in Nara fielded a fighting force of three thousand
men. By 1113, Kofukuji and Enryakuji were said to have mobilized
"tens of thousands" of armed monks in a dispute over the appointment of a priest trained at Enryakuji as chief administrator {betto) of
Kiyomizudera, a branch temple of Kofukuji. Armed monks of the
Enryakuji (overlooking the capital on the northeast) and Onjoji (in
present Otsu), being close to Kyoto, were especially prominent in the
history of the eleventh century, when their large bands marched into
the capital in order to press petitions and demands (goso), backed by
the threat of imminent violence, on the court or on individual
courtiers.39 The belligerent intrusions of the monks were usually supported and protected by displaying symbols of the temples' guardian
shrines, the Kofukuji monks carrying at the head of their procession
the "divine tree" (shimboku) of the Kasuga Shrine, and the Enryakuji
monks descending from Mount Hiei bearing the sacred palanquin
(mikoshi) of Hie Shrine. Armed monks descended on the capital sixty
times during the years of the retired emperors Shirakawa, Toba, and
Go-Shirakawa. To counter the threat of these religious institutions,
39 Tsuji Zennosuke, Nihon bukkyoshi (Tokyo: Iwanami shoten, 1944), vol. 1, pp. 765-920.
Cambridge Histories Online © Cambridge University Press, 2008
CONDITIONS IN THE CAPITAL
683
the court relied on the Imperial Police, the hokumen warriors, and, increasingly, on the toryo of the Minamoto andTaira clans. Heike monogatari quotes Shirakawa as lamenting: "Three things refuse to obey
my will: the waters of the Kamo River, the fall of backgammon dice,
and the monks of Enryakuji Temple."40 (For more on the actions of
the armed monks, see Chapter 7.)
The heads of the military branches of the Minamoto and Taira
were regarded as proper chieftains (toryo) only if they had demonstrated superior martial prowess. Of equal importance were connections with the highest levels at court. Hence they maintained a base
in the capital, where they and their kin were appointed to a succession of court titles related to guard offices and they served as private
protection for members of the imperial and noble families. The most
successful served a succession of province governorships. They built
their family's economic strength by acquiring shoen in the provinces,
some of which they commended to their highest patrons at court.
These acts of commendation and gifts of land and horses to their patrons were essential for advancement in court position. The toryo
welcomed imperial appointments to lead punitive expeditions to
punish enemies of the court, even though they had to recruit most
of the military force, because a successful campaign could bring promotion in court rank and position as well as high esteem in the eyes
of other tsuwamono. Generation after generation, the toryo of the
Seiwa Genji were highly successful in this process: Yorinobu settling
the Tadatsune uprising, his son Yoriyoshi victorious in the Earlier
Nine Years'War, and the latter's son Yoshiie in the Later Three Years'
War. Yoshiie was acknowledged as the peerless warrior (musha) of
the realm. In the words of a popular song, "Of all the Minamoto,
Hachiman Taro [Yoshiie] is the most feared." Yet he remained a
samurai of the retired emperor.41
Shirakawa, who became the first cloistered emperor (in) in 1087
(see Chapter 9), was extremely effective both in employing the warrior leaders to his advantage and in balancing rivals against each
other to deflect the danger inherent in their increasing strength.
40 Takaki Ichinosuke, Ozawa Masao, Atsumi Kaoru, and Kindaichi Haruhiko, eds., Heike
monogatari, vols. 32—33 in Nihon koten bungaku taikei (1959-60), vol. 1, kan 1.4, p. 129.
Translation by Helen Craig McCullough, The Tale of the Heike (Stanford, Calif.: Stanford
University Press, 1988), p. 50.
41 Chuyuki (or Nakauki), vols. 8-14 in Shiryo taisei (Tokyo: Naigai shoseki kabushiki kaisha,
1934), vol. 3, p. 129, 16/7/Kajo 1 (1106); and Ryojin hisho, compiled by Go-Shirakawa, kan
2, poem 444, vol. 19 of Nihon koten zensho (Tokyo: Asahi shinbunsha, 1953), p. 137- The
nickname derives fromYoshiie's coming-of-age ceremony at Iwashimizu Hachiman Shrine.
Cambridge Histories Online © Cambridge University Press, 2008
684
T H E RISE
OF T H E
WARRIORS
Yoshiie, after gaining a reputation as a courageous fighter as a young
man in the Earlier Nine Years' War, spent many years doing Shirakawa's bidding. However, he was not reappointed to a lucrative
position after the Later Three Years'War, from which he emerged as
the preeminent warrior. The compensation he gave to his men out
of his own resources won the admiration of the warrior class. The
court was much concerned to check his power and influence, and the
opportunity came when there was the confrontation in 1091 between
his powerful bushidan and the strong force of his brother Yoshitsuna.
The prospect of their forces clashing near the capital alarmed the
government, but the regent, Morozane, was able to mediate the dispute. To control Yoshiie he obtained an imperial order forbidding
provincial troops from coming up to the capital and also prohibiting
landowners from commending additional land to Yoshiie.42 The prestige of this powerful toryo among provincial landholders posed a
threat to the balance of power at court, for the elite had long derived
most of its revenue from the possession and commendation of shoen.
As commander of the next punitive expedition two years later,
Shirakawa turned to Yoshiie's brother and rival and rewarded him
with elevation to the Fourth Rank.43 Five years later Shirakawa mollified Yoshiie by granting him the privilege of entrance to his private
quarters. The aristocrats were appalled that a tsuwamono was allowed
to enter their world on an equal footing.44
In 1106, Yoshiie died at the age of sixty-seven, occasioning a struggle to succeed to the position of the Minamoto toryo. By this time the
bushidan no longer dissolved on the death of the leader, and it became customary for the eldest son to succeed his father. However,
Yoshiie's eldest son died young, his second son, the powerful warrior
Yoshichika, was exiled for murder while governor of Tsushima, and
when another son was killed, Yoshitsuna and his son were rumored to
be implicated. Yoshichika's thirteen-year-old son, Tameyoshi (10961156), was commissioned to arrest Yoshitsuna, and he was credited
with this feat. He succeeded his grandfather as the Minamoto toryo.
However, there was little aristocratic support for a young toryo, and
the Minamoto bushidan weakened significantly. Many of the leaders
left the capital for their bases in the provinces. The Minamoto of the
42 Hyakuren sho, (Shintei zoho) Kokushi taikei, vol. 11 (1929), p. 41; and Go-Nijd Moromichi ki,
in Dainihon kokiroku, series 7, vol. 2 (Tokyo: Iwanami shoten, 1957), p. 132, 12/6/Kanji 5
(1091).
43 Takeuchi, Bushi no tojo, pp. 214-15.
44 Chuyuki, vol. 2, pp. 123-24, 23/io^otoku 2 (1098).
Cambridge Histories Online © Cambridge University Press, 2008
CONDITIONS IN THE CAPITAL
685
generation of Yoshiie and his sons and nephews left a record of violence, lawlessness, and mutually destructive behavior that cost the
family its reputation as the leading samurai. Shirakawa, who had
fomented Minamoto factionalism as a means of checking their
strength, had already chosen the Ise Heishi (Heishi is the Sino-Japanese reading of "Taira clan") as their replacement. Despite a less developed bushidan, with Shirakawa's backing the Taira would rise
higher and more quickly than the Minamoto.
Although the Taira had originally considered the vast eastern
plains their own, they had gradually been absorbed into the Minamoto network after the Tadatsune rebellion. Taira no Korehira was
an exception. The ancestor of the Ise Heishi (a branch of the Kammu
Heishi) and probably a son or grandson adopted by Sadamori (see
Figure 9.7, Genealogy of the Taira), Korehira is believed to have established an estate in Ise when he was appointed governor in 1006.
However, his connection with Ise predated this appointment. In 998,
Fujiwara noYukinari's diary Gonki records a large-scale dispute over
land in Ise between the two renowned warriors, Korehira and Taira
no Muneyori (d. 1011), and their many roto.45 Both men were summoned to the capital for interrogation by the Imperial Police and detained there. Korehira apologized and, on the recommendation of
the Minister of the Right, was appointed governor of Ise. Michinaga,
who was allied with the Minamoto, unsuccessfully opposed the appointment.46 However, he ensured that Korehira's descendants
would not be appointed to higher office in the capital than the Imperial Police.
Three generations later, Taira no Masamori was appointed governor of Oki. Unable to derive much revenue from the small island, he
developed the family land bequeathed to him by his ancestors in Ise
and Iga. In 1098 he commended twenty cho (66 acres) of his holdings in Iga to the Rokujo-in, formerly the residence in Kyoto of Shirakawa's favorite daughter, Teishi (Ikuhomon-in). When she died at
the age of twenty, the grieving Shirakawa became a priest and turned
the residence into a temple dedicated to her salvation. The act of
commendation was probably suggested by Shirakawa's female attendant, the Gion Consort, who was a neighbor of Masamori. This
provided Masamori with political backing in his disputes over land
with the powerful Todaiji and Ise Shrine. It was also the first step by
45 Gonki, Shiryo taisei, vol. 35 (1939), vol. 1, p. 6o, 14/12/Chotoku 4 (998).
46 Takeuchi, Bushi no tojo, p. 222.
Cambridge Histories Online © Cambridge University Press, 2008
686
THE RISE OF THE WARRIORS
the Ise Heishi toward employment in the capital and political power.
After further gifts to Shirakawa, Masamori was employed as a guard
in the hokumen. This was the first time that a Taira had been employed by a retired emperor.
The fact that the Taira had not been favored by the Fujiwara made
them attractive to Shirakawa. Masamori was appointed to lead an attack onYoshiie's son, the powerful warrior Minamoto noYoshichika.
Yoshichika, who had been accused of murder and plunder while
serving as governor of Tsushima and had been exiled to Oki, left his
island exile and in 1107 crossed over to Izumo, where he killed a provincial deputy and looted government stores. Masamori, then in the
vicinity as governor of Inaba, was commissioned to attack Yoshichika. He distinguished himself by killing Yoshichika and four of his
chief retainers, and Shirakawa rewarded him and his sons lavishly.
Shirakawa's trusted aide, the Consultant (sangi) and literatus Fujiwara no Munetada, complained of the favoritism shown by the retired emperor to such a low-ranking warrior, granting him the governorship of the prized province of Tamba.47 With Masamori's
triumphal return to the capital, the martial reputation of the Taira
began to rival the Minamotos'. However, some doubted that a relative unknown could have succeeded so easily against a seasoned warrior. It was suggested that the severed head so prominently displayed
had not belonged to Yoshichika.
Masamori received other important commissions. In 1113, he and
Minamoto no Tameyoshi fought to defend the capital against armed
monks, and Masamori allowed his young sonTadamori (1096—1153)
much of the credit. In 1119, while governor of Bizen, Masamori defeated brigands who had threatened the capital, and he also suppressed rebels in Kyushu, for which he was awarded the Fourth Rank.
Tadamori was granted many opportunities to demonstrate his
prowess and speed his rise through the ranks. According to Ima
kagami (1170), Tadamori's wife, the sister of the Gion Consort,
spoke to Shirakawa often on his behalf.48 In 1129, he moved against
the pirates of the Inland Sea.49 These pirates were coordinating raids
involving dozens of ships to prey on the shipments by shoen of the
annual dues to their guarantors and central proprietors in the capi47 Chuyuki, vol. 3, p. 322, 24/1/Tennin 1 (1108).
48 Ima kagami, (Shintei zoho) Kokushi taikei, vol. 21.2 (1940), kan 4, pp. 100-102. Tadamori's
wife, who is referred to in Chuyuki as "one who is close to the retired emperor," had been
one of Shirakawa's concubines before she was given to Tadamori in marriage.
49 On conditions in the provinces, see Ishimoda, Kodai makki seijishi josetsu.
Cambridge Histories Online © Cambridge University Press, 2008
CONDITIONS IN THE CAPITAL
687
tal.5° To judge from tales in Konjaku monogatari-shu, however, there
does not seem to have been sufficient pirate activity to warrant a
court-ordered military campaign against them. It was more likely
that Tadamori, as governor of Bizen, had requested permission to
confront the pirates.
In 1135, Tadamori led a procession of prisoners through the streets
of the capital. According to Choshuki, "Tadamori had taken seventy
pirates prisoner. . . . Everyone came to see them. He displayed the
head of the pirate Hidaka Zenshi, but most of the captives were not
pirates. It is said that Tadamori had taken men and presented them
as 'pirate captives'."51 The forty "pirates" who were not immediately
turned over to the Imperial Police were presumably released after the
victory procession had served its propagandistic function.
As related in Heike monogatari, Tadamori was ridiculed at court
for his provincial origins. He realized that military successes were
not enough to gain acceptance in elite society, and that he needed
also to develop the skills of a courtier. His dancing at the Kamo Festival in 1119 was admired. But in 1130 he was invited to a poetry
composition party in the retired emperor's residence, where his poetry was judged "not good."52 In 1132, when he built the Sanjiisangendo for Retired Emperor Toba, he was granted the extraordinary
privilege of entry into the ex-emperor's private quarters. Tadamori
managed to narrow the distance between the worlds of the warrior
and the aristocrat. His courtly skills and financial resources compensated in part for the lesser degree of development of the Taira
bushidan.
In their encounters with the pirates, Masamori and Tadamori
had become aware of the profits to be made in foreign trade. Sung
and Koryo traders negotiated under the jurisdiction of the Dazaifu.
By the eleventh century, many of the Dazaifu officials had begun to
expand private trade at the expense of official trade. There was a
great market for foreign goods. Appointed as an official on Shirakawa's Kanazaki shden in Hizen, Tadamori drew up documents to
try to convince the Dazaifu officials that the retired emperor had
granted him permission to trade directly with Chinese ships that
arrived in 1133.53
50
51
52
53
Takeuchi, Bushi no tojd, p. 247.
Choshuki, Shiryo taisei, vol. 7 (1934), vol. 2, p. 301, 19/8/Hoen 1 (1135).
Choshuki, vol. 2, p. 301, 5/9/Daiji 5 (1130).
Mori Katsumi, Nisso boeki no kenkyu, rev. ed., vol. 1 of Mori Katsumi chosaku senshu
(Tokyo: Kokusho kankokai, 1975), p. 176.
Cambridge Histories Online © Cambridge University Press, 2008
688
THE RISE OF THE WARRIORS
By the time of his death in 1153, Tadamori had secured a strong
economic base for the Ise Heishi in landholding and trade, as well as
a leading role among the warrior nobles at court. It was from this
privileged position that his son Kiyomori (1118-81) began his spectacular career as warrior and politician, exploiting the factional rivalries among emperors, retired emperors, and Fujiwara families to
rise to the highest offices at court.
HOGEN DISTURBANCE
The military clash that erupted in the capital in the first year of
Hogen (1156) resulted from a quarrel that deeply divided the imperial family and from rivalries within the Fujiwara regental family
(sekkanke). The administrative organization of the retired emperor's
household, the inset or cloister government (see Chapter 9), shaped
by Shirakawa, was perpetuated by Toba and Go-Shirakawa, who
successively wielded power as senior retired emperor (hon'in) until
the latter's death in 1192. This unusual system of administration was
effective in improving the economic resources of the imperial family
and in reducing the influence of the Fujiwara regents, but it led to
tensions within the imperial family. The dominant role of the cloistered emperor in governmental matters as well as family affairs was
resented by the reigning emperor and his entourage, and by junior
retired emperors who were excluded from the powers of the senior
position.
Shirakawa controlled the court during the reigns of his son
Horikawa, his grandson Toba, and Sutoku, supposedly his greatgrandson. (According to the Kojidan, a thirteenth-century collection
of tales and anecdotes, it was common knowledge that Sutoku, the
son ofToba's consort, Shirakawa's "adopted daughter" Taikemmonin [Fujiwara no Shoshi], was actually fathered by Shirakawa.) Shirakawa required Toba to abdicate in favor of Sutoku.54 Upon Shirakawa's death in 1129, Toba, succeeding as senior retired emperor,
immediately began to undo much of what Shirakawa had accomplished. He replaced Shirakawa's closest retainers, but Taira no
Tadamori managed to retain his position as the main military retainer. Toba recalled the former regent, Fujiwara no Tadazane, who
had been banished from the court by Shirakawa. Tadazane made
plans to reestablish the fortunes of the regental family by reunifying
54 Takeuchi, Bushi no tojo, pp. 318-19.
Cambridge Histories Online © Cambridge University Press, 2008
HOGEN DISTURBANCE
689
its divided estates, which had diminished under Go-Sanjo and Shirakawa, and by expanding existing shoen. In 1150 he also attempted
to remove his son Tadamichi from the position of regent in order to
entrust leadership of the clan to his favored younger son, Yorinaga.
However, Yorinaga was outmaneuvered by his half brother Tadamichi andToba's consort Bifukumon-in (Fujiwara noTokushi). Accused of causing the illness of which Emperor Konoe died in 1155
by sticking needles into the eyes of a statue of the emperor, Yorinaga
was excluded from the court and a bitter rivalry divided the regental family.
Toba had forced Sutoku to abdicate in 1141 in favor of his younger
"brother" Konoe. Sutoku then pinned his hopes on his own son becoming crown prince, but when Konoe died in 1155, Toba selected
another of his own sons, Go-Shirakawa, as emperor and the latter's
son as crown prince (the future emperor Nijo). Sutoku was still determined that his son should gain the throne, and that, as a consequence, he himself would become the senior retired emperor as father of the reigning emperor. He was supported in this ambition by
Yorinaga, who aimed to wrest the leadership of the regental house
from Tadamichi. When Toba fell critically ill the next year, Sutoku
sought military supporters.
The fraternal quarrels between Go-Shirakawa and Sutoku and between Tadamichi and Yorinaga would likely have been resolved in
earlier centuries by maneuver, political manipulation, and a judicious use of exile, but the ready availability now of military forces
was a temptation too great for either side to resist.55 It was rumored
that Sutoku and Yorinaga, who were backed militarily by the Minamoto toryo Tameyoshi, and by Taira no Tadamasa (younger
brother of Tadamori), were about to "mobilize troops and overthrow
the state."56 Taira no Tadamori's son, Kiyomori, and Tameyoshi's
son, Yoshitomo, who was on bad terms with his father, followed by
55 On the Hogen and Heiji disturbances, see Uwayokote Masataka, Gempei no seisui, vol. 6 of
Nihon no rekishi bunko (Tokyo: Kodansha, 1975), pp. 51-81; and Yasuda Motohisa, Gempei
no sokoku, vol. 7 of Kokumin no rekishi (Tokyo: Bun'eido, 1967).The narrative related here
generally follows Hogen monogatari and Heiji monogatari, semifictional accounts that in
their earliest versions seem to date not later than the thirteenth century. The texts used are
the Muromachi-period versions printed in vol. 31 of Nihon koten bungaku laikei. A partial
English translation of another, somewhat different, text of Heiji is included in Edwin O.
Reischauer and Joseph K. Yamagiwa, TranslationsfromEarly Japanese Literature (Cambridge, Mass.: Harvard University Press, 1951), pp. 396-446. A complete translation of
Hogen, also based on a different, fifteenth-century or later text, is in William R. Wilson,
"Hogen monogatari": Tale of the Disorder in Hogen (Tokyo: Sophia University Press, 1971).
56 Takeuchi, Buke no tojo, pp. 344—45, quoting Heihanki, the diary of Taira no Nobunori, an
important source on the Hogen clash, in Shiryo taisei, vols. 15—18 and 37 (1934-39) •
Cambridge Histories Online © Cambridge University Press, 2008
690
THE RISE OF THE WARRIORS
most of the Minamoto row, aligned themselves with the imperial faction of Go-Shirakawa. The dying Toba issued orders to his military
chiefs, Kiyomori and Yoshitomo, to mobilize their retainers for the
defense of his palace near the junction of the Kamo and Katsura
rivers south of the city, and also Go-Shirakawa's Takamatsu palace.
With the death of Toba on the second day of the seventh month, Sutoku became concerned for his own personal safety and summoned
such military support as he could command: Yoshitomo's father,
Tameyoshi, and two of Tameyoshi's junior sons; Minamoto warriors
from Yamato province; some Taira warriors, most notably Sutoku's
favorite Tadamasa; Yorinaga and the armed men he had mustered
from his shoen; and the armed monks of the Kofukuji Temple at
Nara. The warriors responding to Sutoku's call included formidable
fighters, but among them theYamato Minamoto were taken prisoner
by Go-Shirakawa's men before they reached the capital, and the
Kofukuji monks never arrived, so that the total number of fighters
defending Sutoku's Shirakawa palace remained small.
The Shirakawa palace was a large establishment east of the Kamo
River at the northwest corner of what is now the site of the Heian
Shrine. The palace was about a mile and a quarter from Go-Shirakawa's residence and base at the Takamatsu palace north of Sanjo
Avenue just southwest of the Greater Imperial Palace. On the night
of the eleventh, fighting erupted in a furious attack, launched at
nightfall on Shirakawa palace by the forces of Go-Shirakawa. The attacking force included Kiyomori with about three hundred mounted
men,Yoshitomo with two hundred, andYoshiyasu with one hundred.
They set fire to the palace, and Sutoku's men fled. Within a few
hours both the battle and the war were over.
Sutoku succeeded in escaping to the Ninnaji Temple just beyond
the northwest corner of the capital, where he was captured and sent
into exile in Sanuki on Shikoku, dying there a bitter man. Hit by a
stray arrow during the fighting at Shirakawa, Sutoku's supporter
and accomplice Yorinaga died a few days later. Men linked with
Yorinaga at court were exiled, and some seventeen of Sutoku's warriors were - against three and a half centuries of bloodless precedent — condemned to execution by imperial order. Despite Yoshitomo's plea for his father Tameyoshi's life, he was ordered to
execute him, as was Kiyomori his uncle Tadamasa. This tsuwamonolike disposition of prisoners stunned the court, a decision that was
urged by Go-Shirakawa's eminence grise, Lesser Counselor Fujiwara no Michinori, known by his Buddhist name Shinzei (1106Cambridge Histories Online © Cambridge University Press, 2008
HEIJI DISTURBANCE
691
60). Although he was of low rank, he enjoyed the trust of the emperor because his wife had been Go-Shirakawa's wet-nurse, presumably creating the kind of emotional bond that the fiction of the
period suggests.
