THE
ORIGINS
OF
YOGA A N D T A N T R A
Indie Religions to the Thirteenth Century
GEOFFREY SAMUEL
C ambridge
T H E O R I G I N S O F Y O G A A N D TA N T R A
Yoga, Tantra and other forms of Asian meditation are practised in
modernised forms throughout the world today, but most introductions to Hinduism or Buddhism tell only part of the story of how they
developed. This book is an interpretation of the history of Indic religions up to around 1200 CE, with particular focus on the development
of yogic and Tantric traditions. It assesses how much we really know
about this period, and asks what sense we can make of the evolution
of yogic and Tantric practices, which were to become such central and
important features of the Indic religious scene. Its originality lies in
seeking to understand these traditions in terms of the total social and
religious context of South Asian society during this period, including the religious practices of the general population with their close
engagement with family, gender, economic life and other pragmatic
concerns.
ge of f rey samue l is Professorial Fellow at the School of Religious and Theological Studies at Cardiff University. His publications
include Mind, Body and Culture: Anthropology and the Biological Interface (2006).
T H E O R I G I N S O F YO G A
A N D TA N T R A
Indic Religions to the Thirteenth Century
G E O F F R E Y SA M U E L
c a m b r i d g e u n i ve r s i t y p re s s
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo, Delhi
Cambridge University Press
The Edinburgh Building, Cambridge cb2 8ru, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521695343
c Geoffrey Samuel 2008
This publication 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 2008
Printed in the United Kingdom at the University Press, Cambridge
A catalogue record for this publication is available from the British Library
Library of Congress Cataloging-in-Publication Data
Samuel, Geoffrey.
The origins of yoga and tantra: Indic religions to the thirteenth century /
Geoffrey Samuel.
p. cm.
Includes bibliographical references and index.
isbn 978-0-521-87351-2 (hardback) – isbn 978-0-521-69534-3 (pbk.)
1. India–Religion. 2. Yoga–History. 3. Tantrism–History. I. Title.
bl2005.s26 2008
294.09′ 01–dc22
2007046053
isbn 978-0-521-87351-2 hardback
isbn 978-0-521-69534-3 paperback
Cambridge University Press has no responsibility for the persistence or
accuracy of URLs for external or third-party internet websites referred to
in this publication, and does not guarantee that any content on such
websites is, or will remain, accurate or appropriate.
Contents
List of figures
Preface
page vi
ix
1 Introduction
1
2 Stories and sources
15
part one meditation and yoga
39
3 The Second Urbanisation of South Asia
41
4 Two worlds and their interactions
61
5 Religion in the early states
94
6 The origins of the Buddhist and Jaina orders
119
7 The Brahmanical alternative
153
8 Interlude: Asceticism and celibacy in Indic religions
173
part t wo tan tra
191
9 The classical synthesis
193
10 Tantra and the wild goddesses
229
11 Subtle bodies, longevity and internal alchemy
271
12 Tantra and the state
291
13 The later history of yoga and Tantra
324
14 Postlude
339
References
Index
354
402
v
Figures
1.1. ‘Proto-Śiva’ Seal (M-304, after Parpola 1994: 188)
1.2. Seal Mohenjo-Daro M-1186 (after Parpola 1994: 260,
Fig. 14.35)
3.1. Map: Painted Grey Ware and Northern Black Polished
Ware (after Schwartzberg’s Historical Atlas of South Asia)
3.2. Map: Areas of Early Urbanisation (after Allchin)
3.3. Map: The Mahājanapadas (after Schwartzberg)
3.4. Map: Nanda and Mauryan states (after Schwartzberg)
4.1. Map: The Two Worlds (c. 500 BCE)
4.2. The Candravam
. śa (Lunar Dynasty): Outline Diagram
4.3. The Sūryavam
. śa (Solar Dynasty) and the Kings of
Mithilā: Outline Diagram
4.4. The Early South Indian Caste System according to
George Hart
4.5. The Brahmanical Varn.a System
5.1. Unidentified Yaks.a. Vidisha Museum (photo by author)
5.2. Indra disguised as a woodcutter offers grass to Śakyamuni
(accompanied by Vajrapān.i), Gandhara, 1st cent CE,
Peshawar Museum (photo by Sylvia Sax)
5.3. Vaiśravan.a. Modern Tibetan wall-painting, Tongsa
Gompa, Kalimpong (photo by author)
5.4. Gajalaks.mı̄ on Stūpa 2, Bharhut Stūpa, Indian Museum,
Calcutta (photo by author)
6.1. Śākyamuni prior to his enlightenment (accompanied by
Vajrapān.i) consults a jat.ila Brahmin. Gandhara, late first
century CE, Peshawar Museum (photo by Sylvia Sax)
6.2. Stūpa as depicted on Bharhut Stūpa, Indian Museum,
Calcutta (photo by author)
vi
page 4
5
44
46
57
59
62
65
66
86
87
103
105
106
108
121
141
List of figures
6.3. Two Deities. Bharhut Stūpa, Indian Museum, Calcutta
(photo by author)
6.4. Tambiah’s Diagram of the Value System of Northeast
Thai Village Society (Tambiah 1970: 338)
7.1. Veena Das’s Diagram of the Value System of Indian
Society
9.1. Map of Kus.ān.a and Sātavāhana States, first to third
centuries CE (after Schwartzberg)
9.2. Map of Gupta, Vākāt.aka and Pallava States, fourth to
sixth centuries CE (after Schwartzberg)
9.3. Early states in SE Asia to c. 650 CE (after Schwartzberg)
9.4. Anointing of Kumāra as Divine General (India, Gupta
Period) (photo from Theresa McCullough)
10.1. Bhairava, Vidisha Museum (photo by author)
10.2. Cakreśvarı̄. Mount Abu (photo by author, 1972.)
10.3. Guhyasamāja, probably 16th century, Vajrabhairava
Temple, Tsaparang, Western Tibet (photo by Rob
Linrothe)
10.4. Hevajra. Abu Cave, Piyang, Western Tibet, probably
17th or 18th century (photo by Rob Linrothe)
11.1. Meditator with Cakras and Kun.d.alinı̄ (Wellcome Library,
Sanskrit & Prakrit MSS No.640)
12.1. Map of South Asia, eighth to tenth centuries
(after Schwartzberg)
12.2. Map of Southeast Asia, eighth to tenth centuries
(after Schwartzberg)
12.3. (left to right). Indrayan.ı̄, Brahmayan.ı̄, Camun.d.a
(from Vergati 2000) (photos by Suzanne Held)
12.4. ’Cham dancer at Zangs mdog dpal ri dgon pa, Kalimpong
(photo by author)
12.5. Bhuta dancer, South Kanara District (photo by Mark
Nichter)
vii
143
148
168
196
198
200
206
244
258
261
263
272
294
296
316
318
320
Preface
This book is based on the Wilde Lectures in Natural and Comparative
Religion, which I delivered at the University of Oxford in November and
December 2002 under the title ‘Indic religions to 1200 AD: a critical and
anthropological approach’. Those who were present at the lectures will
realise that this book differs from the lectures in other respects besides the
title. Most of the text of the lectures is here, in one form or another, but I
have taken the opportunity to rethink and extend the argument in many
places. Unfortunately, the extensive visual material presented in the lectures
has had for practical reasons mostly to be excluded from the book.
This is a relatively short book, however, on a very large subject, and
there has been no attempt to be comprehensive. The book focuses on
the development of the yogic and Tantric tradition in Indic societies, and
while I have discussed the wider context in which these events happened
in considerable detail, I have not attempted to provide a comprehensive
history of Indian religion.
It is difficult to deal with language transcription consistently and systematically in a book that ranges over several bodies of scholarly literatures with
different conventions. The omission of diacritics is nevertheless a major irritant and often deprives the reader of vital information, quite apart from
rendering it impossible to know how words might be pronounced. My
general strategy has been to give only modern place names and words that
are thoroughly Anglicised without diacritics. I have generally given Sanskrit
forms in preference to Pali or other Prakrits, though have employed the
latter in contexts where it would seem clumsy or inappropriate to do otherwise (e.g. when I am citing the Pali Canon). I beg the reader’s indulgence
for remaining errors and inconsistencies; I am not a Sanskritist.
I wish to express my appreciation to the Electors of the Wilde Lectures
for allowing me to give the lectures, and in particular to Richard Gombrich,
who was a most kind and gracious host during my stay in Oxford, as indeed
during my previous stay in 1999, and who helped me in very many respects
ix
x
Preface
in relation to this book. I would also wish to thank those who attended
the lectures and provided valuable discussion and insight, and to many
others with whom I have discussed some or all of these issues in recent
years. The list is a long one, but I wish to mention at least Naman Ahuja,
Nick Allen, Robert Beer, Jim Benson, Marieke Clarke, Lance Cousins,
Max Deeg, Gill Farrer-Halls, Gavin Flood, Will Tuladhar-Douglas, David
Gellner, Sanjukta Gupta, Adam Hardy, James Hegarty, Saunaka Rishi Das,
Will Johnson, Klemens Karlsson, Kim Chong-Ho, Elizabeth de Michelis,
Mogg and Kym Morgan, Ruth Rickard, Rob Mayer, Cathy Cantwell, Brian
Bocking, Kate Crosby, Brenda McGregor, Ted Proferes, Robert Pryde, Julia
Shaw, Andrew Skilton and Michael Willis. I apologise to others whom I
have undoubtedly omitted. I particularly thank Thomas J. Hopkins for his
graciousness in allowing me to read and refer to his unpublished work on
the early history of Indian religions, Gunnar Haaland for allowing me to
use a photograph of a thang-ka in his possession for the cover, and Rob
Linrothe, Theresa McCullough, Mark Nichter, Asko Parpola, Sylvia Sax
and Anne Vergati for providing photographs and for assistance in obtaining
permission to use photographs. I also thank the National Museum of India,
the Department of Archaeology and Museums, Government of Pakistan,
for permission to use images and Peshawar Museum, Vidisha Museum, and
the Indian Museum, Calcutta for permission to photograph objects in their
collections. I thank Kate Brett and Gillian Dadd, of Cambridge University
Press, for their friendly, helpful and efficient assistance with producing the
book.
I also thank the University of Newcastle, New South Wales for allowing
me to undertake two periods of study leave during which much of the
research for the book was undertaken, the Leverhulme Trust and Brian
Bocking for a visiting professorship at the School of Oriental and African
Studies in 2003–4 which gave time for valuable further work on this project,
and Cardiff University for appointing me to the Professorial Fellowship
which has allowed its completion. I also wish to acknowledge the participants in the May 2004 workshop at SOAS on the politics of Asian religions,
among them Saunaka Rishi Das, Madhu Kishwar, Rajiv Malhotra, Hiroko
Kawanami, Chakraborty Ram Prasad and Ursula King, who helped greatly
in formulating some of the ideas in Chapter 1 and elsewhere in the book. I
do not think that at this point in time there is any fully satisfactory answer
to the questions raised on that occasion, but I hope that this book will be
in its way a positive contribution to the ends towards which that workshop
was directed.
chap t e r 1
Introduction
The ‘Indic Religions’ of my subtitle are early forms of what we now know
as Hinduism, Buddhism and Jainism. Their development, even with an
arbitrary end-date of 1200 CE,1 is a large topic, and only some aspects are
covered in this book. I am particularly concerned with the growth of one
of their central and most characteristic features, the group of traditions of
mental and physical cultivation that developed into what we now know as
‘yoga’, ‘Tantra’, and ‘meditation’. The indigenous terms vary, and do not
correspond neatly to modern Western uses of these terms, but practices
involving mental and physical cultivation, mostly directed towards the
achievement of some kind of liberating insight,2 are found in all the major
religions originating in the Indian sub-continent.
The main body of the book consists of five chapters (3 to 7) focusing
on the early growth of Buddhism, Jainism and the renunciate traditions
within Brahmanical religion, roughly from the fourth to second centuries
BCE, and three chapters (10 to 12) discussing the period from the fifth to
twelfth centuries. The first of these periods corresponds, as far as we can
tell, to the initial development of yogic and meditational techniques; the
second period covers the growth of Tantric practices and the relationship
between yoga and Tantra. The remaining chapters provide introduction
and commentary, and sketch developments before, between and after these
two key periods.
At the beginning of the twentieth century, these practices were scarcely
known outside of South Asia and the Buddhist societies of Southeast and
East Asia, a few specialist scholars and esoteric practitioners aside. By the
beginning of the twenty-first century, millions, if not tens of millions, of
1
2
I use BCE (before the common or Christian era) and use CE (common or Christian era) in place of
the specifically Christian terms BC and AD, as is the general convention in religious studies.
‘Liberating insight’ is a generic term that I use in this book, following Johannes Bronkhorst (1993),
for the various goals of the renunciate traditions of India (nirvān.a, moks.a, bodhi, kevala/kaivalya,
etc.).
1
2
Origins of yoga and Tantra
people around the world were practising Hindu yoga, Buddhist meditation
and related traditions, and ideas, concepts and practices deriving from yogic
and Tantric contexts had become a familiar part of global society. Although
this massive social development is not dealt with in detail in these pages,
it helps to explain why it is worth understanding these practices and their
origins. Like many other people, I have lived through aspects of these
developments in my own life, as a scholar of Tibetan and Indian religions
and of social and cultural anthropology, as a friend and acquaintance of
numerous people involved with the global spread of yogic and Tantric
practices, and as an intermittent practitioner myself of several varieties of
these traditions.
The impetus behind this book is the desire to understand what these
developments mean, and what yoga, meditation and Tantra have become
and might still become within their new global context. Part of the answer
to that question has to come from a study of contemporary developments in
their own terms. There have been quite a few studies of Western adoptions
of Asian spiritual techniques and approaches, including some of my own,
and there is plenty more to be done along these lines. Another part of the
answer, though, involves re-examining the history of these practices within
the Indic religions from which they originated. That is the focus of the
present work.
Yoga, meditation and Tantra are complex and problematic labels, and
rather than attempting to define them in detail at this point, I shall leave
the scope and meaning of our investigation to emerge in the course of the
book. Perhaps it is enough at this point to say that we are concerned with
disciplined and systematic techniques for the training and control of the
human mind-body complex, which are also understood as techniques for
the reshaping of human consciousness towards some kind of higher goal.3
In an earlier work, I have made some suggestions about the anthropological
analysis of mind-body processes in human life (Samuel 1990). This book
only occasionally ventures into such areas; it is primarily an attempt at the
historical understanding of the development of a particular set of techniques
and practices within Indic religions.
The most usual starting point for a history of Indic religions is the
religion of the Indus Valley cultural tradition in what is now Pakistan and
3
For any readers who are familiar with yoga primarily as a physical exercise, as one often encounters
it today, it is important to appreciate that the physical aspects of yoga were historically a secondary
part of a set of techniques that was aimed at training mind and body as a whole, and that (given
some quibbles about exactly what is meant by ‘religion’) had a specifically religious orientation. See
e.g. Alter 2005; de Michelis 2004.
Introduction
3
North-West India, best known from the extensive remains of the early
urban societies at Mohenjo-Daro, Harappa and elsewhere. These cities
correspond to what is now known as the ‘Integration Era’ of the Indus
Valley cultural tradition, and dated to around 2600 to 1900 BCE. The
large body of imagery found on the seals at these urban sites has been
particularly significant for scholars seeking to understand the religious life
of the Indus Valley peoples. In particular, ever since Sir John Marshall’s
suggestion in the 1930s that one of the Indus Valley seals represented an early
version of the god Śiva in a specifically yogic posture, it has been common
to trace the origins of yoga and of various other aspects of Indian religion
back to the Indus Valley cultural tradition (Marshall 1931: vol. I, 7).
I begin my account of the history of Indic religions by discussing some of
these interpretations, but I should make it clear from the start that I do not
feel that we can learn very much from this early material. Given what we
now know about cultural continuities in the archaeological record between
the Indus Valley cultural tradition and succeeding populations in the region,
it is certainly possible that there was some continuity in the area of religion.
The difficulty is that the early evidence is far from unambiguous, and that
it is almost always interpreted by reading later religious forms into it.
Consider the well-known image (Fig. 1.1) that Marshall regarded as
depicting a three-headed god, seated in a yogic posture, and saw as a
prototype of Rudra or Śiva as Lord of the Beasts (Paśupati). This story
has been widely accepted and the presence of a ‘proto-Śiva’ figure in the
Indus Valley is perhaps the most frequent assertion made about religion in
that period. However, on closer examination, the case for a ‘proto-Śiva’
interpretation of this image is far from conclusive. To begin with, Śiva is
not shown in this posture in later iconography.4 Nor is he ever shown with
a horned headdress in later times. Nor is it clear that the image has three
heads. Nor is it self-evident that the animals are to be read in terms of the
main figure being a ‘Lord of the Beasts’.
In fact, this image has been read in a variety of other ways. Alf Hiltebeitel
has suggested that the head represents a buffalo (others have preferred a
bull), and that the four surrounding animals correspond to the Vedic gods
of the four directions (Hiltebeitel 1978). For Bridget and Raymond Allchin
the image is ithyphallic (1982: 214). Herbert Sullivan and Shubhangana
Atre have both argued that it does not depict a god at all, but a female deity
(Sullivan 1964; Atre 1998).5 As for the posture of the figure on this and similar
4
5
For early Śiva iconography, see N. Joshi 1984 and Srinivasan 1984.
For further interpretations, see Dhyansky 1987: 90–1.
4
Origins of yoga and Tantra
Figure 1.1. ‘Proto-Śiva’ Seal (M-304)
seals, for Yan Dhyansky (1987: 94–9) and Thomas McEvilley (2002: 104),
it is clearly the yogic posture mūlābandhāsana, and has to be understood in
terms of proto-Tantric techniques aimed at driving ‘the sperm-marrow-soul
fluid up the spinal channel’ (McEvilley 2002: 110; cf. Dhyansky 1987: 100).
Sullivan notes that the posture ‘seems to us a natural enough one and need
not be a yogic posture at all’ (1964: 120), while Asko Parpola suggests that
this ‘so-called “yoga” posture may simply imitate the Proto-Elamite way
of representing seated bulls’ (1994: 250, caption to fig. 14.16)! The only
reasonable conclusion is that we do not actually know how to interpret the
figure, nor do we know what he or she represents.
Another possibly more explicit piece of imagery is the famous ritual scene
shown on the Mohenjo-daro seal M-1186.
It seems reasonably safe to interpret this as a ritual scene, since a simpler
version of the same scene is shown on several other seals, such as Harappa
H-177 (illustrated in Parpola 1994: 110, fig. 7.13) and Mohenjo-Daro M-488
(illustrated in Farmer, Sproat and Witzel 2004). All three of these seals show
a divine or human figure standing in a tree-frame of some kind (Parpola
Introduction
5
Figure 1.2. Seal Mohenjo-Daro (M-1186)
reads this as a fig-tree (1994: 256–61), another human figure kneeling or
sitting before him/her, and an animal with long wavy horns. One might
read the figure in the tree as a standing version of the ‘proto-Śiva’, and
the kneeling or seated figure as a priest or priestess who is worshipping
him/her, though it may be noted that in the two Mohenjo Daro seals both
wear the horned headdresses (if that is what they are), and in the Harappan
seal neither do.
For Parpola, the worshipper on M-1186 is ‘probably the chief priest of
the deity who possessed this seal’ and the animal is a human-faced markhor
goat. In front of the priest is a low table ‘on which is placed a human head,
identifiable as that of a warrior from its “double-bun” hairstyle which recurs
elsewhere in fighting scenes and is of Mesopotamian origin’ (Parpola 1999b:
249). The seven figures at the bottom are apparently female, since they wear
their hair in a plait, so the rest of the interpretation is straightforward:
The tree is probably the banyan fig, and the deity inside it a predecessor of Durgā,
the goddess of victory and love, to whom a human sacrifice of a brave warrior
has been made. The decapitated victim is likely to have been the groom in a
‘sacred marriage’ performed at the new year festival, and to have personified the
6
Origins of yoga and Tantra
predecessor of Rudra/Skanda/Rohita/Agni. [. . .] The seven females at the bottom
probably represent the ‘Seven Mothers’ of this war-god, the stars of the Pleiades,
which became the constellation of the new year when the naks.atra calendar was
compiled around the 23rd century BC. The markhor goat (śarabha in Sanskrit) is a
symbol of Agni in Vedic texts; according to the Kālikā-Purān.a, the Goddess most
appreciates man as a sacrificial victim, but next to a human victim she likes best
the śarabha. The human face of this beast in the seal may indicate that a ritual of
head exchange was practised [. . .]. (Parpola 1999b: 249–50; see also 1994: 259–61)
Now, I must admit that all this could be true. What is happening here,
however, is that Parpola, like other interpreters of these seals, is straining
to interpret objects that are far from clear, and is reading them in terms of
his knowledge of a wide range of texts and practices dating from a much
later period. If we did not have these parallels in mind, it is unlikely that we
would read the figure in the tree-frame as Durgā or Kālı̄, let alone see the
scene as implying a human sacrifice of a god impersonating Rudra/Skanda
etc.6 It is hardly obvious that the object on the table is a human head, for
one thing, though Farmer, Sproat and Witzel are apparently convinced,
since they read the equivalent object on Mohenjo Daro M-488 as another
human head (2004: 46, Fig. 13). It should also be noted that the Kālikā
Purān.a is generally dated to the eleventh or twelfth century CE (Urban
2001), so that Parpola is assuming that the Goddess’s tastes in sacrificial
meat remained unchanged for around three thousand years.7
However, as with the ‘proto-Śiva’ figure, a variety of other interpretations
are possible. Jayakar, looking at the same seal (M-1186), reads the figure in
the tree as a yaks.ı̄, and the seven figures at the bottom as apsarās or virgins
rather than the Seven Mothers. The kneeling figure is an alchemist-priest,
and the trunk of the tree is ‘shaped like the garbha yantra, the womb vessel,
wherein the ultimate secrets of alchemy were revealed’ (Jayakar 1989: 73). In
her interpretation, the object that Parpola reads as the head of a decapitated
warrior has been transformed into the Śrı̄ Cakra, the ‘mark and altar of
the goddess’ (1989: 73). Shubhangana Atre reads the kneeling figure as a
‘High Priestess’, noting that this ‘is obvious from her attire which exactly
resembles that of the deity’ (Atre 1998: 168), and the seven figures, rather as
in Jayakar’s version, as ‘vestal virgins’; she neglects to mention the alleged
warrior’s head, but interprets the scene as part of a sequence illustrating the
6
7
Note that Parpola needs to conflate Rudra and Skanda, because Rudra (Śiva) is the goddess’s consort
in later Brahmanical mythology, but it is Skanda who is associated with the ‘Seven Mothers’.