Although the fighting had demonstrated for the first time the effectiveness of the warrior families in determining events in the capital, their rewards were meager. Go-Shirakawa was no more generous in his treatment of Kiyomori andYoshitomo after the coup than
most previous emperors and regents had been in rewarding warriors. From the court's point of view, the warriors who had defended
the emperor and stormed the Shirakawa palace were simply useful
servants of the throne who could be employed in time of need and
otherwise mostly ignored. The rewards publicly granted the two
great warriors were insignificant in comparison with their political
and economic strength. Kiyomori, the senior warrior representative
at court and the heir to a long family tradition of court service, received merely the governorship of Harima; Yoshitomo, the Minamoto chieftain and the architect of the victory at Shirakawa, had
to be content with lowly appointment as Provisional Head of the
Left Horse Guards, a post of minimal distinction and probably no
power or remuneration whatever.
The Hogen Disturbance resolved sharp conflicts within the imperial and Fujiwara leadership, placing control of the court firmly in
the hands of the reigning emperor, but it left unanswered the question of military supremacy as between Go-Shirakawa's two chief
warrior leaders, Kiyomori andYoshitomo, and the further and more
fundamental issue of the role of powerful warriors vis-a-vis the imperial court that remained, fatally to the regime, unaddressed.
HEIJI DISTURBANCE
Court politics at Heian in 1159, the first year of the Heiji era, were
guided by the complex web of rivalries, ambitions, personal relationships, and alliances that had arisen at court following the Hogen
conflict three years earlier. Go-Shirakawa had come to the throne in
1155 as a stopgap successor transitional to the accession of his son
Nijo. When he abdicated in favor of Nijo in the fall of 1158, the relationship between the two was for that reason even more tense than
usual in the case of an ex-emperor who would rule as senior retired
emperor and a reigning emperor. Bifukumon-in, former mistress of
the late ex-emperor Toba and an original supporter of Nijo in the
Cambridge Histories Online © Cambridge University Press, 2008
692
THE RISE OF THE WARRIORS
succession struggles, was still alive, exercising a powerful influence
at court as a result of her relationship with Nijo and also because of
the vast shoen wealth that had come to her fromToba.
Emperor Nijo was still young, only fifteen years old, but he had a
reputation for quick intelligence, and he was intent on ruling the
court in his own right, after the manner of his father, whose wishes
he tended to ignore. He was supported in that ambition and advised
by two skilled Fujiwara politicians: his maternal uncle, Grand Counselor Tsunemune (1119-89), and his wet-nurse's son, Korekata (fl.
1125-66), the Superintendent of the Imperial Police (kebiishi no
betid). Warriors in the emperor's entourage included Minamoto no
Yoshitomo's senior kinsmen Mitsuyasu (1095-1160), the father of
ex-emperor Toba's last mistress, who had served as a wet-nurse to
Nijo, and Minamoto noYorimasa (1104-80), who had long been in
the employ of imperial family members and remained loyal now to
his sovereign, even eventually turning his back on his own clan
leader to follow Nijo into the embrace of the Taira.
The control of the court that had been exercised by Shinzei since
the Hogen Disturbance was now endangered by the strength of the
emperor's party and more particularly by the rapid growth in influence of a new favorite of Go-Shirakawa, Fujiwara no Nobuyori
(1133-60), who seemed to be on the verge of usurping Shinzei's formerly dominant role. Shinzei was alive to the danger and endeavored
to bring Nobuyori into disrepute with the ex-emperor, denigrating
him at every opportunity, but the result was only to incense Nobuyori
further against himself. The relations between Nobuyori and Shinzei
reached their nadir when Shinzei succeeded in blocking the appointment of Nobuyori to a key office that was usually a stepping-stone to
high ministerial rank. Pleading illness, the enraged Nobuyori refused
to attend court, staying in seclusion "to practice the military arts."
Toward the end of the first year of the Heiji era, on a day corresponding to a date in January 1160, Taira no Kiyomori, in fulfillment
of a vow, left the capital accompanied by his eldest son, Shigemori
(1138-79), and a small body of followers on a pilgrimage to Kumano, a well-known and popular sacred site of important shrines,
mystics, and mysticism near the west coast of Ise Bay in the southern part of modern Mie Prefecture. The shrines were about 175
miles distant by the circuitous route around the mountains that usually required a minimum five days to traverse.57
57 Omori, Bukejidai no kenkyu, vol. 2, p. 188.
Cambridge Histories Online © Cambridge University Press, 2008
HEIJI DISTURBANCE
693
Just five days after Kiyomori's departure, his chief rivals at court,
Nobuyori and the Minamoto chieftain Yoshitomo, began to move
against him and his patron Shinzei. Despite Yoshitomo's critical services at the time of the Hogen conflict, his fortunes had been checked
by Kiyomori acting in league with Shinzei, and he was quick now to
join a cause that might offer remedy for that injustice. The NobuyoriYoshitomo alliance was joined by other disaffected warriors and by
courtiers seeking to strengthen Emperor Nijo's personal control of
the court, notably Tsunemune and Korekata. Its military core and
most of its strength was Minamoto.
On the ninth of the twelfth month of the Heiji year, a body of
mounted men under the command of Yoshitomo and Nobuyori
forced ex-emperor Go-Shirakawa to vacate his Sanjo palace and
move to the Greater Imperial Palace, where he and Emperor Nijo
were held under house arrest. Fire was set to the palace at Sanjo and
to Shinzei's residence next door with great loss of life. On the following day, Nobuyori exercised his new power at court to rid the
government of Shinzei's sympathizers and to obtain the dismissal of
Shinzei and Kiyomori from their official posts. He also obtained for
himself a coveted appointment to a high ministerial post and for
Yoshitomo the governorship of Harima that had just been snatched
from Kiyomori. Shinzei fled to Nara, where a few days later he was
caught and executed following a suicide attempt interrupted by his
executioners. His severed head was brought back to the capital for
public display like that of a common criminal, and the revenge of
Nobuyori seemed complete.
Kiyomori, whose absence from the capital had created the opportunity for the conspirators, was at a post station on the Pacific coast
toward the tip of the Kii peninsula when a messenger reached him
with news of the untoward events at court. After some vacillation, he
set out for the capital at die head of his small retinue, now grown
with local reinforcement to one hundred mounted men. Picking up
the support of three hundred additional men en route, he entered
his Rokuhara residence in the city just a few days later, on the seventeenth of the month. He made an initial show of submission to the
newly triumphant regime, but when on the twenty-sixth Emperor
Nijo, disenchanted with Nobuyori's highhanded ways and advised
by Tsunemune and Korekata, managed to slip out of the imperial
palace disguised as a woman to join the Taira at Rokuhara, Yoshitomo and Nobuyori were clearly defined as rebels and the Rokuhara
Taira acquired the legitimacy of an imperial army. On the same day,
Cambridge Histories Online © Cambridge University Press, 2008
694
THE RISE OF THE WARRIORS
Go-Shirakawa also escaped custody at the imperial palace and fled
to the Ninnaji Temple.
The day after the flight of Nijo and Go-Shirakawa, on the twentyseventh of the month, the Taira force under Shigemori launched an
attack on the forces ofYoshitomo and Nobuyori at the Greater Imperial Palace. The attackers withdrew after fierce fighting, luring the
Minamoto out of the palace, which the Taira were then able to occupy and secure. The dispossessed Minamoto advanced on Rokuhara, where Kiyomori waited to do battle with them, crushing them
and sending them into headlong retreat toward their homelands in
the east. After only a few hours of actual fighting between troops of
several hundred horsemen, the conflict was over.
The badly frightened Nobuyori, fleeing for his life before the final
battle, was caught and executed near Rokuhara on the bed of the
Kamo River at Rokujo. His age was just twenty-six. Yoshitomo attempted to flee to the east, but he was betrayed and killed at a refuge
in Owari Province. Yoshitomo's two older sons were executed, and
the thirteen-year-old Yoritomo would have suffered the same fate but
for the plea of Kiyomori's stepmother, Ike no Zenni. He was banished to Izu, placed in the custody of a trusted Taira ally, H6j6 Tokimasa.The infant Yoshitsune was also spared and was assigned to the
priesthood. The Taira victory over the Minamoto appeared to be
complete and the Minamoto toryo line at an end. But in a remarkable
reversal of fate, the brothers Yoritomo andYoshitsune survived to lead
the Minamoto to victory over the Taira twenty-five years later.
The Heiji Disturbance had its beginnings in the rivalries and jealousies among confidants of Retired Emperor Go-Shirakawa and
Emperor Nijo, but unlike the Hogen incident, its action was determined not by commands issued by emperors or court officials, but
by warrior chieftains acting on their own personal ambitions. The
events revealed to the warriors the powerlessness of the court and
the first hints of the political realities of the warriors' military power.
Nevertheless, it is probably premature to speak of this time as the
beginning of warrior rule, as some historians have. It is at about this
time, when the military and political landscape was changing and
warriors acted more on their own initiative, that the term bushi came
into use in referring to warriors. (The term existed in Nara and early
Heian times, when it referred neither to warriors or military officials
[bukan], but rather to literati who were skilled in military arts.)58
58 Shoku Nihongi, (Shintei zoho) Kokushi taikei, vol. 2 (1935), p. 84, 27/1/Yoro 5 (721).
Cambridge Histories Online © Cambridge University Press, 2008
TAIRA RISE TO POWER
695
TAIRA RISE TO POWER
The political history of the two decades following the Heiji Disturbance until the beginning of the war between the Minamoto and
Taira (the Gempei War, 1180-85) is discussed more fully in Chapter
9. We are concerned here specifically with the rise to power at court
of Taira no Kiyomori and his kinsmen.
In the Hogen and Heiji conflicts, the Taira mobilized retainers in
particular from Bizen and Bitchii provinces (most of what it now
Okayama Prefecture) on the Inland Sea and from Iga and Ise
(northern Mie Prefecture) southeast of the capital. Those were regions where Taira leaders since the time of Kiyomori's grandfather
Masamori had been custodial governors (zuryd). Following the decimation of the Minamoto leaders in the Heiji fighting, Kiyomori, as
the triumphant hegemon of warriors, had no difficulty in recruiting
numbers of the Minamoto's roto as well as others to his organization.
In the following years the Taira developed their network of allies, not
only in western Japan and in the capital area, where they had long
been prominent, but also in the east. The commission given by GoShirakawa to Kiyomori's son, Shigemori, to be prepared to provide
military security and police functions wherever needed throughout
the country, gave him the opportunity to extend Taira military influence. The economic position of the Taira was also strengthened by
their seizing control of some shoen of Minamoto and courtier families. The appointment of Taira relatives to more and more provincial
governorships also strengthened the financial position of the Taira.
More impressive than Kiyomori's military and economic gains
were his rapid promotions in the court hierarchy. In 1160 he received
the Third Rank, the first of a warrior family to join the senior nobles
(kugyd). In 1167 he rose to the post of Chancellor (daijo daijiri), and
the First Rank, the pinnacle of the bureaucratic structure. He
arranged the promotion of several Taira kinsmen to the ranks of
the senior nobility and membership in the Council of State. He left
the chancellorship after a few months, speaking of poor health, and
took holy orders; but he had made his point, and continued to assert
his influence on court appointments and policy decisions. An adroit
politician, Kiyomori took full advantage, after the Heiji affair, of the
rivalry for power between the senior retired emperor and his son,
Emperor Nijo, both of whom courted him with rewards as they vied
for his military support. He also played the game of marriage politics in the Fujiwara style to advance his personal relationships in the
Cambridge Histories Online © Cambridge University Press, 2008
696
THE RISE OF THE WARRIORS
highest court circles. His eight-year-old daughter was married to the
future regent Motozane in 1164. His connection with Go-Shirakawa
became familial when his Taira sister-in-law, Shigeko (Kenshummon-in), bore the retired emperor a son, later enthroned as Emperor
Takakura (1161-81, r. 1168-80). Subsequently Kiyomori's daughter
Tokushi became Takakura's consort, and their son, the infant Antoku, was placed on the throne by his grandfather, Kiyomori, in
1180, at the height of his control of the court.
Contemporary sources do not, for the most part, comment critically on the rapid advancement of Kiyomori, as a person of warrior
lineage, to such high position at court. The explanation of his initial
acceptance may lie in prevalence of the rumor, which appears in
several sources, that his natural father was actually Emperor Shirakawa. The source considered the most reliable says that the emperor, upon the death of Kiyomori's mother, gave the two-year-old
child for adoption to Taira no Tadamori, who thus became his genealogical father. 59
The Tale of Heike {Heike monogatari) gives a dramatic account of
the "flowering fortunes" of Kiyomori at the height of his success:
"Not only did Kiyomori himself attain the pinnacle of worldly success, but his entire family shared his prosperity." The account enumerates the high offices held by three of his sons and the good marriages of seven daughters. "Sixteen Taira ranked as senior nobles,
more than thirty were courtiers, and more than sixty held appointments as provincial governors, guards officers, or members of the
central bureaucracy. It was as though there were no other people in
the world."60
Although the key to Kiyomori's ascendancy was ultimately his military strength, he climbed the ranks of court position and advanced
his kinsmen by employing the familiar techniques of the civil courtiers rather than by intimidation, at least until 1177. In time, however,
Kiyomori's growing power at court caused increasing resentment
among courtiers, and Go-Shirakawa in particular sought to check his
strength, leading to an organized attempt in 1177 (the so-called
Shishigatani plot) by officers of the retired emperor's household, acting in league with some military figures, to move against the Taira
59 The statement in the Heike monogatari that Kiyomori's mother was Shirakawa's favorite
concubine, Gion no Nyogo, is thought to be incorrect. Rather, it was her younger sister
who bore Shirakawa's child, according to documents discovered in modern times.
Takeuchi, Bushi no to jo, pp. 388-89.
60 Heike monogatari, vol. 1, kan 1.5; in H. McCullough, The Tale of the Heike, pp. 28-29.
Cambridge Histories Online © Cambridge University Press, 2008
TAIRA RISE TO POWER
697
leader. The plot was discovered and crushed before it could be put
into action, however, leaving Kiyomori more powerful than ever but
with a heightened enmity toward Go-Shirakawa and his court.
One issue of contention was over control of the landholdings of
the regental house. When Regent Motozane, who was married to
Kiyomori's daughter, Moriko, died in 1166, Kiyomori arranged for
her to inherit most of the estate, passing over Motozane's eldest son,
Motomichi, the explanation given being that Motomichi was too
young to assume such a responsibility. He was only six years old, it
is true, but Moriko herself was just ten. When Moriko died in 1179,
the Fujiwara leaders demanded that the holdings revert to Motomichi. At that point Go-Shirakawa intervened in the dispute and
confiscated the property. He also confiscated the holdings of Shigemori, Kiyomori's heir, when he died the next month.
Kiyomori countered this threat to his power with a military coup
in the eleventh month of 1179. He led several thousand warriors from
Fukuhara and paraded them in the capital. He demanded the dismissal of thirty-nine officials, retainers (kinshiri) of Go-Shirakawa,
and replaced them with Taira adherents. The retired emperor was
placed under house arrest and moved to the Toba-dono. Kiyomori
was declared a traitor by his distraught imperial enemy, but all power
was now in his hands. Three months later, Go-Shirakawa's son,
Takakura, was forced to abdicate and was replaced as emperor by
Kiyomori's infant grandson. Kiyomori's climb to full control of the
court had taken two decades, but his victory was short: he died of illness early in 1181.
Takakura, upon his abdication, instead of paying his respects at
the nearby shrines of Iwashimizu Hachimangu and Hie (or Hiyoshi), as was the recent custom, went on pilgrimage to a distant
shrine under Taira patronage on the small island of Itsukushima off
the Inland Sea coast near what is now the city of Hiroshima. The
Iwashimizu Hachimangu and Hie shrines were linked by religious
and other ties to the powerful Onjdji Temple at Otsu, and the temple's angry reaction to the Taira-inspired imperial slight contributed
to the formation of a potentially powerful coalition of anti-Taira interests among the armed monks of that temple, their sometime enemies at the Enryakushi on Mount Hiei, and the monks of Kofukuji,
the Fujiwara clan temple at Nara.
Open, military opposition to Taira rule began, however, not in the
temples but within the imperial court itself. Among the few Minamoto chieftains to survive the Heiji Disturbance was a remote relCambridge Histories Online © Cambridge University Press, 2008
698
THE RISE OF THE WARRIORS
ative of the clan leader Yoshitomo named Yorimasa (1104-80), an
aged warrior-courtier connected also with the Fujiwara through his
mother. In the court's neglect of deserving Minamoto warriors following the Heiji fighting,Yorimasa's career had not prospered, but in
his old age he had been permitted a considerable degree of court
success, rising to the ranks of the senior nobility (the first Minamoto
warrior to do so). What Yorimasa's specific motives may have been is
not known, but whether they were resentment of pastTaira injustice
and highhandedness or simply opportunistic personal ambition, in
the fourth month of 1180 he persuaded Go-Shirakawa's son Mochihito (1151-80) to issue a call to warriors everywhere for the chastisement of the Taira.
Mochihito's call to arms was a bold and dangerous step, one that
soon cost him his life. But in agreeing to take it the prince may have
believed he had more than sufficient grievance against the Taira to
justify the risk, as well as an adequate prospect for success among
the disgruntled, ambitious warriors in the provinces. He may have
shared his father Go-Shirakawa's hostility toward Kiyomori, encouraged by his Fujiwara connections, but he also had his own personal
complaints against the Taira dispensation. At the time of Kiyomori's
coup d'etat in 1179, Mochihito suffered the confiscation of a rich
landholding. The following year, 1180, his hope of succeeding his
cousin Takakura on the imperial throne was dashed when Kiyomori
obtained the succession for his grandson. Deprived of his property
and disappointed in his fondest hope by the all-powerful Taira
leader, Mochihito must have been a willing auditor for any proposal
aimed at the destruction of the Taira.
The plot against the Taira became known, however, before the
conspirators had their supporters in place, and under threat of arrest
and exile, the prince fled the capital to the protection of the Onjoji
Temple. The temple responded to his plea for help and attempted to
rally also to his cause the armies of the Enryakuji and the Kofukuji,
but when that attempt failed to bring to the prince the hoped-for
support, he set out accompanied by Yorimasa and a body of warriors
to obtain the backing of the monks at Nara. Overtaken by a large
Taira force at Uji, he was slain in the ensuing battle and rout, and
his wounded commander Yorimasa was seized and allowed at the age
of seventy-six to commit suicide. The first stage of the TairaMinamoto conflict thus came to an abrupt conclusion the month
after the prince's fateful order, but that call to arms had galvanized
Minamoto warriors in many regions, preparing the way for the great
Cambridge Histories Online © Cambridge University Press, 2008
TAIRA RISE TO POWER
699
battles that achieved the prince's aim of removing the Taira from the
court.
During the tense and uneasy days just after the deaths of Mochihito andYorimasa, Kiyomori abrogated the administrative authority
of Go-Shirakawa's household office (in-no-cho) and announced the
transfer of the emperor and two ex-emperors to what he intended to
be a new imperial capital at Fukuhara near the Inland Sea coast. Taira
men had been the "beneficial holders" (chigyo) of the district in
which Fukuhara was located since 1162, and after Kiyomori took
holy orders in 1168 he had made his residence there, continuing to
direct affairs, and receiving visits at his hillside retreat from at least
one Chinese trader and also, before their falling-out, from the exemperor Go-Shirakawa. It was from Fukuhara that Kiyomori had descended on the capital in 1179 to carry out the palace coup that year.
Now, in 1180, after Mochihito's reliance on temple armies in the
vicinity of the capital during his attempt to overthrow the Taira, and
after a long earlier history of disruption in the city caused by those
monks, Kiyomori may have been seeking to escape the threat of the
temple forces by the move to Fukuhara some fifty miles distant.61
Kiyomori seems to have had a particular interest and involvement
in seagoing trade, and in that connection he turned his resources to
the improvement and development of an old port near Fukuhara
called 6wada-no-tomari. Little is known about the nature or the volume of the traffic passing through the port, but it lay only about a
day's voyage from the entrance to theYodo River, the shipping route
into Heian, and it would have been a convenient port of call for both
domestic and foreign bottoms navigating the Inland Sea. Through
its waters, one may assume, passed trading ships bearing the Chinese books, the Zen Buddhist texts, and the copper coins that were
helping to transform the contemporary intellectual and economic
life of Japan.
The emperor and the ex-emperors were temporarily housed in
residences of Kiyomori's sons at Fukuhara, and planning for the new
capital proceeded apace. After only a few months, however, toward
the end of the year of die court's arrival at Fukuhara and at a time
when construction of an imperial residential palace had just been
completed, the entire project was abandoned, the imperial party and
the courtiers and officials all hastening back to Heian. The reasons
61 Heike monogatari, vol. 1, kan 3.11, pp. 379-80; in H. McCullough, The Tale of the Heike,
P- 193Cambridge Histories Online © Cambridge University Press, 2008
700
THE RISE OF THE WARRIORS
for the abandonment of Fukuhara after such a short time remain obscure, but the cramped space of the site, its windswept, dismal location, the ill health of Takakura, and the outbreak at just this time of
Minamoto uprisings in the east were possibly key considerations.
The removal of the court and its government to Fukuhara left
Heian tattered, dazed, and desperate. The Enryakuji, Heian's guardian temple, was moved to declare itself an enemy of the Taira tyranny,
and the many followers of the court, unable to obtain living accommodations at Fukuhara, also had cause to resent the Taira leadership.
The city had suffered a devastating fire just three years earlier, and it
had become more desolate than ever when people dismantled buildings in order to ship the materials to Fukuhara.62
GEMPEI WAR
The several years of warfare following the failed attempt of Mochihito andYorimasa to overthrow the Taira in 1180, usually known as
the Gempei (i.e., Minamoto-Taira) War, has been popularly portrayed as an epic struggle between two great warrior clans for control of the imperial court and its institutions. It is dramatized as a
rousing tale of political intrigue and heroic battle, with victory going
ultimately to the Minamoto because of the political adroitness and
military genius of the Minamoto overlord Yoritomo and the superhuman fighting abilities of the Minamoto generals, aided by the
sometimes fumbling cowardice of Taira generals, whose martial
valor is supposed to have been corrupted by the genteel life of the
court.