For a more detailed presentation of Parpola’s overall perspective on the origins of ‘Śākta Tantrism’,
see Parpola 2002a.
Introduction
7
retirement of an older High Priestess and the consecration of a replacement
from among the seven vestal virgins (1998: 167–8).
We note that both Parpola and Jayakar read ‘Tantric’ themes into the
material (the sacred marriage and human sacrifice for Parpola, the priestalchemist and Śrı̄ Cakra for Jayakar), but the Tantric themes are completely
different from each other and have nothing in common. Atre’s interpretation is not ‘Tantric’ at all. Clearly, these seals are not self-explanatory. In all
three cases, the reading of the seal depends on a whole set of assumptions
about the nature of Indus Valley religion.8
I find similar difficulties with other readings of the seal-images, such as
Jayakar’s interpretation of what she regards as a series of seals depicting
a goddess and a tree (Jayakar 1989: 71–3 and pls. 5–8). Here, again, the
possibility of onward continuity of tree and goddess cults is quite tempting,
but completely unproven. One has to strain quite hard to find continuities,
and there is ample scope for fantasy.
In any case, we have little or no idea what these so-called ‘seals’ were used
for, which makes it difficult to be sure that the images represent scenes of
religious significance. They are too fragile for use for labelling consignments
of goods, which seems to have been the function of apparently similar seals
in the Mesopotamian context. Farmer, Sproat and Witzel apparently assume
that they have a ritual function, and that the ‘inscriptions’ (which for them
are non-linguistic) are collections of symbols of deities or celestial forces.
Again they could be right, but there is little to inspire any degree of certainty
in this or any other interpretation (Farmer, Sproat and Witzel 2004: 41–3).
Parpola’s reading of a ‘sacred marriage’ theme into M-1186 (Fig. 1.2)
is premised on his assumption of a linkage with sacred marriages in
Mesopotamian religion (1994: 256), a suggestion also made by the late Prof.
D. D. Kosambi. Kosambi suggested that the Great Bath at Mohenjo Daro
served as an artificial ‘lotus pond’ constructed for ritual purposes, and was
surrounded by rooms in which visiting men took part in ritual sexual union
with ‘female attendant representatives of the mother goddess to whom the
citadel complex belonged’. ‘This is not far-fetched’, Kosambi continues.
8
One could easily provide further possible interpretations. If the kneeling figure is seen as female and
a high priestess, for example, as argued by Atre, the standing animal could represent her consort
and Sacred King, giving another variant of Parpola’s (Frazer-style) sacred marriage. This would link
with the well-known seal from Chanhujo-daro which may (or may not) represent a ‘bison bull about
to have intercourse with a priestess lying on the ground’ (Parpola 1994: 256, fig. 14.32), not to say
with Biardeau’s work on goddesses and buffalo-gods in modern South India (Biardeau 1989). One
could then interpret the numerous representations of individual bulls and other animals on seals as
representing Sacred Kings of particular communities. But the real point is that all these interpretations,
Parpola’s, Jayakar’s, Atre’s and my own, have to be regarded as speculative and unproven.
8
Origins of yoga and Tantra
‘The temples of Ishtar in Sumer and Babylon had similar practices in which
girls of the leading families had also to participate’. These priestesses were
the origin of the later Indic mythology of the apsaras, ‘irresistibly beautiful
women who would entice men to consort with them and eventually lead
the heroes to destruction’ (Kosambi 1965: 68).
Kosambi was often a sensitive and insightful scholar, and such an interpretation might, like some of the other interpretations we have been considering here, enable us to take back the origins of sexual yoga in India
by another 2000 years, but it does, like the work of Parpola, Jayakar and
others, seem to go rather a long way beyond the evidence.
Other attempts to make positive assertions about Indus Valley religion
strike me as equally conjectural (e.g. Jairazbhoy 1994). I am in no position
to say that any of these interpretations are incorrect, but they certainly
cannot all be right. Since there is no obvious way to choose between them,
they do not actually take us very far. At the end of the day, we know quite a
lot about the daily life of the people of the Indus Valley urban civilisation,
but little or nothing for certain about their religious practices. In particular,
it seems to me that the evidence for the yogic or ‘Tantric’ practices is so
dependent on reading later practices into the material that it is of little
or no use for constructing any kind of history of practices. I am therefore
taking a more cautious view in the present work, and assume that we do
not have conclusive evidence for yogic or ‘Tantric’ practices in the Indus
Valley cultural tradition.
I find myself equally unpersuaded by attempts to see yogic or ‘Tantric’
practices in their developed forms in the R.gveda or Atharvaveda. There
are certainly indications both of magical ritual for pragmatic purposes, and
of ecstatic religious practices, ‘shamanic’ if the reader wishes to use the
term. I shall discuss some of these in later chapters (7 and 9). There is also
conceptual material, such as the role of ‘breath’ ( prān.a) within the body,
which is taken up and reworked by later yogic and ‘Tantric’ theory.
There is nothing, however, to imply yogic practice, in the sense of a
developed set of techniques for operating with the mind-body complex.
Our best evidence to date suggests that such practices developed in the
same ascetic circles as the early śraman.a movements (Buddhists, Jainas and
Ājı̄vikas), probably in around the sixth and fifth centuries BCE.
It is for these reasons that Part One of my book is focused in this period,
which follows on what has been called the ‘Second Urbanisation’ of South
Asia (the first being the growth of the Indus Valley cities in the third
millennium BCE). The growth of cities and early states in the sixth and
Introduction
9
fifth centuries BCE was the context for the early śraman.a movements,
and this is where, as far as we can tell, the new techniques of spiritual
development were first developed and propagated.
As for ‘Tantra’, much depends on what we mean by that much-contested
term, which has a wide variety of meanings within the Indic traditions
themselves. The central issue with which I deal in Part Two, however, is
the development of the relatively coherent set of techniques and practices
which appears in a more or less complete form in Buddhist and Śaiva
texts in the ninth and tenth centuries CE. This comprises a number of
elements: elaborate deity visualisations, in which the practitioner identifies
with a divine figure at the centre of a man.d.ala or geometrical array of
deities; fierce male and particularly female deities; the use of transgressive
‘Kāpālika’-style practices associated with cremation-grounds and polluting
substances linked to sex and death, and internal yogic practices, including
sexual techniques, which are intended to achieve health and long life as
well as liberating insight.
The various components of this set of spiritual techniques appear to
come from different sources. I attempt to trace their growth, and to make
sense of their adoption within relatively mainstream Buddhist, Jaina and
Brahmanical contexts.
Thus my main narrative runs approximately from about 500 BCE to
about 1200 CE. These limits are somewhat arbitrary, particularly given
the uncertainties of dating for the early part of this period. ‘About 500
BCE’, however, represents a point at which I assume that an early form
of ‘Brahmanical’ culture using an Indo-Aryan language had been firmly
established in parts of Northern India (present day Punjab, Haryana and
Western UP), but had not yet reached dominance over the North-Eastern
areas (including present-day Bihar, West Bengal and Bangladesh), or over
the remainder of South Asia. The year 1200 CE represents a point at which
Muslim rule had been established over most of North and Northeast India,
and the main remaining centres of Buddhist culture in these regions had
been destroyed. This was far from the end-point of Indic religious developments, but it forms a convenient point at which to close a narrative that is
already seeking to encompass a very large range of cultural time and space.
Islam, of course, had an impact on India and other areas of Indic culture
well before 1200 CE. The story of the early stages of its incorporation
within Indic societies is an important and significant one, but it goes well
beyond my personal competence, and I have not attempted to tell it here.
As for Buddhism, it continued to develop both within and beyond South
10
Origins of yoga and Tantra
Asia after 1200 CE, but that development entered a new phase, and took
place in very different conditions. Part of that story, the part that involved
the people of Tibet, was the subject of one of my earlier books, Civilized
Shamans (Samuel 1993), and in some ways the present work is a kind of
prequel to that book.9
The study of Indic religious traditions has to be approached today with
some sensitivity and care. During the main period discussed in this book,
Buddhism, Jainism and forms of Brahmanical religion ancestral to modern
Hinduism were all vital parts of the South Asian religious scene, alongside
a basic level of folk and village religious practice that still exists today
in various forms throughout South Asia and is best regarded as neither
Buddhist, Jaina nor Brahmanical.10 Today’s Hindus, Buddhists and Jains
have their own traditions of scholarship and study, and their own ways
of understanding themselves and their religions. Much of this may be a
modern development, a reaction to the unequal dialogue with Christianity
and other Western forms of knowledge during colonial times (cf. Lopez
1995; King 1999; Viswanathan 2003). The modernist self-understandings
that have resulted from this dialogue, however, are a reality in the lives of
thousands of millions of people in Asia and the Asian diasporas. These
people have a legitimate concern with how their religions are portrayed. At
times, however, this can lead to problematic attempts by pressure groups to
control what is said about Indic religions and eliminate features that do not
fit neatly into a spiritually sanitised and benign picture. This is unfortunate,
since attempts at excessive purity in the religious field generally backfire.
At the same time, Western societies have developed their own modes
of understanding Asian religions, both popular and academic, and these
undoubtedly have their own flaws and limitations. Critics of Western
academia often fail to appreciate the deep and positive engagement with
Indic religions that underlies much ‘Orientalist’ and more recent scholarship. This aside, I would hardly want to suggest that all of this work is
beyond criticism.11
9
10
11
Three substantial chapters (Twenty to Twenty-Two) of Civilized Shamans discussed the history of
Buddhism in India up to around 1200 CE. I realised at the time that an adequate treatment of the
material in those chapters was a much larger project than I could then undertake. In subsequent
years, I have become increasingly involved with the early history of Indic religions, particularly of
Tantric Buddhism. The present book is to some degree an attempt to take stock of my work in this
area.
We can see these, for example, in the women’s rituals of marriage (strı̄-ācār) practiced by Hindus,
Muslims and followers of other religions in West Bengal and Bangladesh today and in similar rituals
elsewhere in South Asia (Fruzzetti 1990; Good 1991).
Much of this discussion in its modern form goes back to the late Edward Said’s Orientalism (1978).
Said’s argument developed primarily in relation to the Arab context. The Indian and various
Introduction
11
The story of Indic religions is nevertheless an important one, and part of
the world’s heritage. It would be unfortunate if scholars, Asian or Western,
were to feel that it has become too politically sensitive to give as honest as
possible an account of that story.
The central concern of this book is to sketch the development of the
techniques of mental and physical cultivation that came to form a key
element of the religious traditions of the Indic world. This is, however, part
of the wider story of the growth of Indic civilisation and of societies in
South and Southeast Asia, and this story, like that of other civilisations and
societies, contains its fair share of warfare, destruction, human exploitation
and suffering. The lotus of spiritual enlightenment, as Indian traditions
themselves so often remind us, grows out of the mud of everyday life. I
have tried to include both sides of this picture, the sophisticated spiritual
culture and the solid ground of ordinary life out of which it grows. A
prettified and unrealistic picture of a religious tradition is of little use to
anyone.
Thus I hope that this book will be found useful by Westerners interested
in Asian religions (including scholars working in this and related areas) and
also by Asian members of the religious traditions that I am discussing. I hope
too that it will be found of value by a third group of people: Westerners and
other non-Asians who have become personally involved in the practice of
one or another of these religious traditions. It is particularly difficult to get a
clear view of a religious tradition when one encounters it in a cultural setting
very different from its own, mediated perhaps through English or another
Western language, and without the social and cultural context that gives
sense to many of its practices. Indeed, that context has changed radically in
the societies where these religions grew up as well. I feel that there are many
important things that the world can learn from Asian religious traditions,
but that our ability to integrate their knowledge into the evolving body of
understandings on which our now global civilisation conducts its affairs
will be greatly aided by knowing more about the origins of these traditions.
That is one of the principal aims of this book.
It is important in approaching subject-matter as large and complex as
that considered here to recognise the limits of our knowledge. A striking
feature of the last few decades of Indological research has been not only
the growth of new knowledge but also the gradual realisation that much
of what scholars thought that they already knew was far less secure than
Southeast Asian contexts have both similarities and differences, reflecting in part different schools
of colonial scholarship (British, French, Dutch, German) and different periods. Cf. Inden 1992;
Mackenzie 1995; King 1999; Viswanathan 2003; D. Smith 2003.
12
Origins of yoga and Tantra
had been assumed. This is perhaps most striking in the area of chronology,
where a series of conjectural datings adopted as working hypotheses by
the great nineteenth-century Indologists and Buddhologists had become
a kind of received doctrine. It is now clear that many of the details were
wrong and that the scheme as a whole is quite shaky and problematic (see
Chapter 2). This is hardly a surprise, given the scanty evidence on which
our chronological understandings were initially constructed, but certainly
for someone approaching the field from the outside it is striking how unsure
we are about the dating of many crucial events, people or texts.
Our knowledge of the texts is also less secure than might appear at first
sight in a world where works like the Vedas, the Upanis.ads and large parts
of the Pali canon have been translated repeatedly into European languages.
The Vedas were not written down until a fairly late stage, and I do not
believe myself that they can be taken as a literal and accurate witness for
any specific stage of early Indic religion. It is far from clear exactly what
the established text of the Pali Canon amounts to in terms of its relation
to early Buddhism (see e.g. Collins 1990), or when to date any particular
Upanis.ad or section of an Upanis.ad. Similar problems attend texts such as
the Rāmāyan.a or the Mahābhārata, the Purān.as, or the Jaina canon, though
the precise ways in which our previous certainties have been eroded varies.
As a totality, this uncertainty is part of the corrosive effect of a pluralistic intellectual universe on the established assumptions of a field, a
phenomenon that can be observed in many parts of today’s intellectual
universe. In the case of Asia, there is also the specific issue of post-colonialist
critiques, and the complex ways in which Westerners and Asians position
themselves in relation to the many tense and difficult political issues which
have affected the study of Asia in the Western academy in recent years.
I will say a little more about what might be called the ‘standard view’
of Indian religions and their history in Chapter 2, but I do not intend to
rehearse its problems here at length. This is a job that has been done with
commendable enthusiasm – some people clearly think a little too much
enthusiasm – by a variety of scholars in recent years under the impact of
post-modernist and post-colonialist thought.12
On the whole, I feel that there is a good deal to be said for these critiques. My job in this book, however, is somewhat different: it is to ask
what we could put in the place of the problematic conceptions which have
come to dominate this field of study. After all, terms such as ‘Hinduism’,
12
See King 1999 for a good summary; also Viswanathan 2003 and D. Smith 2003 for two contrasting
recent perspectives.
Introduction
13
‘Buddhism’, and the like, and definitions of those religions in ways ultimately derivative from Christian theology, for all of their questionable
origins, have now become very real parts of contemporary discourse, and
central to the self-definition of present-day Asian nations. Is there some way
in which we can move beyond these damaging and limiting perspectives?
To an anthropologist like myself, at least, it comes fairly naturally to
argue for a wider and more inclusive understanding of religion, and to see
Brahmanical Hinduism, Buddhism and Jainism as variant developments
from a shared basis of relationship to the problems of everyday life. That is
essentially the view from which I work in this book. However, I would not
want my approach to be read as ‘anthropological’ in a narrow sense, and it
should be plain to most readers that it is not. As far as I am concerned, no
point of view – textual, anthropological, internal or external – can be taken
as absolute or primary. All may have truth-content (using those words in a
general rather than a technical philosophical sense). We are almost always
better off integrating as many different perspectives as possible into our
understanding. Encounters between disciplines can be bruising at times,
with a sense on each side that the other’s presuppositions, if accepted, will
undermine one’s own position. However, there are usually ways of going
forward which lead us to greater understanding by integrating the key
points of both perspectives.
In arguing for the fundamental commonality of the Indic religious background (a very different position, it should be added, from seeing Buddhism
merely as a development from or reaction to ‘Hinduism’, which is a position that by now has hopefully lost any scholarly respectability it may once
have had), I am aware of the political dimensions of what I am saying.
In any case, one can scarcely speak of investigating the history of Indic
religions today without being aware of the contemporary political implications of any such inquiry. As the conflict over the site of the former mosque
of Babur at Ayodhyā – now claimed as the birth-site of Lord Rāma –
has demonstrated, religion nowadays can be an immensely divisive and
destructive force within Indian society. The Ayodhyā conflict, which led to
several hundred deaths in Gujarat in 2002, is only one of a number of such
conflicts of varying degrees of intensity. Elsewhere in India, Hindus and
Sikhs, and Hindus and Kashmiri Muslims have been embroiled in bitter
and long-standing conflicts in which religion plays a major role. Nor are
things much better in other parts of South Asia: Sinhalese religious nationalism has been a major contributor to the Sri Lankan conflict, while both
Pakistan and Bangladesh have seen violence against religious minorities in
recent times.
14
Origins of yoga and Tantra
Yet this religious nationalism, or ‘communalism’, as it is known in India,
is in many ways a modern imposition. It has to be understood in terms of
the colonial and post-colonial social transformations of the sub-continent,
and the globalising forces to which South Asia, like the rest of the world, is
now subject. The similarities between the religious nationalisms of South
Asia and those of West Asia, the Philippines, Indonesia or the Balkans – or
even Northern Ireland – are scarcely coincidental, though I will not argue
the point further here (cf. Fox 1996; Castells 1997: 1–67). What it is more
important to appreciate is that the societies of South Asia before, say, the
mid-nineteenth century, while not entirely free from religious conflicts,
had much more fluid and less categorical conceptualisations of religious
identity than we now see.
Thus the creation of a Sikh religion and identity distinct from ‘Hinduism’
is a fairly recent and still-contested development (e.g. Oberoi 1994). As for
Hindus and Muslims, in many parts of India they regularly took part in
each other’s festivals and visited each other’s temples and shrines, while
some communities were quite difficult to categorise as either ‘Hindu’ or
‘Muslim’. There has been pressure from both sides for clearer definition
and more puristic forms of observance, but even nowadays the boundary
between ‘Hindu’ and ‘Muslim’ is often hazier and more complex than one
might think at first sight (see e.g. Salomon 1991; Openshaw 1997; Saheb
1998; Mayaram 2000; Gottschalk 2000). I do not intend to idealise the
past here, but the changes have been real, and not for the better, and
interpretations of the Asian past have gained a new salience as justifications
for contemporary political positions.
Thus it is that a study of the past, even of the relatively distant period
discussed in this book, is far from irrelevant to the present. We are, I think,
at a point in time when stressing the commonalities and the deep mutual
implications of the religious traditions of South and Southeast Asia, from a
perspective sympathetic to, but by no means identified with, any of them,
may be a useful exercise for reasons that go beyond the purely scholarly.
If this book can, in any way, help to point to ways of going beyond some
of the entrenched hostilities between Asian communities, then the exercise
will have been well worthwhile.
chap t e r 2
Stories and sources
The standard story of the development of Indic religions was developed
in the mid to late nineteenth century, in a collaboration between Western
scholars on the one side, and Hindu and Buddhist scholars and intellectuals
on the other. This story was essentially that of the development of a number
of separate religions, principally Hinduism, Buddhism and Jainism, with
Hinduism being seen as the earliest and Buddhism and Jainism as reactions
against it. The narrative began with the hymns of the R.gveda, thought to
be the oldest existing texts in an Indian language. The R.gveda was seen as
representing the earliest stage of Hinduism and subsequent stages as a series
of developments from it. The R.gveda was treated as, in effect, a foundational text with a somewhat similar role to that of the Five Books of the
Jewish Torah in relation to traditional accounts of Judaism, Christianity and
Islam.
The story continued through the remaining texts of the Vedic corpus,
down to such ‘late Vedic’ texts as the Upanis.ads, Śrauta Sūtras, Gr.hya Sūtras
and Dharma Sūtras. All these texts were placed in a historical sequence, and
seen as defining an early Hinduism, composed of sacrificial rituals, legal
prescriptions and philosophical speculations, against which the śraman.a
movements (Jainism and Buddhism) reacted.
The great epics (Mahābharata and Rāmāyan.a) formed the defining texts
of the next period on the Hindu side, and the Purān.as, along with the
writings of the Vedānta philosophers and the bhakti poets, of the next,
with the Tantras as a rather uncomfortable parallel development. Buddhist
and Jain histories, after the initial break with Hinduism, were presented as
largely separate stories. The Buddhist story focused on the preservation of
a rational and philosophical ‘early’ Buddhism by the Theravādins, and the
development of new philosophical schools by the Mahāyānists. Jainism, like
Theravada Buddhism, was seen primarily in terms of the preservation of
rational and philosophical early traditions. Tantra again formed a somewhat
unwelcome parallel development on both sides, though the influence of
15
16
Origins of yoga and Tantra
Tibetan lamas on Buddhist scholarship in recent years has meant that
Vajrayāna Buddhism has gained a new respectability.
The above story may be something of a parody, and most contemporary
scholars give a more nuanced and tentative version, but it is still recognisable
behind modern accounts of the development of Indic religions prior to the
arrival of Islam in the subcontinent. Yet, as has been repeatedly pointed
out by numerous scholars, most parts of this story are problematic, and the
story as a whole is largely untenable.
Many of the problems here derive from the tendency of past Western
scholars, whether or not themselves Protestant Christians, to see religion
in terms of Protestant Christianity, a religion that defined itself against
its rivals in terms of a return to the ‘authentic’ texts of the Bible. The
Protestant polemic against Catholicism was largely carried out in terms of
accusations of deviation from the ‘original’ teachings of Jesus as seen in
the New Testament. These deviations consisted in the growth of magical
and superstitious practices, unnecessary theological complexities, and a
subordination of religion to political and economic purposes. Religions in
this model were founded by an inspired teacher who created a body of
texts that were then systematically misinterpreted and distorted through
succeeding generations. There was little point in studying what people
actually did, since it was only valid if it reflected the texts.1
This Protestant model formed a template that was repeatedly applied
by Western scholars to Asian religious traditions. At the same time, it
provided a model in terms of which nineteenth and twentieth century
Hindu and Buddhist reformers, from the Brahmo and Arya Samaj down to
the Mahabodhi Society, attempted to reshape their own religious traditions.
Buddhism, which had its own narrative of teaching and decay, fitted the
model particularly neatly. The Buddha could be seen as a humane religious
reformer on the model of Jesus, teaching through parables and other simple
and straightforward means. In this perspective the Buddha became a kind
of Christ figure reacting against a legalistic and caste-bound Brahmanical
priesthood, the Hindu equivalent of the Sadducees and Pharisees of the
New Testament account.
Theravāda Buddhism, with its claims to be a pure, uncorrupted tradition, fitted neatly into this view of religion. The Pali Canon was seen by
both Sinhalese and many Western scholars as the original version of a body
of scriptures later expanded and corrupted by later Mahāyāna or, worse,
1
There was also an interest on the part of Catholic scholars, especially Jesuit and other missionary
orders working in Asia, with local traditions of ritual practice. This led to some important early
studies of Asian ritual life, though it arguably had little effect on a dominant view of religion as based
around ‘belief’, at least in the English-speaking world.
Stories and sources
17
Tantric clerics. Sinhalese Buddhist reformers in the late nineteenth and
early twentieth century emphasised the rational and ‘scientific’ nature of
Theravāda Buddhism, so that seeing the Buddha in this light also provided
a useful weapon against the Christian dominance of European society.
This also had the advantage that the Asian Buddhism actually being practised by most Asian Buddhists (including most Buddhists in Sri Lanka and
Southeast Asia) could be dismissed as a superstitious and degenerate development of the real teachings of the Buddha. This left a scholarly elite as
the true interpreters of the teachings of the Buddha, as well as opening the
way to racist interpretations of Buddhism as an ‘Aryan’ teaching corrupted
by its Asian environment (cf. Deeg 2006). Such interpretations are out of
fashion, but the idea of a ‘real’ Buddhism that can be found in the Pali
Canon rather than the practice of historical or contemporary Buddhists
remains alive and well.
The model worked rather less well on the Hindu side, since the Vedic
hymns are not much like the Pentateuch either in content or in the way in
which they were used by later Indian religious traditions. Hindu religious
reformers did not necessarily use the early Vedic texts as their primary
‘scriptural’ basis, since these were poorly adapted to such a function; Rām
Mohan Roy looked rather to the Upanis.ads and the Brahma Sūtra.
Increasingly, scholars have seen ‘Hinduism’ as a problematic term for
the pre-modern period, at least before the mid-first millennium CE, when
a religious tradition with recognisable similarities to modern Brahmanical
Hinduism took shape under the patronage of the Guptas and subsequent
dynasties. The terms ‘Buddhism’ and ‘Jainism’ are still widely used for the
early period, but they too tend to carry questionable Western assumptions
about the nature of religion along with them.2
In particular, for Christianity, exclusivity has always been a critical issue.
One cannot be a Christian and a member of another religion at the same
time.3 Yet virtually all Buddhists in Southeast Asia have also been involved
in the cult of local gods and spirits, just as most East Asian Buddhists have
2
3
Both ‘Jaina’ and ‘Hindu’ as terms used to designate religions date from the second millennium CE
(see Fluegel 2005 for ‘Jaina’), though terms relating to ‘Buddha’ were used as self-designations by
Buddhists from relatively early times.
In part this is linked to the misleading and confusing Western identification of religion with ‘beliefs’
(in the sense of ‘statements held to be true’). ‘Belief’ and the related word ‘faith’ in religious statements
(‘I believe in/have faith in Jesus Christ’ etc.) have shifted historically from meaning something like ‘I
have trust in, I have made a personal commitment to the Christian deity’ to meaning ‘I affirm that the
Christian deity exists and that some version of Christian theology is a correct account of the nature
of the universe’. If religion is identified with theology and belief statements, then being involved
in two religions at the same time implies being committed to two irreconciliable sets of statements
about the nature of the universe. This, however, is a problem for philosophers and theologians, not
for most ordinary human beings. See e.g. Southwold 1979, 1983.
18
Origins of yoga and Tantra
also been involved in Confucian, Taoist, Shinto or shamanic cults. We will
see more of this relationship in later chapters. As for Jainism, the question
of whether it is a part of ‘Hinduism’ or a separate ‘religion’ is far from
resolved even for many contemporary Jains. The terms ‘Buddhism’ and
‘Jainism’ also tend to suggest that these ‘religions’ are primarily modes of
understanding the universe (the first question for a naı̈ve Western student
even today tends to be ‘What do they believe in?’) rather than ways of living
in relation to the universe (including one’s fellow human beings).
The implicit (and occasionally explicit) parallels with the development
of Western religions in the ‘standard story’ are dated historical baggage,
but they still have a real presence in and beyond the Western academy. At
the same time, the kind of multi-dimensional approach that I undertake
here, where textual, epigraphic, iconographic and archaeological material
are seen as mutually-illuminating bodies of evidence, is widely accepted
in theory, if not always in practice, and there are plenty of studies that
have explored such combined approaches for particular periods or topics
(cf. Schopen 1997a; Kosambi 2002a).
Yet there are still problems. The histories of the Buddhist and Jain traditions remain far from adequately integrated into the general history of
Indic religions, and the development of specialist fields of Buddhist, Hindu
and Jaina studies, dealing with distinct bodies of texts, has meant that this
division has become increasingly entrenched in academic approaches. Even
to write a general history of Indic religions is at this point in time an undertaking that is somewhat against the grain of academic work.
A recurrent issue is in effect a derivation from the Protestant model and
the primarily textual approach to religion that it encourages. This is the
question of how much of the ongoing development of Indic religions can be
traced back to the texts of the Vedas, meaning both the hymns of the R.gveda
and the extensive body of early material that grew around them, down to
the Brāhman.as, Āran.yakas, Upanis.ads and the ritual and legal manuals
(Śrauta Sūtras, Gr.hya Sūtras, Dharma Sūtras), and how much needs to
be assigned to ‘external’ influences. Approaches that emphasise continuity
are sometimes called ‘orthogenetic’ (as opposed to ‘heterogenetic’). Such
approaches can be found among Buddhist scholars as well as scholars of
Tantra and of Brahmanical Hinduism.
By necessity, most serious nineteenth-century approaches to Hinduism
and Buddhism were in effect orthogenetic. Western scholars of this period
were working primarily with texts and their primary tools were philological.
A significant part of their task was to arrange the texts in a presumed
historical sequence of development. The hymns of the R.gveda were clearly
Stories and sources
19
the most archaic texts around in terms of their linguistic features and were
therefore assumed to be the oldest, especially in view of their evident affinity
with the early texts of Iranian religion.
Much of this changed in the course of the twentieth century, and for
a variety of reasons, both political and academic. On the political side,
‘racial’ theories of Aryan origins become popular among a variety of groups:
racially-motivated Western scholars, Indian groups for whom the myth of
‘Aryan’ racial identity provided a positive self-image, Tamil nationalists who
developed the mythic corollary of an indigenous ‘Dravidian’ racial identity.
The racial foundations of such ideas were largely discredited by the midtwentieth century. The assumed linkage between culture and genetics is now
evident nonsense, and in any case recent developments in genetic research
have made it clear that, as with most or all modern human populations,
the genetic variety and complexity of the population of South Asia today
is far too great to be encompassed in such simplistic models.
Other, more academic reasons, included the discovery of the early urban
cultural tradition of the Indus Valley. This provided evidence for an indigenous cultural tradition that appeared both distinct from the Vedic material
and to have some continuities with later Indian religion, as with Sir John
Marshall’s identification of a ‘proto-Śiva’ image on an Indus Valley seal
(Chapter 1). Initially, these discoveries fed into the racially-based models of
‘Aryan’ invasions and conquest of ‘Dravidian’ indigenes and the like, but
they also raised intellectual issues that could not be dealt with so crudely,
and are still far from resolved today. More generally, the gradual uncovering
of the archaeological record, for all of its inadequacies by modern scientific
standards, made it clear that South Asia had a complex and varied prehistory, and that this involved a number of cultural complexes that could
not be mapped at all easily onto a simple model of an expansionist Vedic
civilisation.
The growth of social scientific approaches to culture also began to bring
about more awareness of the many aspects of culture that are not transmitted
textually. Social and cultural anthropology in particular mostly developed in
the context of studies of non-literate societies and so developed approaches
that emphasised actual cultural and ritual practices and their variations
and transformations rather than elite textual models. In the Indian context,
this meant that a large body of material was progressively revealed with often
only a remote relationship to the Vedic texts.
Indologists in the late twentieth and early twenty-first century are
still divided on these issues. In part, this is a methodological question.
Textual scholars are at home working with texts, and are not necessarily
20
Origins of yoga and Tantra
at ease dealing with other kinds of material. The revival of ‘orthogenetic’
approaches provides a way to bypass the problem. Textual studies are far
more sophisticated today, if only because we have a far greater body of
textual material at our disposal and the substantial body of work creates a
background in relation to which today’s scholars must operate.
It is tempting to comment here, as Olivelle has done, that it is time ‘to
move beyond this sterile debate and artificial dichotomy’ and to reject ‘the
untenable conviction that we can isolate Aryan and non-Aryan strands in
the Indian culture a millennium or more removed from the original and
putative Aryan migrations’ (Olivelle 2003: 273). His own suggested focus
on ‘social, economic, political and geographical factors’ is also very much in
tune with my own sympathies. Yet the dispute is not so easy to dismiss, in
part because there is undoubtedly a political edge to many of the arguments
involved. The texts, some of them at least, are part of the way in which
modernist versions of Hinduism, Islam and Jainism see themselves, and
both Asian and Western adherents to these modernist versions of Asian
religions clearly prefer to present their traditions in ways that emphasise
‘orthogenetic’ textual elements and downplay both ‘external’ influences and
folk and popular ‘deviations’.
In practice, the orthogenetic approach, combined with a central emphasis
on the textual tradition, leaves a number of problems unanswered, particularly for the earlier part of the period covered in this book (the second half
of the first millennium BCE):
r It is clear that large parts of later Indic religions do not derive in any
simple way from the R.gveda or other early Vedic texts. But how does
one fit the non-Vedic components of Indic religions into the picture?
What kind of historical, cultural or social location did they have?
r Given the late date at which much of the textual material was committed
to writing, and the even later date of most of our surviving texts, how
representative are the texts we have of religion in any particular area or
period?
r Our texts undoubtedly reflect urban elite and court religious practices
much more than those of villages and of non-elite strata of society. What
kind of relationships do we assume between urban religion and village
religion, between elite and non-elite religious practices?
r The geography of early Indic religions is only beginning to be charted.
When we are talking about early Indic religions, where and when specifically are we talking about? At what stage were various parts of South
and Southeast Asia drawn into the orbit of Brahmanical and Sanskritic
culture (to the extent that one can talk of such a thing as a unity at all)?
Stories and sources
21
Similar uncertainties surround the major religious developments of the
second half of the period. The growth of the cults of Brahmanical,
Buddhist and Jain deities (including the deified Buddhas, bodhisattvas
and tirthankaras) is far from fully understood. The same can be said for
the growth of the cult of fierce deities, male and female, of the origins of
Tantric religion and its effects on the Indic religious field, issues that will be
of considerable prominence in our account. Another major development,
that of the bhakti cults, will be treated in much less detail, but is of equal
significance.
I do not intend to be over-dismissive of work based on ‘orthogenetic’
assumptions. Much of the research carried out by scholars working in this
way has been of very high quality, and all scholars of Indic religions are
indebted to it. Conceptually and intellectually, however, a strictly orthogenetic and textual view of Indic religions seems to me to be fundamentally
flawed. A religious tradition is not just a body of texts. It is, above all, something that lives and is maintained through the lives of human beings. A
text, unless we assume that certain texts are indeed divinely inspired and so
beyond academic analysis or criticism, can only be the product of one or
more human beings and has to make sense in terms of their lives and their
understandings of their situation (which may of course include concepts of
divine revelation through dreams or other visionary processes). The onward
transmission of a text, often complex in the South Asian context where our
written versions can be quite recent and can reflect a long textual history,
similarly has to make sense in the same terms.
In reality, then, even textual scholarship cannot be divorced from a reconstruction of the intellectual, emotional, social and political context of the
people who produced those texts, however difficult that might be to achieve.
I would go further and state that it is that intellectual, emotional, social
and political context which is the real object of study of scholars of Indian
religion. Ultimately, it is people and their specific life-worlds that we are
attempting to understand.
Viewed from within this perspective, it is not surprising that much Indian
textual material can be represented in orthogenetic form, in terms of the
continual reworking of intellectual frameworks laid down in the early years
of the Brahmanical tradition. This material was, after all, almost all written
by scholars who were trained in the Brahmanical tradition of scholarship
or in the closely related Buddhist and Jaina variants of that tradition. In the
case of specifically Brahmanical scholars, demonstrating continuity with
Vedic texts was, as it were, part of their job description. For the Buddhists
and Jainas this was less true, but they were still working within the same
22
Origins of yoga and Tantra
intellectual universe, with common intellectual tools, and shared many of
the same basic assumptions. In such a world it is to be expected that, for
example, the Brahmanical and Buddhist scholars who wrote down the early
Tantras would do so in terms that replicated Vedic models and presented
the material in terms of familiar intellectual categories.
Yet to assume that this is all that was going on would be to neglect that
we all live in several intersecting cultural contexts. These same scholars were
involved in folk and domestic ritual, or its monastic equivalents. If living
in urban contexts, as most of them were, they had access, if they chose, to
people and ideas from a wide range of intellectual traditions. They were
almost certainly deeply involved in Tantric practice at an experiential level.
They might, as the great Nāgārjuna was said to have been, be specialists
in ritual magic, erotics and medicine as well as in Tantric philosophy and
practice. Above all, whatever they wrote had to make sense in terms of their
everyday lives, including the court and elite cultures of which they might
be a part, as well as in terms of the textual tradition to which they belonged.
These assumptions underlie much of what I write in this book.
Thus I proceed, where possible, by trying to reconstruct what we can
about the life-world within which a particular religious development took
place, including its social, cultural and political aspects, as an essential
background to what we can see in the surviving texts and other material
(iconography, archaeology). Such an approach tends to militate against
orthogenetic explanations, and indeed I try where possible to highlight
the complex and multidirectional influences that may underlie particular
development. This often makes for a more conjectural and provisional result
than some writers on these topics, but perhaps also for a more realistic one.
After all, our evidence for many of the developments I shall be considering
in this book is fragmentary and partial, and much of what scholars can say
is necessarily guesswork and conjecture.
the n ature of the evidence: archaeology,
n umismatics, iconography
In the following sections, I look at the nature of the evidence on which
our account is built. I begin with the archaeological evidence. While the
‘standard story’ was essentially constructed from texts, and archaeology was
called on, if at all, to support the texts, scholarship is increasingly moving in
the direction of taking the archaeology as primary and reading the texts in
relation to that account. This, it seems to me, is an essential move, though it
should be noted that it would scarcely have been possible for nineteenth- or
Stories and sources
23
early twentieth-century scholars, since it is really the development of modern scientific approaches to archaeology (e.g. radiocarbon dating) that has
begun to give us a moderately reliable chronological sequence for the prehistory and history of South and Southeast Asian societies. A large part of
the surviving textual material cannot be located at all securely in time and
space through textual criteria alone, and any attempt to derive a historical
sequence primarily on the basis of texts seems to me to be doomed to failure,
particularly for the pre-Gupta period (up to early fourth century CE).
At the same time, it is worth retaining some caution in relation to the
archaeological record. Relatively few sites in the geographically extensive
territories of South and Southeast Asia have as yet been subjected to modern
scientific archaeological techniques. Many major sites were excavated before
the development of such techniques, while many others have not been
excavated at all.
In addition, there are massive issues of interpretation with the archaeological material, particularly when it is attempted to bring it into relation
with information from textual and other sources. For example, the recent
move from seeing a radical break between the Indus Valley urban cultures
and subsequent developments towards stressing continuities between them
is undoubtedly significant, but it is far from clear precisely what it may mean
in relation to cultures on the ground (e.g. Kenoyer 1995). Cultural complexes defined in terms of pottery and other artefacts do not necessarily bear
a simple and straightforward relationship to actual social, cultural or linguistic groups. There is thus always a risk of over-interpreting the evidence.
As far as radiocarbon dating is concerned, there are also both general and
specific problems. Dates can be over-interpreted and taken too literally;
more specifically, the C14 calibration curve is more or less flat for a substantial and critical part of the first millennium BCE (c. 750–450 BCE;
cf. Magee 2004: 38), making it impossible to arrive at C14 datings for
this period with any degree of precision. The archaeological record, however, combined with radiocarbon and other modern dating methods, is the
nearest we have to solid ground in understanding the social and cultural
developments of which Indic religions formed part, and it needs to be
primary in our understanding of the situation.
Epigraphic and numismatic evidence also adds to our basis of information from the third and second century BCE onwards. Inscriptions are
important but unfortunately scanty and not always easy to date clearly.
There are nevertheless some very important bodies of inscriptions (for
example, land grants to Brahmins in Bengal, Orissa and South India are
all vital clues to the expansion of Brahmanical culture (e.g. Morrison 1970;
24
Origins of yoga and Tantra
Singh 1993). Numismatic evidence is also valuable, as with the early religious iconography on Indo-Greek and Kus.ān.a coinage (e.g. Srinivasan 1997:
pls. 16.6, 16.7, 16.9; Cribb 1997).
Iconographical evidence is also important and again provides a fairly
continuous record from around the third century BCE onwards. There
have been some impressive recent studies on the basis of iconographical
material: instances include Doris Meth Srinivasan’s Many Heads, Arms and
Eyes (1997) and Rob Linrothe’s recent book, Ruthless Compassion (1999),
both impressive examples of what can be done by combining iconographical
and textual material, though there are certainly outstanding issues in either
case (cf. Bakker 1999 on Srinivasan). But iconographical evidence has its
problems too. The dates for iconographical material are often uncertain,
since relatively few images are unambiguously dated. Also, many things
may be missing from the iconographical record, either because they were
created in perishable materials (images of wood, clay, cowdung, etc., which
may have been deliberately destroyed after the ritual as is still frequently
done) or simply because they have disappeared in the vagaries of history.
the nature of the evid ence: texts
In relation to texts, we undoubtedly have a very large body of texts relating
to the subject-matter of this book, Indic religions up to 1200 CE. In comparison, however, with European or for that matter Chinese history of the
same period, there is a great deal of uncertainty attached to the origins and
chronology of much of this material. The lack of early manuscripts and
the general fluidity of the textual tradition means that it is often difficult
to know exactly when large parts of many of these texts were written. In
the following section, I examine some of these issues, since they form an
important background to much of what follows.
Vedic and Brahmanical material
I start by looking at textual sources associated with the Brahmanical traditions (i.e. modern ‘Hinduism’).4 Our oldest substantial texts are undoubtedly, in some sense, the Vedas, both the hymns of the R.gveda itself, and the
associated material, including the Atharvaveda, Yajurveda, Brāhman.as and
Āran.yakas. The R.gveda Sam
. hitā (Sam
. hitā = textual collection) consists
4
For concise introductions, see Olivelle 1998: xxiv–xxxiii; Jamison and Witzel 2002; Kulke and
Rothermund 1990.
Stories and sources
25
of a body of 1028 ritual texts or hymns5 in the form of invocation to various deities, attributed to various inspired sages (r..sis) who are associated
with several identified priestly families. It is generally, no doubt correctly,
assumed by modern scholars to have been written over a period of several
hundred years, although its precise textual history is far from clear. The
Sāmaveda Sam
. hitā and Yajurveda Sam
. hitā are rearrangements of parts of
the R.gvedic hymns in the context of specific sacrificial rituals, including
additional ritual and explanatory material in the case of the Yajurveda,
while the Atharvaveda Sam
. hitā consists of a largely similar but mostly separate body of ritual hymns, many of them intended for healing and other
pragmatic purposes.
A further large body of Vedic material (the Brāhman.as and Āran.yakas)
indicate how the Vedic hymns were employed in ritual, how they were
performed, and how they were understood. This material, along with the
later hymns of the R.gveda itself and the Atharvaveda, is often interpreted
by modern scholars as deriving from a reconstruction of an original body of
Vedic religious practices, associated with a primarily pastoral and tribal way
of life, in the context of a new, settled agricultural, context (e.g. Heesterman
1985, 1993; Witzel 1995a, 1995b, 1997, 2003). It is possible that some of the
hymns of the R.gveda Sam
. hitā itself go back to the earlier period. In the
view of Witzel and others, these would reflect ritual practices of the IndoAryan-speaking peoples before and in the earliest stages of their arrival in
the Indian subcontinent.
The Brāhman.as and Āran.yakas were again clearly written over a long
period of time, and merge into another body of texts, the Upanis.ads, which
include further speculation on the meaning of the sacrificial ritual as well
as independent philosophical and metaphysical speculation. While some
of the material in the earliest Upanis.ads evidently predates the earliest
Buddhist texts, works which described themselves as Upanis.ads were written over a very long period, perhaps as late as the sixteenth century CE
(Olivelle 1996: xxxiii).
A further body of ‘Vedic’ texts are the Śrauta Sūtras, Gr.hya Sūtras, and
Dharma Sūtras, ritual and legal manuals that were probably written in
the second half of the first millennium BCE and are attributed to specific
Brahmanical authors. As we will see, these are often interpreted by modern
scholars as part of a second reconstruction of Vedic religion, parallel to
the development of the śraman.a traditions of the Buddhists, Jainas and
5
The term ‘hymn’ in this context is a Christian borrowing and while it has become widely accepted it
has misleading connotations. These texts were and still are used as ritual invocations, not occasions
for congregational singing.
26
Origins of yoga and Tantra
Ājivikas. All of this material can be referred to as ‘Vedic’, which leads to
a certain blurriness as to how the term is used by some modern writers.
The term is, however, significant, among other reasons because modern
Hinduism is often defined through its reference back to the Vedic texts
(e.g. B. Smith 1989: 13–20).
The hymns of the R.gveda Sam
. hitā have been given a wide variety of
dates, with a group of mostly Indian revisionist scholars nowadays arguing
for an extremely early dating (e.g. 4000 BCE; cf. Rajaram and Frawley
1995: 142–3). The Western scholarly consensus would probably place them
somewhere between 1200 BCE and 1000 BCE. While the oldest parts of
this body of ‘Vedic’ material undoubtedly go back in some form to before
my notional starting point of 500 BCE (the striking similarities between
the R.gveda and the parallel body of early textual material from Iran, written in a closely related language, make this clear), it is important to be
aware that our actual textual evidence for the Vedas, and for everything
else, is relatively late. The oldest surviving manuscripts for any of the
Brahmanical textual corpus, including the epics and the Purān.as, date from
around the eighth century CE, and many of the texts we are using are much
later.