That one-dimensional picture of clan rivalry is at once muddied
by the presence of Taira and Minamoto warriors on both sides of
the conflict and by major intraclan fighting, especially among the
Minamoto. Unlike earlier wars in Japan, which were mostly military
expeditions sent by the court, the Gempei War was a national civil
war, the fighting extending from the Kanto and northeast coast to
the western end of Honshu. The war was also part of a broader political, economic, and social upheaval, a revolution that eventually
resulted in a changed society, a society that may loosely be called
feudal. Dominated by a new class of warriors supported by land
62 Kamo no Chomei, Hojoki, in Nishio Minoru, ed., Hojoi. Tsurezure gusa, vol. 30 of Nihon
koten bugaku taikei (1957), pp. 26-28; Helen McCullough, trans., "An Account of My Hermitage," in her Classical Japanese Prose: An Anthology (Stanford, Calif.: Stanford University
Press, 1990), pp. 382-84.
Cambridge Histories Online © Cambridge University Press, 2008
Map 10.2. The Gempei War.
Cambridge Histories Online © Cambridge University Press, 2008
7O2
THE RISE OF THE WARRIORS
rights conferred or confirmed at least in part by military chiefs, the
country under the new regime was governed in large part by a
warrior-controlled political structure, at the top of which stood the
bakufu shogun in Kamakura.
The first stage of the war followed some months after Mochihito's
call to arms in the fourth month of 1180, when warrior uprisings
began to challenge and in some instances overthrow Taira authority
in many parts of the country. The greatest of the challenges came
from two Minamoto cousins, Yoritomo (1147-99) and Yoshinaka
(1154-84), grandsons of the Tameyoshi who had been executed after
the Hogen Disturbance. They had escaped the slaughter at the time
of the subsequent Heiji Disturbance, and had been living obscurely
in the east and north.
Following defeat in the Heiji fighting, Tameyoshi's sonYoshitomo
had been killed as he fled eastward; his thirteen-year-old son and
heir, Yoritomo, in an act of clemency, was exiled to the distant wilds
of Izu Province (the Izu peninsula south of Tokyo). By the time news
of Mochihito's call to arms reached Izu in 1180, Yoritomo, now adult
and married to Hojo Masako (1157-1225), a daughter of one of his
custodians, Tokimasa (1138-1215), was on such terms with local
warriors that, with his father-in-law's backing, he was able in the
tenth month to raise a force that attacked and killed the Taira deputy
in Izu.
The warriors who rallied to Yoritomo's cause were often not so
much partisans of his clan as rebels against the old regime of the
imperial court and its control of land rights. The conflict is accordingly best described as a struggle between the military usurpers of
that regime, the Taira, and warriors striving to gain secure access to
the management of land resources, rallying to the Minamoto. Insofar as the Minamoto fought merely for clan supremacy, or military
overlordship, they were fortuitous leaders of a revolution long in the
making.
Yoritomo accepted all warriors, whatever their earlier affiliations,
if they would pledge allegiance to him. In return he issued, on his
own authority, confirmation of their rights to land and office. To
usurp the powers of the imperial government in this way was a radical departure. Basing his authority on Mochihito's call and his own
ancestral association with the Kanto region, he claimed jurisdiction
over all public and private land in the east. He claimed, in essence,
to be the ruler of the region. His message to local provincial officials
was to oust officials and agents of the central government and seize
Cambridge Histories Online © Cambridge University Press, 2008
GEMPEI WAR
703
control of their estates and provinces. In the first two precarious
months of Yoritomo's bold program, he was joined by some Taira
adherents and opposed by some Minamoto competitors. The
promise of secure and permanent confirmation of hitherto uncertain
tenures was compelling. Province by province the tide turned in
Yoritomo's favor. Some who at first opposed him or would not commit to him now joined his ranks. They were drawn less by the mystique of the Minamoto lineage than by calculations of how their interests could best be served. Yoritomo established a military and
administrative headquarters at Kamakura, a site northeast of Izu on
Sagami Bay associated with his family for five generations since the
time of Yoriyoshi. His offer of confirmations to eastern warriors
brought more than troops and the provinces' resources needed for
the forthcoming military struggle; it also built a new political organization essential for the long-term goal of ruling the now autonomous eastern region.
In the tenth month the Taira at court dispatched an army under
the command of Kiyomori's grandson, Koremori, to subdue the
outlaw Yoritomo. Yoritomo advanced to meet the enemy force of
70,000 that waited at the Fuji River, just west of Mount Fuji, with a
force of "200,000 horsemen," according to the wildly exaggerated
account in the Tale o/Heike.6s In any case,Yoritomo may have had a
numerical advantage that persuaded the Taira to withdraw without a
fight and return to the capital, further encouraging rebellious warriors in the east and in the provinces around the capital to rise
against the reeling Taira. Yoritomo decided to consolidate his position in the east rather than to pursue the Taira.
The Taira cause was by no means lost. Shortly after Koremori's retreat, Kiyomori returned the court to Kyoto from Fukuhara, reinstated Go-Shirakawa's administrative authority as senior retired emperor, and moved decisively against the great temples, attacking
Onjoji and sending his fifth son, Shigehira, to burn Todaiji and
Kofukuji in Nara in retaliation for their armed support of Minamoto
forces. That piece of sacrilege, which occurred at the beginning of
1181, deeply offended court society. The Kofukuji was the clan tem63 Heike monogatari, vol. i, kan 5.9, pp. 371; in H. McCullough, The Tale of the Heike, pp.
187-88. Azuma kaganti (Shintei zoho) Kokushi taikei, vol. 32 (1932), kan 1, p. 51, also says
200,000. A selected translation of Azuma kagami for the years 1180-85 ' s in Minoru Shinoda, The Founding of the Kamakura Shogunate 1180-1185 (New York: Columbia University Press, i960), see p. 187. For a discussion of the size of Heian armies and expeditionary
forces, and the tendency of most sources to inflate the numbers, see Farris, Heavenly Warriors, pp. 291-93, 300-302, 392-93.
Cambridge Histories Online © Cambridge University Press, 2008
704
THE RISE OF THE WARRIORS
pie of the Fujiwara; the Todaiji served a similar role for several other
noble families, and it was also revered by the imperial family as the
historic Buddhist protector of the state. A month later Kiyomori died.
In the aftermath of the debacle at Fujigawa, the Taira leaders undertook to mobilize military manpower and resources in the nine
provinces surrounding and to the east of the capital. Kiyomori's
third son, Munemori, was appointed to the new post of commanderin-chief (sdkari), and three months later the Taira gained a victory
overYoritomo's uncle, Yukiie, at the Sunomata River (near modern
Nagoya).That battle was followed by a two-year lull in the fighting,
the effect in part of a famine that hampered the recruiting and provisioning of troops, especially in the west. During this period of
preparation for the inevitable confrontation of the military forces,
Yoritomo had an advantage in his independence from central authority, being able to seize control of provincial headquarters, lands,
and shoen in the east to gain the men and supplies he needed. The
Taira, as administrators of the central government offices they had
usurped, were constrained by the bureaucratic procedures of the
court system. They requisitioned troops and supplies from the provincial administrators and shoen holders, but the process was slow
and cumbersome and they encountered resistance. There was strong
resentment of the Taira at court and among the temples, but perhaps
more serious was the alienation of many of the provincial warrior
class. The new Taira proprietorships in the provinces often reduced
land rights and income of the local warriors, the very group that the
Taira needed for their armies. The Taira at court did little to champion the interests of the warrior class, while provincial warriors
viewed with distrust the rise of the Ise Taira to high court position
and intermarriage with court nobility.
The next challenge to the Taira that revealed the weakness of their
forces came not from Yoritomo, who continued to consolidate his administration and military strength in the Kanto, but from his cousin
Yoshinaka (commonly called KisoYoshinaka), who had responded to
Prince Mochihito's call a month after Yoritomo. Emerging from his
refuge in the Kiso Mountains of Shinano (Nagano Prefecture), he
continued to raise troops and won a series of small battles against
Taira adherents. The Taira sent a large army led by Koremori against
him, but Yoshinaka scored a crushing victory in the fifth month of
1183 at Kurikara in Etchu (near the present boundary of Ishikawa
andToyama prefectures near the Sea of Japan). This victory opened
the road to Kyoto. Just three days before Yoshinaka's seizure of the
Cambridge Histories Online © Cambridge University Press, 2008
GEMPEI WAR
705
capital at the end of the seventh month, the Taira leader Munemori
fled with his nephew. Emperor Antoku, to Kyushu, traditionally a
center of Taira strength. Go-Shirakawa, who had evaded the Taira to
join Yoshinaka at Mount Hiei the day before Munemori's flight, ordered him andYukiie to pursue thefleeingTaira, referring now to the
Minamoto forces as the "imperial army."64 Ignoring the absence of
the imperial regalia with Antoku in Kyushu, the ex-emperor also installed his three-year-old grandson, Go-Toba (1180-1239, r. 118398), on the throne to replace Antoku (although the latter, as the possessor of the regalia, is usually considered to have been the emperor
until his death in 1185).
Yoshinaka's occupation of the capital left the country divided into
four satrapies, each dominated by a different and hostile military authority: (1) Kyushu, Shikoku, and the Inland Sea provinces ruled
mostly by the Taira themselves under Munemori; (2) the capital,
nearby provinces, and most of the provinces on the Sea of Japan from
Wakasa Bay north of Kyoto northeastward nearly as far as the
Mogami River beyond the modern city of Niigata, all held by Yoshinaka; (3) the interior and Pacific seaboard provinces east of the capital to the northern borders of present day Tochigi, Ibaraki, and
Gumma prefectures, ruled by the head of the main Minamoto lineage, Yoritomo; and (4) the domain of theOshu Fujiwara occupying
the upper tip of Honshu north of Yoshinaka's andYoritomo's spheres.
Yoritomo, at his seat in Kamakura, was the head of the main line
of Minamoto warriors, but he was in danger of being relegated to a
peripheral status by Yoshinaka's victories. Yoritomo moved to buttress
his position, taking several steps to gather the support of principal
powers in the capital region. He made friendly gestures toward the
great temples whose prayers were accepted as having contributed to
the success of Minamoto arms; he returned landed holdings in the
east confiscated by the Taira to their rightful temple and noble owners; he followed a policy of leniency in his treatment of captured Taira
warriors; and, perhaps most significantly, he established a secret understanding with Go-Shirakawa that led in the tenth month of 1183
to court recognition of both his civil and his military authority in the
east. (It is a matter of debate among historians, as to whether the
court's recognition of Yoritomo's authority at this point in 1183 was
a strengthening or an attempted invasion of his position.)
64 Gyokuyo, kan 35, in Kokusho kankokai sosho, series 1, vols. 53-55 (Tokyo: Kokusho
kankokai, 1906-7), vol. 2, p. 609, 25/7/Juei 2 (1183).
Cambridge Histories Online © Cambridge University Press, 2008
706
THE RISE OF THE WARRIORS
Following his occupation of the capital, Yoshinaka advanced down
the Inland Sea coast in pursuit of his Taira foes, but with his army
somewhat diminished by the return of warriors to their northern
homes and the remainder weakened by months of hard marches and
fighting, Yoshinaka was roundly defeated by his enemy midway down
the western leg of Honshu at Mizushima Bay (southeast of the modern cities of Okayama and Kurashiki), and he returned to Kyoto.
There, the hostility that his arrogance and his unruly warriors'
plundering and looting had provoked throughout Kyoto society coalesced into a rising against him by armed monks from Enryakuji and
Onjoji, miscellaneous court warriors, nonwarrior ruffians, and city
riffraff, all gathered under the command of a Taira warrior in the service of Go-Shirakawa at the ex-emperor's Hojuji-dono palace and
numbering, it was rumored, twenty thousand men. Yoshinaka's relations with Go-Shirakawa had already been strained by the favoritism
the ex-emperor had shown Yoritomo, but his discovery of secret communication between them, the discovery, too, of an assassin dispatched by Yoritomo to the capital to kill Yoshinaka, and the open
armed defiance of his rule by the ragtag force at the Hojviji-dono were
no doubt final straws in his burden of resentment. Shortly after the
beginning of 1184 he attacked and burned the ex-emperor's palace
and easily routed its inexpert defenders. Confining Go-Shirakawa at
another palace in the city, he also dismissed from court office the exemperor's courtiers and confiscated many of their landholdings. At
the same time, he obtained for himself appointment to the long unused court title of "Barbarian-subduing Generalissimo" (seii taishoguri), which originally had been conferred directly by the emperor on
the leader of an expedition to suppress the Emishi in the northeast.
In Yoshinaka's case, it was directed, of course, against Yoritomo as a
declaration of war.
Yoshinaka's forceful actions at the beginning of 1184 gave him a
nearly complete stranglehold on the Kyoto government, but it also
left him isolated and exposed, with only a small army (perhaps as
few as six or seven thousand men) to confront enemies on three
fronts: the Taira in the west; Yoritomo from the east; and the forces
of the Enryakuji nearby.
The unchastened and ever-scheming Go-Shirakawa issued an
order to Yoritomo calling on him to punish his wild northern cousin,
a task that the Kamakura lord was ready for and probably only too
willing to perform. Rather than lead an army against Yoshinaka himself, however, Yoritomo prudently remained as usual ensconced in
Cambridge Histories Online © Cambridge University Press, 2008
GEMPEI WAR
707
his eastern stronghold, dispatching to the capital instead a punitive
force under the command of his brothers Noriyori and Yoshitsune
(1159-89), the latter a general of peerless, if largely legendary, ability and accomplishment. Yoshinaka, his forces outnumbered and still
suffering from their recent defeat at Mizushima Bay, went down to
defeat in battles at Uji and Seta on the outskirts of Kyoto, and he
took his own life. The victory of the Minamoto brothers resulted in
the consolidation of anti-Taira forces under the single command of
the Kamakura lord. At lastYoritomo turned his attention to fighting
the Taira, and the war entered its final stage.
The Taira had regrouped, following their flight from the capital to
Kyushu, and had established a base on the small island of Yashima
at the northeastern corner of Shikoku (near Takamatsu), from
which, with their maritime superiority, they were able to control the
Inland Sea. Following their success at Mizushima, they advanced toward the capital and constructed fortifications at Ichinotani near
Fukuhara. Less than a month after Yoshinaka's defeat, still in the
early part of 1184, Yoshitsune outflanked the Taira at Ichinotani in a
daring mounted charge down what had been thought to be an impassable cliff-like route behind the camp, routing the surprised Taira
forces with heavy loss of life among the clan's leaders.
In the fall,Yoritomo sent Noriyori's army west from Kyoto down
along the Inland Sea coast, and shortly after the beginning of 1185
it reached the province of Bungo in Kyushu behind the Taira island
base of Hikoshima in the Shimonoseki Strait. Yoshitsune was ordered to resume his assault on the Taira, which he did in a daring sea
attack on their stronghold at Yashima in the second month. Yoshitsune pursued his Taira quarry through the Inland Sea, at the western extremity of which, near the Hikoshima base, the remaining
Taira, effectively bottled up by Noriyoshi's army on the Kyushu
shore, were finally annihilated in the spring of that year (1185) in a
sea battle at Dannoura. Emperor Antoku was drowned at the end of
the battle in the arms of his grandmother, it is said, a little more than
one month past his sixth birthday.
Yoshitsune's brilliant victories brought the war to a sudden conclusion. But these successes, as well as his fame and popularity at
court, fanned Yoritomo's suspicions concerning the loyalty of his half
brother. They had been at odds since the previous year when GoShirakawa, ever the troublemaker, conferred court title and rank on
Yoshitsune in violation of Yoritomo's specific instructions that court
offices were not to be granted to his vassals, or accepted by them,
Cambridge Histories Online © Cambridge University Press, 2008
JO8
THE RISE OF THE WARRIORS
without his authorization. The retired emperor had also granted
Yoshitsune the rare privilege of entry to his private quarters. Reports
continued to reach Yoritomo accusing Yoshitsune of betraying private ambitions. Yoritomo concluded that Yoshitsune andYukiie, who
were reported to be in collusion, should be destroyed. When Yoritomo's attempt to assassinate Yoshitsune in Kyoto failed, Yoshitsune
prevailed on Go-Shirakawa to issue an imperial command to attack
Yoritomo. The retired emperor appointed Yoshitsune steward (Jito)
of Kyushu andYukiie steward of Shikoku to empower them to raise
troops. This plan was clearly hopeless. Yoshitsune went into hiding
and finally made his way with a few companions to the far northeast,
where he found refuge with the Oshii Fujiwara whose leader, Hidehira, had been his protector during his youth. Unhappily for Yoshitsune, however, his patron died toward the end of the year of his arrival in Hiraizumi (1187). Hidehira's successor, Yasuhira, quailing
before Yoritomo's threats, attacked and killed Yoshitsune in the
fourth month of 1189 in a desperate attempt to save himself from the
wrath of the Kamakura overlord. Yasuhira's treachery failed to secure his own safety, however; he fell before the large conquering
army of an outraged and unappeased Yoritomo a few months later.
Yoritomo did not rest until he had eliminated the last military power
that could challenge his hegemony. The Oshu Fujiwara's lands in
Mutsu and Dewa, which Minamoto no Yoshiie had coveted in vain
in the Later Three Years' War more than a century earlier, at last
came into the hands of his great-great-grandson.
Yoritomo was the final victor to emerge from the series of military
contests that, for the first time since the seventh century, decided
political supremacy. Beginning with the Hogen coup in the capital in
1156, initiatives taken by warriors increasingly overwhelmed civil authority, culminating in the great battles of the Gempei War. Society
was fundamentally altered as the warrior gentry of the provinces created a new political structure independent of the imperial court that
supported a role for warriors in the administration of land. The
court became mostly subservient to the will of Yoritomo, and its economic foundations were much eroded in favor of the warriors. The
key relationship in society came to be that between military lord and
his kinsmen and retainers.
The Lord of Kamakura (Kamakura-dono) chose to retain his administrative headquarters (the bakufu) in Kamakura, maintaining
an identity well separated from the machinations of imperial and
noble cliques in Kyoto. Yoritomo seems to have had little need of
Cambridge Histories Online © Cambridge University Press, 2008
GEMPEI WAR
709
court honors and titles, often resigning them soon after they were
conferred. Following the death in 1192 of Retired Emperor GoShirakawa, who opposed the appointment, the emperor gaveYoritomo the title seii taishogun, signifying chief of the warriors, but only
two years later Yoritomo expressed his wish to resign the title, a proposal that was not accepted by the court. Yoritomo's bakufu did not
replace, for the most part, the ancient civil offices of the court. But
he kept under his own control the political and judicial functions he
considered essential for his purposes, building until his death in 1199
the institutions and precedents that regulated the military and landadministration functions of the warrior class.65
65 OnYoritomo's governance of the Kamakura regime, see "The Kamakura bakufu" in Kozo
Yamamura, ed., Cambridge History ofJapan, vol. 3 (1990); and Jeffrey P. Mass, Warrior Government in Early Medieval Japan: A Study of the Kamakura Bakufu, Shugo, and Jim (New
Haven, Conn.: Yale University Press, 1974).
Cambridge Histories Online © Cambridge University Press, 2008
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
Abe Akio P°TSi$fK£. Chuko Nihon bungaku gaisetsu 4 1 * H -fcX&WM. Tokyo:
Shuei shuppan M^tiif^., 1977.
Abe Takeshi fflffiML. "Heiankyo no keizai kozo" ^ M O g S ^ g J g . In Ito
Tasaburo §*iC^Hgp, ed. Seikatsu to shakai keizai £ g t &£&&. Vol. 2 of
Kokumin seikatsu shi kenkyu HR4.fi§£P5E. Tokyo: Yoshikawa kobunkan 'a
Jll&£tg. 1959Abe Takeshi TOS. Heian zenki seijishi ¥%ffM&%3%.. Tokyo: Shinseisha
a , 1974Abe Takeshi WWSi. Nihon shoen shi B # £ [ H £ . Tokyo: Shinseisha
1972.
Abe Takeshi P5fSi5ffi. Oroan wo kunigebumi no kenkyu g3§B8?:£€>5f3u. Tokyo:
Shinseisha ff£&:, 1971.
Abe Takeshi W^ffi. Ritsuryo kokka kaitai katei no kenkyu W^
W^E. Tokyo: Shinseisha f f ^ t t , 1966.
Abe Takeshi TOffi. "Sekkanki ni okeru chozei taikei no kenkyu"
5 W M © I % . In Kodaigaku kyokai ^ft¥ffl,£,
ed. Sekkan jidaishi no
kenkyu fSMB^ft^'OP^. Tokyo: Yoshikawa kobunkan W;il&;£gg, 1965.
Akimoto Kichiro &3f^ffi>, ed. FM^O^I a ± f E . Vol. 2 of Nihon koten bungaku
taikei B#~£!&'%'¥JZ%. Tokyo: Iwanami shoten S K t i J S , 1958.
Akiyama Terukazu fylUibftl. Emakimono !£##>. Vol. 8 of Genshoku Nihon no
bijutsu JKfeB#©§!*. Tokyo: Shogakukan 'h^W,, 1968.
Akiyama Terukazu. Japanese Painting. New York: Rizzoli International Publications, 1977.
Andrews, Allan A. The Teachings Essential for Rebirth: A Study of Genshin's
Ojoyoshu. Tokyo: Sophia University, 1973.
Araki, James T. The Ballad-Drama of Medieval Japan. Berkeley and Los
Angeles: University of California Press, 1964.
Aston, W. G., trans. Nihongi: Chronicles of Japan from the Earliest Times to A.D.
697. 1896. Reprint. London: George Allen and Unwin, 1956.
Azuma kagami IFiMit. Vol. 32 of (Shintei zoho) Kokushi taikei, qv.
Backus, Robert L. The Riverside Counselor's Stories: Vernacular Fiction of Late
Heian Japan. Stanford, Calif.: Stanford University Press, 1985.
Besshu fusensho MW.ffitjfy • Vol 27 of (Shintei zoho) Kokushi taikei, qv.
Boardman, John, Jasper Griffin, and Oswyn Murray, eds. Oxford History of the
7"
Cambridge Histories Online © Cambridge University Press, 2008
712
WORKS CITED
Classical World. Oxford and New York: Oxford University Press, 1986.
Bock, Felicia G. Classical Learning and Taoist Practices in Early Japan: With a
Translation of Books XVI and XX of the Engi-shiki. Tempe: Arizona State
University Center for Asian Studies, Occasional Paper No. 17, 1985.
Bock, Felicia, trans. Engi-shiki: The Procedures of the Engi Era. 2 vols. Tokyo:
Sophia University Press, 1970-72.
Bommokyo $£H18E (Ch. Fan-wang ching). In Taisho shinshu daizokyo JzJE$ft$£JZ
M&. Vol. 24. Tokyo: Taisho issaikyo kankokai ;*dE—iSSfUff^r, 1925.