The oldest layer of the parallel Iranian material consists of the seventeen
hymns or gāthās attributed to the semi-legendary figure of Zarathustra,
along with a short liturgy, the Yasna Haptanghāiti, from around the same
period or slightly later, both in the Old Avestan language. Both this and later
material in the Avestan language appears to have been written in Eastern
Iran (Boyce 1997; Skjærvø 1995).
Zarathustra’s historicity was generally taken for granted in Avestan scholarship until fairly recently, but it is by no means certain that he was the
author of the gāthās or the Yasna Haptanghāiti, or that he was a historical
figure at all. The Old Avestan texts, which were not written down until
the Sasanian period (third to seventh century CE), are nevertheless in a
language close to Vedic Sanskrit, and the gāthās are ritual hymns similar to
those of the R.gveda (Boyce 1997; Skjaervø 1997).
The state religion of the Achaemenid empire (sixth to fourth century
BCE), based in western Iran, was evidently closely related to the early
Zoroastrian texts, though it is not known to what extent, if at all, they
would have considered themselves as followers of Zarathustra. This religion
is represented by a body of inscriptional material from the sixth century
onwards. This provides evidence for a state ideology based on Zoroastrian
principles, including the idea of the king as a maintainer of justice and
divine order (Skjaervø 2005).
Stories and sources
27
The dating of the Old Avestan texts is as controversial and unsettled
as that for the Vedas, with figures for the lifetime of Zarathustra (and/or
the composition of the gāthās) ranging between 1400 BCE and 500 BCE.
The earlier dates are largely based on parallels with the Vedic material, the
dating of which as we have seen is also uncertain, but seem more plausible
than the later figure, based on the ‘traditional’ date for Zarathustra.
The one secure early date for something recognisably ‘Vedic’ comes from
West Asia, where the Mitanni archives preserve some names of recognisable
Vedic deities and other terms in a language identifiable as an early form of
Vedic Sanskrit in texts dating from around 1360–80 BCE (Thieme 1960).
Exactly what this early proto-Vedic group was doing in the region is a puzzle
to which a variety of speculative answers has been proposed (Parpola 1994:
126, 145–8, 1995, 1999a; Witzel 1995a, 2005). In fact the whole question of
the early history of the Indo-Aryan and Indo-Iranian speaking peoples is
both heavily contested and, at least at this point in time, largely undecidable
(see Erdosy 1995a; Bryant and Patton 2005). I discuss the Mitanni evidence
further in Chapter 5.
Returning to the Vedic material itself, the absence of early textual material
is worth underlining, since there is still a tendency to assume that the Vedas
represent early Indian religion in a relatively unproblematic manner.6 Many
of the texts of European religious history are of course equally based on
quite late manuscripts, but the European textual tradition does have many
early texts from Egypt and Mesopotamia, with relatively secure dates, and
significant parts of the Greek and Latin corpus survive in manuscript sources
from Roman and Hellenistic times. It also seems that there was considerably
more textual stability in Europe than in South Asia, so that we are not
generally in serious doubt, except with very early or primarily legendary
figures such as Homer, as to how much dates back to the original author
or compiler. The conditions of textual transmission in India, however, are
such that we rarely know with certainty which parts of a text date from
which period. This is not only true for the Vedic corpus, but also for later
texts such as the Ramāyan.a, the Mahābharata or the Purān.as, or the works
attributed to well-known authors such as Nāgārjuna or Śaṅkara.
It is possible to work on internal evidence in such matters, as Max
Müller and more recently Michael Witzel have for the Vedas, or John
Brockington for the Ramāyan.a. As Arvind Sharma pointed out some years
ago, however, we could be completely wrong about some of our basic
6
It would be more correct to say, as Olivelle does, that ‘[t]hese texts [. . .] offer us only a tiny window
into this period, and that too only throws light on what their priestly authors thought it important
to record’ (2003: 273).
28
Origins of yoga and Tantra
assumptions, including the periodisation of the major components of the
Vedic corpus, which after all were clearly edited and arranged in their
present form well after the time when many of the component items were
first composed (A. Sharma 1995).
In fact, dating is a major issue throughout a project such as this and I
will say more about it in a little while. An even more critical issue, and it
is one that I shall attempt to say something about later, is the question of
who used the Vedas. Our evidence increasingly suggests that the context for
the hymns of the R.gveda and other presumably ‘early’ stages of the Vedic
corpus was geographically localised in North-West India (roughly presentday Punjab, Haryana and Western Uttar Pradesh, as well as adjoining parts
of modern Pakistan). This raises the obvious question of what was going on
elsewhere in South (and Southeast) Asia at the time, and how Vedic culture
related to this wider context. Here it is important, as Parpola and others have
noted, to separate the question of the specifically Vedic culture associated
with the R.gveda from that of the usage of Indo-Aryan languages, which
may have been much more widespread. The worship of Vedic deities may
also have been present elsewhere in forms other than those of the R.gvedic
tradition.
Michael Witzel and others have suggested that the highly sophisticated
techniques used for memorising and transmitting Vedic texts, which have
a great deal of internal redundancy, have ensured that we have an extremely
accurate representation of the original – ‘[w]e can actually regard presentday R.gveda-recitation as a tape recording of what was first composed and
recited some 3000 years ago’ (Witzel 1995a: 91). While I do not have any
personal expertise with Vedic texts, there seem to me to be some problems
here.7 The elaborate mnemonic devices used, in which syllables are repeated
and permutated repetitively in complex formal patterns, doubtless did lead
to a high degree of textual stability, but since we have no independent
evidence as to how far back these techniques go, this does not necessarily
help us in determining the antiquity of any particular part of the textual
corpus, or for that matter in knowing how representative the surviving
material is of what might have been around at an earlier period. To me, some
at least of the Vedic hymns themselves look as if they originated in a society
considerably earlier than that which produced the mnemonic system; they
might have been passed on and transformed through many generations of
oral transmission before the mnemonic techniques were elaborated to their
7
Witzel is undoubtedly right that the R.gveda itself, and the early Vedic material in general, contains
numerous traces of its complex history, regional origins, substrate languages and the like. As a nonspecialist, I am unsure how precise a reconstruction can be made on the basis of these traces.
Stories and sources
29
present degree of complexity. Certainly other authors have taken a much
less sanguine position than Witzel regarding the antiquity of these texts,
with a number of authors arguing for the eighth century BCE or later (e.g.
Wright 1966; Pirart 1998).8 In any case, we need to be aware that the early
Vedic texts represent at best a narrow and selective window into a restricted
part of north Indian society in the first millennium BCE.
Apart from the various texts in the Vedic corpus, three texts associated
with named authors have been used as significant witnesses for early Indian
religion and religious history: the grammatical texts attributed to Pān.ini
and Patañjali, and the Arthaśāstra or treatise on statesmanship attributed
to Kaut.ilya. Unfortunately, yet again, none of these can be dated with
much reliability. I shall discuss the case of Pān.ini and Patañjali below,
as an example of the kinds of uncertainties prevalent within Indological
scholarship. As for Kaut.ilya, he is traditionally associated with the court
of Candragupta I (c. 300 BCE) but the linkage is far from certain and
the Arthaśāstra as we know it today is almost certainly a compilation from
somewhere between the mid-second and fourth century CE (Willis 2004:
25 n. 114).
Other very important sources include the two major Indian Epics
(Mahābhārata and Rāmāyan.a, to which may be added the Mahābhārata’s
appendix, the Harivam
. śa) and the Purān.as. In all these cases, we again
have to bear in mind the existence of a complex and uncertain textual history. Current scholarship includes respected authorities arguing that the
Rāmāyan.a and the Mahābhārata were each written over a period of three to
four centuries (e.g. Brockington 1985 for the Rāmāyan.a), and other equally
respected authorities arguing for authorship by a single person or group
of people over a few years (e.g. Hiltebeitel 2004). To some degree this is
a question of exactly what is meant by the authorship of the text, since in
either case the existence of prior oral versions is highly probable, but there
are also genuine disagreements. In any case, there is wide variation between
the existing manuscripts, particularly in relation to the Mahābhārata, suggesting that the texts as we know them developed over a long period of time
and represent material from a variety of periods and places. While there
8
‘La R.gvedasam
. hitā n’est pas aussi ancienne qu’on le dit. Les Perses, qui séparait les Mèdes des
Indiens et qui ont pu causer le déplacement de ces derniers d’Iran vers l’Inde, sont nommés. D’autre
part, le contenu des hymnes suppose la renaissance du mouvement urbain en Inde, si bien que la
date à retenir doit se situer apres la premiere mention des Perses dans les documents assyriens et la
reapparition des villes en Inde: le VIIIe siècle?’ (Pirart 1998: 521). Witzel on the contrary regards the
R.gveda as composed up to c.1200 BCE and redacted in the ‘Brāhman.a period’, perhaps around 700
BCE. The redaction involved selecting texts from existing collections; apart from minor details of
pronunciation, the text remained the same (Witzel 1995a: 91 n.13).
30
Origins of yoga and Tantra
has been a lot of very interesting and creative scholarship on both epics in
recent years, we are far from being able to give a precise historical location to
any individual section of either text (cf. Fitzgerald 2004; Hiltebeitel 2001,
2004; Brockington 1985, 1998, 2003a).
The Puran.as raise similar issues, more acute because they have received
less detailed attention. As far as I know, the oldest Purān.ic manuscript
we have is the early Skanda Purān.a MS on which Hans Bakker, Harunaga
Isaacson and others have been working. This manuscript dates from around
the eighth century, and the original text may go back to the sixth. This
version of the Skanda Purān.a is quite different in textual content from
the modern text of that name, although it was later incorporated as a
component of it under a different name (Adriaensen, Bakker and Isaacson
1994, 1998).
Overall, the dating of Purān.ic texts seems to be an extremely conjectural
area, which is hardly surprising since many of the Purān.as were being added
to and revised until quite recent times. Some of these variations can be quite
striking: does the Devı̄māhātmya, a relatively short section of one of the
Puran.as which is the classic early textual basis for the worship of a fierce
warrior goddess, date from the third or fourth century (M. Joshi 2002: 47)
or the eighth (Yokochi 1999: 89–90)? What are the implications for the
development of the cult of the fierce warrior goddesses in India if one
makes one choice or the other?
Other important bodies of texts (e.g. the Śaiva Tantric texts) raise similar
issues, with the dating, localisation and authorship both of surviving texts
as a whole and of individual parts of texts being still largely conjectural.
By the middle of the first millennium CE, many texts are attributed to
known authors, but this does not necessarily resolve issues of dating. The
dates of authors may be far from certain, and the attribution of texts to
authors is often equally problematic. Even as late and central a figure as
Śaṅkara raises major problems in this respect: his dates are far from certain
and many of the texts attributed to him are probably apocryphal.
All in all, we have a lot of material for the Brahmanical and associated
traditions but it often tends to be heavily rewritten and edited in the forms
we have, we do not know for certain when it dates from and we are not
always too sure what it means.
Buddhist and Jaina material
When we turn to Buddhist texts, the situation is in some ways considerably
better. The earliest Buddhist texts consist mostly of sūtras (Pali suttas),
Stories and sources
31
which are held to be accounts of the actual words of the historical Buddha
Śākyamuni, inserted in a short narrative describing the place and occasion of
their first delivery and those present at their delivery. These texts were held
to have been recited immediately after the death of the Buddha, which is
now dated by the majority of scholars (see below) to around 400 BCE. The
texts were eventually compiled into a number of collections: long sūtras (the
Dı̄rghāgama, Pali Dı̄ghanikāya), medium-length sūtras (Madhyamāgama,
Majjhimanikāya), two collections of shorter sūtras, and in some traditions
a group of miscellaneous texts (Ks.udrakāgama, Khuddakanikāya). All this
made up the Sūtra Pit.aka. Two further groups of texts, the Vinaya Pit.aka
and Abhidharma Pit.aka, contained the Vinaya or monastic disciplinary
code and the Abhidharma texts, presenting an early school of Buddhist
philosophical analysis.
The Buddhist tradition divided into a variety of schools over the next
few centuries, and what we actually have are versions of these collections
of texts as transmitted by particular schools, notably the Theravadins (in
Pali) and the Sarvāstivādin and Dharmaguptaka (in Chinese), along with
some mostly fragmentary texts in Sanskrit and Prakrit, again probably from
either the Dharmaguptaka or the Sarvāstivādin school. The collections as
transmitted by different schools do not correspond precisely to each other,
with different sūtras being included in some cases, and the texts of individual
sūtras also do not correspond precisely to each other, but changes are for
the most part relatively minor (e.g. Choong 2005).9
The existence of Chinese translations means that we have fairly reliable
early datable versions of parts of the textual corpus from around this period,
though this refers mostly to the Mahāyāna sūtras (eg. Harrison 1987a, 1993);
the Chinese versions of the āgamas date from around the fourth century
CE onwards.10 In some cases, in fact we have a series of dated translations
representing different Sanskrit texts from different periods. The Pali textual
tradition also seems to have been relatively stable from about the first century CE onwards, and some of the Sanskrit and Prakrit fragments contain
manuscript material from about the same period. We also have far more in
the way of historical writing from Sinhalese, Tibetan and Chinese authors
than from the subcontinent itself. However, as Stephen Collins pointed
out in his ‘On the Very Idea of the Pali Canon’, there are still unanswered
questions about exactly how and when the texts came to be written down
9
10
The recent discovery of a Sanskrit text of the Dı̄rghāgama is likely to throw considerable light on
early textual history (Cousins 2005).
Max Deeg (personal communication, February 2007). For translations of Tantric texts, see
Matsunaga 1977a.
32
Origins of yoga and Tantra
and how they came to be assembled into the relatively closed textual canons
we know today (Collins 1990). The Gandhārı̄ material suggests that there
was a substantial period when there were short anthologies of sūtras rather
than the larger and more systematic collections we know today, which may
not have taken shape until well into the first millennium CE.
Altogether, even if we accept the new chronologies for the life of the
historical Buddha, which would reduce the period in which versions of
the sūtras were being transmitted orally to perhaps only two or three centuries, we do not know and cannot know what the historical Buddha might
really have said in more than a very approximate way. It also seems likely,
as Richard Gombrich has pointed out, that people may have lost touch
quite soon with much of the necessary interpretive background of the texts
(Gombrich 1990, 1992, 1996). If Gombrich is right that we should read a
fair amount of humour and irony into the voice of the Buddha as represented in the early sūtras (the suggestion seems to me both attractive and
plausible), then this makes it even more difficult to interpret these texts as
literal accounts of the social and religious life of their times.
Jaina textual material has often received less attention than Brahmanical
or Buddhist texts, since relatively few scholars have worked on Jaina material, but it is of considerable importance. Early Jaina scriptural material
survived only in the Śvetāmbara tradition, and it is not clear when the
surviving texts were first written down; the existing manuscripts date back
only to the eleventh century CE. The standard version of the canon today
contains 45 texts, some of which undoubtedly go back in some form to
quite early times (Dundas 2002: 73–5). Thus the situation here is more
like that of the Brahmanical tradition than the Buddhist, in that the texts
contain quite early material but the surviving texts are all late.
The various issues with the textual material, much of which falls generally
into the category that we could refer to as scriptural, demonstrate why
a viable history of Indic religions has to look elsewhere for much of its
evidence, particularly for the first millennium BCE. The texts are vital, but
they are not self-explanatory.
chronological issues
It is worth looking briefly at what the textual material provides us with
on the chronological front. As noted above, our one secure textual date
for the second millennium BCE is provided by the Mitanni archives, and
this is essentially an isolated item of information which has not yet been
brought into any meaningful relationship with South Asian history. For the
Stories and sources
33
whole of the first millennium BCE there is only one reliable fixed point,
the invasion of Alexander in 329 to 325 BCE, which coincided with the rise
to power of Candragupta, the founder of the Mauryan empire. Everything
else – the datings for Aśoka (including the dates of the Aśokan inscriptions),
the Buddha, Mahāvı̄ra, the Upanis.ads and the guesstimates for the Vedic
texts – is inference and guesswork on the basis of this one figure.
The earlier Persian presence in Northwest India (Gandhara) in the sixth
century BCE left little that can be linked directly to later developments, but
there is some room here for speculation about the possible impact on stateformation and concepts of the state in the sub-continent (see Chapters 3
and 4). The relationship between earlier Iranian and early Indian (‘Vedic’)
material has received relatively little attention, despite the obvious connections, though there has been some work in this area of late, particularly
in relation to the vexed question of Vedic soma/Iranian haoma (cf. Parpola
1995; Houben 2003a, 2003b; Thompson 2003).
The general social and political chronology of South Asia becomes more
secure after the start of the common era, particularly once we reach the
Guptas (c. 320 CE onwards).11 The datings of many of the major religious
texts and figures of the first millennium CE however still remain less than
certain.
Given this overall situation, it is not entirely surprising that there is some
circularity to much Indological argument, particularly in relation to dates.
To quote Patrick Olivelle on the dating of the early Upanis.ads, ‘in reality,
any dating of these documents that attempts a precision closer than that
of a few centuries is as stable as a house of cards’ (Olivelle 1998: xxxvi).
This is one of the reasons why radical revisionist authors such as David
Frawley, N.S. Rajaram, Subhash Kak and others can advance alternative
chronologies in which, for example, the hymns of the R.gveda date back to
4000 BCE (e.g. Rajaram and Frawley 1995: 142–3) rather than the 1200–
1000 BCE suggested by Max Müller and broadly accepted by most modern
scholars, at least for the older hymns.
I do not find the contentions of the radical revisionists persuasive. The
astronomical datings that provide their most significant support are unconvincing, and such early dates would require a radical and implausible
rethinking of our understanding of both global history and of human
cultural processes. However, the fact that it is possible to advance such
contentions and for them to be taken seriously by many people, at least
11
The chronology of events in the preceding centuries is still somewhat unclear, although the dating
of the Kus.ān.as has now probably been placed on a reasonably secure basis (see Sims-Williams and
Cribb 1996).
34
Origins of yoga and Tantra
outside the academy, stands as something of a warning of the weakness of
the conceptual edifice that has been built up over the last couple of centuries
of Indological and Orientalist research.
In any case, while the redating of the R.gveda to 4000 BCE may not be a
real issue, there have been less dramatic but still quite substantial upheavals
in other areas. One of the most significant of these is the recent controversy
over the date of the historical Buddha. The traditional Theravadin date for
the Buddha’s death or parinirvān.a of 543 BCE had already been generally
brought forward by sixty years or so, to c. 486 BCE. In a series of conferences
and associated publications in the early 1990s, it became clear that the
scholarly consensus had moved considerably further forward again. Many
scholars now support a date for the Buddha’s death of c. 400 BCE or
somewhat later (Bechert 1991–1997; Cousins 1996a).
This is a move that has considerable implications not only for our understanding of the origins of the Buddhist tradition but also for the Jain and
Brahmanical traditions, since much of their early chronology too was determined in relation to the date of the historical Buddha. What has happened
here, though, as in many parallel cases, is not so much that the scholarly
community now knows that the Buddha lived 150 years longer than it had
originally thought. It is more that an accepted date has been replaced by
a question mark; we now know that what appeared to be solid knowledge
was in fact conjecture.
A case study of the chronological problem: Pān.ini
One has only to look at the uncertainties surrounding the dating of major
figures such as Pān.ini or Nāgārjuna, and the indirect nature of much of the
evidence on which our datings depend, to see that similar rethinkings might
well be likely in a variety of areas. It is worth giving an illustration of the
kind of situation we are in. Consider the evidence for the dating of Pān.ini,
the semi-legendary figure associated with the first surviving grammar, as
summarised some years ago by George Cardona (Cardona 1976: 260–8).12
To begin with, there is no direct evidence at all for the date of Pān.ini.
However, Cardona notes, Pān.ini has to be dated substantially before the
grammarian Kātyāyana, since Kātyāyana both commented on Pān.ini’s work
and had predecessors who had also commented on Pān.ini. Cardona suggests
that a period of one or two centuries would be appropriate.
12
I omit a whole series of arguments advanced by other scholars that Cardona found irrelevant or
inconclusive.
Stories and sources
35
Unfortunately, we have no substantial evidence for the date of Kātyāyana
either. Cardona cites a couple of arguments for his being in the mid-third
century, but clearly regards these as quite weak. However, Kātyāyana in his
turn has to be substantially before Patañjali (another one or two centuries)
for similar reasons. This Patañjali is the grammarian Patañjali, who may or
may not be the same person as the Patañjali who wrote the Yogasūtra.
Consequently, everything depends on the dating of the grammarian
Patañjali. Unfortunately, Patañjali is also not easy to date reliably. However, he provides in his grammar, as a grammatical example illustrating the
present tense, the sentence iha pus.yamitram
. yājayāmah., ‘We are officiating
here at Pus.yamitra’s sacrifice’. This may mean that he was writing during
the reign of Pus.yamitra, the first Śuṅga king (c. 178–c. 142 BCE), though
it is also possible that he is talking about a different Pus.yamitra or that the
sentence should be seen as merely a grammatical example. Patañjali also
gives grammatical examples suggesting that a yavana (Indo-Greek) king
attacked the towns of Sāketa and Madhyamikā during his lifetime. This
might be a reference to the Indo-Greek king Menander, between 140 and
120 BCE, though other dates are possible.
Cardona considers various arguments for later dates for Patañjali, which
he also finds inconclusive, and concludes ‘The evidence is thus not absolutely probative but sufficient to warrant considering seriously that Patañjali
lived in the second century B.C.’ (Cardona 1976: 266).
At this point we can return to Pān.ini, but now everything turns on how
long an interval needs to be assumed between Pān.ini and Kātyāyana, and
between Kātyāyana and Patañjali. Cardona outlines two possibilities:
(1) If Patañjali dates from the mid-second century BCE, and we allow
two centuries between him and Katyāyana and two centuries between
Kātyāyana and Pān.ini, then Kātyāyana would be mid-fourth century
BCE and Pān.ini would be mid-sixth century BCE.
(2) If we accept the mid-second century BCE dating for Patañjali, but
allow only one century between Patañjali and Kātyāyana and between
Kātyāyana and Pān.ini, then Kātyāyana would be mid-third century
BCE and Pān.ini would be mid-fourth century BCE. This would fit
with the other (very weak) evidence for Kātyāyana being mid-third
century BCE.
Cardona’s overall conclusion is:
The evidence for dating Pān.ini, Kātyāyana, and Patañjali is not absolutely probative
and depends on interpretation. However, I think there is one certainty, namely
that the evidence available hardly allows one to date Pān.ini later than the early to
mid fourth century B.C. (Cardona 1976: 268)
36
Origins of yoga and Tantra
But is this really the case? Cardona admits that Patañjali might be using stock
grammatical examples from an earlier period, so that he might not actually have been a contemporary of either the yavana attack or Pus.yamitra’s
sacrifice. Even if he was, the reference might also have been to another
yavana attack or to another Pus.yamitra, since there were several of each. It
hardly seems beyond question that further evidence might lead to significant further movement. In addition, the whole of this discussion assumes
that the present text of Pān.ini (or of Patañjali) was in origin the unified
work of a single author and was not substantially modified at a later period.
In reality, we do not have anything like a solid date for Pān.ini or even for
Patañjali.13
Now, I could have chosen any out of quite a long list of the significant
dates of South Asian history and gone through a similar exercise. I chose this
one primarily because Cardona lays out his arguments with such exemplary
clarity. It is also significant for another reason, however. The dates of Pān.ini
and Patañjali are themselves one of the most important sources for dating
many other developments – for example, references in Pān.ini and Patañjali
are important evidence for the growth of the Bhagavata cult and early
Vais.n.avism. It makes quite a lot of difference to our understanding of
early Indian religion whether Pān.ini lived in the fourth or the sixth century
BCE – to stay within the range that Cardona suggests. But a high proportion
of the dates we deal with, including many of those I give in this book, have
to be treated as equally uncertain.
It is striking how far Cardona’s arguments, and for that matter those
he cites and dismisses, are textual in nature. In this he is not unusual.
Indologists have tended to take the texts first – and often a fairly small
selection of texts – often treating what they said rather literally, and have
tried to fit everything else into the framework this produces.
conclusion
Thus any coherent account of the first millennium BCE has to depend
heavily also on the archaeological evidence. The archaeological record for
South Asia has its own problems, with many major sites either excavated at a
period before the development of modern scientific techniques in archaeology, or still unexcavated, but with the advent of modern dating techniques
13
Sheldon Pollock recently noted that ‘[a]rguments placing [Patañjali] as late as the middle of the
second century CE are entirely credible’ (2007: 80). This would allow for a considerably later dating
for Pān.ini than those suggested by Cardona; Pollock places him in the third or fourth century BCE
(2007: 45).
Stories and sources
37
it is at last becoming possible to construct a sensible account of the development of Indic civilisation on the basis of the archaeological record, and
to bring the various textual sources into some kind of relationship to it. It
seems to me that this is the only way in which to place the earlier stages of
the historical record on a reasonably firm basis, and to get a real sense of
what we know and what we do not know.
I do not want to be too discouraging here, because in some respects
this openness of the evidence seems to me to represent an opportunity as
well as a limitation. We have room now in a way that we did not perhaps
have fifty years ago to look for models considerably different from those we
have previously accepted. But I think that the situation requires a certain
tentativeness about our approach, and a willingness to accept that any one
kind of evidence is unlikely to be decisive. Instead, we need to work towards
a mutual coherence of our different kinds of evidence. The aim should be
to work gradually towards something which fits as well as possible all round
but which does not privilege any particular material – particularly textual
material, which is the most tempting, but in some ways the least reliable
in the South and Southeast Asian context.
part one
Meditation and yoga
In spite of a life devoted to meditation, prayer and books, Brahmagupta was like many other monks I have come to know and admire
who are perpetually cheerful, who laugh easily and loudly, the laughter
not springing from a sense of humour but from an evolved spirit of
mischief and playfulness. Sometimes, watching the frequent and obvious merriment of these monks, I have wondered whether the Buddha’s
message is indeed about the world being full of pain and sorrow; or
perhaps, the Enlightened One has left a secret message for his monks,
a cosmic joke which never palls with any number of re-tellings, which
makes them laugh so much.
Sudhir Kakar, The Ascetic of Desire (Kakar 1998: 157)
chap t e r 3
The Second Urbanisation of South Asia
The term ‘Second Urbanisation’ in the title of this chapter refers to the
growth of cities in South Asia during the first millennium BCE. The process is described as a ‘Second Urbanisation’ because of the existence of a
‘First Urbanisation’, that which was associated with the Indus Valley cultural
tradition. The major cities of the Indus Valley, including Mohenjo-Daro
and Harappa, are now generally dated to the period from c. 2500 to c. 1900
BCE, the so-called ‘Integration Era’ of the Indus Valley cultural tradition,
which followed on a long period of more localised, small-scale settlement
in the Indus Valley region (e.g. Shaffer 1992; Shaffer and Lichtenstein 1995;
Kenoyer 1995). ‘Integration’ here is used in a cultural sense, since the evidence does not at present suggest a high degree of political integration
for the region (Possehl 1998; Samuel 2000). In around 1900 BCE, the
major Indus Valley cities ceased to exist, and settlement size became much
smaller. This may be linked with the drying up of the Ghaggar-Hakra
river system, generally identified with the River Sarasvatı̄ mentioned in the
R.gveda, and the gradual desertification of much of the area which this
river used to irrigate. Around 75 per cent of the Indus Valley settlements
from the Integration Era are in fact along the banks or close to this river
system.
Substantial urban settlements did not appear again until about 500 BCE.
It is clear from the archaeological record however that there was considerable
cultural continuity after the end of the Integration Era, with localised
versions of the Indus Valley tradition continuing to exist in the Punjab and
elsewhere until about 1200 BCE.
It is I think important to establish as detailed a picture as possible of the
social and cultural aspects of the period of the Second Urbanisation as a
background to the development of religion in South Asia. An important
starting point is an understanding of the regional differences in North and
Northeast India at the period when the Second Urbanisation got under way.
41
42
Origins of yoga and Tantra
This will then give us a context within which we can place the dramatic
religious developments of this period.
Here there is evidence from a variety of sources, of three main kinds.
In the first place, there is the archaeological evidence, which is by now
quite extensive and detailed, and provides the basis for a reasonably secure
chronology through carbon-12 datings and other modern techniques. However, many important sites have still not been excavated in detail, and there
are technical issues with carbon dating, in particular the flat C14 calibration
curve for the period from c. 750 to c. 450 BCE, that mean that radiocarbon
dates give only very approximate dates for some critical periods (e.g. Erdosy
1995c: 100; Magee 2004: 38).
None of our textual material dates from as early as the start of the
Second Urbanisation, but some of it probably represents traditions that
go back close to this period, so that there is information here which needs
to be reconciled with the archaeological evidence. In particular, traditions
regarding the early Indian states or mahājanapadas, while legendary rather
than historical for this early period, still convey useful information, and
I will look at some of this material in Chapter 4. Another kind of textual
evidence comes from close textual analysis of the early Vedic corpus, as with
the work of Michael Witzel. Witzel’s material bears in part on underlying
language patterns in the surviving texts, including indications of ‘substrate
languages’, that is traces of languages that were spoken before the adoption
of Indo-Aryan languages in various parts of north and east India. Language
in general, both from early textual material and from the modern linguistic
record, provides a third body of material that needs to be reconciled with
archaeology and textual sources. Other sources of evidence, for example
iconography and numismatics, also provide useful clues at later stages in our
enquiry.
the second urbanisation in south and
southeast asia as a whole
First, however, it is useful to look at the whole question of the Second
Urbanisation in a slightly wider context than that of North India alone.
At the beginning of this period, both South and Southeast Asia were
very different from today, with a very much lower population density and
large areas still covered with forest. There were, however, agricultural communities throughout the region, with most of the major crops used today
already being grown. These were wheat and barley in the northwestern
areas, rice in the Central Gangetic and Bengal Delta regions, South India
The Second Urbanisation of South Asia
43
and Southeast Asia,1 and various millets throughout the region. While iron
did not form part of the Indus Valley civilisation or, apparently, of early
Vedic society, its use was also becoming fairly common by this period.
What did not yet exist in either South or Southeast Asia at this time is
anything that could be called a medium-sized town, or a medium to largescale state organisation. It is unclear in fact how far the First Urbanisation
in South Asia, the Indus Valley civilization involved a centralised state
organisation. Kenoyer, Shaffer, Possehl and others have suggested in recent
years that the Indus Valley societies were not in fact centralised states as we
know them today (e.g. Possehl 1998). Given the extensive trading networks
associated with the Indus Valley towns, we might think of them more as
trading communities with localised governmental structures, such as existed
in large parts of West Africa in quite recent times. We do not really know
very much about the social and political structures, though Possehl points
to the lack of evidence for war or militarism of any kind (Possehl 1998:
269, see also Samuel 2000). However, there is no doubt that the Indus
Valley civilisation supported relatively large communities: Mohenjo-Daro
and Harappa appear to have covered over 200 and around 150 hectares
respectively (Chakrabarti 1995: 61, 83) with populations of perhaps around
twenty or thirty thousand.
By comparison, the largest settlement sizes in the post-Harappan period
are around 10 hectares. It was from this basis that the so-called Second
Urbanisation began. Chakrabarti places the early growth of substantial fortified urban settlements, initially at Ujjayinı̄, Kauśāmbı̄, Vārān.ası̄, Rājagr.ha
and Campā, in around 600 BCE, though we need to remember that the
flatness of the C14 calibration curve for this period makes precise dating
impossible (Chakrabarti 1995: 247).
Why did it happen at this time? As usual, a complex development
like this probably had multiple causes, but one component may well
have been a significant rise in agricultural production resulting from
the introduction of multi-cropping techniques. This meant that the old
combination of agriculture and pastoralism typical of the Indus Valley
cultures and their successor states was shifted decisively in the direction of
agriculture. This could have provided the surplus in production that could
support a centralised state. It could also have set off the progressive clearing of forests, with consequent further rises in production and in population
1
Pejros and Shnirelman 1998 cite 7500–5450 BCE as the earliest date for Oryza indica in the Central
Gangetic valley, 4600 BCE for South China, 3250 BCE for Thailand (1998: 384, fig. 16.1). They also
note that ‘the Dravidians had borrowed rice-growing technology from a Sino-Tibetan group by the
end of the third millennium BC’ (1998: 385).
Origins of yoga and Tantra
44
PAINTED
GREY
WARE
NORTHERN BLACK
POLISHED WARE
EASTERN
GANGETIC
CULTURES
MALWA
AND
JORWE
CULTURES
Painted Grey Ware
Northern Black Polished Ware
Figure 3.1. Map: Painted Grey Ware and Northern Black Polished Ware
The Second Urbanisation of South Asia
45
(cf. e.g. Erdosy 1995c: 119–22). The introduction of iron has been suggested
as a major factor in the growth of cities, but Chakrabarti argues against
this, noting that
the village-life and the crop patterns of the entire subcontinent were laid down in
the earlier chalcolithic stage and that not a single major crop was added to the list
after iron was introduced [. . .] Given the firm-rooted agricultural base all over
the country, especially in the fertile alluvium of the Gangetic valley, the catalytic
factor in the early historic urban growth in the region is more likely the formation
of clear and powerful regional kingdoms which confront us right in the first phase
of our documented political history. (1995: 169)
A critical marker for early urbanisation has generally been taken to be
the advent of a new pottery style, known as Northern Black Polished
Ware (NBPW) and associated with the early urban settlements, initially
in the Central Gangetic region (Kauśāmbı̄, Sāketa etc.) and later elsewhere
(see Fig. 3.1). Dates for the beginning of NBPW again run up against the
flat C14 calibration curve problem: the majority of the radiocarbon dates
appear to cluster around 400–450 BCE and a plausible starting point for
this period might be around 550 BCE (e.g. Erdosy 1995c: 102–5; see also
Possehl and Rissman 1992), but Chakrabarti and others have suggested
earlier dates (Chakrabarti 1995: 169, 170; Magee 2004: 43). There are preceding pre-NBPW phases at some of these sites, going back perhaps to
around 1000 BCE.
These earlier phases are characterised by Painted Gray Ware (PGW),
which has been associated, in the Indo-Gangetic divide and the Doab, and
by Black Slipped Ware (BSW) further east. PGW, which appears from
around 1100 BCE in the Kuru-Pañcāla Region, has generally been linked
with the early Vedic-Brahmanical culture of that area. One should perhaps not put too much reliance on pottery styles. Erdosy suggests that the
emphasis on PGW in relation to the early Vedic material is a result of the
general poverty of material culture in the archaeological record; there is
little else to go by. He nevertheless comments
While it may be a mistake to equate the distribution of this ware [i.e. PGW] with an
effective social group, the coincidence of the territory of madhyadeśa, representing
the heartland of ārya orthodoxy, with it is striking. (Erdosy 1995b: 95, referring to
Witzel 1989: 243)
Dates for the early use of iron in the subcontinent are again contested,
though dates of around 700 BCE have been suggested for the north-eastern
region, with earlier occurrences (around 1000 BCE) in the south of India
(Chakrabarti 1995: 167–9). It has been plausibly argued that iron-smelting
may have been an indigenous development within the Indian subcontinent
(cf. Erdosy 1995b: 83–4).
Origins of yoga and Tantra
46
N O RTHWE S TERN
BO RDE RLANDS
CENTRAL
GANGES
PLAIN
W E STE RN
CO ASTAL
AREA
CENTRAL
INDIA
BENGAL
DELTA
MAHANADI
DELTA
KRISHNA
DELTA
KAVERI
DELTA
SRI
LANKA
Figure 3.2. Map: Areas of Early Urbanisation
The Second Urbanisation of South Asia
47
While the Second Urbanisation attained its first major development in
the central Ganges plain, it was not confined to this region. Bridget Allchin
notes that the new cities were located in a ‘a wide range of different regions’
(B. Allchin 1995: 13). Figure 3.2 shows the principal areas of early urbanisation in South Asia. Similar developments in Southeast Asia took place
somewhat later. The use of iron became common in the northern parts of
Southeast Asia from around 500 BCE onwards, and there is growth in settlement size and some indications of political centralisation in areas such as
northern Vietnam, which was the centre of the Dong-son culture (Bellwood
1999: 94), from this period. The first real Southeast Asian polities, however,
seem to date from around the second to fourth centuries CE.2 These states
were closely linked to the developing maritime trade with India and China,
and some of them became early centres of Indic culture.
To return to South Asia, the early political units in the northern part of
the subcontinent are the so-called mahājanapadas. The term derives from
an earlier Vedic term janapada for the territory of a Vedic tribe. There are
classical lists of sixteen mahājanapadas (see below and Fig. 3.3), and they
are described as having a variety of different political structures.3 Further to
the south, the river deltas and coastal plains of South India and Sri Lanka
were also early centres of urbanisation and political centralisation.
It is useful to have this larger picture in mind. While the Ganges plain
was the location of the majority of the mahājanapadas, and later became the
centre of the first large Indian state, Magadha, and the first Indian empire,
that of the Mauryas, India’s Second Urbanisation in India took place over a
number of different regions during more or less the same period. These were
not independent of each other or separate from the outside world. Thus,
for example, the question of Iranian influences on political and religious
developments in the Gangetic plain, which has been raised on and off over
the years (Wright 1966; Boyce 1997), is by no means implausible, especially
given the Achaemenid control over the North-Western region from around
520 BCE onwards.4
2
3
4
McGee 1967: 29–41; Wheatley 1979, 1983; Bronson and White 1992; Hall 1999: 192; K. Taylor 1999.
Here again a critical feature seems to be agricultural productivity. (‘The area around Oc-eo is the
only place near the coast west of the Mekong where a particular kind of topography and soil enabled
its people to grow significant amounts of rice with the existing pre-canal irrigation wet-rice methods
of cultivation’, Hall 1999: 192). The capital of the Oc Eo state, a few miles inland, was known as
Vyadhapura.
As Thapar points out, the original meaning was something like ‘the place where the tribe places its
foot’. The janapada was normally named for the ruling clan within the tribe (Thapar 2003: 121).
The suggestion surfaced again in early November 2002 on the INDOLOGY e-list: see contributions
by Michael Witzel and others (http://listserv.liv.ac.uk/cgi-bin/wa?A1=ind0211&L=indology).
48
Origins of yoga and Tantra
The Central Gangetic plain was one of the two most significant regions
in relation to early religious movements. This was where the early ascetic
movements, including the Buddhists and Jains, took shape, and it was also
a very important area for the Upanis.ads and developments in Brahmanical
traditions. The other significant region in relation to early religion was the
area to the immediate west and north-west, including present-day Western
UP, Haryana and Punjab. This was the region associated with the early
Vedic texts and was clearly the heartland of the early development of the
Vedic and Brahmanical tradition in India. By contrast, this region initially
took a much more marginal role in relation to the Second Urbanisation.
The population (or at least the ruling elite) in this region, which I refer
to, following Witzel, as the Kuru-Pañcāla Region, after the names of the two
classical political units (mahājanapadas) most closely associated with it, saw
itself as ‘Aryan’ (ārya) by contrast with surrounding peoples and used Vedic
Sanskrit for ritual purposes. We do not have any meaningful information
about the ethnic composition of the region at this stage, but ‘Aryan’ here
is best interpreted as a cultural rather than an ethnic or linguistic label (cf.
Deshpande 1995). As for language, it seems likely that by 500–400 BCE
the population of the fertile lowland areas of most of present-day North
India and Pakistan were already speaking Indo-Aryan languages, so that
the contrast between the Aryan heartland and the areas to its south and
east was not primarily language-based (Deshpande 1995; Parpola 1988).
t wo different worlds: kuru-pa ñc āl a and the
central gangetic region
In a recent unpublished work (Hopkins 1999), the American Indologist
Thomas Hopkins has emphasised the cultural differences between these
regions. Rather than seeing Vedic-Brahmanical culture as expanding outwards into areas which were culturally and technologically at a much lower
level, he develops a picture in which the Kuru-Pañcāla region in around
500–400 BCE was surrounded to the east and south by a group of related
cultures with a complex and sophisticated cultural adaptation as well as
probable links to the Iranian world via Gandhara. A detailed presentation
of Hopkins’ arguments would be inappropriate here, and in any case I hope
that Hopkins will publish his own work in this area in the near future.5
5
Hopkins was in Oxford in 1999 as director of the Oxford Centre for Vaishnava and Hindu Studies.
He was kind enough to lend me a copy of his unpublished manuscript at that time, and we have had
further discussions regarding this material (Hopkins, personal communication, 2006).
The Second Urbanisation of South Asia
49
However I will sketch some of the main features of his model, since it
brings together much of the archaeological and textual material in a useful
synthesis. In addition, as we will see, there is further material in support of
several parts of his picture.
In agreement with most recent authorities, Hopkins regards the PGW
pottery, beginning in around 1100 BCE, as associated with the VedicBrahmanical complex in the Kuru-Pañcāla region, and as overlapping
in some areas at least with late settlements of the Indus Valley Cultural
Tradition. He notes however that if the PGW culture ‘defines the region of
Aryan control and primary influence, it also marks its limits’ (1999:14). More
specifically, it indicates that this culture did not expand past the GangaYamuna Doab, stopping before the ‘heavily forested eastern tip where the
Ganges and Yamuna converge’ (1999: 33).
What then was happening in other parts of north and central India?
Looking first at the northern and western Deccan areas, to the south of the
Ganga-Yamuna valleys and the PGW region, Hopkins stresses the different
ecological conditions here from that in the ‘western and northern alluvial
plains’ (1999: 17). A variety of subsistence strategies developed in these
upland areas during the period of the Indus Valley traditions, coalescing
into the so-called Malwa and Jorwe Cultures (beginning around 1700 BCE
and 1400 BCE respectively) in the northern and western Deccan (cf. Figure
3.1, and Possehl and Rissman 1992: I, 486–7). These peoples grew barley and
millet, lentils, rice, and a variety of fruits, and the archaeological material
from these sites shows no traces of influence from the ‘Aryan’ areas to the
north. The region remained important (as the kingdom of Avanti) in the
time of the Buddha, and its major cities, including Vidiśā and Ujjayinı̄,
were significant centres during the following centuries.
The Central Gangetic region was characterised by a distinct but related
cultural complex. This was the area of the earliest known cultivation of
rice in South Asia and by 1800 BCE was the location of an advanced
Neolithic population associated with the sites of Chirand and Chechar who
appear from their subsistence repertoire to be associated with the Malwa
culture and the Deccan. Hopkins notes that this culture ‘had reached the
Chalcolithic stage by the time the Aryans came into northwest India [. . .]
and entered the Iron Age at least as early as any known sites in the Doab –
i.e. by around 700 BCE’ (Hopkins 1999: 32–33).
Thus, Hopkins argues, we should revise our model from one of an Aryandominated expansion into the Central Gangetic region, bringing with it
Iron Age culture and urbanisation, to a model of
50
Origins of yoga and Tantra
two cultural processes moving more or less concurrently toward the use of iron
and urbanization from two separate sources: one in the eastern Punjab, Rajasthan,
the Doab, and northward to the Himalayas west of 81◦ longitude, identified with
the Painted Grey Ware culture and the Aryans; the other – based on the Eastern
Gangetic culture with its apparent initial connection to the Malwa-type cultural
complex – in the region of Patna, in the valleys of the Ghagara and Gandak rivers
northwest of Patna, and westward to the region around the lower Doab. (Hopkins
1999: 34)6
Hopkins goes on to note specifically that the urban sites of Śrāvastı̄ and
Sāketa (Ayodhyā) were outside the PGW region ‘and had only marginal
contact with PGW culture before their inclusion in the iron-using Northern
Black Polished Ware culture around 700/600 BCE’ (Hopkins 1999: 36).7
This later (NBPW) culture was associated with the expanding urban culture
of the Central Gangetic region and the early states of Kosala and Magadha.
Hopkins notes the association of this region with the story of the Rāmāyan.a,
a theme to which I shall return later. He also points to the contacts from
the early stages between the NBPW culture and the Achaemenid-ruled
territories in the Northwest (1999: 48–52). These are probably associated
with the early adoption of coinage in the Central Gangetic region and with
trade routes between the Central Gangetic region to the Deccan and via
Mathurā to Gandhāra and the Northwest.8 Hopkins stresses the difference
between the ‘new political-economic culture’ being created in these regions,
a culture within which the Buddhists and Jains played an active role, and
the Vedic-Brahmanical culture of the Kuru-Pañcāla region:
To say that there was tension between these two worlds – the non-Aryan/
Buddhist/Jain world on the one hand, and the Aryan/Brahmanical/Vedic world on
the other – is to understate the case. In the sixth century BCE they were really two
different worlds, at least as perceived by their main representatives, or perhaps –
as seen especially by the Brahmans – two opposite worlds. (1999: 53–4)
If the Kuru-Pañcāla region was dominated in religious terms by the progressive reworking of Vedic material represented by the Brāhman.as, the
6
7
8
Hopkins refers to the ‘Eastern Gangetic’ culture in the following quote and elsewhere; I have preferred
to speak of the ‘Central Gangetic region’ throughout this book, partly for consistency with other
authors, and partly to distinguish it from the Bengal Delta region further to the east. We are however
referring to the same area.