Borgen, Robert. "The Case of the Plagiaristic Journal: A Curious Passage from
Jojin's Diary." In Aileen Gatten and Anthony Hood Chambers, eds. New
Leaves: Studies and Translations of Japanese Literature in Honor of Edward G.
Seidensticker, pp. 63-88. Ann Arbor: Center for Japanese Studies, University
of Michigan, 1993.
Borgen, Robert. "Oe no Masafusa and the Spirit of Michizane." Monumenta
Nipponica 50 (Autumn, 1995): 357-84.
Borgen, Robert. Sugawara no Michizane and the Early Heian Court. Harvard
East Asian Monographs, 120. Cambridge, Mass.: Council on East Asian
Studies, Harvard University, 1986.
Bowring, Richard. Murasaki Shikibu: Her Diary and Poetic Memoirs. Princeton,
N.J.: Princeton University Press, 1982.
Bowring, Richard. Murasaki Shikibu: The Tale of Genji. Cambridge, England:
Cambridge University Press, 1988.
Brower, Robert Hopkins. "The Konzyaku tnonogatarisyu: An Historical and
Critical Introduction, With Annotated Translations of Seventy-Eight Tales."
Ph.D. diss., University of Michigan, Ann Arbor, 1952.
Brower, Robert H., and Earl Miner. Japanese Court Poetry. Stanford, Calif.:
Stanford University Press, 1961.
Bunka shiirei shu '$CM^W^. In Kojima Noriyuki ' h i f c i i i , ed. Kaifuso. Bunka
shurei shu. Honcho monzui S H I S . ~ScMJ§W9i. ~fcM~XW. Vol. 69 of Nihon koten bungaku taikei RJfc~&$3c¥JZ?fc. Tokyo: Iwanami shoten S&HJ&, 1964.
Cambridge History of Japan. Vol. 1. Cambridge, England: Cambridge University Press, 1993.
Cambridge History of Japan. Vol. 3. Cambridge, England: Cambridge University Press, 1990.
Charlier, Michael. Das Dairi-shiki: Eine Studie zu seiner Entstehung und Wirkung. Wiesbaden: Otto Harrassowitz, 1975.
Choshuki J§:#Cl2. Vol. 7 of Shiryo taisei ^M^fSL. Tokyo: Naigai shoseki kabushiki kaisha rt:?Hlrll$c5£#£t, 1934.
Choshuki ftfcK. 2 vols. In Zoho shiryo taisei tBM£$4^C)S!. Kyoto: Rinsen shoten KsJIIltJE, 1965.
Choya gunsai $li?i¥<ic. Vol. 29.1 of (Shintei zoho) Kokushi taikei, qv.
Chuyuki (or Nakauki) ^ f E . Vols. 8-14 of Shiryo taisei tPiJztfL. Tokyo:
Naigai shoseki kabushiki kaisha
rt^Hiift^^ih,
1934.
Chuyuki t t ^ f E . Vols. 9-15 of Zoho shiryo taisei i^W$.^iZf$C. Kyoto: Rinsen
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
713
shoten Ei£Jll!tJ£, 1965.
Cranston, Edwin A. "Atemiya: A Translation from the Utsubo Monogatari."
Monumenta Nipponica 24, 3 (1969): 289-314.
Cranston, Edwin A. The Izumi Shikibu Diary: A Romance of the Heian Court.
Cambridge, Mass.: Harvard University Press, 1969.
Daianji garan engi narabi ni ruki shizai cho ^^c^fftnlIiiSSkJffilfiStltfil. In Dai
Nihon Bukkyo zensho JZ 0 i1$ifc:£li. Vol. 118. Tokyo: Yuseido shuppambu
*mstB)&a$, 1932.
Daibirushanakyo shiki -X&MMWWJfe^i. In Taisho shinshu daizokyo ioE§\^i\
Mf&, vol. 58. Tokyo: TaishS issaikyo kankokai ±IE—iJUgflJfT^, 1930.
Dai Nihon Bukkyo zensho JZ B #0M&ifeli. 151 vols. Tokyo: Bussho kankokai $>
HfllfT^, 1912-22.
de Bary, Wm. Theodore, ed. The Buddhist Tradition: in China, India and Japan.
New York: The Modern Library, 1969.
Dengyo Daishi zenshu W$ki\$i'£M. 5 vols. Sakamoto: Hieizan tosho kankosho
J t S O J H O T f T ^ . 1926.
Denryaku WM. 5 vols. In Tokyo daigaku shiryo hensanjo
comp. Dai Nihon kokiroku ~k 0 ^ ^ 1 2 ® . Tokyo: Iwanami shoten
1969-70.
Duquenne, Robert. "Daigensui." Hobogirin 6 (1983): 610-40.
Duquenne, Robert. "Daiitoku my5o." Hobogirin 6 (1983): 652-70.
Earhart, Byron. A Religious Study of the Haguro Sect of Shugendo. Tokyo: Sophia
University, 1970.
Eigaku yoki fK^Uffi. In Gunsho ruiju W9W&. Vol. 15. Tokyo: Keizai zasshisha S ^ H I & t t , 1905.
Eizan Gakuin fXlil^K, ed. Ganzan Jie Daishi no kenkyu
Kyoto: Dobosha ISUWtt, 1984.
Ema Tsutomu QHSSg. Yusoku kojitsu ^frfliiftsi. Kyoto: Kawara shoten
1965.
Enchin Y\%. Yuisei MM, included in Yoho hennen zasshu Ife3§gl^#iifc. In Dai
Nihon Bukkyo zensho iz 0 ^•^WL-^.%. Vol. 28. Tokyo: Yuseido shuppambu
^ffiilttiJig^, 1933Engi shiki ^M^,. Vol. 26 of (Shintei zoho) Kokushi taikei, qv.
Ennin H t . Jusan-ju kanjo hiroku -pHfifSlBf&^c, included in Tendai kahyo ^
•&WM, kan 5, part 1. In Dai Nihon Bukkyo zensho -k 0 -&B%L£M.
Vol. 126.
Tokyo: Bussho kankokai BSfflff^, 1914.
Ennin P i t . Nino guhojunrai koki Ag^fe^^Lfffe. In Dai Nihon Bukkyo zensho * B # $ * £ * . Vol. 113. Tokyo: Yuseido shuppambu *ffi^i±i)S^,
1932.
Ennin R C . Nino shingu shogyo mokuroku AM%\~$-1M3k @ ®. In Taisho shinshu
daizokyo JZiE$i1£JzM& Vol. 55. Tokyo: Taisho issaikyo kankokai JziE—ty
g f l j f f ^ , 1928.
Enryaku kotai shiki MM^W^. Vol. 26 of (Shintei zoho) Kokushi taikei, qv.
Farris, Wayne. Heavenly Warriors: The Evolution ofJapan's Military, JOCTIJOO.
Cambridge Histories Online © Cambridge University Press, 2008
714
WORKS CITED
Cambridge, Mass.: Harvard East Asian Monographs, 1992.
Field, Norma. The Splendor of Longing in The Tale of Genji. Princeton, N.J.:
Princeton University Press, 1987.
Flannery, Kent. "The Cultural Evolution of Civilizations." Annual Review of
Ecology and Systemics, 3 (1972), 399-426.
Frank, Bernard. Histoires qui sont maintenant du passe. Paris: Gallimard, 1968.
Frank, Bernard. "Kata-imi et kata-tagae: Etude sur les interdits de direction a
l'epoque Heian." Bulletin de la Maison Franco-Japonaise, Nouvelle Serie,
Tome 5, No. 2-4. Tokyo and Paris: 1958.
Friday, Karl F. Hired Swords: The Rise of Private Warrior Power in Early Japan.
Stanford, Calif.: Stanford University Press, 1992.
Fujioka Kenjiro jUlBHt&gP. Kokufu HJfr. Tokyo: Yoshikawa kobunkan ^Jl|&
-Xm, 1974.
Fujiwara no Morosuke WW-W$. "Kujo Ujosho yuikai" A#:£2<ffljSf$. In
Yamagishi Tokuhei Ul^t&T, Takeuchi Riz6 VifaW=., Ienaga Saburo fCzkH
gp, Osone Shosuke ± t ffi*^, eds. Kodai seiji shakai shiso #ftigtfe
Vol. 8 of Nihon shiso taikei U^$M~fc%>. Tokyo: Iwanami shoten
1979Fujiwara no Nagako. The Emperor Horikawa Diary, translated by Jennefer
Brewster. Honolulu: University of Hawaii Press, 1977.
Fujiwara no Sukefusa WW-'SW. Shunki #12. In Tankaku sosho fHBjftH. Vol. 1.
Tokyo: Kokusho kankokai B O T f ^ , 1912.
Fujiwara no Yasunori den HHtSIlHS. In Kodai seiji shakai shiso 'S'f^jgciintt^S
iffi. Vol. 8 of Nihon shiso taikei B^SSE^C^. 67 vols. Tokyo: Iwanami shoten
S S » J £ , 1979.
_
Fujiwara no Yoshifusa WW.&M et al. Daisozu Kukaiden ^rfifP^Sfi. In Kobo
Daishizenshu ^ffi^Sfiikll, 3rd ed., rev. and enlarged, shukan 1lf#. Koyasan:
Mikkyo bunka kenkyujo ^WcXitW(%Wi, 1967.
Fukunaga Mitsuji H^cjtB], Ueda Masaaki _hffl&Rg, and Ueyama Shumpei ±
|JJ#T, eds. Dokyo to kodai no tennosei i t S i ' ^ ' f t O ^ M W . Tokyo: Tokuma
shoten HFfflH/S, 1978.
Fukunaga Mitsuji MM^t^, Dokyo to Nihon bunka jtfft t B ^ i f b . Kyoto: Jimbun shoin KXWt^., 1982.
Fukuyama Toshio ^§114®!?. Chusei no jinja kenchiku ^WoWQMSk. Tokyo:
Shibundo Hii^CM., 1977.
Funao Yoshimasa ^BSf IE. "Suiko no jittai ni kansuru ichikosatsu: Bitchii no
kuni taizei 0 otaru shibonin cho o chushin to shite" Hj^oHSPi-Kl-f 2>— #
^ t d ^ B ^ S i ^ A ^ t A J i ^ + ' l l - t L"C. Shirin %M 56.5 (September 1973):
74102.
Furushima Toshio 'S'&feJI. Nihon nogyo gijutsu shi B^JtHS^rSt. Vol. 1.
Tokyo: Jichosha B%ffi&, 1947.
Fuso ryakki £fe#|fl8I2. Vol. 12 of (Shintei zoho) Kokushi taikei, qv.
Futaba Kenko ZM'MM. Kodai Bukkyo shiso-shi kenkyu
Kyoto: Nagata bunshodo 7kffl:$:l!^, 1962.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
715
Gempei seisuiki f ^ F ^ a K . In Yuhodo bunko i l J i M i W . Vols. 16-17. Tokyo:
Yuhodo shoten %JiM*J£, 1911-17.
Genshin WM. Ojo yoshu ft£gft. In Taisho shinshu daizokyo ~X'SM^^M&.
Vol. 84. Tokyo: Taish5 issaikyo kankSkai ^ I ^ S M f l J f T ^ , 1931.
Gifli ken ft-*-®, ed. Gifu kenshi K-i-lr*. Tsushihen: kodai MR.M : # t t . Gifu:
Taishu shob5 JZ^MW, 1971.
Gifu ken tfc&m, ed. Gifu kenshi &-$•*£. Tsushihen: chusei MR.M : 41it. Gifu:
Taishu shob5 X^MW, 1969.
Go-Nijo Moromichi ki ^ n ^ f f i a f B . In Dainihon kokiroku JZ 0 *"fi"l5ll. Part 7.
Tokyo: Iwanami shoten ^ K H / S , 1956-58.
Gonki IfilE. Vol. 3 5 of Shiryo taisei ^^f^cfiK. Tokyo: Naigai shoseki kabushiki
kaisha
rt^t-^Stt^^tt.
1934.
Go sannen kassen ekotoba ^H^-a'ife^ff. Vol. 15 of Nihon emaki taisei B ^ ^ ^
^T^c. 27 vols. Tokyo: Chuo koron sha 'P^k-Qlimit, 1977.
Goto Akio ^ ^ B S H , Heiancho kambungaku ronko ¥-<£mMJC¥lm%. Tokyo:
Ofusha fe'Wft, 1981.
Grapard, Allan. "Patriarchs of Heian Buddhism: Kukai and Saicho." In Great
Historical Figures of Japan. Tokyo: Japan Culture Institute, 1978.
Grapard, Allan. "Kukai: Stone Inscription of the sramana Shodo, who Crossed
Mountains and Streams in His Search for Awakening." In The Mountain
Spirit, edited by Michael Tobias and Harold Drasdo, pp. 50-59. New York:
Overlook Press, 1978.
Grapard, Allan. The Protocol of the Gods. Berkeley and Los Angeles: University
of California Press, 1992.
Grapard, Allan. "Japan's Ignored Cultural Revolution: The Separation of
Shinto-Buddhist Divinities and a Case-study: Tonomine." History of Religions 23.3 (1984): 240-65.
Grapard, Allan. "Lotus in the Mountain, Mountain in the Lotus." Monumenta
Nipponica 14.1 (1986): 21-50.
Grapard, Allan. "Enmountained Text, Textualized Mountain: The Lotus Sutra
in the Kunisaki Peninsula." In The Lotus Sutra in Japanese Culture, edited
by George Tanabe and Willa Tanabe, pp. 159-89. Honolulu: University of
Hawaii Press, 1989.
Groner, Paul. Saicho: The Establishment of the Japanese Tendai School. Berkeley,
Calif.: Asian Humanities Press Buddhist Studies Series, 1985.
Gukansho S f %P. Vol. 86 of Nihon koten bungaku taikei B * ^ A i ^ ^ C ^ .
Tokyo: Iwanami shoten SSStlJiS, 1967.
Gyokuyo :SM. 3 vols. Tokyo: Meicho kankokai ^ilffjffax, 1971.
Gyokuyo 3E^. Vols. 53-55 of Kokusho kankokai sosho BUflJfirxsiiilr. Tokyo:
Kokusho kankokai BUfllfT^, 1906.
Gyonen Wkf&. Sangoku Buppo denzu engi H @ $ f f i f g a # g . In Dai Nihon Bukkyo zensho XR^-ffiWL^kM. Vol. 101. Tokyo: Bussho kankQkai flMflJfTtK,
1913.
Hakeda, Yoshito S. Kukai: Major Works. New York: Columbia University
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
jl6
Press, 1972.
Hall, John W. Government and Local Power in Japan, 500 to 1700. Princeton,
N.J.: Princeton University Press, 1966.
Hall, John W. Japan from Prehistory to Modern Times. New York: Delacorte
Press, 1970.
Hamada Takashi ^EBI^. Zuzo 0SL Tokyo: Shibundo ^ £ g , 1969.
Hanawa Hokinoichi ijSftcL— et al., eds. Shinko: Gunsho rujiu 0ft£lfitSi#£. 24
vols. Tokyo: Naigai shoseki F*3^HilI, 1983.
Hanayama Shinsho TEUlfllit, trans. Ojoyoshu £E4.1fli, Iwanami bunko
M- Vols. 2992-96. Tokyo: Iwanami shoten S K S / S , 1942.
Harashima Reiji W-Bi%\Jk. Nihon kodai shakai no kiso kozo
ttji. Tokyo: Miraisha *3f5tt, 1966.
Hashimoto Yoshihiko ffij^jijit. Heian kizoku shakai no kenkyu
#F3S. Tokyo: Yoshikawa kobunkan ^IW'iU'XM, 1976.
Hashimoto Yoshihiko Wi^WB-. "Kizoku seiken no seiji kozo
^g." In Iwanami koza Nihon rekishi g $ ^ f f i 0 * E 5 & . Kodai * f t 4. Tokyo:
Iwanami shoten SSSIUS, 1975.
Hayami Tasuku 3l7Mt. Heian jidai kizoku-shakai to bukkyo ^'k^fiKmM.^c.^ t
iLWL. Tokyo: Yoshikawa kobunkan $ J I | & £ t i , 1975.
Hayashi Rokuro # ^ f 3 , ed. Ronshu: Nihon rekishi liail B *MS&, 15 vols. Tokyo:
Yuseido ^i@C^, 1976.
Hayashi Rokuro #W&1. Jodai seiji shakai no kenkyu _kftftfe&:£; ®W2E. Tokyo:
Yoshikawa kobunkan ^ J l l ^ ^ t l , 1969.
Hayashiya Tatsusaburo # S S H g p . "Insei" K&. Zusetsu Nihon bunkashi taikei
mMU*Xit£.JZ%;.
Vol. 5. Tokyo: Shogakukan ' J ^ t l , 1957.
Hazama Jiko ffiMlt. Tendaishu shigaisetsu Ji.1$m%MWL. Rev. by Okubo Ryojun ±X$k&M. Tokyo: Daizo shuppan ±M&WL, 1969.
Heihanki TIEIE. 5 vols. In Zoho shiryo taisei MM<kPtXl$.. Kyoto: Rinsen
shoten K&JIKrJS, 1965.
Heiji monogatari W-fa^Sm. Vol. 31 of Nihon koten bungaku taikei
%. Tokyo: Iwanami shoten S S ^ / S , 1961.
Heike monogatari ^•M'tidWi. Vols. 32-33 of Nihon koten bungaku taikei B
3t^;*:^. Tokyo: Iwanami shoten S3£tlJ£, 1959-60.
Herail, Francine. "Un lettre a la cour de l'empereur Ichijo: Oe no Masahira
(953-1012)," in Melanges offerts a M. Charles Haguenauer en I'honneur de son
quatre-vingtieme anniversaire. Etudes japonaises. Paris: L'Asiatique, 1980, pp.
369-87.
Herail, Francine. Notes journalieres de Fujiwara no Michinaga, ministre a la cour
de Heian (P9J-1018): Traduction du Mido kanpakuki. Hautes etudes orientales, II, 23. Institut des hautes etudes japonaises. Geneva: Librairie Droz,
1987.
Herail, Francine, Yodo no Tsukai ou le Systeme de Quatre Envoyes. Paris: Presses
Universitaires de France, 1966.
Hirabayashi Moritoku ^Fft^ff. Ryogen fi$L Jimbutsu sosho K%o9M. Vol.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
717
173. Tokyo: Yoshikawa kobunkan ra)\\m±'%f§L, 1976.
Hiraide iseki chosakai ^ F f f i J t ^ H * ^ , ed. Hiraide ¥ffi. Tokyo: Asahi shimbunshalHBfrHtt, 1955.
Hirakawa Minami Tillffi. "Zoto to seii" ^MtWk.
In Hashimoto Yoshihiko
t , ed. Komonjo no kataru Nihon shi ^ i H o H Z > H * ^ . Vol. 2.
. Tokyo: Chikuma shobo jfc^USF, 1991.
Hirano Minoru ¥£??&. JCw/riw *AIH*O J|i£fi|#. Tokyo: Kadokawa shoten
Hj£, 1969.
Hirata Koji ^ H l k — . "Heian jidai no koseki ni tsuite" ^ B t f t O F f f i e o i , - t .
In Toyoda Takeshi kyoju kanreki kinenkai ^ E H S l f i c S S B E ^ ^ , ed. Nihon
kodai chuseishi no chihoteki tenkai B * * f t 4 I 1 t i < O i f e ^ ^ J S W . Tokyo: Yoshikawa kobunkan ^ l l ^ i t l , 1973, PP- 59~96Hisaki Yukio Xft^B.
Nihon kodai gakko no kenkyu B * * f t ^ K € > W ^ .
Tokyo: Tamagawa daigaku shuppambu IS-lWJZ^h&WiM, 1990.
Hochstedler, Carol. The Tale of Nezame: Part Three of Yozva no Nezame Monogatari. Ithaca: Cornell University East Asia Papers 22, 1979.
Hogen monogatari i5kjt%>Wi • Vol. 31 of Nihon koten bungaku taikei B ^ ' S ' A J t ^
'X^h. Tokyo: Iwanami shoten SJ&Hr/S, 1961.
ed. Nihon
Hogetsu Keigo sensei kanreki kinenkai IS.F\^M%^MM^^.^,
shakai keizai shi kenkyu B * t t ^ g ^ r ^ W 5 E . 3 vols. Kodai chusei hen isK^W
%§, Chusei hen tpWM, and Kinsei hen jfiifir^ll. Tokyo: Yoshikawa kobunkan ^
IW&Xm, 1967.
Hojo Hideki it0k^ffi. "Heian zenki chozei kiko no ichikosatsu"
—#^.
In Vol. 3 of Inoue Mitsusada hakushi kanreki kinenkai
MU±'k,
ed. Kodaishi ronso #ft5&l&*. Tokyo: Yoshikawa kobunkan ^)\\^UX1&, 1978.
Hojoki 7?3tf2. Vol. 30 of Nihon koten bungaku taikei B ^ " t f t ^ C ^ : * : ^ . Tokyo:
Iwanami shoten Sz£i§j£, 1957.
Hori Daiji MJzM- "Ryogen no Nijurokkajo kisho seitei no igi" U$i*<0 I t + 7 " ^
ffi&mmi ®m<oMM. Shiso £ & 25 (March 1967): 12-39.
Hori Ichiro iS—fi[S. Kuya ?&-&. Jimbutsu sosho AtydM^. Vol. 106. Tokyo:
Yoshikawa kobunkan ^ J l l ^ i t t l , 1963.
Hori Ichiro. "On the concept of Hijiri(Holy-Man)." Numen 5 (1958): 128-232.
Horyuji garan engi narabi ni ruki shizai cho ffi^^FftnS#©3£MEl2Sftt<!R. In Dai
Nihon Bukkyo zensho JZ.R'$-W>%L£M. Vol. 117. Tokyo: Yuseido shuppambu
**Si:ffiJiSSiS, 1932.
Hoshino Tsune MifM. "Yo no iwayuru Seiwa Genji wa Yozei Genji naru kangae" X£<OlfmittW&KlSil8M&,tsL**2&_.
In Shigaku sosetsu 5&^*tfc.
Tokyo: Fuzambo S l i l S , 1909, pp. 108-47.
Hurst III, G. Cameron. Insei: Abdicated Sovereigns in the Politics of Late Heian
Japan, 1086-1185. New York: Columbia University Press, 1976.
Hyakuren sho HWfa. Vol. 11 of (Shintei zoho) Kokushi taikei, qv.
Ienaga, Saburo. Painting in the Yamato Style. New York and Tokyo: John
Weatherhill and Heibonsha, 1973.
Cambridge Histories Online © Cambridge University Press, 2008
718
WORKS CITED
Iinuma Jiro SxfS—£P. Nihon nogyo gijutsu ron B :£JtHftWlw. Tokyo: Miraisha
* ^ t t , 1971.
Iinuma Jiro gg?B—£P. "Nihonshi ni okeru sviki to kuwa"
(Kyoto Daigaku jimbun kagaku kenkyujo) Jimbun gakuho
Wfeffi) AX3MS 32 (March 1971).
Ikeda Kikan ftbfflflyg, Kishigami Shinji j#±W—, and Akiyama Ken
eds. Makura no soshifcfc^f-,Murasaki Shikibu nikki $?^nl$B!£. Vol. 19 of
Nihon koten bungaku taikei B 'fc~&$;'%.'¥• JZ%>. Tokyo: Iwanami shoten
m, 1958Ima kagami 4"Ht. Vol. 21.2 of (Shintei zoho) Kokushi taikei, qv.
Inagaki Yasuhiko W&M1B-. "Ritsuryoseiteki tochi seido no kaitai"
Vife<vmW. In Vol. 5 of Takeuchi Rizo ttfilH et al., ed. Taikei Nihonshi
sosho # ^ B * £ M t . Tokyo: Yamakawa shuppansha UjJIIttJliStt, 1973.
Inagaki Yasuhiko W&M%- "Shoki myoden no kozo: Yamato no kuni Ota no Inumaru myo ni tsuite" %)ffl%>&l<DMm. • ^cfOH^BB^A^t-o^-t. In Inagaki
Yasuhiko WsM^ and Nagahara Keiji ^<ISK—. Chusei no shakai to keizai 41
ttdtt^ig^r.
Tokyo: Tokyo daigaku shuppankai ^M^^tiifig^, 1962.
Inoue Mitsusada #J:itM. Nihon Jodo-kyo seiritsu-shi no kenkyu
^^>W^E. Tokyo: Yamakawa shuppansha LUJHttlliSli:, 1956.
Inoue Mitsusada # ± * i l . Nihon kodai kokka no kenkyu
Tokyo: Iwanami shoten SSfcUJi!!, 1965.
Inoue Mitsusada #_h3t^l. Nihon kodai no kokka to Bukkyo B ^ ^ " f ^ o B ^ t {$
Wt.. Tokyo: Iwanami shoten S2JHJ2!, 1971.
Inoue Mitsusada #±3fcj| et al., ed. Nihon no rekishi 0 $ O l £ , 26 vols.
Tokyo: Chuo koron sha t t ^ ^ l l t t , 1973-74.
Inoue Mitsusada # ± * ^ , Nagahara Keiji *SCfir., Kodama Kota
eds. Nihon rekishi taikei
^ , and Okubo Toshiaki ±XmM,
Vol. 1. Genshi. Kodai K#i. *f^. Tokyo: Yamakawa shuppansha
1984.
Inoue Mitsusada #_k3£^[ and Osone Shosuke -fcitWiMft, eds. Ojo den. Hokke
genki. ffi£R &^i^IE. Vol. 7 of Nihon shiso taikei B*Siffi^:*. Tokyo:
Iwanami shoten SKtt/iS, 1974.
Inoue Mitsusada hakushi kanreki kinenkai RJiJtMW-zbMWiii&TS, ed.
Kodaishi ronso 'S'ft^iwS. 3 vols. Tokyo: Yoshikawa kobunkan
1978.
Inoue Tatsuo #±M.ffli. Kodai oken to shukyo-teki bemin * R S t t
Tokyo: Kashiwa shobo ffi#M, 1980.
Ishida Mizumaro 5BB^B. Jodo-kyo no tenkai &±|fc©Klfl. Tokyo: Shunjusha
#^Ctt, 1967.
Ishida Mosaku 5EBi^fF. Shakyoyori mitaru Nara-cho Bukkyo no kenkyu ^S<t
>)&tz%$i$kMffiWL<OW$t. Tokyo: Toyo bunko tLWIlM, 1930.
Ishida Shiro 5ffl*6P. "Kyoto bonchi hokubu no senjochi: Heian-kyo sento ji
no Kyoto no chisei"
ffM&^t^<OM^m
Kodai bunka ~&foXit 34.12 (December 1982): 1-14
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
719
Ishii Ryosuke 5 # f t l f t , Nihon hoseishigaisetsu R^-&V\$ZMWL. 2nd. ed. Tokyo:
Sobunsha flJ^Ctt, i960.
Ishii Susumu 5 # j t . "Insei jidai gcMMit." In Rekishigaku kenkyukai and
Nihonshi kenkyukai K £ ^ 5 r 9 £ £ • S # £ S F 3 E £ , eds. Koza Nihonshi MS. B
* £ , 10 Vols. 1. Tokyo: Tokyo daigaku shuppankai ^ M ^ ^ t t i l i K ^ , 1970.
Ishimaru Hiroshi SA^R. "Inseiki chigyokokusei ni tsuite no ichikosatsu" KBc
gS&fffflffelU-o^TcD—#^. In Kurokawa Takaaki *JIISiM and Kitazume
Masao ;|bJRKfe:£, eds. Kamakura seiken WL^WM.. Vol. 4 of Ronshii Nihon
rekishi iwifcB#K£. Tokyo: Yuseido %ffi^ 5 1976.
Ishimoda Sho S ^ B I E , "Kodai ho" * f t f t . In /Codai # f t . Vol. 4 of Ienaga
Saburo ic7kH£P, Ishimoda Sho 5SEBIE, Inoue Kiyoshi #_k?f, Inoue
^^M&B^-M
Mitsusada # ± : ) £ j | , et al., eds. Iwanamikoza: Nihon rekishi
$_. Tokyo: Iwanami shoten S$£(!J£, 1962.
Ishimoda Sho ESBBIE. Kodai makki seijishi josetsu #ft*$3$tfe5&J¥:I&. Tokyo:
Miraisha T^JfStt, 1964.
Ishimura Teikichi 5 W ^ ^ . Yusoku kojitsu kenkyu iffli&^IW^u. 3 vols. Tokyo:
Yusoku kojitsu kenkyu kankokai ^^ScHSf^ETlJff^, 1957Iyanaga Nobumi. "Daijizaiten." Hobogirin 6 (1983): 713-65.
Iyanaga Teizo <&7kjlH. "Ritsuryoseiteki tochi shoyu" W^flJKJittfeHf^. In
Kodai Sit. Vol. 3 of Ienaga Saburo i?acHg|$, Ishimoda Sho 5 S H I E , Inoue
Kiyoshi #_hiW, Inoue Mitsusada #_h^£^ et al., eds. Iwanami koza: Nihon
rekishi Sj&l^ffi B # J S ^ . Tokyo: Iwanami shoten Si&dJS, 1962.
Izumi Shikibu shu M&i^U^k. In Matsushita Daizaburo K T ^ H ^ P , ed.
Kokka taikan Wi^WJtM.. Tokyo: Kadokawa shoten fil)\\tgl£, 1963.
Izumiya Yasuo M§H*. "Choyosei no henshitsu ni tsuite" HJf 1SI©?g:gR
"C. In Kodaigaku kyokai ^ " f t ^ © ^ , ed. Engi Tenryaku jidai no kenkyu S
mmitoWR.
Tokyo: Yoshikawa kobunkan WJH!^i:fi|, i960.
Izumiya Yasuo 7S&JR*. "Zuryo kokushi to nin'yo kokushi" S ^
W|. In Hayashi Rokuro #l^gp, ed. Ronshii: Nihon rekishi I H # : B * K ^ . Vol.
3. Tokyo: Yuseido shuppan %ffi^tlj)iS, 1976.
Jinten ainosho Mfi&WM'®. In Dai Nihon Bukkyo zensho ~k B '%-$<WL£M. Vol.
150. Tokyo: Yuseido shuppambu iffif^ttifiSpC, 1933.
Johnson, Wallace trans. The Tang Code: Volume I. General Principles. Princeton, N.J.: Princeton University Press, 1979.
Joseishi Sogo Kenkyukai ^ r t t ^ i ^ i f ^ , ed. Nihon josei shi 0 * ^ 1 4 * . Vol.
1. Genshi. Kodai W-1&. 'S'ft. Tokyo: Tokyo daigaku shuppankai ^M^C^UjfiS
£ , 1982.
Kagawa Mitsuo J p J I I * * and Fujita Seiichi HfflBf—, eds. Usa ? f e . Tokyo:
Mokujisha ^.mt, 1976.
Kaiei ^.%. Koya shunju hennen shuroku i t j i ^ ^ C i i ^ P l i f . In Dai Nihon Bukkyo zensho ;*; B # 1 $ i f c £ ^ . Vol. 131. Tokyo: Bussho kankokai ^Wfflfir^,
1912.
Kakunyo %!&. Gaijasho afcJR5#. In Shinshu shogyo zensho M^FMWL^M.
Kyoto: Oyagi kQbundo izAif.^r$Ci£, 1964.
Cambridge Histories Online © Cambridge University Press, 2008
Vol. 3.
J2O
WORKS CITED
Kameda Takashi Hffllil^. Nihon kodai yosuishi no kenkyu
3£. Tokyo: Yoshikawa kobunkan ^/IISAiCgl, 1973.
Kamens, Edward. 77ie TAree Jewels: A Study and Translation of Minamoto
Tamenori's Sanboe. Ann Arbor: Center for Japanese Studies, University of
Michigan, 1988.
Katsumata Shunkyo P3£g?$c, ed. Kobo Daishi zenshu ^ f f i ^ f f i ^ ^ . 3 vols.
Tokyo: SankibS OlSM, 1968.
Katsumata Shunkyo ffi~X.&M. Mikkyo no Nihon-teki tenkai ^ifc<OH*#}Jg|lfl.
Tokyo: Shunjusha # # ; t t , 1970.
Katsuno Ryushin PBHitfi. Hieizan to Koyasan Jtf?O41 MW\M, Tokyo: Shibundo M ^ t ^ , 1959.
Katsuno Ryushin MW&im. Sohei f i ^ . Tokyo: Shibundo mjC% 1955.
Katsuyama Seiji Ulililf#;, "Fukosei no saihen to kaitai: ju-juni seiki no fiikosei"
MPMonmtMW—-A-^r-ytfcnmPVlNihonshikenkyu 0 * S i F ^ 194
(October 1978): 1-41.
Katsuyama Seiji UliliW'A, "Koden kammotsu rippo no seiritsu to sono zentei"
^rffl'g'tJ^ftOBStiii^rOlii^. Shirin $M 70 (February 1987): 1-43.
Kawaguchi Hisao Jll P XW. and Nara Shoichi S?&iE—. Godansho chu UM&&..
Tokyo: Benseisha %&Mit, 1984.
Kawaguchi Hisao ill P # J t . Heiancho no kambungaku T ^ ^ O ^ i : ^ . Tokyo:
Yoshikawa kobunkan $JI|&:£g§, 1981.
Kawaguchi Hisao ill n X $ l . Oe MO Masafusa "XiUMM. Tokyo: Yoshikawa k5bunkan WW^Xte,
1968.
Kawasaki Tsuneyuki )\\^MZ.. Jimbutsu Nihon no rekishi
Vol. 3. Ocho no rakujitsu 3 E ^ O ^ B. Tokyo: Yomiuri shimbunsha
1966.
Kawase Kazuma \\\fflr-M. Kojisho no kenkyu * j M © W 2 E . Tokyo: Kodansha
MBit, 1955Keene, Donald. Seeds in the Heart. New York: Henry Holt, 1993.
Keene, Donald. "Taketori monogatari." In J. Thomas Rimer, Modern Japanese
Fiction and Its Tradition. Princeton, N.J.: Princeton University Press, 1978.
Keirstead, Thomas. The Geography of Power in Medieval Japan. Princeton, N.J.:
Princeton University Press, 1992.
Ken'i MM. Kobo Daishi go-den ^ i S ^ S W R In Zoku gunsho ruijii
W&^Wt.
Vol. 8. Tokyo: Keizai zasshisha g$rHi££fc, 1904.
Kikuchi Kyoko HfifaM^P. "Tokoro no seiritsu to tenkai" EffOJ&iEiSIf!. In
Vol. 3 of Hayashi Rokuro #KSP, ed. Ronshu: Nihon rekishi I&4ftB*K£.
Tokyo: Yuseido. shuppan WffiSiijfiS, 1976.
Kiley, Cornelius J. "Estate and Property in Late Heian Japan." In John W. Hall
and Jeffrey P. Mass, eds. Medieval Japan: Essays in Institutional History. New
Haven: Yale University Press, 1975.
Kinda Akihiro ^fflj|:#?. "Heian-ki no Yamato bonchi ni okeru jori jiwari naibu
no tochi riyo" ¥^KS<0*^S%K*>H-2,^aifeflJrt^O±%flJffl. Shirin
61.3 (May 1978): 75-112.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
721
Kinda Akihiro ^ f f l $ & . "Nara Heian-ki no sonraku keitai ni tsuite"
o H I f I C ? l ^ t . Shirin %M 543 (May 1971): 80-108.
Kinoshita Tadashi * T & . "Nogu" ftA. In Wajima Seiichi ft^lS—, ed. Nihon
no kokogaku B * € > # ^ " ^ . Vol. 3. Tokyo: Kawade shobo MttillM, 1966.
Kishi Toshio MM£f!. Nihon kodai seiji shi kenkyu B *#ftigt)££ffl3E. Tokyo:
Hanawa shobo JjliltHF, 1966.
Kishi Toshio ^ ^ ^ . Tojo no seitai UM<O^.M. Vol. 9 of Nihon no kodai B * ©
•£ft. Tokyo: Chuo koronsha 4"*£riaS:, 1987.
Kitanishi Hiromu sensei kanreki kinenkai ^bffi^5fe4.SBIB^:^, ed. Chusei
shakai to ikko ikki ^ t i i & ^ t — ft—gt. Tokyo: Yoshikawa kobunkan
£ t l , 1985Kiuchi Hiroshi ;£(*!:&. Dengyo Daishi no shogai to shiso {x%ckffii<o£M
Regurusu bunko ^ Psi'^'ScM. Vol. 56. Tokyo: Daisanbummeisha
1976.
Kluge, Inge-Lore. Miyoshi Kiyoyuki: sein Leben und seine Zeit. Berlin: Institut
fur Orientforschung, 1958.
Kluge, Inge-Lore. "Fujiwara Morosuke und seine 'Hinterlassene Lehre."
Mitteilungen des Instituts fur Orientforschung 1 (1953); 178-87.
Kobayashi, Hiroko. The Human Comedy of Heian Japan: A Study of the Secular
Stories in the Twelfth-Century Collection of Tales, Konjaku Monogatarishu.
Tokyo: The Centre for East Asian Cultural Studies, 1979.
Kochi Shosuke fflFWfS. "Chokushiden ni tsuite" i&gfflKlo^T. In Vol. 2 of
Tsuchida Naoshige sensei kanreki kinenkai iEBjS^jfc4J^BfS^^, ed.
Nara Heian jidaishi ronshu S.S^F^cBtftS&lm^. Tokyo: Yoshikawa kobunkan
WUItfCte, 1984Kodaigaku kyokai 'HK'^Wi^, ed. Engi Tenryaku jidai no kenkyu
W3E. Tokyo: Yoshikawa kobunkan ^ J l l ^ j t t l , i960.
Kodaigaku kyokai -&R^W,£, ed. Sekkan jidaishi no kenkyu J
Tokyo: Yoshikawa kobunkan ^JH^^Ctl, 1965.
Kodansha Encyclopedia of Japan. 9 vols. Tokyo: Kodansha, 1983.
Koho S S . Toboki ^ S f S . In Zokuzoku gunsho ruiju ffi^&tSWifc. Vol. 12.
Tokyo: Kokusho kankokai S t f f J f f ^ , 1907.
Koizumi Yoshiaki 'hM.3L&. "Todaijiryo Akanabe no sho" Miz^f^M^>J±. In
Tsilshihen: chusei &$_%% • ^W
of Gifu ken Jg-$.!^, ed. Gifu kenshi
Gifu: Taishu shobo ^ t f i l , 1969.
Kojo ^ t ^ . Denjutsu isshinkai tnonfijifi—<t?l$.~%..In Dengyo Daishizenshu
. Vol. 1. Shiga-ken, Shiga-gun, Sakamoto: Hieizan tosho kankosho
, 1926.
Kokan Shiren ^ H S B ^ . Genko Shakusho x ^ f ^ l l . In Dai Nihon Bukkyo zensho
* H # $ » £ » . Vol. 101. Tokyo: Bussho kankokai 1$mWr5£, 1913.
Kokke no on-tame ni shuho sen to kou hyd ^faMWM^fe&M. Included in Henjo
hakki seirei shu J l M H S t t B * , kan 4. In Sango shiiki. Seirei shu ^.Meffi. tt
9M. Vol. 71 of Nihon koten bungaku taikei B * * A 5 : ^ * ^ . Tokyo:
Iwanami shoten StStfJS, 1965.
Cambridge Histories Online © Cambridge University Press, 2008
722
WORKS CITED
Kokushi daijiten I $ ± f f i . 15 vols. Tokyo: Yoshikawa kSbunkan
I979-97Kokushi taikei Illfc*:^. See (Shintei zoho) Kokushi taikei.
Kon Toka 3 ^ * , Kobayashi Takeshi 'hflfllj, and Gorai Shigeru EJfcJI, eds.
Rokuharamitsuji A S I I # . Kyoto: TankSsha $3£St, 1969.
Kondo Yoshirfi jfi^^gp and Okamoto Akio Pl:£Wgp. "Nihon no suits n6ko
gijutsu" B ^O^fSSiiKgffr. In Ishimoda Sho 5SHIE, ed. Kodai no koza 1$
ftiome.. Vol. 3. Tokyo: Gakuseisha #:£&, 1962.
Kongobuji ^E^IJS^F ed. Koyasan senhyakunen shi rtjifllj^fii'^-^!. KSyasan:
Kong5buji &B!I*3F, 1942.
Konishi, Jin'ichi, "The Genesis of the Kokinshu Style." Trans. Helen C.
McCullough. Harvard Journal of Asiatic Studies 38, 1 (June 1978): 61-170.
Konishi, Jin'ichi. /4 History of Japanese Literature. 3 vols. to date. Princeton,
N.J.: Princeton University Press, 1984-91.
Konjaku monogatari shu ^t"W%^%. In Yamada Yoshio UL|BB#S6, Yamada
Tadao \UB3!£M, Yamada Hideo Hlffl3?£8, and Yamada Toshio (Jjffl^^, eds.
Vols. 22-25 of Nihon koten bungaku taikei B^^^^C^iz^.
Tokyo: Iwanami
shoten !G$&9I£, 1959-63.
Kono Fusao ffiWtW^. Heian makki seijishi ^-^%M&.^$.. Tokyo: Tokyodo H
&&, 1979.
Kubota Osamu XtikmU. Chusei Shinto no kenkyu ^Wffl^.<OW$t. Kyoto:
Shinto-shi gakkai # i t £ ^ £ , 1959.
Kudo Yoshiaki HM^M and Nishikawa Shinji ffiJIIff^t. Amidado to Fujiwara
chokoku W'foft&kmW&M. Vol. 6 ofGenshoku Nihon no bijutsu
3S^l5. Tokyo: Shogakukan ' J ^ l l , 1969.
Kukai ~SM. Go-shorai mokuroku 'W&% @ %k. In Taisho shinshu daizokyo
&X&&. Vol. 55. Tokyo: TaishS issaikyo kankokai *IE—%}&W{i£, 1928.
Kukai *SM. Go-yuigofflM%.In Taisho shinshu daizokyo ±JE$i®J:M&. Vol.
JJ. Tokyo: Taisho issaikyo kankokai ^cIE—-gJSflJfir^, 1931.
Kukai ^M. Takao kanjo ki MMBMB, included in Shui zasshu teSHi*. In
Kobo Daishi zenshu 5^ffi±Sfi^:^. 3rd ed., revised and enlarged. Vol. 3.
Koyasan: Mikkyo bunka kenkyujo ^^XWM^fh,
1965.
Kurata Bunsaku 'kWX'ft. Mikkyo jiin to Jogan chokoku mWL^tMMMM.
Vol. 5 ofGenshoku Nihon no bijutsu HRfe 0 :&©§|#r. Tokyo: Shogakukan 'h^
t&, 1967-
Kuroda Toshio HBB^«. "Chinkon no keifu ffi&<O%W In Kuroda Toshio.
t "mW~. Tokyo: Iwanami shoten S
Nihon chusei shakai to shukyo Bft^Wit^
$£•/£, 1990.
Tokyo: Iwanami shoten
Kuroda Toshio Hfflg?8£. Jisha seiryoku ^tt^t).
9fS, 1980.
Kuroda Toshio HBB^tt. 065 to buppo I f f i i ^ S ; . Kyoto: Hozokan
1983.
Kuroda Toshio Sfflg?g£. 5Aoe«ie»' ^a*a»" ffiHflja:^. Vol. 2 of
£ . Tokyo: Nihon hyoronsha B*ffFiiit, 1967.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
723
Kuroda Toshio MES$?#£. "Historical Consciousness and Hon-jaku ;fc2£ Philosophy in the Medieval Period on Mt. Hiei," trans. Allan Grapard. In The
Lotus Sutra in Japanese Culture. Edited by George Tanabe and Willa Tanabe, pp. 143-58. Honolulu: University of Hawaii Press, 1989.
Kuroda Toshio HffliSlSt. "Shinto in the History of Japanese Religions." Journal of Japanese Studies 7-1 (1981): 1-21.
Kuroita Katsumi j§#i§l§t. Kotei: kokushi no kenkyu jgfT : H^<o5f55. 3 vols.
Tokyo: Iwanami shoten 5ai£f}J2j, 1931-36.
Kuroita Nobuo WS-WJz. "Fujiwara no Tadahira seiken ni taisuru ichikosatsu"
MW.&¥WmzM-tZ>—3i%k. In Kodaigaku kyokai # f t ^ & £ , ed. Engi Tenryakujidai no kenkyu 3£S^J§B$ft<OW3E. Tokyo: Yoshikawa kobunkan ^jll
%3cm, i960.