Unfortunately, this is the period for which radiocarbon dating is least precise (see e.g. Magee 2004).
Hopkins does not regard the new model of kingship itself as a borrowing because he is still using the
‘old’ dating of the life of the Buddha and consequently dates the growth of Magadha to around 540
BCE. If we move this forward to 420–400 BCE (cf. Bechert 1991–7; Cousins 1996a) direct influence
on adminstrative structures becomes more plausible, though as Kulke and Rothermund note this
still does not explain the success of Magadha in relation to its rivals (Kulke and Rothermund 1990:
57).
The Second Urbanisation of South Asia
51
Central Gangetic region and the northern Deccan was by contrast a ‘world
of female powers, natural transformation, sacred earth and sacred places,
blood sacrifices, and ritualists who accept pollution on behalf of their community’ (Hopkins 1999: 55). Here Hopkins’ evidence derives in part from
the mythology and legends associated with the Central Gangetic region,
above all the Rāmāyan.a, and in part perhaps on later ritual patterns that
would not appear to derive from Vedic sources. We can also point here to
extensive iconographical evidence for a religion of fertility and auspiciousness throughout this region (Central Gangetic, Mathurā, Gandhāra) from
the Mauryan period onwards. I shall discuss both this material, and the
evidence of the Rāmāyan.a and the Mahābhārata, which developed out of
the corresponding Kuru-Pañcāla legends, in subsequent chapters.
th e vedic m aterial: the view f rom kuru-pa ñc āl a
Working from a very different body of material, the Vedic texts themselves,
Michael Witzel, Jan Gonda, Jan Heesterman, Harry Falk, Patrick Olivelle
and others have produced a picture of the growth and expansion of the
Kuru-Pañcāla cultural complex. Some of this will be best considered in
subsequent chapters, but an introductory sketch will be given here, with
particular reference to Witzel’s work.
Witzel’s argument about the Kuru-Pañcāla region and its relationship to
the ‘Vedic’ element within Indian society has several components. Basic to
them is a revised periodisation of the so-called Vedic literature, the body
of literature which includes the hymns of the R.gveda and Atharvaveda and
a substantial body of ritual commentaries, philosophical texts, and ritual
and legal manuals and treatises, which are generally regarded as the oldest
layer of surviving Indian literature.
Witzel divides the language in these texts into five sub-periods, which fall
within three main periods, Old Vedic, Middle Vedic and Late Vedic (e.g.
Witzel 1989, 1995a, 1995b, 1995c, 1997; Jamison and Witzel 2003). The most
important contrast from our point of view is between Old Vedic, which
corresponds to the hymns of the R.gveda, and Late Vedic, represented by
the later Brāhman.as, Āran.yakas, early Upanis.ads and most of the Sūtras. So
far, this may seem relatively conventional, but Witzel’s intention is to give
this material a historical placing in the social and political developments of
the time.
Specifically, Witzel argues that the geographical context of the Old Vedic
material is Afghanistan, the Punjab and the surrounding areas up to the
Yamunā. In other words, this material goes back to the period of and
52
Origins of yoga and Tantra
immediately before the arrival of the Vedic peoples in North India. The
society of that time, in Witzel’s view, consisted of around fifty small tribes in
constant conflict with each other and with other peoples, whom he refers to
as ‘aborigines’ (Witzel 1995c: 4–5). ‘Aborigines’ is a problematic term in this
context; Witzel uses it as equivalent to ‘servant/slave (dāsa, dasyu, purus.a)’
and evidently assumes a model in which an invading elite has enslaved an
indigenous population (1995c: 5).9
The leading role in the tribes was taken by chieftains (rājan), belonging
to an aristocratic group (the rājanya or ks.atriya) who dominated the viś or
mass of the people. In Witzel’s account, it is under the leadership of one
of these tribes, the Pūru subtribe of the Bharata tribe, that the Kuru tribal
federation in the Haryana-West UP region was formed. Witzel dates the
coming of the Kuru to power during the period from c. 1200 to c. 900 BCE.
In the subsequent period, from c. 900 to c. 500 BCE, leadership in the
region shifted to their allies, the Pañcāla, and expansion continued to the
south and west.
The period of dominance of the Kuru and then of the Pañcāla corresponds to Witzel’s Middle Vedic period and the establishment of the four
Vedas in more or less the form we know them today. Witzel sees this latter
process as closely linked to the evolving structure of the Kuru state. It is in
this period, too, that we would need to place the transformation of the Vedic
sacrifice from a competitive performance by tribal chieftains to the royal
ritual of the Kuru kings (cf. Heesterman 1985, 1993; Smith 1989: 36–45).
Witzel’s Late Vedic period corresponds to the somewhat tentative ‘500
BCE’ that forms the starting point for my own narrative. The compass of the
Late Vedic texts is all of Northern India from Gandhāra to Bengal, and south
to what is now Northeast Maharashtra and Andhra. I think that we need
to bear in mind however that this region would have been far from evenly
‘Āryanised’ at this time. This issue will be explored in later chapters in relation to the role of Brahmins in Kuru-Pañcāla and elsewhere in north India.
The Kuru state according to Witzel was actually a dual state made up
of the Kuru and Pañcāla tribes. He regards it as the first large state in
the post-Indus Valley period (1995c: 9), with its origins going back to the
immediately post-R.gvedic period.10 The Kuru people regarded Kuruks.etra
as the sacred centre of their realm. As Kuru-Pañcāla power increased to
9
10
‘Aborigines’ implies a population that is in some sense indigenous. The local population is likely to
have included Indo-Aryan-speaking elements, if we assume a two- or multi-wave migration model
(e.g. Parpola 1988), and Dravidian speakers may also be relatively recent immigrants (McAlpin 1981).
For Parpola’s later thoughts on the multi-wave model, see Parpola 2002a.
See the Kuntāpa rituals, carried out at the winter solstice on the mahavrāta day at the end of the
one-year rite. This occasion was a fertility ritual as well as a ritual to help bring back the sun and
some of the hymns are also about royal fame and power (1995c: 7).
The Second Urbanisation of South Asia
53
absorb R.gvedic tribes in the eastern Punjab, Haryana and Western UP,
Kuruks.etra became the ‘center of the newly emerging Vedic orthopraxy
and “orthodoxy”’ (1995c: 8). The Kuru-Pañcāla state undertook a wellplanned process of Sanskritisation11 (1995c: 9), including the creation of the
new complex Śrauta ritual, which allowed various ranks of society to claim
status through appropriate levels of ritual performance, and which also
included some major new royal and ‘national’ rituals (1995c: 18). Another
result of this process was that non-Aryans,12 who had previously been able
to follow Aryan religion, were now only admitted to Vedic society as Śūdras,
unable to perform ritual (and so unable to enter heaven). ‘The effect was
the creation of a permanent . . . artificial boundary between Aryans and
non-Aryans’ (1995c: 10).
The Kuru-Pañcāla state was also responsible for the collection of the
old ritual hymns of the R.gveda, until then still associated with individual
Brahmin families, and their compilation into a standard anthology that
forms the basis of the R.gveda as we know it today. They also produced
the original forms of the other three Vedic collections. All of these were
edited and compiled, according to Witzel, in forms that suited the new
Kuru political programme (1995c: 14–15).
The Kuru idealogues developed a new mythology for the Kuruks.etra
region (1995c: 15–16), in which the Sarasvatı̄ river, equated with the goddess
Sarasvatı̄ and the Milky Way, flowed down from heaven in Kuruks.etra at
the time of the winter solstice. Kuruks.etra was conceived of as the centre of
the world. Essentially, the Kuru state created the Aryan sense of ritual and
cultural superiority to surrounding peoples that was to become a central
component of later Brahmanical culture.
linguistics
Apart from the limited information that can be derived from possible
traces of ‘substrate languages’, the history of the replacement of indigenous
languages by Indo-Aryan languages in North India has left little by which
11
12
It should be noted that Witzel is using the term ‘Sanskritisation’ here in a very different sense from that
well-established in the anthropological literature, which refers to a voluntary adoption of high-status
cultural practices by low-status groups in an attempt to raise their status. This anthropological usage
goes back to MN Srinivas (Srinivas 1952). Witzel appears to be referring instead to an imposition
of a cultural pattern from above. Witzel also uses the term to refer to the expansion of Brahmanical
culture to other regions; I discuss this usage in Chapter 4.
I should stress again that this did not necessarily mean ‘indigenous’ people as opposed to IndoEuropean speaking populations. If we accept something like the Parpola model of several waves of
Indo-European-speaking migration (Parpola 1988), and assume that this was followed by a gradual
replacement of non-IE languages by IE languages, then non-Aryans by this period could be IE or
non-IE speaking, and of ‘indigenous’, ‘immigrant’ or (most likely) mixed origins.
54
Origins of yoga and Tantra
it can be traced (Deshpande 1995).13 What is striking about the Upanis.adic,
Buddhist and Jaina texts is that there is little or no suggestion of non-IndoAryan languages being spoken in the Eastern Indian regions where much
of these narratives are placed.14 Arguments from silence are always risky,
but one would imagine that if Munda, Dravidian or other non-Indo-Aryan
languages were still being widely spoken in the major centres of population
at the time of the Buddha or Mahavı̄ra (say c. 425–400 BCE), they would
have left more significant traces. Languages do not disappear overnight,
and it seems likely that the shift to Indo-Aryan languages had got going
several centuries before the time of these texts.15
There is little real evidence in support of the suggestions that the IndoAryan languages originated in India, and strong arguments against the
idea (see Bryant and Patton 2005 for a presentation of arguments on both
sides). Thus, if one assumes that the Indo-Aryan languages were brought
into South Asia from the outside, as linguistic and other evidence strongly
suggests, then this process was presumably well underway by 1200 BCE,
and had been largely completed by the time that the Second Urbanisation
got under way. I would have to agree, however, that we do not have any
very precise knowledge of how and when the population of large parts of
the Indian sub-continent became speakers of Indo-Aryan languages. The
most plausible scenario would have Indo-Aryan-speaking groups moving
into South Asia from the Iran-Afghanistan region from around 1500 BCE.
13
14
15
To the extent that we can trace pre-Indo-Aryan substrate languages in North and North-East India,
they do not seem to be Dravidian. They either show Austroasiatic features, as do many of the
surviving tribal languages, or in a few cases relate to no currently known language group.
As opposed to occasional suggestions of mispronunciation of Indo-Aryan languages.
Whether the population of the main Indus Valley cities spoke an Indo-Aryan language during the
Integration Era has not been decided conclusively and is unlikely to be resolved without a convincing
decipherment of the Indus Valley seal inscriptions. Probably the most scholarly attempt to decipher
the inscriptions is that of Asko Parpola, who assumes a Dravidian basis for the language, but it
can hardly be regarded as decisive (Parpola 1994). Steve Farmer has produced arguments recently
for the seal-inscriptions being non-linguistic (Farmer, Sproat and Witzel 2004). It is also entirely
possible that the Indus Valley tradition encompassed groups speaking a variety of languages. Given
the presence of a Dravidian language (Brahui) in the region, a Dravidian-speaking population would
nevertheless seem plausible. David McAlpin’s Proto-Elamo-Dravidian hypothesis, which assumes a
common ancestor for the Elamite languages of Western Iran and the Dravidian languages would
make this more likely, but this hypothesis is again far from universally accepted (McAlpin 1981;
Parpola 1994: 128). An interesting aspect of McAlpin’s argument is that he argues on lexical grounds
that Dravidian and Elamite separated from each other at a relatively late period (perhaps in the
fifth millennium BCE) and that this took place in West Asia. This would suggest that Dravidian
speakers were not indigenous to the Indian sub-continent, and may indeed have arrived there not
very long before Indo-Aryan speaking populations. McAlpin suggests that the reason why Sinhalese
is an Indo-Aryan language rather than a Dravidian one is that Indo-Aryan speakers, moving down
the east coast, in fact got to Sri Lanka before the Dravidians, moving down the west coast (McAlpin
1981).
The Second Urbanisation of South Asia
55
This would seem to have been a gradual process over several hundred years
in which various tribal groups progressively trickled in. The starting point
for the process is less clear, though there are arguments for identifying
sites in Swat towards the end of the Indus Valley Localisation period with
Indo-Aryan speaking populations (Stacul 1992; also see Erdosy 1995b).
By the time of the Second Urbanisation, however, the general distribution of Indo-Aryan languages in North India and Pakistan was perhaps not
all that different from the distribution in modern times. There was also
considerable differentiation among the Indo-Aryan-speaking population.
While there may have been some presence of Brahmins and of Vedic deities
throughout the region, the specific body of ritual hymns and practices
associated with the R.gveda grew up among a self-consciously distinctive
group in the Punjab-Himachal Pradesh-Western Gangetic region, in the
region which was at a later stage the location of the Kuru and Pañcāla
peoples. This community saw itself as ‘ārya’ and as distinct from surrounding peoples, especially those to the east, whom it regarded as outside the
‘Aryan’ fold. These were areas where Brahmins from the Kuru-Pañcāla
region should not venture. If they did, they had to be purified upon return.
There are a number of ways in which one can read this evidence, assuming
that it is accepted that Indo-Aryan languages did not begin to reach South
Asia until the end of the Indus Valley urban period and the succeeding
centuries. The principal alternatives are to assume that a number of already
distinct waves of immigrants arrived in South Asia, so that the distinction
between ‘Aryans’ and others reflects pre-existing divisions, or to suppose
a relatively homogenous Indo-Aryan population moving into South Asia
and subsequently differentiating into ‘Aryan’ and ‘non-Aryan’ components.
The former position has been taken by Asko Parpola, who has produced
a series of studies postulating increasingly complex migratory patterns to
make sense of the archaeological and textual material (Parpola 1988, 1995,
1999a; Parpola and Carpelan 2005).16 F. R. Allchin has argued for a more
homogenous population, subsequently differentiating in South Asia, while
suggesting that the difference between the models is not all that significant
(F. R. Allchin 1995b: 50–1; cf. also Erdosy 1995b).
As far as our present purposes are concerned, we can agree with
Allchin that the difference between the models is not critical. The difference between the self-consciously ‘Aryan’ region and the surrounding
populations is however of considerable importance. On an Allchin-type
model, this may reflect differing relationships between Indo-Aryan speaking
16
According to Michael Witzel 1995b, the Inner and Outer Wave idea goes back to Hoernle in 1880.
56
Origins of yoga and Tantra
immigrants and local populations, with a higher proportion of Indo-Aryan
speakers in what was to become the Kuru-Pañcāla region and lower proportions elsewhere. It doubtless also reflects specific local and regional histories.
In the period described by the early Buddhist texts, as we will see, certain
fundamental assumptions were shared by the Aryans and the populations
to the south, east and west. Elite groups in both regions saw themselves
as ks.atriyas and traced their descent to a common body of early mythic
figures. Brahmins of some kind were present throughout both regions, and
were providing priestly services to ruling groups. Indra and Brahmā were
recognised deities throughout the region. Yet there were also notable differences, and the Brahmins had clearly achieved or maintained a much higher
status in the Kuru-Pañcāla region than in surrounding areas.
the m a h ā j a n a pa d a s
The political map of this period according to the textual material consists
of the so-called mahājanapadas. The political situation is described in texts
of slightly later periods from Brahmanical, Jain and Buddhist traditions
(see Fig. 3.3). The mahājanapadas were medium size political units, mostly
consisting of one or two major towns and a rural hinterland. Traditionally
there were sixteen of them, though somewhat different lists survive in
various Buddhist, Jain and Brahmanical texts, perhaps reflecting different
times or the perspectives from various parts of the subcontinent. In the time
of the historical Buddha, these had been absorbed into a small number
of expansionist states, which were to be incorporated shortly after into
the first large-scale empire in Indian history, that of the Mauryas. Thus
by the time that the Buddhist sūtras were committed to writing in the
second and first centuries BCE, the mahājanapadas had long disappeared
into history. Much the same is true in regard to Jaina references to this
period.
The Upanis.ads say little directly about the mahājanapadas, reflecting
their historical mise-en-scène in a largely legendary period of early Vedic
sages, though the earliest texts in this corpus do refer to a couple of kings,
Janaka of Mithilā and Ajātaśatru of Kāśı̄. Pān.ini’s grammar also contains a
considerable amount of geographical information, including references to
many of the mahājanapadas. As I noted in Chapter 2, Pān.ini’s dating is very
uncertain, but his references are generally treated as describing a somewhat
earlier situation than the Buddhist material.
Erdosy refers to the mahājanapadas as ‘city states’ (Erdosy 1995c), and like
the Greek city-states they exhibited a variety of political structures, though
The Second Urbanisation of South Asia
57
OJA
MB
KA
U
Pus.karāvatı̄ G A N
D H ĀRA
Taks.aśı̄lā
LA
Mathurā
Śrāvastı̄
VIDEHA
A LA
OS
K
SE
L A VAJJI
Sāketa
NA
Vaiśālı̄
VA
A
T ¯S´ I¯ Vārān.āsı̄ Pāt.aliputra
Campā
Kauśāmbı̄
HA ANGA
.D
. RA
AD Rājāgr ha PUN
.
AG
MA
ŚŪ
L
K
A
C H
RA
A
M
RĀ
AVANTI
S
CETIY
A
C
S
RU
A
N
PAN̄CĀ
H
KU
M
I
D
A
DH
Ujjayinı̄
Ā S
M
AK
A
Figure 3.3. Map: The Mahājanapadas
VAN.
GA
58
Origins of yoga and Tantra
the analogy with the city states of classical Greece at around the same time
may be misleading. Certainly they were quite varied in political structure
and economy. In the Buddha’s time, they included four substantial kingdoms (Magadha, Kosala, Vatsa and Avanti), though all of these were incorporated into Magadha during or shortly after his lifetime. So were a variety
of smaller states that employed a system known as gān.a-saṅgha, generally
referred to in English as ‘oligarchies’ or ‘republics’. These were governed
by a council of clan heads or members of leading families (J. Sharma 1968;
Thapar 2002: 146–50). The largest of these, the Vajjian confederacy, was
actually a league of several such units and its capital, Vaiśālı̄, was one of
the main cities of the Central Gangetic region in the Buddha’s time. The
Buddha is described as commending the gān.a-saṅgha system and it was
evidently a significant influence on the early development of the Buddhist
and Jaina orders.
A brief survey of the individual mahājanapadas may be useful. Kuru and
Pañcāla together made up the core region of the self-consciously ‘Aryan’
region associated with the compilation of the R.gveda and other early Vedic
texts (Brāhman.as, etc.). For Witzel, the Kuru state was the original focus
for the creation of the Vedic Aryan model of society, in the early part of the
first millennium BCE. By 500 BCE, this area seems to have been something
of a backwater politically and economically, though it may have continued
to be an important centre of Brahmanical influence.
Kosala and Magadha were the major kingdoms in the Central Gangetic
region in the time of the Buddha. Kosala’s main towns were Śrāvastı̄ and
Sāketa (the later Ayodhyā), Magadha’s were Rājagr.ha and later Pāt.aliputra
(modern Patna). Kosala had recently conquered and absorbed the kingdom
of Kāśı̄, previously an important kingdom with its capital at the city of Kaśı̄
(Vārān.ası̄). During the lifetime of the Buddha, the neighbouring state of
Aṅga (capital Campā) and Kosala itself were absorbed into Magadha. A
number of smaller states in the region were ruled by assemblies of local rājas
or other notables. The largest of these was the Vajjian confederacy, with
its capital at the large town of Vaiśālı̄. This confederacy, which included
the Buddha’s own home territory of the Śākyas and the old kingdom of
Videha, was absorbed into Magadha shortly after the Buddha’s death, as
was the nearby Malla territory.
To the south of the Kuru-Pañcāla region were a number of other early
states. These included Vatsa (Vam
. sa), with its main city of Kauśāmbı̄ to
the southwest of Kaśı̄, Śūrasena to its west with its capital at Mathura,
Maccha (Matsya) and Cetiya (Cedi). Further south, on the dakśin.āpatha
or southern route, were the important early kingdom of Avanti, with its
The Second Urbanisation of South Asia
59
s
k
e
e
I
n
d
o
-
G
r
Approximate maximum
extent of Maurya rule
Area under
Nanda rule
M
A
Pāt.aliputra
U
R
N
A
Y
Figure 3.4. Map: Nanda and Mauryan states
N
A
D
A
S
S
60
Origins of yoga and Tantra
capital at Ujjayinı̄, and Āśmaka (Assaka). This whole region, stretching
down to the northern part of the Deccan plateau, included important early
centres of commercial activity and of support for the śraman.a movements
(the Buddhists, Jainas and Ājı̄vikas).
Further to the east, in the Ganges Delta, the regions of Pun.d.ra, Vaṅga
and Rādha were not listed among the sixteen mahājanapadas, but included
significant commercial centres, certainly by Mauryan times (Jahan 2004).
In the northwest, along the uttarapatha or northern route, lay the
mahājanapadas of Gandhāra and Kamboja. The major cities in this area
were Taks.aśı̄lā (Taxila) and Pus.karāvatı̄. Gandhāra and Kamboja, along
with the Indus valley (Sindhu), had been incorporated into the Achaemenid
empire and provided an important channel through to Iran and the West.
Taks.aśı̄lā itself was both the location of the first major academic centre of
the subcontinent, and an important early centre of Buddhism.
The conquests of Anga, Kosala and the Vajiian confederacy by Magadha, which took place towards the end of the fifth century BCE,17 made
Magadha much the largest and strongest state in North India. The dynasty
of Bimbisāra and his son Ajātaśatru, the kings of Magadha during the
Buddha’s life, was followed by that of Śiśunāga and that of the usurper
Mahāpadma Nanda, founder of the Nanda dynasty. Mahāpadma Nanda
was of śūdra background, not ks.atriya, as were many subsequent dynasties;
his dynasty was replaced by that of another usurper, Candragupta Maurya,
in around 321 BCE, shortly after Alexander’s invasion of India (Thapar
2002: 154–6, 175–6). The Mauryas established the first large-scale Indian
empire (see map, Fig. 3.4).