Kushida Ryoko SSEB &&. Kukai no kenkyu ^ S © I 5 L . Tokyo: Sankibo Busshorin [UMWB9W, 1981.
Kyohan gifii. Daishi on-gyojo shuki ^Sfi^ffttUfB. In Zoku gunsho ruiju W&M
m&. Vol. 8. Tokyo: Keizai zasshisha g8f3fIH±, 1904.
Kyokai jftjR, comp. Nihon ryoiki B ^ f i ^ l S . Edited by Endo Yoshimoto j i H K
S and Kasuga Kazuo # B fPH. Vol. 70 of Nihon koten bungaku taikei B $^T
$k-%^W&. Tokyo: Iwanami shoten £ $ » J £ , 1967.
Kyoto-shi SSPTfJ, comp. Kyoto no rekishi ~^.M<OM$.. Vol. 1. Tokyo: Gakugei
shorin ^ S ^ # , ' 1971.
Lammers, Wayne P. "The Succession (Kuniyuzuri): A Translation from Utsubo
Monogatari." Monumenta Nipponica 37, 2 (1982): 139-78.
Lammers, Wayne P. The Tale of Matsura: Fujiwara Teika's Experiment in Fiction. Ann Arbor: Center for Japanese Studies, University of Michigan, 1992.
Lu, David John. Sources of Japanese History. New York: McGraw Hill Book Co.
1974Manabe Shunsho Mitikikffk. "Seishuku bijutsu—hoshi mandara Jl?ii§t#r—£#
^ S . " Kobijutsu 35 (1971): 1-48.
Man'yoshu TiW:%. Vols. 4-7 of Nihon koten bungaku taikei B '^•^^'X^-'X^h.
Tokyo: Iwanami shoten S R U / S , 1957-62.
Marra, Michele. Review of// Dio incatenato: Honcho shinsenden di Oe no Masafusa, trans. Silvio Calzolari. Monumenta Nipponica 41 (Winter 1986): 49597Mass, Jeffrey P. Warrior Government in Early Medieval Japan: A Study of the
Kamakura Bakufu, Shugo, andjito. New Haven, Conn.: Yale University Press,
1974Matsumura Hiroji ^MW-^, ed. Okagami ;*;£S. Vol. 21 of Nihon koten bungaku
taikei B '&~&#$:3c¥JZ%. Tokyo: Iwanami shoten S R t r J £ , 1960.
Matsumura Hiroji &tiW-^ and Yamanaka Yutaka lil11)3^, eds. Eiga monogatari
9£WfcM. Vols. 75-76 of Nihon koten bungaku taikei B # * f t : £ 3 M ; ^ .
Tokyo: Iwanami shoten S2£1tJ2i, 1964-65.
McCullough, Helen Craig, trans. "An Account of My Hut." In Helen Craig
McCullough, comp. and ed. Classical Japanese Prose: An Anthology. Stanford,
Cambridge Histories Online © Cambridge University Press, 2008
724
WORKS CITED
Calif.: Stanford University Press, 1990.
McCullough, Helen Craig. Brocade by Night: 'Kokin Wakashu' and the Court
Style in Japanese Classical Poetry. Stanford, Calif.: Stanford University Press,
1985.
McCullough, Helen Craig, trans. Classical Japanese Prose: An Anthology. Stanford, Calif.: Stanford University Press, 1990.
McCullough, Helen Craig, trans. Genji and Heike: Selections from The Tale of
Genji and The Tale of the Heike. Stanford, Calif: Stanford University Press,
1994.
McCullough, Helen Craig, trans. Kokin Wakashu: With 'Tosa Nikki'and 'Shinsen Waka.' Stanford, Calif.: Stanford University Press, 1985.
McCullough, Helen Craig, trans. Okagami, The Great Mirror: Fujiwara Michinaga and His Times. Princeton, N.J.: Princeton University Press, 1980.
McCullough, Helen Craig, trans. The Tale of the Heike. Stanford, Calif.: Stanford University Press, 1988.
McCullough, Helen Craig, trans. "A Tale of Mutsu." Harvard Journal of Asiatic Studies 25 (1964-65): 178-211.
McCullough, Helen Craig, trans. Tales of Ise: Lyrical Episodes from Tenth-Century Japan. Stanford, Calif: Stanford University Press, 1968.
McCullough, William H., and Helen Craig McCullough. A Tale of Flowering
Fortunes: Annals of Japanese Aristocratic Life in the Heian Period. 2 vols. Stanford, Calif.: Stanford University Press, 1980.
McCullough, William. "Japanese Marriage Institutions in the Heian Period."
Harvard Journal of Asiatic Studies, 27(1967): 103-67.
Miho Tadao E i f e ' S ^ and Miho Satoko Hft-f h :f-. Unshu orai Kyoroku bon
kenkyu to sosakuin S W t t ^ ¥ ^ * W ^ E i ^ 5 ^ ? l . Osaka: Izumi shoin H t t K ,
1982.
Minamoto, H. An Illustrated History of Japanese Art. Kyoto: K. Hoshino, 1935.
Minamoto no Tamenori W&W.. Kuchizusami P j ^ . In Vol. 32.1 oiZoku Gunsho
ruiju mmmmfe. Tokyo: Keizai zasshi sha. i&ffiMUtt, i960.
Minamoto no Tamenoriffl.J&M.Sambo ekotoba E g ^ P l . In Dai Nihon Bukkyo
zensho ± B # < $ * : £ # . Vol. i n . Tokyo: Bussho kankokai $Wf(lfT£, 1913.
Tokyo: SSbunsha flj
Misaki Ryoshu ^m&M. Taimitsu no kenkyu is^^Wc^.
:£fit, 1988.
Misaki Ryoshu =*f &J1. Mikkyo to jingi shiso ^ # £ *WS;ffl. Tokyo: Sobunsha I K f l : , 1992.
Misaki Ryoshu =M$fkM. Nihon • Chugoku Bukkyo shiso to sono tenkai B ^ • •t'H
UMBMtZVfeffi. Tokyo: Sankibo OlSS, 1993.
Misaki Ryoshu =M&M. Taimitsu no riron to jissen ^^(DMMst^^.
Tokyo:
Sobunsha flJlfctt, 1994.
Miura Keiichi Hffl^—. "10 seiki-13 seiki no Higashi Ajia to Nihon iotSr|fi-i3
Vol. 2. Tokyo: Tokyo
1fl:iiS£>^7i>7 t B * . "Koza Nihonshi M&B^$..
daigaku shuppankai ^JfCT^tfjJiSIs, 1970.
Miyagawa Mitsuru 1S)\\M. Oyama sonshi ^ l l l f t •£. 2 vols., Hommon hen
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
725
and Shiryo hen %_PrW*. Tokyo: Hanawa shobo *HitM, 1964.
tftK.
MiyaharaTakeo "UW&lz. Nihon kodainokokka to nomin B^^iK<oSM
Tokyo: Hosei daigaku shuppankyoku ftBc^^HttliS/l}, 1973.
Miyai Yoshio 'U^rWiiWt. Fujiwara-shi no ujigami-ujidera shinko to sobyo saishi M
HftOftflft^MSMiffiSi&JE. Tokyo: Seiko shobo J S W M , 1978.
Miyai Yoshio H # S * § . Jodai no shimbutsu shugo to Jodo-kyo ± f t © WAI?£-1
& ± t t Tokyo: Seiko shobo j £ ¥ » S , 1980.
Miyamoto Tasuku ^ * | S c . "Ritsuryosei-teki tochi seido" ^#Jtfj±Jfe$JlK. In
Takeuchi Rizo frfilH et al., eds. Taikei Nihonshi sosho
ft^B#5&MI.
Vol. 6. Tokyo: Yamakawa shuppansha UUJIIttJKSth, 1973.
Miyoshi Tameyasu =M%s>Wk. Go-shui ojo den f&J'niiiifiE^.fK. In Inoue Mitsusada
# ± * * and Osone Shosuke ^ : # f f i ¥ ^ , eds. Ojo den. Hokkegenki tt£R ffiH
i^lS. Vol. 7 of Nihon shiso taikei H * S i ® ± ^ . Tokyo: Iwanami shoten
• ^ , 1974_
In Inoue Mitsusada
Miyoshi Tameyasu =.W%sM- Shut ojo den iailtt^ft.
* ^ and Osone Sh5suke ± # © ^ ^ > , eds. 0;5 den. Hokkegenki tt£g. &
IB. Vol. 7 of Nihon shiso taikei 0 * S i B ^ : ^ . Tokyo: Iwanami shoten
)£, 1974.
Momo Hiroyuki W&ft. Jodai gakusei no kenkyu ±ft^$J©W3E. Tokyo: Meguro shoten gSflll^!, 1947.
Montoku jitsuroku ~$CMiM.£k. Vol. 3 of (Shintei zoho) Kokushi taikei, qv.
Mori Katsumi S ^ 3 . Nisso boeki no kenkyu. BSRg^Off^S. Rev. ed. Vol. 1 of
Mori Katsumi chosaku senshu S^LfillfFJlllS. Tokyo: Kokusho kankokai B ^
Wtt&, 1975^^^PffiU^^
Morita Tei ^ E B ^ , "Heian chuki gunji ni tsuite no ichikosatsu"
o ^ t © — # ^ . In Vol. 3 of Hayashi Rokuro W&$0>, ed. Ronshu: Nihon
rekishi l m * B # K £ . Tokyo: Yuseido shuppan *ffi^tB)S, 1976.
Morita Tei Mm®, Heian jidai seijishi kenkyu ^Bttt&feS&ffl^E. Tokyo:
Yoshikawa kobunkan $JII3Z>:£tI, 1978.
Morita Tei ^EB®, Nihon kodai ritsuryoho shi no kenkyu B ^ ' S ' f t ^ ^ p f t ^ o P 1
3E. Tokyo: Bunken shuppan ^CS^ttifig, 1986.
Morris, Dana. "Peasant Economy in Early Japan, 650-950." Ph.D. diss.,
University of California, Berkeley, 1980.
Morris, Ivan, trans. As I Crossed a Bridge of Dreams: Recollections of a Woman in
Eleventh-Century Japan. New York: Dial Press, 1971.
Morris, Ivan, trans. The Pillow Book ofSei Shonagon. 2 vols. New York: Columbia University Press, 1967.
Morris, Ivan. The Tale of Genji Scroll. Tokyo: Kodansha International, 1971.
Morris, Ivan. The World of the Shining Prince. New York: Alfred A. Knopf,
1964.
Murai Yasuhiko tt#B&. Heian kizoku no sekai ¥ ^ f t K o U t l ^ . Tokyo:
Tokuma shoten W.fe\9JS, 1972.
Murai Yasuhiko tt#JS?t. Kodai kokka kaitai katei no kenkyu
L. Tokyo: Iwanami shoten S K l l J S , 1965.
Cambridge Histories Online © Cambridge University Press, 2008
J26
WORKS CITED
Murai Yasuhiko tt#flt^. "Shoensei no hatten to kozo" £H$J©3SJg t Slig. In
Kodai isit, Vol. 4 of Ienaga Saburo f?7k=£|$, Ishimoda She SSBBJE, Inoue
Kiyoshi #-h?!t, Inoue Mitsusada #_L3fc^l et al., eds. Iwanami koza: Nihon
rekishi istiS.MMB&M'JS.. Tokyo: Iwanami shoten S$£l!/£, 1962.
Murao Jiro I W I i W . Ritsuryo zaiseishi no kenkyu ft^ftfi&StOSf^L. Tokyo:
Yoshikawa kobunkan ^IW^UXW,, 1967.
Murayama Shuichi ftlH^-. "Heian Bukkyo no tenkai" ¥^B^<om^.
In
Ienaga Saburo ^acHgp, ed. Kodai hen 'S'ftM. Vol. 1 of Nihon Bukkyo-shi 0
*#ifc5&. Kyoto: Hozokan feiStl, 1967.
Murayama Shuichi tflU^—. Heian-kyo ¥ ^ J i U Tokyo: Shibundo '£.'%% 1957.
Murayama Shuichi fctLUI^-". Honji suijaku ^i&IlM:. Tokyo: Yoshikawa kobunkan -^jwm&.w,, 1974.
Murayama Shuichi Wlllift—. MAow ommyddo-shi sosetsu
B^^^M$.I^M.
Tokyo: Hanawa shobo ^HtB?, 1981.
Murayama Shuichi falllift—. Shimbutsu shugo shicho W A i ^ K i i l . Kyoto:
Heirakuji shoten ^ F ^ ^ t t / S , 1957.
Murayama Shuichi fallltf—. Yamabushi no rekishi |ilt*t<£>Bi#I, Tokyo: Hanawa
shobo f$MW, 1970.
Murayama Shuichi Wlilft— and Kageyama Haruki ^lil#ffl. Hiei-zan ittilil.
Tokyo: NHK Bukkusu NHK 7 y ^ ^., 1970.
Murdoch, James. ^4 History of Japan. Vol. 1. 1910. Reprint. London: Routledge
and Kegan Paul, 1949.
Vol. 8 of
Mutsu waki HUfglB. In Kodai seiji shakai shiso ^iWifhit^lSM.
Nihon shiso taikei B^fSM'X^.
67 vols. Tokyo: Iwanami shoten
1979Nachi Taisha W®i3t, ed. Kutnano Sanzan to sono shinko tiEpHULj t
3 vols. Tokyo: Meicho shuppan ^SiftBHK, 1984.
Nagahara Keiji TKI%M~. "Shoensei ni okeru 'shiki' no seikaku" ffli
rUsJ Ottt&. In ^oiaj chusei hen fttt&mfc. Vol. 2 of Hogetsu Keigo sensei
kanreki kinenkai 'g.B 3zi!lrfc£.S.Ma£ik.£, ed. Nihon shakai keizai shi kenkyu
H * i t # g ^ * W ^ . Tokyo: Yoshikawa kobunkan £JII&;£ffi, 1967.
Nagahara Keiji 7KKJI— et al., eds. Taikei Nihon no rekishi
ft^a$©H.
15
vols. Tokyo: Shogakukan ' J ^ t l , 1987-89.
Nagahara Keiji, with Kozo Yamamura. "Village Communities and Daimyo Power." In John Whitney Hall and Toyoda Takeshi, eds. Japan in the Muromachi
Age. Berkeley and Los Angeles: University of California press, 1977.
Nagayama Yasutaka ft|ilS#. Ritsuryo futan taikei no kenkyu
ft^Affift^oW
%. Tokyo: Hanawa shobo *iii1IrJi, 1976.
Nagazumi Yasuaki TKWB.^ and Shimada Isao JlsEBilfcS, eds. Kokonchomonju ~$5
4 - # M S . Vol. 84 of Nihon koten bungaku taikei B ^f-1S^C^iz^h.
Tokyo:
Iwanami shoten 3§S£(I/3J, 1966.
Nakada Kaoni ^ISM. Shoen no kenkyu /±H<OW^E. Tokyo: Shok5 shoin # # H
%t, 1949Nakamura, Kyoko Motomichi. Miraculous Stories from the Japanese Buddhist
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
727
Tradition: The Nihon Ryoiki of the Monk Kyokai. Cambridge, Mass.: Harvard
University Press, 1972.
Nakano Hatayoshi ffffSHi. Hachiman shinko-shi no kenkyu A<tfl#^OSf^E.
2 vols. Tokyo: Yoshikawa kobunkan ^i\\%.~£.W,, 1975.
Nakano Hatayoshi 4"IH8Mi!i. Usa-gu ^feUa. Tokyo: Yoshikawa kobunkan ^Jll
&;£li, 1985.
Nakano Hideo #WS£Ji. Ritsuryosei shakai kaitai katei no kenkyu
#MM<®W9L.
Tokyo: Hanawa shobo tfaWK, 1979.
Nakao Shumpaku f+B^W. Dengyo Daishi Saicho no kenkyu i s f ^
Kyoto: Nagata bunshodo a c E B i e ^ , 1987.
Needham, Joseph, ed. Science and Civilization in China. Vol. 2. History of Scientific Thought. Cambridge, England: Cambridge University Press, 1962.
Nihon kiryaku 0^I2SS. Vols. 10-11 of (Shintei zoho) Kokushi taikei), qv.
Nihon koki 0 ^ ^ : f c . Vol. 3 of (Shintei zoho) Kokushi taikei, qv.
Nihon Montoku Tenno jitsuroku 0 ^ j t ^ . ^ i i l H I S . Vol. 3 of (Shintei zoho)
Kokushi taikei, qv.
Nihon sandai jitsuroku 0 ^ H f t H H . Vol. 4 of (Shintei zoho) Kokushi taikei, qv.
Nihon shoki B^tflfi. Sakamoto Tar5 K # * £ P , Ienaga Saburo icjkHgp, Inoue
Mitsusada #Jb^fe^, and Ono Susumu JzWW, eds. Vol. 6y of Nihon koten bungaku taikei 0^•^$k'X% 1 i\ : &. Tokyo: Iwanami shoten S K t r / S , 1967.
Nihonshi kenkyukai B ^ ^ W ^ E ^ , ed. Chuseino kenryoku to minshu
t&M. Osaka: Sogensha f y x t t , 1970.
Nihonshi kenkyukai B * ^ W ^ E ^ , ed. Chusei shakai no kihon kozo
*ffija. Tokyo: Ochanomizu shobo fflI^O*^M, 1958.
Ninchu tl!&. Eizan Daishi den fJlll^cSlig. In Dengyo Daishi zenshu
£M. Vol. 5. Shiga-ken, Shiga-gun, Sakamoto: Hieizan tosho kankosho
mmWnBf, 1926.
Nishibeppu Motoka i5#U/ff7C 0 . "Kyu seiki chuyS ni okeru kokusei kicho no
tenkan ni tsuite" XW^.'PM^^'&^SWl.MM'O^^kK.^y.^X.
Nihonshi
kenkyu 0 ^ ^ W ^ E 169 (September 1976): 30-54.
Nishimoto Yutaka H ^ S . Sumiyoshi taisha &l£jzit. Tokyo: Gakuseisha
tt, 1977Nyiiya Tetsuichi fl-^S©—. "Zaichi tone no keisei to rekishi-teki ichi"
M<OMf&tm%.W)&S.. In Osaka rekishi gakkai * K E ^ ^ # , ed. Chusei
shakai no seiritsu to tenkai *¥¥£&£ CXfeiiSilW. Tokyo: Yoshikawa kobunkan ^iW^UXW,, 1976.
Oe no Masafusa ^iLSW. "Bonen no ki" S^FlH. In Yamagishi Tokuhei lU&W>
¥ , Takeuchi Rizo ttfllH, Ienaga Saburo ^ * H g p , Osone Shosuke ~k^U
m-ft, eds. Kodai seiji shakai shiso iSrftBcfett^iglL Vol. 8 of Nihon shiso
taikei H ^ S i S ^ C ^ . Tokyo: Iwanami shoten S & t t / £ , 1979.
Oe no Masafusa -kUMW. Godansho UM&. In Gunsho ruiju S¥tf W&. Vol. 17.
Tokyo: Keizai zasshisha SSISIiS?t, 1905.
Oe no Masafusa izUMW. "Zoku honcho ojoden" S t * ^ t t 4 ^ . In Inoue
Mitsusada #_h3t^| and Osone Shosuke JziaWkMft, eds. Ojoden. Hokkegenki
Cambridge Histories Online © Cambridge University Press, 2008
728
WORKS CITED
mmmii. Vol. 7 of Nihon shiso taikei U*!SM*.%. Tokyo: Iwanami
shoten ?e$.1iJ£, 1974Ogi Mitsuo $JtJ*5fc. Nikon kodai ongaku shiron B ^ f t ^ =&£!&. Tokyo:
Yoshikawa kobunkan 'a JUSttflt, 1977.
Ohase Keikichi ' h U S S T a . Nihonkoku genzai shomokuroku kaisetsu ko 0 ^ @ ^ ,
&9BMfflM$Si. Tokyo: Komiyama shuppan, 'M^LUttiliK, 1976.
Oka Kazuo PI—!%, ed. Heiancho bungaku jiten T ^ U K ^ ^ f t . Tokyo: Toky6d5
lit,
1972-
Okada, H. Richard. Figures of Resistance: Language, Poetry, and Narrating in
'The Tale of Genji' and Other Mid-Heian Texts. Durham: Duke University
Press, 1991.
Okamoto Akio flSI^Wgp. "N6gyo seisan" ftH£g. In Kondo Yoshiro ifi:J§t£P
and Fujisawa Choji WR&fti, eds. Nihon no kokogaku B # ® # # # . Vol. 5.
Tokyo: Kawade shobo MIH9W, 1966.
Okudaira, Hideo. Narrative Picture Scrolls. New York and Tokyo: Weatherhill
and Shibundo, 1973.
Okuno Nakahiko H S F ^ 1 ^ . "Hachi kyii seiki ni okeru shiteki tochi shoyusei no
rekishiteki seikaku" A%1&fcii£totfZ&tft±tel%ji'®\V>m£#}&fa.
Nihon
rekishi B^ft£$L 279 (August 1971): 53-69.
Okuno Nakahiko H B ^ ^ . "Rinji zoyaku no seiritsu to tenkai" &?%WSi?)tf£±L
iJSIH. Nihon rekishi U^Wt'fc. 255 (August 1969): 32-49.
Okuno Takahiro HEPiSJK. "Go-Sanjo tenno goryo" ^ H ^ ^ M f f l l ® . In Vol. 4
of Nihon rekishi daijiten 0#B££;*c¥W. 11 vols. Tokyo: Kawade shobo ffltti
mm, 1970Omi no Mifune Genkai g^iSH^S7cW. To Daiwajo tosei den 0 * f P ±^ t i E f e . In
Dai Nihon Bukkyo zensho ~k B if-®%k£M. Vol. 113. Yuseido shuppambu 'ft'S
#m«gifi5,1932.
Omori Kingoro ^C^S^SgP. Buke jidai no kenkyu K^B#f^©W^. Tokyo:
Fuzambo SOiM, 1923
Omori Kingor6 -XM&£.$$. "Ocho jidai no shigakko ni tsuite"
&HZ.Wt\<*X. Rekishi chiri E £ % S 53 (1929): 330-37.
Ono Katsutoshi ' h f F i ^ . Nitto guho junrai koki no kenkyu ASf
W^E. 4 vols. Tokyo: Suzuki gakujutsu zaidan I ^ ^ ^ ^ B t H , 1964-69.