The varn.a status of the Mauryas was contested (Thapar 2002: 176), with
Buddhist writers according them ks.atriya status and Brahmanical writers
describing them as śūdra, but in any case they became supporters of the
śraman.a traditions. Candragupta is said to have been a convert to Jainism,
and his grandson Aśoka was a strong supporter of Buddhism. The Mauryans
were also on friendly terms with the Seleucid (Indo-Greek) successor states
to Alexander’s empire in the north-west (Thapar 2002: 177). Thus India
at this stage was relatively open to outside influences. In this and other
respects, the Mauryan state was to be a successor to the Central Gangetic
world much more than to that of Kuru-Pañcala.
In Chapter 4, however, we return to the fifth century BCE, and look
more closely at the two contrasting worlds which Hopkins suggests existed
at that time.
17
On the basis of the ‘new’ chronology for the Buddha’s life.
chap t e r 4
Two worlds and their interactions
Chapter 3 introduced the wider context within which early developments
in Indian religion and society were taking place at around 500 to 400 BCE.
In it I mentioned Tom Hopkins’ suggestion that Northern India at that
time contained two contrasting and in some ways opposed worlds, the
Brahmanical and Vedic-derived world associated with the Kuru and Pañcāla
kingdoms, and the non-Vedic world of the Central Ganges and northern
Deccan (Fig. 4.1).
While these worlds were in many ways opposed, it is important to
remember that they also had some important things in common. These
included a shared Indo-Aryan language. They also shared at least some
other elements of Indo-Aryan cultural heritage, including the presence of
Brahmins and of the basic social divisions of Indo-Aryan society, as well
as the worship of some Indo-Aryan deities such as Indra. This is not an
‘Aryan invasion’ model; I assume that populations in both regions were
mixtures of earlier Indo-Aryan and non-Indo-Aryan speaking populations.
The distinctiveness of the Kuru-Pañcāla region was cultural, not racial.
In addition, we should be aware that the two worlds were not in any
way sealed from each other; the early Upanis.ads describe Kuru-Pañcāla
Brahmins travelling to Kāśi and Mithilā, and Buddhism and Jainism rapidly
made inroads into areas like Śūrasena with close links to the Kuru-Pañcāla
region. In fact Mathurā, the capital of Śūrasena and an important early
urban centre with trading connections to the Northwest, Deccan and
Central Gangetic regions, was a major point of intersection of the two
worlds, and would become an important location for all of the newly
developing religious traditions (Srinivasan 1989).
Yet the two worlds were different, and saw themselves as different. In
this chapter, I look at the bodies of myths and legends through which two
main regions (Kuru-Pañcala and Kosala-Videha) saw themselves, with a
view to gaining some insight into the social and political orientations of
these societies. I will also look more briefly at two other areas of significance
61
NORTHWEST
(GANDHĀRA)
KURU
PAÑCĀLA
MATHURA
KOSALA-VIDEHA
CENTRAL
GANGETIC
REGION
GUJARAT
NORTHERN &
BENGAL
MAGADHA DELTA
WESTERN DECCAN
Figure 4.1. Map: The Two Worlds (c. 500 BCE)
Two worlds and their interactions
63
in the cultural and religious history of this period. These are the Bengal
Delta region, and the Northwest Region (Gandhāra and Kamboja), which
formed part of the Achaemenid empire through the sixth, fifth and early
fourth centuries BCE, and then became incorporated into the Indo-Greek
states that resulted from the conquests of Alexander the Great. Finally, I look
at South India, an area which probably did not come into direct contact
with North India until a considerably later period, but which may have
some important clues about the nature of early Indic society and religion.
lunar and sol ar d ynasties
An important clue to the self-understanding of the Kuru-Pañcāla and
Kosala-Videha regions is provided by the body of legends and myths surrounding their origins and the origins of their ruling dynasties. The Kuru
kings of the Kuru-Pañcāla belonged, at least according to later Indian
sources, to the Lunar Dynasty (Candravam
. śa), ultimately tracing their
descent to the Moon (Candra) through a complicated series of mythological events that was set in motion by a sexual relationship between the
Moon and the goddess Tārā, wife of Br.haspati (the planet Jupiter). The
son of the Moon and Tārā was Budha (the planet Mercury). His son, a
king named Purūravas, formed a liaison with the apsaras (celestial nymph)
Urvaśı̄.1 Their descendant, King Yayāti, had five sons of whom the youngest
was Pūru, originator of the Paurava clan in which the great king Bharata
was born.
Bharata is the eponymous ancestor of the Bharata tribal group and the
figure from whom the modern name of India (Bhārat) derives.2 Witzel
reads the R.gveda as providing evidence for a struggle between the main
body of the Pūrus and the Bhāratas, who originally formed part of the
same tribal federation.3 This is the point of the ‘Battle of the Ten Kings’
1
2
3
The story is the subject of a presumably late R.gvedic hymn in the form of a dialogue between
Purūravas and Urvaśı̄ (10.95; see O’Flaherty 1981: 252–6).
While the Mahābhārata account of Bharata’s birth, as represented in Figure 4.2, makes Bharata the son
of King Duh.s.anta and Śakuntalā, who was the daughter of the sage Viśvāmitra and another apsaras,
Menakā, he was a sufficiently important figure to be the subject of competing genealogies. There is
a particularly interesting Jaina version, which was later incorporated into the Bhāgavata Purān.a, in
which Bharata is the son of the first Jain tı̄rthaṅkara (enlightened teacher) of the present time cycle,
R.s.abha, by R.s.abha’s twin sister Sumaṅgala (see Jaini 1977). Bharata’s struggle for world domination
leads to a fight with his brother Bahubali. Bahubali wins but chooses to become a renunciate rather
than accept the kingdom.
Note that Witzel includes the Yadu-Turvasa and Anu-Druhyu groupings as part of the peoples who
produced the R.gveda. The descendants of Ayu (i.e. the Bharatas) are latecomers who defeat them
and take control.
64
Origins of yoga and Tantra
in which the Bhārata king Sudas defeats all the other groups and gains
supremacy (Witzel 1995b: 329, 331).
The subject matter of one of the two great Indian epics, the Mahābhārata,
is the war between the Pān.d.avas and the Kauravas, two groups of descendants of Bharata, in which the Pan.d.avas, aided by Kr.s.n.a, were victorious,
but virtually the entire ks.atriya population was destroyed. The central battle of the war took place at Kuruks.etra, at the heart of the Kuru-Pañcāla
realm. As we saw in Chapter 3, Kuruks.etra, according to Witzel, was the
central point in the state mythology devised for the Kuru realm in the early
Vedic period.
The Kuru kings claimed descent from one of the great heroes of the
Mahābhārata, Arjuna, via his son Abhimanyu, whose mother was Kr.s.n.a’s
sister Subhadrā. Abhimanyu was killed at Kuruks.etra, but his posthumous
son Pariks.it continued the Pān.d.ava lineage; it is Pariks.it’s son, Janamejaya,
who held the snake sacrifice that provides an outer frame story for the
narration of the Mahābhārata. It seems likely that the basic text of the
Mahābhārata as we know it today was assembled at a relatively late date,
probably in response to the Sanskrit Rāmāyan.a and to a period of Buddhist
and Jaina dominance over North India (e.g. Hiltebeitel 2001, 2004;
Fitzgerald 2004) although the period is still open to dispute (whether the
time of the Guptas or a couple of centuries earlier). However, the core of
the epic presumably derives in large part from pre-Mauryan stories about
the Lunar Dynasty and the mythology of the Kuru-Pañcāla region.
Figure 4.2 shows the basic Lunar Dynasty genealogy in diagrammatic form.
The legends and mythology of the Central Gangetic region are based
around a separate series of legends and myths, those of the Solar Dynasty
(Sūryavam
. śa). The central figure here is Rama, King of Ayodhya, and the
other major Indian epic, the Rāmāyana, is built around his marriage to
Sı̄tā, adopted daughter of the King of Mithilā. The dynasty is represented
as originating from a figure named Iks.vāku, who is a son of Vaivasvata
Manu, the sun of Vivasvat (Sūrya, the Sun) and the first human being in
this world-age. The name Iks.vāku is mentioned in the R.gveda (10.60.4) and
the Atharvaveda (19.40.9) but neither reference is particularly informative.
Michael Witzel has suggested that the Iks.vāku may have been a sub-tribe of
the Pūru people, but there is little evidence either way.4 Figure 4.3 presents
the principal relationships here.
4
See Witzel 1995b: 319, 329, 339. Iks.vaku’s position in the genealogy is that of common ancestor, not
of founder of a Pūru sub-tribe. If he were the latter, one would expect to find him presented as a
descendant of Pūru.
Two worlds and their interactions
65
Candravamśa (Lunar Dynasty)
Brhaspati
(Jupiter)
=
Vivasvant
Tara
(Star)
Candra
(Moon)
Budha
(Mercury)
Manu
Vaivasvata
Ila (‘son’ of
Manu in Mbh.
but takes
female form)
=
Urvaśi
=
Pururavas
Two intermediate
generations
´
Devayani
Sarmistha
=
=
Yayati
Puru
Yadu
Sixteen intermediate
generations
´
Sakuntala
Duhsanta
=
Bharata
Paraśara
Krsna
Vyasa
Satyavati
=
´
Samtanu
Ambika Ambalika
= (1)
= (2)
Vicitravirya
(by Vyasa)
(by Vyasa)
Dhrtarastra
Kauravas
Pandu
(by various
deities)
Pandavas
~
Kings of Kuru-Pancala
Figure 4.2. The Candravam
. śa (Lunar Dynasty): Outline Diagram
These dynastic lists and traditions are preserved in a variety of sources,
none of them particularly early, although fragments of the stories on which
they are based occur in the R.gveda and the Brāhman.as. The earliest complete versions are those in the Mahābhārata and Rāmāyan.a, while the
Purān.as include a series of later reworkings. All this is much later than
the period of which we are speaking here (500–400 BCE), but it has generally been assumed that these stories preserve early traditions, and there
is sufficient consistency between the accounts, allowing for different viewpoints, to support such an assumption.
Origins of yoga and Tantra
66
Suryavamśa (Solar Dynasty)
and the Kings of Mithila
(as described in the Valmiki Ramayana, I. 69–70)
Brahma
Two intermediate generations
Vivasvant (Surya, the Sun)
Kings of
Mithila
Manu Vaivasvata
Iksvaku (first
king of Ayodhya)
Nimi
(first king
of Mithila)
Mithi
Eight intermediate
generations
Janaka
Prasenajit
Dhruva-sandhi
Bharata
Seven intermediate
generations
Nineteen
intermediate
generations
Raghu
Twelve intermediate
generations
Daśaratha =
Rama
Kausalya =
Kaikeyi
Janaka
Kuśadhvaja
´
Laksmana Bharata Satrughana
Sita Urmila two
(adopted)
daughters
Figure 4.3. The Sūryavam
. śa (Solar Dynasty) and the Kings of Mithilā: Outline Diagram
At the same time, with the exception of some of the most recent figures
in the Purān.ic king lists, these accounts are clearly not literal history. There
have been attempts to interpret them as such, but for the most part they
need to be read in other ways.
As was classically demonstrated in British social anthropology through
the work of Sir Edward Evans-Pritchard and his colleagues on the Nuer
and other lineage-based societies in East and West Africa, such genealogical accounts are generally, at one level of meaning at least, narratives of
the political relationships between the groups that claim descent from the
figures described (cf. Evans-Pritchard 1940). These myths also have a considerable amount to say about the nature of kingship and polity within the
realms they refer to. What is striking is the extent to which these stories,
particularly the Lunar group as a whole when contrasted with the Solar
Two worlds and their interactions
67
group as a whole, represent several quite distinct mythologies, which have
been only very loosely fitted together by later tradition, and which encode
different concepts of kingship and ideas about the nature of the state.
The Lunar Dynasty stories, as I noted, are primarily focused around the
Kuru-Pañcāla region, but also involve other groups to their south and west
that evidently played an important part in the early period. Most conspicuous are the Yadavas, who are descended in the genealogy from Yadu, an
elder half-brother of Pūru and so are a Bhārata clan closely related to the
Kauravas and Pān.d.avas. The Yādavas are the people associated with Kr.s.n.a.
Kr.s.n.a, as is well known, reveals himself in a key episode of the
Mahābhārata, the Bhagavadgı̄tā, to be a divine emanation, but this is
unlikely to be an early feature of the mythology, and the Bhagavadgı̄tā itself
has often been regarded as a late addition to the Sanskrit Mahābhārata.
Kr.s.n.a as a heroic rather than divine figure may go back further, as an early
mythological leader of the Yādava people who were allied to the KuruPañcala rulers. The Mahābhārata does not include the well-known legends
of Kr.s.n.a’s childhood, although a version of these forms part of an early
‘appendix’ to the Mahābhārata, known as the Harivam
. śa.
By the time that the Sanskrit text of the Mahābhārata was compiled
there clearly was very little left of these ‘Lunar Dynasty’ ks.atriya descent
groups, and the ruling dynasties in North India were no longer claimants to
ks.atriya descent.5 A dominant role in the story, and presumably also in the
society in which it was written, had been taken over by a particular group of
Brahmins, the Bhārgava (i.e. descendants of Bhr.gu). The stories specifically
associated with the Bhārgavas are centred around warrior-type Brahmins
such as Rāma Jāmadagnya (Paraśurāma) and represent them as rightfully
taking over dominance from the ks.atriya (Sukthankar 1936; Goldman 1977;
Hein 1986a; Fitzgerald 2002). Thus the ‘Lunar Dynasty’ myths in the
Sanskrit Mahābhārata were no longer functioning as legitimation charters
for specific dynasties.6 However it is reasonable to assume that the basic
mythological material from which it is compiled goes back considerably
earlier, and represents, in effect, the mythological charter of the KuruPañcāla region.
The Solar and Lunar Dynasty stories are connected together somewhat
arbitrarily through the figure of Ilā, wife of Budha (Mercury), who is
5
6
The Śuṅgas were not ks.atriyas and neither were the Guptas, although the Guptas married into the
Licchavi royalty, who claimed descent from the Solar Dynasty.
A number of later dynasties in the first millennium CE claimed Solar Dynasty ancestry, including the
Iks.vāku kings of Central India and several Rajput clans. Other Rajput clans claimed Lunar Dynasty
affiliation or descent from Agni.
68
Origins of yoga and Tantra
described as the daughter of Manu Vaivasvata and so sister of Iks.vāku.7
In some accounts, Ilā’s identity shifted between female and male, so s/he
is also credited with a series of descendants in his/her male identity of
Sudyumna.8
the sol ar d ynast y and the r ā m ā ya n. a
I turn now to look at the Solar Dynasty in more detail. If the Mahābhārata
is centred around the mythology of the Lunar Dynasty kings, the Rāmāyan.a
represents the story of the Solar Dynasty kings. This is the dynasty to which
Rāma, ruler of the city of Ayodhyā and king of Kosala, belongs.
These Solar Dynasty kings are also important in Buddhist and Jain
history. The Śākyas, the family of the historical Buddha, claimed descent
from the Solar Dynasty founded by Iks.vaku (Pali Okkaka9 ). The Iks.vakus
are also the lineage from which Mahāvı̄ra, the twenty-fourth and last Jain
tı̄rthaṅkara or enlightened teacher of the present age, and all but two of
his twenty-three predecessors are held to derive. While the Solar Dynasty
kings are represented as Aryan by descent there are, as we shall see, notable
and characteristic differences between the two bodies of stories that suggest
a separate political history and cultural identity.
As I mentioned earlier, the name Iks.vāku is mentioned in the R.gveda
and Atharvaveda, but this does not really establish whether the Central
Gangetic people associated with Solar Dynasty mythology had or claimed
a common origin with the Kuru-Pañcāla group (for example, as a continuation of the Iks.vaku sub-tribe which Witzel postulates as part of the
Pūru group who were defeated by the Bharatas at the Battle of the Ten
Kings). They could equally well be a separate group who found Iks.vāku a
convenient ancestral figure. The genealogy of Iks.vāku in the Rāmāyan.a is
not particularly elaborated, particularly in contrast with the complex Lunar
Dynasty genealogies in the Mahābhārata, and this may suggest that it is a
later invention, though it may also reflect the different style of the two texts
and the different context of narration (Brockington 1985, 1998). In either
case, it would seem that the group claiming descent from Iks.vāku were
recognised as Aryan, but were distinct from the Aryans of the Kuru-Pañcāla
region.
7
8
9
Ilā is a complex figure who appears in the middle vedic Śatapatha Brāhman.a as the goddess Id.ā, a
personification of the domestic fire-offering ritual or id.ā, one of whose purposes is to bring about
progeny (e.g. Gonzalez-Reimann 2006: 222; B. Smith 1989: 162–6). See also Hartzell 1987: 114–15,
which discusses the possible relationship between Ilā/ id.ā and the Tantric channel id.ā.
In the Skt. Mahābhārata 1.70 s/he is described as a son of Manu and as both father and mother of
Purūravas (van Buitenen 1983: 174).
Dı̄ghanikāya 3 (Ambat..tha Sutta), 1.15–16.
Two worlds and their interactions
69
mithil ā and the wisdom king model
The contrast between the Lunar and Solar Dynasty mythologies emerges
most clearly when we look at one of the key figures in the Solar Dynasty
mythology, King Janaka of Videha. Videha is generally placed by modern
scholars in northern Bihar and its capital city, Mithilā or Janakpur, is identified with the modern city of Janakpur in southern Nepal. This location was
not identified with the city of king Janaka, however, until the eighteenth
century, when the town was taken over by the Ramanandi order (Burghart
1978; cf. Fuller 1992: 166–9). It seems likely that the original Mithilā was a
legendary location as much as a real one, as indeed was Rāma’s own city of
Ayodhyā (see below).
The most familiar item in the mythology of Videha is undoubtedly the
marriage of Sı̄tā, adopted daughter of Janaka, King of Videha, with Rāma,
a central episode in the Rāmāyan.a. Thus the Rāmāyan.a is not just the
account of the Aryan descendants of Iks.vaku; it is also centrally about an
alliance between these kings and the Kings of Mithilā, which is sealed by
a marriage – four marriages, in fact, in the Sanskrit Rāmāyan.a of Valmı̄ki,
since Rāma’s three brothers are also married to daughters of Janaka and
Janaka’s brother.
The kings of Mithilā are occasionally presented with an Aryan genealogy
but they seem to contrast strongly with the Aryan kings of the Kuru-Pañcāla
region, and to represent a significantly different pattern of kingship. In brief,
this is a model of the wisdom king rather than the warrior king; its ideal
figure is a sage, not a warrior, and he shows strong renunciate tendencies. Vālmı̄ki’s Rāmāyan.a can in fact also be seen as about the relationship
between these two kinds of kingship, a theme to which we will return
in Chapter 9. Rāma is a warrior, but it is no accident that he spends an
extended period of renunciation in the forest. Before considering Vālmı̄ki’s
Rāmāyan.a in more detail, however, it is worth looking at some of the other
contexts in which the kings of Videha appear.
The most striking element of this kingdom is the strong association
between this dynasty and the idea of the ruler who is a figure of wisdom and/or renunciation. Wiltshire discusses various wise or just kings of
Videha in Brahmanical sources, some of them called Nami or Nimi and
others by other names, including ‘Janaka’, the name of Rāma’s father-inlaw and also of the patron of Yājñavalkya in the Br.hadāran.yaka Upanis.ad
and the Śatapatha Brāhman.a. Wiltshire traces a range of references to these
renouncer or quasi-renouncer kings of Videha, including a number of
Jātaka stories, and four distinct episodes in the Mahābhārata (Wiltshire
1990).
70
Origins of yoga and Tantra
Nami or Nimi is the type of the renunciate king in early Indian religious
literature, and is described as a royal renouncer in Brahmanical, Buddhist
and Jain sources. In Buddhist and Jain sources he is one, and much the most
significant, of a set of four renunciate kings who became pratyekabuddhas,
a type of enlightened being who is regarded by the later Buddhist and
Jain tradition as somewhat less in stature than that of a fully enlightened
Buddha or tı̄rthaṅkara. This set occurs in both Buddhist and Jain sources,
an indication of its early date. In Wiltshire’s book, Ascetic Figures before
and in Early Buddhism (1990) he analyses and discusses these sources.10
The names of the kings are essentially the same in both Buddhist and Jain
versions, although the brief accounts of how they came to renounce their
kingdoms for the life of a śraman.a are slightly different. The four kings are
King Nimi or Nami of Videha, King Dummukha or Dummuha of Pañcāla,
King Karan.d.u (or Kakaran.d.u) of Kālinga and King Naggaji (or Naggai) of
Gandhāra (Wiltshire 1990: 120).11
Three of these kings are known essentially only from this list of four kings,
but King Nimi or Nami of Videha is a more substantial figure who occurs
in a variety of other sources, Buddhist and Jain. He is in fact included by
the Jains as one of the twenty-four tı̄rthaṅkaras (No. 21 – not to be confused
with No. 22, Nemi).
Returning to the central myth of the Rāmāyan.a, this tells the story of
the marriage between a prince of Kosala, Rāma, and a princess of Videha,
Sı̄tā, whose name means ‘furrow’, is born of the earth and, Hopkins has
suggested, can be considered as a personification of the earth goddess. By
implication at least, Janaka could be regarded as married to the earth, so
that both Rāma and Janaka are kings who are seen to rule as consorts of
the earth goddess.
When Sı̄tā’s father Janaka recites his lineage, as Sı̄tā’s adoptive father,
in the ‘Bālakān.d.a’, Book I of Vālmı̄ki’s Sanskrit Rāmāyan.a, he begins it
with Nimi, describing Nimi’s son Mithi, the founder of the city of Mithila,
as the first Janaka, and himself as a distant descendant. Janaka describes
three of his ancestors as royal sages (rājar..si) and describes how his father
renounced the kingdom and went to the forest (I, 70, vv.1–16 = Goldman
10
11
The Buddhist version occurs in the Kumbhakāra Jātaka, and the Jain version in the Uttarādhyāyana
Sūtra. Wiltshire also cites a later Jain commentary on this sūtra. Nimi is also mentioned in the
Majjhimanikāya, Sutta 83 (Makhādeva Sutta = Ñan.amoli and Bodhi 1995: 692–7). In this sūtra, he
is described as the last of a long series of 84,000 kings of Mithilā, all of whom became renunciates
when their first grey hairs appeared, passing on their kingdoms to their sons. The first of the series,
Makhadeva, is described as a previous rebirth of the Buddha.