Ono Tatsunosuke izWM^M. Jodai no Jodo-kyo ±ft<Djf ±%k. Tokyo: Yoshikawa kobunkan ^ill5^XI§, 1972.
Osabe Kazuo ftpflfPffi. Ichigyo zenji no kenkyu — ffPffiOiF^E. Kobe: Kobe
shoka daigaku WPMPCX^,
1963.
Osaka rekishi gakkai izWM.^P-'^, ed. Chusei shakai no seiritsu to tenkai 't'iSrtt
£<Df&iLtmffl. Tokyo: Yoshikawa kobunkan ^ i l l ^ ^ t l , 1976.
Osaka rekishi gakkai ^ E H ^ ^ ^ , ed. Ritsuryo kokka no kiso kozo W^WB-<o3$s
_ S l ^ a i . Tokyo: Yoshikawa kobunkan ^)U&$.g&, i960.
Ota Seiroku ^BB^T^. Shindenzukuri no kenkyu 3Si£jg€>W9E. Tokyo: Yoshikawa kobunkan ^ J l l & ^ l i , 1987.
Oyama KSjun ± | i | & ^ . Shimbutsu kosho-shi W{^$^S&. Kyoto: Rinsen shoten
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
729
£, 1975Ozawa Masao 'h^RIE^, ed. Kokin waka shu ^ 4 " ^ ^ ^ . Vol. 7 ofNihon koten
bungaku zenshu B # # * : £ ^ : £ : ^ § . Tokyo: Shogakukan ' h ^ t g , 1971.
Ozawa Masao 'h^RIE*, Goto Shigeo t&$§J|gP, Shimazu Tadao g ^ , & * , and
Higuchi Yoshimaro H n ^ f ^ S . FM^MTO SO$AI chushaku W^-$,&M. Tokyo:
Hanawa shobo, ijS^M, 1974.
Paine, Robert Treat, and Alexander Soper. The Art and Architecture of Japan.
Pelican Historv of Art. Baltimore: Penguin Books, 1955.
Pearson, Birger A., ed. Religious Syncretism in Antiquity. Missoula: Scholars
Press, 1975.
Pekarik, Andrew, ed. Ukifune: Love in the Tale of Genji. New York: Columbia
University Press, 1982.
Piggott, Joan. "Chieftain Pairs and Co-rulers: Female Sovereignty in Early
Japan." In Hitomi Tonomura et al., eds. Women and Class in Japanese History. Ann Arbor: Center for Japanese Studies, University of Michigan, 1999.
Pollack, David. The Fracture of Meaning. Princeton, N.J.: Princeton University
Press, 1986.
Ponsonby-Fane, R. A. B. Kyoto: The Old Capital of Japan, 794-1869. Kyoto:
The Ponsonby Memorial Society, 1956.
Rabinovitch, Judith N., trans. Shomonki: The Story of Masakado's Rebellion.
Monumenta Nipponica Monograph 58. Tokyo: Sophia University, 1986.
Reischauer, A. K. "Genshin's Ojo Yoshu: Collected Essays on Birth into Paradise." Transactions of the Asiatic Society of Japan. 2nd series. 7(1930): 16-97.
Reischauer, Edwin O. Ennin's Diary: The Record of a Pilgrimage to China in
Search of the Law. New York: Ronald Press, 1955.
Reischauer, Edwin O. Ennin's Travels in T'ang China. New York: Ronald Press,
1955Reischauer, Edwin O., and Joseph K. Yamagiwa, trans. Translations from Early Japanese Literature. Cambridge, Mass.: Harvard University Press, 1951.
Reischauer, Robert Karl. Early Japanese History (c. 40 B.C.-A.D. 1167). 2 vols.
Princeton, N.J.: Princeton University Press, 1937.
Renondeau, Gaston. LesMoines Guerriers du Japan. Paris: Imprimerie Nationale,
1965.
Renondeau, Gaston. Le Shugendo: Histoire, Doctrines et Rites des Anachoretes
Dits Yamabushi. Paris: Imprimerie Nationale, 1965.
Rheinstein, Max. Max Weber on Law in Economy and Society. Cambridge,
Mass.: Harvard University Press, 1954.
Ritsu. Ryo no gige W. ^Wtffl. Vol. 22 of (Shintei zoho) Kokushi taikei, qv.
Rohlich, Thomas H. trans. A Tale of Eleventh-Century Japan: Hamamatsu Chunagon monogatari. Princeton, N.J.: Princeton University Press, 1983.
Rotermund, Hartmut O. Die Yamabushi: Aspekte Ihres Glaubens, Lebens und
Ihrer Sozialen Funktion im Japanischen Mittelalter. Hamburg: Kommissionsverlag Cram, De Gruyter & Co., 1968.
Ruiju kokushi S S S £ . Vols. 5-6 of {Shintei zoho) Kokushi taikei, qv.
Cambridge Histories Online © Cambridge University Press, 2008
730
WORKS CITED
Ruiju sandai kyaku. W&&=^fiMs. Vol. 25 of (Shintei zoho) Kokushi taikei, qv.
Ryogen &ffl,. Nijurokkajd kisho H+^iffi^:gst. In Takeuchi Rizo frfilE ed.,
Heian ibun ¥ ^ 3 1 ^ . Vol. 2. Tokyo: Tokyo-do JSM% 1964.
Ryogen fkW- Tengen sannen Chudo kuyo gammon ^ x H ^ + S ^ f i t . In Koji
ruien # ^ S ? S . Vol. 12. Tokyo: Yoshikawa kobunkan ^IW^UXW,, 1969.
Ryojin hisho %£M%&$. Go-Shirakawa ?£6M, comp. In Vol. 19 of Nihon koten
zensho B * " £ " l l i K . Tokyo: Asahi shimbunsha ^ B f f M t t , 1953.
Ryo no shuge ^^M. Vols. 23-24 of (Shintei zoho) Kokushi taikei, qv.
Saeki Arikiyo feteHit. Dengyo Daishi den no kenkyufc%ki;ffiifc<D$tfZ.Tokyo:
Yoshikawa kobunkan "a)W-jL~3cf$, 1992.
Saeki Umetomo feftSS, ed. Kokin waka shit ~&$-%mkM. Vol. 8 of Nihon koten
bungaku taikei B^-^^X^iz^.
Tokyo: Iwanami shoten Sz&trJS, 1959.
Saicho MM. Dengyo Daishi shorai Esshu roku gttt&IiifJt^'Hgl. In Taisho
shinshii daizokyo -kJE%\1fe-kM&. Vol. 55. Tokyo: Taisho issaikyo kankokai iz
jE-$jmWa£,
1928.
Saicho ft®. Dengyo Daishi shorai Taishu roku fcifc^&Iiif Jfz&HM. In Taisho
shinshii daizokyo ~XiEM^'XM&. Vol. 55. Tokyo: Taisho issaikyo kankokai
^lE-iJO&PJfr^, 1928.
Saicho ftS. Kansho Tendai-shu nembun gakusho shiki W)¥kz£.a'miE-/fr¥-&£<, included in Sange gakusho shiki l l l S ^ ^ . ^ . In Taisho shinshii daizokyo JZlEM^
±W&. Vol. 74. Tokyo: Taisho issaikyo kankokai *IE—^JgfJtT^, 1931.
Saicho ftg. Kenkairon m$M. In Dengyo Daishi zenshu ^Wi±U^M. Vol. 1.
Shiga-ken, Shiga-gun, Sakamoto: Hieizan tosho kankosho
1926.
Saicho S @ . Kenkairon engi P ^ l m & S . In Dengyo Daishi zenshu
Vol. 1. Shiga-ken, Shiga-gun, Sakamoto: Hieizan tosho kankosho
Wrrffi, 1926.
Saicho ft@. Naisho Buppo sojo kechimyaku fu rtlEffffifflSjfilMlg• In Dengyo
Daishi zenshu fiifc^SR^ft. Vol. 1. Shiga-ken, Shiga-gun, Sakamoto: Hieizan
tosho kankosho SXM'MmWuTft, 1926.
Saicho ft@. Tendai Hokke-shu nembun tokudo gakusho myocho z%.i3$i1^'m^-/fr%
ffi^^^gitfi, included in Jo Kenkairon hyo ± P ^ ^ S . In Dengyo Daishi
zenshu &.Wiiz%fo£.%. Vol. 1. Shiga-ken, Shiga-gun, Sakamoto: Hieizan tosho
kankosho itWLWim^W-nfft, 1926.
Saito Tsutomu U H I i . Ocho-jidai no ommyodo If8R$ft©JiH§it. Tokyo: Sogensha 10x3:, 1947.
Sakamoto Shozo WL%%=~ Nihon ocho kokka taisei ron
Tokyo: Tokyo daigaku shuppankai ^ M ^ ^ t B f i S ^ , 1972.
Sakamoto Taro S * * S P . Nihon kodai shi no kisoteki kenkyu
W3E. 2 vols. Tokyo: Tokyo daigaku shuppankai ^Jj^^tBKfi^, 1964.
Sakamoto Taro S^^cSP. Sugawara no Michizane H I S J I J U . Tokyo: Yoshikawa
kobunkan ^IWmjzM, 1962.
Sakamoto Taro. The Six National Histories of Japan. Trans. John S. Brownlee.
Vancouver: University of British Columbia Press, 1991.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
Sakamoto Taro hakushi kanreki kinenkai S**fiPtf ±MMB&£,
731
ed. Nihon
kodaishi ronsu 0 &~£it%.fmM. 3 vols. Tokyo: Yoshikawa kSbunkan ~£)\\ijL~%
t£, 1963Sakanoue Akikane S ± W # et al., comps. Hossd shiyosho &W i i i l l ^ . In Vol. 4
of Hanawa Hokinoichi ^ f t c i — et al., eds. Shinko: Gunsho ruiju SftftSWSft:.
Tokyo: Naigai shoseki FWUflf, 1983.
Sammon dosha ki UUHS#f2. In Gunsho ruiju ffi^Wt. Vol. 15. Tokyo: Keizai
zasshisha g ^ J t t S i t , 1905.
Sandai jitsuroku jEf^flli. In Vol. 4 of (Shintei zoho) Kokushi taikei, qv.
Sansom, George. A History ofJapan to 1334. Stanford, Calif.: Stanford University Press, 1958.
Sansom, George. A Short Cultural History of Japan. Rev. ed. New York:
Appleton-Century, 1943.
Sasaki Keisuke i$E* JF-Mft. "Dazaifu no kannai shihai henshitsu ni kansuru shiron: omo ni zaisei-teki sokumen kara" %'¥ffi<Dfg]ft^ffiM'S.^M'f%U,Wi
:
. In Vol. 2 of Tsuchida Naoshige sensei kanreki kinenkai
&£> ed. Nara Heian jidaishi ronshu
Tokyo: Yoshikawa kSbunkan ^ i l l & i l g , 1984.
Sasayama Haruo ffili|B#4.. (Kodai0 kangaeru) Heian no miyako
'gcoU. Tokyo: Yoshikawa kSbunkan ^Jll^^tfil, 1991.
Sato, Elizabeth. "Oyama Estate and Insei Land Policies." Monumenta Nipponica
34 (Spring 1979): 73-99.
Sato Shin'ichi &MM~. Nihon no chuseikokka B ^CO'PWS^. Tokyo: Iwanami
shoten %$.Wif£, 1983.
Sato Sojun feB^W. Heian zenki seijishi josetsu ^ f j # 3 B t f e £ f f i £ . Tokyo:
Tokyo daigaku shuppankai ^ M ^ ^ t t i K S ^ , 1977Schafer, Edward H. The Golden Peaches of Samarkand: A Study of T'ang Exotics.
Berkeley and Los Angeles: University of California Press, 1963.
Schafer, Edward H. The Vermilion Bird: T'ang Images of the South. Berkeley and
Los Angeles: University of California Press, 1987.
Scharschmidt, Clemens. "Unshu shosoku oder die Briefsammlung des Unshu."
Ostasiatische Studien 20 (1917): 20-114.
Schipper, Christofer. "The Taoist Body." History of Religions 17 (1978): 35581.
Seidel, Anna. "La Divinisation de Lao-tseu dans le Taoisme des Han." Publications de l'Ecole Franchise d'Extreme-Orient. Vol. 71. Paris: Ecole Franjaise
d'Extreme-Orient, 1969.
Seidensticker, Edward, trans. The Gossamer Years: The Diary of a Noblewoman
of Heian Japan. Tokyo and Rutland, Vermont: Charle E. Tuttle Co., 1964.
Seidensticker, Edward. "Murasaki Shikibu and Her Diary and Her Other Writings." Literature East and West 18.1 (March 1974): 1-7.
Seidensticker, Edward G. The Tale ofGenji. 2 vols. New York: Alfred A. Knopf,
1976.
Seiji yoryaku iWUBS. Vol. 28 of (Shintei zoho) Kokushi taikei, qv.
Cambridge Histories Online © Cambridge University Press, 2008
732
WORKS CITED
Shibata Minoru ^ B a H . Chusei shomin shinko no kenkyu
Tokyo: Kadokawa shoten fii)\\ltl£, 1966.
Shibuya Jigai W&M& ed. Kotei zoho Tendai zasu ki $!Ttf ffi
Tokyo: Daiichi shobo 1^—HM, 1973.
Shigematsu Akihisa Wfc?PiX. Hachiman Usa-gu gotakusen-shu
M. Tokyo: Gendai shichosha mft&ffltt, 1986.
Shiko ]£,%. Jimon denki horoku ^HfilBffi^. In Dai Nihon Bukkyo zensho ~k 0
* # i ! c £ « . Vol. 127. Tokyo: Bussho kankokai ffiSflJfT^, 1915.
Shimada Jiro Irffl&gp. "Shoen-teki 'shiki' taisei no kaitai" £ H W RgiJ WUVffl
# . In Takeuchi Rizo firftSH et al., eds. Taikei Nihonshi sosho folh H * ^ S
^ . Vol. 6. Tokyo: Yamakawa shuppansha (JLjJlI HbiKSth, 1973.
Shimaji Daito %i&k^. Tendai kyogaku shi JiSW&'fc.. Tokyo: Meiji shoin,
1929.
Shimizu, Osamu. "Nihon Montoku Tenno jitsuroku: An Annotated Translation, with a Survey of the Early Ninth Century in Japan," Ph.D. diss.,
Columbia University, 1951.
Shimode Sekiyo TttiSfl. Dokyo to Nihonjin jitfSc t B * A . Tokyo: Kodansha M
Witt, 1975Shimode Sekiyo TffiS-^. "Kukai to Shingon-shu" ^ S i X W ^ ? . In Nihon hen
B ^ i § . Vol. 2 of Nakamura Hajime 4"Wx, Kasahara Kazuo $lM-~!B, and
Kanaoka Shuyu ^M%~K, eds. Ajia Bukkyo shi 7 i>7$,%.%_. Tokyo: Kosei
shuppansha ^lEttiliSS:, 1974.
Shimode Sekiyo TtttSH. Nihon kodai no jingi to Dokyo B ^ ' S ' f t O f W t itfi:.
Tokyo: Yoshikawa kobunkan ^JII^A^ttl, 1972.
Shimode Sekiyo T t t J S # . "Shingon-shu no tenkai" Ka'^ORIfl. In Nihon hen
B ^-M. Vol. 2 of Nakamura Hajime ^¥iit, Kasahara Kazuo 3£)J^—^, and
Kanaoka Shuyu &MJ§~%L, eds. Ajia Bukkyo shi 7 i>T$fc5t. Tokyo: Kosei
shuppansha 6^IEttS)SS:, 1974.
Shimpen Kokka taikan ffilBifc;*:!!. Shimpen Kokka taikan henshu iinkai ffS
H S t t H i m i m ^ , ed. 10 vols. in 20. Tokyo: Kadokawa shoten PlIUWfE,
1983-92.
Shinoda, Minoru, trans. The Founding of the Kamakura Shogunate 1180-1185.
New York: Columbia University Press, 1960.
(Shintei zoho) Kokushi taikei (fffTlS1i)@^±^. 60 vols. in 66. Tokyo: Yoshikawa kobunkan ^ J l l ^ ^ C t i , 1929-64.
Shinto taikei hensankai
ftjt^^S^^,
ed. Shinto taikei ffjt;*;^. Vol. 22.
Tokyo: Seikosha ffi^ft, 1984.
Shioiri Ryodo MA&M. and Kiuchi Gyoo * F * 5 ^ * , eds. Saicho M.W. Vol. 2 of
Nihon meiso ronshu B$;g<il&S. Tokyo: Yoshikawa kobunkan ^ J l l & i f i l ,
1982.
Shirane, Haruo. The Bridge of Dreams: A Poetics of 'The Tale of Genji.' Stanford,
Calif.: Stanford University Press, 1987.
Shochureki S'+'BI. In Zoku gunsho ruiju W^-MW.^.. Vol. 32. Tokyo: Zoku gunsho ruiju kanseikai Wz%tlitM'i£.7Elf$.£, 1923-30.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
733
Shoku Nihongi Sc B ^IS. Vol. 2 of (Shintei zoho) Kokushi taikei, qv.
Shoku Nihon koki Wl B #!£&. In Vol. 3 of (Shintei zoho) Kokushi taikei, qv.
Shomonki #P1f2. In Kodai seiji shakai shiso iift&fetkg&te.
Vol. 8 of Nihon
shiso taikei B&JSM'f:^. 67 vols. Tokyo: Iwanami shoten gj&Ujg, 1979.
Shoyuki ' J ^ I 2 . Part 10 of Dai Nihon kokiroku :£ B #"£"f£3§. Compiled by
Tokyo daigaku shiryo hensanjo Jt£giJ:¥£MM&fli. Tokyo: Iwanami shoten
Si&W£, 1959-69.
Sjoberg, Gideon. The Preindustrial City, Past and Present. Glencoe, 111.: Free
Press, i960.
Sockel, Dietrich. Emakimono: The Art of the Japanese Hand-Painted Scroll.
London: Jonathan Cape, 1959.
Sofu sen'yokai &M,tM<k ed. Rankei yuionshu Writs'M. In Kobo Daishi
zenshu ^fe^cWik^, 3rd ed., rev. and enlarged. Vol. 5. Koyasan: MikkyS
bunka kenkyujo ^WcXitWRTft, 1966.
Sonoda Koyu ®BB»M. "Heian Bukkyo" ^ F ^ i i * . In Kodai 1sR. Vol. 4 of
Ienaga SaburS gtTKHgp, Ishimoda She 5SHIE, Inoue Kiyoshi # ± S et al.,
eds. Iwanami koza: Nihon rekishi ^stikMlSL B *fcM$.. Tokyo: Iwanami shoten
£ « » £ , 1962.
Sonoda Koyu MBBSI4. "Heian Bukkyo no seiritsu" :¥!£$-%L<Dj8.\L. In Ienaga
SaburS g?*=gp, ed. Kodai hen # f t B . Vol. 1 of Nihon Bukkyo-shi 0 ^ffi$L
5&. Kyoto: Hozokan &MM, 1967.
Sonoda Koyu WHl^M. "Saicho to Tendai-shu" W&tJUlm. In Nihon hen B
^ S . Vol. 2 of Nakamura Hajime ^ W x , Kasahara Kazuo 5£Jf!—fj, and
Kanaoka Shuyii ^ K ^ S , eds. Ajia Bukkyo shi T i>T$%.$_. Tokyo: Kosei
shuppansha i&.JEttt¥$i$t, 1974Sonoda Koyu Sfflffli. "Suiko: Tempyo kara Engi made" ffi^ : Wf-frhl&M
4 "C. In Osaka rekishi gakkai ^ c K E ^ ^ ^ , ed. Ritsuryo kokka no kiso kozo W
^SlcOfflUgjg. Tokyo: Yoshikawa kobunkan ^Jll&ifil, i960.
Sontsu ^ S . CAu/10 Daishi nempu §"|E^:gli^lt. In Da» Nihon Bukkyo zensho JZ
R*BWL£M.
Vol. 28. Tokyo: Yuseido shuppambu *ffigffi)ig3l$, 1933.
Staal, Frits. "Indian Concepts of the Body." Somatics (Autumn/Winter 198384): 31-41.
Sugawara no Michizane HJi^jitJC, comp. Kanke bunso 1WW.'$CM-. In Vol. 72 of
Nihon koten bungaku taikei B ^is^X^-W^Tokyo: Iwanami shoten
fS, 1966.
Sumi Toyo M&%FM. Zenkindai Nihon kazoku no kozo gfaifift B *
Tokyo: Kobund5 ^ X ^ , 1983.
Suzuki Shigeo 1 ^ ^ ^ ^ . "Inseiki in-no-cho no kino ni tsuite: in-no-ch6 hakkyu
monjo o tsujite mitaru" &WLMWr<nWtffe^^^-^fT%i1&-$CM-$:M CX%
tz%. Unpublished graduation thesis. Tokyo University, 1961.
Suzuki Tomotar5 ^^^O*gP, Kawaguchi Hisao IWoXM, Endo Yoshimoto j t
H H S , and Nisbishita Kyoichi \S~T%?1—, eds. Tosa nikki ±fe B IE, Kagero
nikki frtfh& B12, Izumi Shikibu nikki ftiM&ffl B12, Sarashina nikki W.B. B
12. Vol. 20 of Nihon koten bungaku taikei B ^ ^ T ^ ^ ^ ^ C ^ . Tokyo: Iwanami
Cambridge Histories Online © Cambridge University Press, 2008
734
WORKS CITED
shoten S S t r / S i , 1957.
Taiki cniS. 3 vols. in Zoho shiryo taisei ifMS&W^ClS. Kyoto: Rinsen shoten
9fS, 1965Tajima Isao ffli|i&. "Bokkai to no kosho" ffl):Bt<D&&. In Hashimoto Yoshihiko Ws^WiHsZ, ed. Komonjo no kataru Nihon shi ~&~$CM<OWk:& B ^ £ . Vol. 2.
Heian W£. Tokyo: Chikuma shobo yiM%W, 1991.
Takada Minoru SBBH. "Chusei shoki no kokka kenryoku to sonraku" 't'ttlflSfl
(DWWfcTltVl'fe. Shicho tM 99 (June 1967): 6-25.
Takada Osamu iSBaft and Yanagisawa Taka MK^. Butsuga {A®. Vol. 7 of Genshoku Nihon no bijutsu JSfeB^'OUffii. Tokyo: Shogakukan ' h ^ t f , 1969.
Takagi Yutaka ig/tdft. "Tendai-shu no tenkai." ^ - ^ © I I K . In Nihon hen B
^ H . Vol. 2 of Nakamura Hajime • t ' t t x , Kasahara Kazuo %iM—¥i, and
Kanaoka Shuyu &W0§1>£., eds. /4;Va Bukkyd shi 7 i>7f$|Sc£. Tokyo: Kosei
shuppansha fiSIEtfcDSit, 1974.
Takahashi Takashi M^^. Emishi: kodai tohokujin no rekishi WM '•
© I S . Chuko shinsho 4 ' ^ f f * 804. Tokyo: Chiio koronsha
1986.
Takahashi Tomvo M^MM. Kodai Ezo: sonoshakaikozo'^^M'%
Tokyo: Gakuseisha ^ £ . £ t , 1974.
Takahashi Toshinori Sffi^SI. Nihon kyoiku bunka shi H^SW^C'fbS. 1933.
Reprint. Tokyo: Kodansha HiSit;, 1978.
Takakusu Junjiro iSifSJII&gP, Watanabe Kaigyoku J&&M1&, and Ono Gemmyo
'bW%W>, eds. Tai'sAo jAimAfi daizokyo ^ c I E f f t f ^ ^ f i . 100 vols. Tokyo:
Taisho issaikyo kank5kai ^TlE—iSHflf?^, 1924-34.
Takeuchi Riz6 ttfil=. SMS/II «O ro/o S£±€>Jg*. Vol. 6 of Nihon no rekishi 0
* < O l J . 31 vols. Tokyo: Chuo kSronsha 4 " * & I ^ t t , 1973.
Takeuchi Rizo ttfiiH, ed. H«fln i'6«w ¥ $ j t ^ : . 15 vols. Tokyo: Tekyodo 3H
M f , 1947-80.
Takeuchi Rizo ftftSH. "Insei no seiritsu" KftJOfiEiL In KOAM * f t . Vol. 4
of Ienaga Saburo ^jKH6P, Ishimoda Sho 5SEBIE, Inoue Kiyoshi
Inoue Mitsusada # ± ^ t ^ et al., eds. Iwanami koza: Nihon rekishi ^ i
fcffit$5_. Tokyo: Iwanami shoten SiSlIJS, 1962.
Takeuchi Rizo ttfllH, ed. Kamakura ibun M'tMX. 41 vols. Tokyo: TSkyodo
KL&.'SL, 1971-90.
Takeuchi Rizo ttf^IH, ed. Nara ibun ^-^'AX. Revised ed. 3 vols. Tokyo:
TokyodS HO&'g, 1962.
Takeuchi Rizo ttrtlH. Nara-cho jidai ni okerujiin keizai no kenkyu ^jSIBBvfft
tcfctf-S^Kg^rOW^. Tokyo: Ookayama shoten iz^lUtfS,
1932.
Takeuchi Rizo 'ftrtSSH. Ritsuryosei to kizoku seiken W^VitfiM&M.
2 vols.
Tokyo: Ochanomizu shobo P ^ O T K H J S , 1957, 1958.
Takeuchi Riz6 ttl^IH, et al., eds. Taikei Nihonshi sosho
ft^B#£«*.
Tokyo: Yamakawa shuppansha UjJIItBUSSh, 1973.
Takeuchi Rizo tt^IH, ed. Tochi seido shi ±*&$ljg£. Vol. 6 of Taikei Nihon
shi sosho W^hB^-^MM. Tokyo: Yamakawa shuppansha t-UJIItBfiSifc, 1964.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
735
Takeuchi Rizo ttrtJIH. "TSgoku no 'tsuwamono' " 3CB© Po;b>fc> ©J. Miura
kobunka H i f ^ - f t 25.5 (i979)Takikawa Masajiro ^JIIBt&gP. Nihon hoseishi B # f e $ J £ . Tokyo: Yuhikaku %
Igffl, i960.
Tamura Koyu ffltt^fe. Saicho MM. Jimbutsu sosho AVOi&tH. Vol. 193. Tokyo:
Yoshikawa kobunkan ^ J l l & i t t , 1988.
Tamuramaro denki fflH^BfiJIE. In Hanawa Hokinoichi lj|fSS—, comp.
(Shinko) Gunsho ruiju (ffft) £ W « & . Vol. 3. Tokyo: Meicho fukyukai £ * #
* £ , 1977Tanabe, George, and Willa Tanabe, eds. The Lotus Sutra in Japanese Culture.
Honolulu: University of Hawaii Press, 1989.
Tanabe, Willa. "The Lotus Lectures: Hokke hakko in the Heian Period."
Monumenta Nipponica 39.4 (1984): 393-407.
Taniguchi Kengo © P I n . "Mino no kuni Oi no sho ni okeru shSkan ichizoku
no ryoshusei: Onakatomi no geshishiki s5ron o megutte" 3ii&
^]±^-m<O^^.Vi
: ^ ^ E o T ^ l i S ^ ^ ^ * <"o-t. Hosei shigaku
28 1976): 24-37.
Tanimori Tomoo S ^ ^ ^ . Kebiishi o chushin to shitaru Heian jidai no keisatsu
jstm #&£&&**'£*t
LtzZ>¥$;ii%R<DWm'tf.W>. Tokyo: Kashiwa shobo ffi
9W, 1921.
Toby, Ronald P. "Why Leave Nara? Kammu and the Transfer of the Capital."
Monumenta Nipponica 40 (1985): 331-47.
Toda Yoshimi P f f l ^ H . Chuyuki ^ ^ E . Tokyo: Soshiete -t L x . t , 1979.
Toda Yoshimi PfflTfU. "Kokuga gunsei no keisei katei"
In Nihonshi kenkyukai B ^ ^ W 5 S ^ , ed. Chusei no kenryoku to tninshu
ffi^JiK*.
Osaka: Sogensha |ij7cit, 1970.
Toda Yoshimi P"EH5j§t. Nihon ryoshusei seiritsushi no kenkyu
<O^%. Tokyo: Iwanami shoten teffiMJU, 1967.
Todaijiyoroku ^ ^ X P H H . In Zokuzokugunsho ruiju %}t*W9W&. Tokyo: Gunsho ruiju kankokai ffiiHW&Wil^, 1909.
Toji choja bunin K ^ S ^ f f i t t . In Zokuzoku gunsho ruiju $z*MltW&. Vol. 2.
Tokyo: Kokusho kankokai M i f l f T ^ , 1907.
Tokoro Isao B?5?J. Miyoshi Kiyoyuki H # f f fj. Tokyo: Yoshikawa kobunkan ^
j n ^ ^ i i , 1970.
Tokoro Isao Eft5ft. "Ritsuryo jidai ni okeru ikenfushin seido no jittai" # ^ B # f t
KL*HrfZMMMMMmom&. In Kodaigaku Ky5kai * f t ^ » ^ , ed. Engi Tenryaku jidai no kenkyu 3£S^BB#ft<OW2E. Tokyo: Yoshikawa kobunkan £JII
StXffi, i960.
comp. Dai Nihon shiryo -fc
Tokyo daigaku shiryo hensanjo ^Miz^^PrWiW^:,
4. Series 3, Revised Edition. Tokyo: Tokyo daigaku shuppankai ^ M
, 1968.
Tokyo National Museum. Painting 6th-14th Centuries. Vol. 1 of Pageant of
Japanese Art. Tokyo: Toto shuppan, 1957.
Torao Toshiya 0 E f t ^ . Engishiki ftMH. Tokyo: Yoshikawa kobunkan
Cambridge Histories Online © Cambridge University Press, 2008
736
WORKS CITED
, 1964.
Torao Toshiya ^ B < £ # t Handen shuj'u ho no kenkyu 5Sffli&Sffi©5f2E. Tokyo:
Yoshikawa ksbunkan l^JI I SZJtti!, 1961.
Toyoda Takeshi JiEB5£, ed. Sangyo shi i t i l ! £ . Vol. 10 of Taikei nihon shi sosho
ft^B#£S*.
Tokyo: Yamakawa shuppansha lilJIIffifigtt, 1964.
Toyoda Takeshi jgfflS;. Tohoku no rekishi ]g.lt<DM%.. Vol. 1. Tokyo: Yoshikawa kobunkan ^) ll&Xlf, 1967.
Toyoda Takeshi kyoju kanreki kinenkai SfflJ§;t!cit«!JtffB;&#, ed. Nihon kodai
chusei no chihoteki tenkai B ^ - S f t + I S ^ ^ ^ ^ J S I f l . Tokyo: Yoshikawa kobunkan ^ J l l & i f f i , 1973.
Tsubo Noritada mWM- Koshin shinko H ^ ® # . Tokyo: Yamakawa Shuppansha
UlJIIffiliga:, 1956.
Tsubo Noritada l i l S t i . Koshin shinko no kenkyu M$tSW®W&. Tokyo: Nihon
gakujutsu shinkokai B ^^ffJPJSl^, 1961.
Tsuchida Naoshige iffljfi®. Ocho no kizoku i f f l O * K . Vol. 5 of Nihon no
rekishi S * o K . Tokyo: Chu5 koronsha <^^&mit, 1981.
Tsuchida Naoshige sensei kanreki kinenkai iEBB^^fe^fejaSIH,^^, ed. Nara
Heian jidaishi no kenkyu i ^ ^ F ^ B t f ^ O U F ^ S . 2 vols. Tokyo: Yoshikawa
kebunkan ^)W^ldCf&, 1984.
Tsuji Zennosuke i ±# £ f t . Nihon Bukkyo-shi: Josei hen B # # t W : ±Mm.
Tokyo: Iwanami shoten SiSH/S, 1944.
Tyler, Royall. The Miracles of the Kasuga Deity. Princeton, N.J.: Princeton
University Press, 1990.
Uda Tenno f ^ ^ .
"Kampyogoyuikai" ^'MMM. In Yamagishi Tokuhei I±J
F, Takeuchi Riz5 ^l^ffiH, Ienaga Saburo ^ * H g | 5 , Osone Shosuke Jz
, eds. Kodai seiji shakai shiso # t t J g t f e f t £ K l L Vol. 8 of Nihon shiso
taikei B^Siffi;*:^. Tokyo: Iwanami shoten 3EJZ&W£, 1979.
Ueda Masaaki iffllEBg, ed. Sumiyoshi to Munakata no kami &.i§ t?RWOffl.
Tokyo: Chikuma shobS vL&9W, 1988.
Ueda Takeshi _hEBffi. Bokkaikoku no nazo ^liSgOSI. KSdansha gendai shinsho WrWdLWSSfi^, 1104. Tokyo: Kodansha WMIXL, 1992.
Uraki, Ziro. The Tale 0} the Cavern. Tokyo: Shinozaki shorin, 1984.
Ury, Marian, "The Oe Conversations." Monumenta Nipponica 48 (Autumn
1993): 359-8o.
Ury, Marian. Tales of Times Now Past: Sixty-Two Stories from a Medieval
Japanese Collection. Berkeley and Los Angeles: University of California Press,
Utsuho monogatari ^ j ^ $:%!!. Vols. 10-12 of Nikon koten bungaku taikei
^X^JZlfc.
Tokyo: Iwanami shoten SJ&fllS', 1959-62.
Uwayokote Masataka ±M¥-W&. Gempei no seisui W^-O^M. Vol. 6 of Nihon
no rekishi bunko B^'OM^.^C.B. Tokyo: Kodansha MWitt, 1975Verschuer, Charlotte von. Les Relations Officielles du Japon avec la Chine aux
Vllle et IXe Siecles. Paris: Librairie Droz, 1985.
Wakamori Taro fP®:S*fiP, Murayama Shuichi PtlUfe—, Gorai Shigeru
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
737
Tokawa Yasuakira FJII3c$, and Sakurai Tokutaro &'#$g*£P, eds. Sangaku
shukyo-shi kenkyu sosho lUI&$KlSc£ffl3u%1l. 18 vols. Tokyo: Meicho shuppan
*¥HJJK, 1977Wakamori Taro WifcgS^gp. Shugendo-shi kenkyu £ifcit£W3u. Tokyo: Toyo
bunko ^ f f ^ m i , 1972.
Wakita Haruko KfflRii 1 . Nihon chusei shogyo hattatsushi no kenkyu B ^'t'ttffl
H H 3 i o W 5 E . Tokyo: Ochanomizu shobo'M^<DyK9W,1969.
Waley, Arthur. The Tale of Genji: A Novel in Six Parts. 2 vols. Boston and New
York: Houghton Mifflin Co., 1935.
Wamyo ruiju sho &%mM®. 3 vols. Tokyo: Kazama shobo MfiH^W, 1954Washio Junkyo SgBIHit Koshitsu to Bukkyo H S £ $ i f c . Tokyo: Daito shuppansha ^c^ttifiSS:, 1939.
Watanabe Minoru W.WM. Nihon shokuseikatsu shi B *A^fSS&. Tokyo: Yoshikawa kobunkan ~n)\\<jk'XW,, 1964.
Watanabe Shoko WMW& and Miyasaka Yusho 'SWL'gWf, eds. Sango-shiiki.
Seireishu ^-Wm%. *&MM. Vol. 71 of Nihon koten bungaku taikei
~Xl¥J\3fc. Tokyo: Iwanami shoten S & t f / S , 1965.
Watanabe Shoko WJDMIZ. and Miyasaka Yushe "gWgW. Shamon Kukai
B. Chikuma sosho # [ # M . Vol. 84. Tokyo: Chikuma shobo VlMtm, 1967.
Watson, Burton, trans. "Chiteiki" (Record of the Pond Pavilion). In Japanese
Literature in Chinese. Vol. 1. New York: Columbia University Press, 1975.
Watson, Burton, trans. Japanese Literature in Chinese: Volume I: Poetry and
Prose in Chinese by Japanese Writers of the Early Period. New York: Columbia
University Press, 1975.
Watson, Burton, trans. Meng ch'iu: Famous Episodes from Chinese History and
Legend. Tokyo: Kodansha International, 1986.
Weinstein, Stanley. "Beginnings of Esoteric Buddhism in Japan: The Neglected
Tendai Tradition." Journal of Asian Studies 34. 1 (1974): 177-91.
Weinstein, Stanley, "Imperial Patronage in T'ang Buddhism." In Arthur F.
Wright and Denis Twitchett., eds. Perspectives on the T'ang. New Haven:
Yale University Press, 1973.
Wheatley, Paul, and Thomas See. From Court to Capital: A Tentative Analysis of
the Japanese Urban Tradition. Chicago: University of Chicago Press, 1978.
Whitehouse, Wilfred, and Eizo Yanagisawa. Ochikubo Monogatari or The Tale of
the Lady Ochikubo: A Tenth Century Japanese Novel. Tokyo: Hokuseido
Press. Rev. Ed. 1965.
Willig, Rosette E. The Changelings: A Classical Japanese Court Tale. Stanford,
Calif.: Stanford University Press, 1983.
Wilson, William Ritchie. "Hogen monogatari": Tale of the Disorder in Hogen.
Tokyo: Sophia University Press, 1971.
Wilson, William Ritchie. "The Way of the Bow and Arrow: The Japanese Warrior in Konjaku Monogatari." Monumenta Nipponica 28, 2 (Summer 1973):
I77-233Yabuki Keiki ^ K f f . Amida Butsu no kenkyu W^FKI^OW^. Rev. and en-
Cambridge Histories Online © Cambridge University Press, 2008
738
WORKS CITED
larged ed. Tokyo: Meiji shoin HJJrnlfK, 1937Yamada Hideo lUFB^M. "Sammi no kenkyu" UcfeoW^S. In Sakamoto Taro
hakushi kanreki kinenkai S-^^^W-dcMMti^t^,
ed. Nihon kodaishi ronshu
B#-£ft5&li#. Vol. 2. Tokyo: Yoshikawa kobunkan ^/l|&;fcfig, 1962.
Yamada Yoshio lilBB^ft. Samboe ryakuchu H^|gB§ft. Tokyo: Hobunkan ^ i
fig, 1951.
_
Yamagami Izumo U-J_h§J.S&. Nihon geino no kigen B :£l!lis©iBiK. Tokyo:
Yamato shobo ^ftHM, 1977.
Yamagishi Tokuhei Uj^^^F, ed. Genji monogatari Mft^H. Vols. 14-18 of
Tokyo: Iwanami shoten
Nihon koten bungaku taikei B ^-IS^X^iz^.
m, 1958-63Yamanaka Toshiji Oj^Sc^ and Sato K8ji fe^^fe. Kodai noyakusho
Hf. Tokyo: Iwanami shoten S^H/E, 1984.
Yamanobe Tomoyuki |il2ft]fr, Okada Jo PB3K, and Kurata Osamu HffliK,
eds. Senshoku, shikko, kinko Ift^ • j ^ I • ^ X . Vol. 20 of Genshoku Nihon no
bijutsu KfeB#©3Sffif. Tokyo: Shogakukan /J^fig, 1969.
Yasuda Motohisa ^EB7cA. "Bushidan no keisei" S i i O M . In /
Vol. 4 of Ienaga Saburo ^TRH^P, Ishimoda Sho CSffllE, Inoue Kiyoshi
ff, Inoue Mitsusada #_h}£^l et al., eds. Iwanami koza: Nihon rekishi ^G&M
S B ^ K . Tokyo: Iwanami shoten Sj&ft/S, 1962.
Yasuda Motohisa ^EHxX. Gempei no sokoku iSTOfflM. Vol. 7 ofKokumin no
rekishi SKOSS&. Tokyo: Bun'eido £ £ £ , 1968.
Yasuda Motohisa ^B07C^. 7»s« to /fewAi K®c i ¥ f t . Vol. 7 of Nihon no rekishi
0 * < O K . Tokyo: Shogakukan 'J^fig, 1974.
Yasuda Motohisa ^ E x A . "Kodai makki ni okeru Kanto bushidan" ^f^SSS
i<cjK-H"SM^S;±H. In Yasuda Motohisa $BB7cA, ed. MAow hokenseiseiritsu
no shozentei B * S ^ f J^3zOfg|tf^. Tokyo: Yoshikawa kobunkan ^JII&X
fig, i960, pp. 1-112.
Yasuda Motohisa $6876^. MAo« ^Aoew shi gaisetsu 0^ffiBSLlSM. Tokyo:
Yoshikawa kobunkan ^fJlliii^tfig, 1957.
Yokan jjcffi. Ojo juin ft^+H. In Jodo-shu zensho W±TK£M. Vol. 15. Tokyo:
Sankibo Busshorin ULjffSff^Stt, 1971.
Yoneda Yusuke *BB«^, "Gigunji ko" WM^.
In Kodaigaku kyokai * f t ^
© ^ , ed. £w^i" Tenryaku jidai no kenkyu 2£1I^JlB#ftoP2E. Tokyo: Yoshikawa kobunkan o ^ l l ^ i t l , i960.
Yoneyama Toshinao ^tlil^H. Dokyumento Gion matsuri-toshi to matsuri to minshu to V**.A's hffiH£-f$ffi£gl£K*£. Tokyo: NHK bukkusu NHK
7*.y ^ ^ , I986.
Yoshida Akira ^ffliPa. "Tato no seiritsu ni tsuite" EB^€>/S;iz:^'-Ol^t:. Historia t x M 7 16 (September 1956): 35-46.
Yoshida Mitsukuni ^ B B * ^ . Hoshi no shukyo fiO^ft. Kyoto: Tankosha
fit, 1970.
Yoshida Sanae ^EB¥la. "Fujiwara no Sanesuke no kazoku" WWMiS:?>'
Nihon rekishi 0^81^330 (November 1975): 69-85.
Cambridge Histories Online © Cambridge University Press, 2008
WORKS CITED
739
Yoshida Sanae ^ffl^-ia. "Fujiwara no Sanesuke to Ononomiya-tei: shindenzukuri ni kan-suru ichikosatsu" BWM'S t'l"&'g% : «ix;i§(e|!|-t 5 — # g .
MAOM rekishi B ^ M i 350 (July 1977): 50-69.
Yoshida Takashi "nffl^. "Ritsuryo ni okeru zoyo no kitei to sono kaishaku" W
^Ki*ittS*ft?g©3£5££*©fi?fR. In Vol. 3 of Sakamoto Taro hakushi kanreki kinenkai Jg#*gP1f±SMB&£, ed. MAo» kodaishi ronshu
§&£. Tokyo: Yoshikawa kobunkan ^JllSZ^tg, 1963.
Yoshida Takashi §ffl#. "Uji to ie" ? ?-"£ < - . In Sasaki Junnosuke fe*
and Ishii Susumu 5 # J S , eds. Shimpen Nihon shi kenkyu nyumon j$?fl B
. Tokyo: Tokyo daigaku shuppankai ^.ff^^mfS.^,
1982. pp.
31-58.
Yoshimura Shigeki ^^tirSJ^. Kokushi seido hokai ni kansuru kenkyu W**\'$\1S.M
SRi§g-t£ffl3E. Tokyo daigaku shuppankai SlM^tBlig^, 1979.
Yoshioka Saneyuki ^WtMi.. "Kizoku shakai no seijuku" Mtfeit£<D$l%&. In
Hashimoto Yoshihiko %fk~%-WM, ed. Komonjo no kataru Nihon shi ~&Jt
Z> B * £ . Vol. 2. Heian ^F$. Tokyo: Chikuma shobo %M^W, 1991.
Yoshioka Saneyuki ~aWlM~2.. "Fuyogeyujo to kageyushi ni kansuru shiron"
M&$k t MM&feiZ-m-tZ&M. In Vol. 3 of Inoue Mitsusada hakushi kanreki
kinenkai #±#£1$±SlgfE;&£, ed. Kodaishi ronso *f^*|^*. Tokyo:
Yoshikawa kobunkan ^ffl%Jt&, 1978.
Yoshishige no Yasutane ^^ftML. Nihon ojo gokuraku ki B ^tt^.@^f£. In
Inoue Mitsusada ^r^JJtM. and Osone Shosuke ^ C # S ¥ ^ 5 eds. Ojo den.
Hokke genki ft£R feUi^lS. Vol. 7 of Nihon shiso taikei
Tokyo: Iwanami shoten SS^#J£, 1974.
Zo Daisojo Kukai Wajo denki lt*{f IE^!gfa_hei2. In Kobo Daishi zenshu
^Cffl^ft. 3rd ed., rev. and enlarged. Shukan H # . Koyasan: Mikkyo bunka
kenkyujo ^WcSC.\m%^y 1967.
Cambridge Histories Online © Cambridge University Press, 2008
Cambridge Histories Online © Cambridge University Press, 2008