At 1990: 120, Wiltshire makes both Nami and Dummuha into kings of Pañcāla in the Jain sources
but this seems to be an error, see 1990: 139 where Nami is king of Videha.
Two worlds and their interactions
71
2005: 360–3). The lineage seems intended to contrast significantly with
that which had been recited immediately before by Vasis.t.ha at the request
of Rāma’s father Daśaratha. This contains far more in the way of warlike
deeds and not a single rājar..si (I, 69, vv.19–36 = Goldman 2005: 358–61).12
The Balakan.d.a of the Rāmāyan.a is a relatively late source, but it is clear
that there was still a distinct image of the Videhan kings even at this stage.
warrior k ings and wisdom kings in the epics
The Sanskrit Rāmāyan.a and Mahābhārata as we know them today are complex texts from a considerably later period, and each has clearly undergone
a considerable amount of internal evolution and reworking. The original
locations of the story are perhaps mythical rather than actual places. As
Hans Bakker has pointed out, the city now known as Ayodhyā, originally
known as Sāketa, only became gradually identified as Rāma’s capital, probably not until well into the Gupta period (Bakker 1986: 11–12). Bakker noted
in 1986, some time before the site of Rama’s birthplace became a major issue
in modern Indian politics, that ‘Ayodhya, like Laṅka was most probably a
creation of the poet’s imagination’ (Bakker 1986: 10). Laṅkā, too, is almost
certainly an imaginative creation, which only much later became identified
with the island of Sri Lanka (Bakker 1986: 10; Brockington 1985: 109–20),
while the small town of Janakpur in Southern Nepal that is now regarded as
the capital of King Janaka was, as we have seen, not identified as such until
the eighteenth century. The capital of Kosala in the Buddhist and Jain texts,
and in Pān.ini, was not in fact at Ayodhyā-Sāketa but at Śrāvastı̄. However,
there seems no reason not to accept that the epic belongs to the Central
Gangetic region, or that its basic core derives from oral legends of that
region.
The Sanskrit Mahābhārata is a longer and more complex text than the
Vālmı̄ki Rāmāyan.a, but again the body of myths that underlie it presumably
derives from oral legends of the Kuru-Pañcāla region and centres around
the history of the Lunar Dynasty. The original story may well be modelled,
as Witzel suggests, on early battles between chieftains of which we can still
see traces in the R.gveda. However, while the epic eventually expanded to
incorporate peoples from virtually the entire subcontinent on one side or the
other of the war, its central events are for the most part clearly located in the
old Aryan heartland around Kuruks.etra, with links to Witzel’s postulated
Kuruks.etra-centred state ideology (Witzel 1995b, 1995c).
12
Except perhaps for Daśaratha himself, who has explicitly been described as a rājar..si in I, 6, v.4.
72
Origins of yoga and Tantra
I have already referred to V. S. Sukthankar’s study of the so-called
Bhārgava element in the Mahābhārata, displayed in characters such as Rāma
Jāmadagnya (Paraśurāma) or Dron.a. The emphasis here is on Brahmins
who are aggressive, often warriors, and on the necessary destruction of
ks.atriya rule. Sukthankar attributed this to a period in which Bhārgava
Brahmins had taken over custodianship and performance of the epic from
the sūtas, the bards who were its original performers. At the same time,
the Mahābhārata shows signs, most visibly in the figure of Yudhis.t.hira,
of a different emphasis, and one that introduces elements of the Central
Gangetic wisdom-king pattern. In an article from 1986, Norvin Hein identified an ‘irenic’ or peaceful element in the Mahābhārata (Hein 1986a). This,
Hein argued, is associated with a distinct group of editors or compilers
from the ‘Bhārgava editors’ responsible for the warrior Brahman emphases.
These people are associated particularly with the recurring phrase sarvabhūtahite ratah. (‘delighting in the welfare of all beings’). This phrase occurs
in some sections of the Mahābhārata, but is on the whole absent from the
major episodes. It is used frequently throughout the Rāmāyan.a. In both
epics, it occurs particularly as a description of how a king should feel and
behave.
James Fitzgerald has suggested that Yudhis.t.hira ‘was designed as a refutation, or at least a rebuttal, of the emperor Aśoka’ (2001: 64). He argues
that the figure of Yudhis.t.hira at the beginning of the Śānti Parvan, in his attempt
to renounce the kingship and go to the forest, was deliberately scripted by the
authors of the epic to represent what they saw to be wrong with the Mauryan
emperor Aśoka, to purge and refute whose rule was, I believe, the principle [sic]
purpose for the creation of the first generation of our written Sanskrit Mahābhārata.
Yudhis.t.hira’s attempted renunciation of the Bhārata kingship was made to allow
the epic poets to show him being corrected and refuted by his family, by the
Brahmins led by Vyāsa, and ultimately by Kr.s.n.a Vāsudeva. (Fitzgerald 2001: 64–5)
It may be noted that Yudhis.t.hira himself is a son of the god Dharma, and
acts as the exemplar of righteous conduct within the epic.13
Thus the Sanskrit texts of the Rāmāyan.a and the Mahābhārata are both
evidently advancing a compromise of some kind between the irenic and
warrior models of kingship. I return to this theme in Chapter 9. For the
present, it is enough to note that we already seem to have two different
models of kingship in the sources from which the Sanskrit Rāmāyan.a and
13
Thus Fitzgerald reads this episode as a rejection of the idea of the renunciate king. On Yayati, the
most conspicuous example in the Mahābhārata of a king who actually succeeds in renouncing his
kingship, see Chapter 9 n. 14.
Two worlds and their interactions
73
the Mahābhārata are drawn. The Rāmāyan.a myth also suggests that the
king is expected to be in harmony with the local powers, as represented
by Rāma’s marriage (and Janaka’s implicit marriage) to an earth-goddess
figure (Hopkins 1999). I shall speak of these in general as the wisdom king
and warrior king models; in Indic terms we might speak of them as the
dharmarāja and cakravartin models.
Despite the relatively late date of the epics, the stories on which they are
based probably go back to earlier sources, and they seem to be part of a
wider body of narratives about kings and their subjects. As I have elsewhere
suggested in relation to the Tibetan Epic of Gesar, and as in fact seems the
case for many epic stories in different cultures, these stories were doubtless
from the beginning, to a significant degree, stories about power and how
it should be properly used (Samuel 1992, 2002a).
I think all this is enough to make it very likely that there was a widespread
stereotype of the wisdom king or proto-dharmarāja model of kingship in
India in the period from 500 BCE onwards, and that it was particularly
associated with the kings of Videha, a principal power in the region before
the rise to dominance of Magadha. A variety of stories describe these kings as
having tendencies towards the śraman.a or renunciate lifestyle, or as actually
becoming śraman.as or renunciates. It also seems likely that this model of
kingship was seen at the time to contrast markedly with the warrior king
or cakravartin model of kingship associated with the Brahmanical reforms
in the Kuru-Pañcāla Region.
The warrior king model was based on the concept of the king as an
exponent of military prowess who had the military force to ensure compliance from the surrounding chiefs or kings. In ritual terms, this is expressed
by the famous aśvamedha or horse-sacrifice. Before the horse-sacrifice can
be performed, a horse is allowed to roam at will for a year and the king’s
warriors have to defend and guard it in whichever territory it may wander
(P. Dumont 1927; Bhawe 1939). Part of what is happening in the Bālakān.d.a
of the Rāmāyan.a, and in the final book (Uttarakān.d.a), is perhaps about
an attempt to combine or reconcile these two models, with Rāma seen as
a synthesis between the two. The first involves an aśvamedha performed
for his father, the rājar..si Daśaratha, while the last involves an aśvamedha
performed for Rāma himself.
It is tempting to go a step further, and ask whether the Kuru-Pañcāla
(Vedic) warrior king variant can be more closely associated with an IndoAryan or Indo-European structure, and the Videha (Central Gangetic)
wisdom king model with something more indigenous. I am not sure that
we have enough information to justify such a step. By 500 BCE, the Central
74
Origins of yoga and Tantra
Gangetic region would have had several centuries of history during which
a mixed population of Indo-Aryan immigrants and hypothetical indigenes
had arrived at a common cultural tradition. The warrior king model does,
though, look like the kind of model of kingship associated with a mobile and
originally horse-riding pastoralist population. The horse-sacrifice model
seems much more about the control over grazing territory appropriate for
a low-density pastoralist society than about the relationship between a king
and a settled population of agricultural communities. The wisdom king
model, particularly if we accept the ‘marriage to the land’ idea as one of its
components, seems more the kind of model one might expect in a settled
agricultural society.14
However, we could also perhaps identify different models of kingship
within the ‘Indo-Aryan’ material itself. Varun.a, although viewed in an
increasingly negative way in later Brahmanical material, retains some of
this sense of a just king (Sutherland 1991: 76–83); originally, Stanley Insler
has suggested, Varun.a corresponded to the role of the tribal ruler or chieftain
in time of peace and Indra, prototype of the warrior king, to the war-leader
elected to take over in time of conflict (Insler 2004). If this is so, then one
could comment further that the Iranian and Indian branches of the IndoAryan peoples took different choices in relation to this dichotomy. In Iran,
the asura Varun.a seems to have been the basis of Zarathustra’s new supreme
deity of Ahura Mazdā, and Indra was in effect demonised in Iran as one of
the rejected daevas (see Chapter 5). In India, especially in the Kuru-Pañcāla
region where the ‘Vedic’ model was developed, Indra took over dominance,
and Varun.a faded into the background as a yaks.a deity (Sutherland 1991:
77–83). Thus the prime identity of the king in this region was that of warleader. The Achaemenid kingship, by contrast, picked up on the ‘Varun.ic’
14
As J. Clifford Wright has pointed out, stories of marriages between rulers and yaks.in.ı̄s, chthonic
female deities, are a staple of the early literature of the period, whether Buddhist, Jain, or Brahmanical.
He instances a Jain text, the Miyāvutta Sutta; a Buddhist jātaka story, the Kinnarı̄ Jātaka (in fact
there seem to be a number of jātaka stories involving marriages or relationships between kings or
princes and yaks.in.ı̄s). and, doubtless the best known, Kālidāsa’s stage play about such an encounter,
the Vikramorvaśı̄ (Wright 1966: 20). The Vikramorvaśı̄ is based on the story of the marriage between
the Lunar Dynasty ancestor Purūravas and the apsaras Urvaśı̄, a legend which occurs in embryonic
form in the R.gveda and the Śatapatha Brāhman.a. These stories have widespread parallels in other
cultures; one might think of the European story of Melusine (also the foundation legend of a dynasty
through marriage between a ruler and a chthonic goddess), or the Javanese myth of the marriage
between the Sultans of Jogjakarta (Jogja = Ayodhya) and the Goddess of the South Seas, Ratu Kidul.
It has been suggested that Kalidasa’s Vikramorvaśı̄ may itself have been written on the occasion of
the installation of the son of one of the Gupta kings as crown prince or yuvarāja (Michael Willis,
personal communication, April 2004). The critical point about the Rāmāyan.a variant, if one wishes
to accept this argument, is that Sı̄tā was born out of the earth; in the case of the Lunar Dynasty
story of Purūravas, the relationship is with a celestial spirit.
Two worlds and their interactions
75
theme of divine righteousness and cosmic order as refigured in the shape
of Ahura Mazdā.15
For my purposes, I think it is best to leave the issue of Indo-Aryan
versus indigenous origins as undecided and probably undecidable. What
is important is that we have a contrast in political style and mythology
between the two regions that helps considerably to explain the religious
and political developments that took place.
The Buddhist texts describe the mahājanapadas in the Central Gangetic
region as mostly operating in the Buddha’s lifetime according to the gān.asaṅgha system of rule by a council of community leaders (Thapar 1984;
J. Sharma 1968). The term saṅgha was of course to become the term for
the Buddha’s own community, and the practices of these gān.a-saṅgha are
clearly regarded in the early Buddhist texts as an important model for the
Buddhist saṅgha. According to the Buddhist texts, however, these citystates were being absorbed during the lifetime of the historical Buddha
into new, larger kingdoms, with the two expansionist kingdoms of Kosala
and Magadha gradually absorbing the others. As mentioned in Chapter 2,
the traditionally-accepted dates for the historical Buddha are now generally
regarded as too late and these events might now be best placed in around
450–400 BCE, a dating which fits reasonably well with the growth of
sizeable, elaborately fortified towns in the region.
What all these accounts make clear is the systematic differences between
the Central Gangetic Region and the Kuru-Pañcāla Region (Fig. 4.1). The
region centring around Kuruks.etra in what is now Northern UP and
Haryana was as mentioned earlier regarded in the Vedic texts as the original
āryavarta or homeland of the people who identified their culture as ārya
(‘Aryan’). It too was divided into mahājanapadas, with the twin kingdoms
of Kuru-Pañcāla at its centre. The gān.a-saṅgha system was not found in
the Kuru-Pañcāla Region or neighbouring areas, and urbanisation seems
to have lagged significantly behind the north-east. While the north-east
was the area of the growth of the ascetic orders, among both non-Brahmins
(Jains, Buddhists, Ājı̄vikas) and Brahmins, the Kuru-Pañcāla Region is associated with conservative forms of Brahmanical culture, though here as at
other points in Indian history ‘conservative’ should perhaps be read more
in terms of self-image than in terms of actual continuity with the past. As
we shall see, it is arguable that Kuru-Pañcāla was the location where a major
15
There is a possibility that Achaemenid kingship was itself a significant influence on the new Central
Gangetic kingdoms (see below). As far as I know, however, the theme of renunciation of kingship
does not form part of the ‘Varun.ic’ or Achaemenid pattern.
76
Origins of yoga and Tantra
reconstruction of Vedic-Brahmanical culture took place, one which in fact
led to the Vedic texts in the form we know them today.
brahmanic expansion and interaction bet ween
the t wo regions
As I noted above, the Central Gangetic region and the Kuru-Pañcāla region
shared a considerable amount of cultural material in the time of the Buddha.
I have suggested above that some of this derived from a shared Indo-Aryan
background. There was also an ongoing interaction. Thus, for example,
we already find in the early Buddhist texts that dynasties are presented as
having Vedic ancestry, elite groups are defining themselves as ks.atriya, and
Brahmins are common figures within the society. These texts were not as
far as we know written down until the first century BCE, and their picture
of social realities may in some respects reflect this period rather than the
time to which they refer, which is now best dated to the second half of the
fifth century and late fourth centuries BCE, but it would nevertheless seem
likely that these aspects date back close to the lifetime of the Buddha. Early
Jain texts also appear to reflect a similar ambience.
It is worth looking at the modes by which Vedic and Brahmanical influences might have spread to these regions.
Certainly, there is substantial textual evidence for the outward expansion of Vedic-Brahmanical culture. Early texts describe the areas to the east
as impure and as unsuitable for Brahmins to live in. The boundary gradually shifts. In a well-known passage, the Śatapatha Brāhman.a I.4,1,14–16
describes the conquest of the Kosala-Videha region through an early prince,
accompanied by Agni, the god of fire (Kulke and Rothermund 1990: 50).
By the Śatapatha Brāhman.a’s time, perhaps the sixth or seventh century
BCE, this region is regarded as having been acceptable for Brahmin residence. Areas further to the east, such as Magadha and present-day Bengal,
remained problematic. However, this referred specifically to where it was
acceptable for Brahmins to settle. It did not imply that Brahmanical religion had become the dominant or universal religion in these regions. Here
we have to be wary of reading the present into the past.
Certainly this process would seem to have been only to a limited degree
a question of the extension of political control. According to Witzel, the
Kuru-Pañcāla realm expanded towards the South and to the East, incorporating the kingdom of Kāśı̄.16 It was eventually overcome by the Śālvas,
16
Witzel speaks of the ‘materially little progressed, chalcolithic cultures of the east’ (1995c: 21).
Two worlds and their interactions
77
a probably non-Vedic group who subsequently coalesced with the Kurus.
Following this, the Pañcāla tribe took over as the centre of Vedic culture
and influence.
There is little or nothing to suggest the existence of a large-scale state
including all of these regions for any length of time. It would seem that the
Kuru-Pañcāla region at this stage did not have the economic resources to
support a large-scale state structure. It is possible that temporary military
conquest was followed by the expanded area thus controlled breaking up
into a number of separate territories controlled by brothers or sons (a
common enough theme in Sanskrit literature). However, the states with
rulers belonging to or affiliated with the Lunar Dynasty rulers would seem
to mark the limits for this mechanism.
Instead or in addition, we should consider a range of other mechanisms:
(1) the settlement of isolated families or clans of cultivators in areas which
were previously uncleared. Such settlement may have involved obtaining permission from a local ruler, where there was one. In some areas it
may have led to the creation of a new local political entity that might
be ‘Aryan’ in style.
(2) Brahmin families might also settle on an isolated basis, as part of ‘Aryan’
settlements of the above kind, or in non-Aryanised regions, in agricultural settlements or in nearby forest areas, relying in whole or part
on their ability to provide ritual services to the local population for
support.
(3) Brahmin families may also have been granted land on the basis of their
providing priestly services to non-Aryan rulers. There is no evidence for
substantial land grants in the early period and in fact it is unclear how
far Kuru-Pañcala kings would have been seen as entitled to alienate and
redistribute land (see Thapar 1984). The first evidence for land grants
refers to the Magadhan period (see Thapar 1984: 110). For this time, the
Buddhist sūtras describe what was in later periods a standard mechanism for the expansion of Vedic-Brahmanical culture: the settlement of
Brahmins on land granted by local rulers (Tsuchida 1991).17 Generally,
these would have been to Brahmins who were serving as court ritualists
or performing other functions for the kings.
17
A typical account is given in the Kūt.adanta Sutta of the Dı̄ghanikāya: ‘Once the Lord was traveling
through Magadha with a large company of some five hundred monks, and he arrived at a Brahmin
village called Khanumata. [. . .] Now at that time the Brahmin Kūt.adanta was living at Khanumata,
a populous place, full of grass, timber, water and corn, which had been given to him by King Seniya
Bimbisāra of Magadha as a royal gift and with royal powers’ (Walshe 1995: 133).
78
Origins of yoga and Tantra
The Kuru-Pañcāla state was, on Witzel’s account, a joint Brahmin-ks.atriya
project, and its achievement involved creating an important new role for
Brahmins where they increasingly formed part of the ideology and practices of kingship. If this ideology spread to the East, the need for court
Brahmins would have arisen, but it is not really clear why or how this
might have happened. There are, however, references to the sponsorship of
Brahmins by kings and to grants of cattle being given to them in the early
Upanis.ads. This suggests that the initial spread may have been on a much
more individualistic pattern.
Witzel has used the term ‘Sanskritisation’ in relation to the expansion
of Vedic-Brahmanical culture, though I am not clear how far he intends
this term to carry the theoretical weight that it has in the contemporary
anthropology of India.18 ‘Sanskritisation’ was initially introduced by the
anthropologist M. N. Srinivas in the context of a so-called ‘tribal’ population, the Coorgs of South India, to describe a process by which the
group gradually adopted characteristics of surrounding ‘Sanskritic’ culture
in order to raise its status in relation to those cultures (Srinivas 1952).
There are some possible analogies with this process. Contemporary
‘tribal’ populations in South India, however, have been in a process of
constant relations and cultural interchange with surrounding Hindu and
Brahmanical cultures for up to fifteen hundred years or more.19 There were
obvious gains for the Coorgs in adopting Sanskritic cultural markers of various kinds. It is not as clear why populations outside the Vedic-Brahmanical
orbit in the first millennium BCE would gain by adopting the Brahmanical
cultural patterns developed in the Kuru-Pañcāla region, since these patterns
would not initially have been associated with a dominant ruling elite or
have provided access to material resources. In fact the process was not a
straightforward or unidirectional one, since, as the Upanis.ads and the Jain
and Buddhist texts indicate, the Central Gangetic region rapidly became
locations of major cultural innovation in its own right.
There remains the possibility, particularly if one assumes that there had
been previous waves of Indo-Aryan-speaking migrants, that some idea of
Vedic-style ritual was already accepted in many of these regions, whether
performed primarily by householders or by local priestly lineages. In this
18
19
In his discussion of the ‘Sanskritisation’ of the Kuru-Pañcala region it clearly does not, since
‘Sanskritisation’ there is imposed from above (see Chapter 3).
I use quote marks around ‘tribal’ to indicate that this is a colloquial term in contemporary Indian
society (and a legal term, in the form of ‘Scheduled Tribes’) rather than a coherent anthropological
category. It covers a very wide variety of peoples speaking many different languages and with varying
degrees of incorporation into mainstream Indic social and religious life.
Two worlds and their interactions
79
case, the Kuru-Pañcāla Brahmins might have been seen as specialists in
ritual performance, and invited by local rulers and elite families primarily
as ritual performers. This might help to make sense of the ambivalent
orientations towards the Brahmins that we see in the Upanis.ads and the
Jain and Buddhist texts.
the northwest region and the bengal delta
Were there other regions that presented important alternative social, political and religious models? There are at least two that are worth considering
for this period, the North-West, with its links to the Achaemenids, and the
Bengal Delta. A third, South India, is probably more significant at a later
period, but will also be discussed here.
The Northwest Region
As I mentioned in Chapter 3, the Northwest Region of Gandhāra and
Kamboja came under the control of the Achaemenid dynasty of Iran from
around 520 BCE onwards. Gandhāra and Kamboja are two of the classic
mahājanapadas. The capital city of Gandhāra was at Taks.aśı̄lā, a major early
centre of urbanisation and trade and a channel through which influences
from the West might, hypothetically, enter into the Punjab and North India.
The region was included in a single large satrapy that stretched from the
Beas to the Hindu Kush. There is very little information regarding this area
during the period of Achaemenid rule. There is little direct archaeological
evidence and little indication that the region was tightly integrated into
the Achaemenid system. If, as has been suggested by Kosambi and others,
the King Porus and his people who were Alexander’s chief opponents in the
region represent the Vedic people of the Pūru, then it would seem that IndoAryan political institutions in the region, probably tribal federations of some
kind, maintained a presence throughout the period, and that there was by
that time a presence of Brahmanical and śraman.a institutions (Kosambi
2002b: 78).
However it is clear that Taks.aśı̄lā was the main city in the region from
early in the period and that it was a major trading city with connections
through to the west (Fussman 1993: 84–6). If nothing else, this region must
have represented a channel through which influences from the Achaemenid
state could flow.
Can we in fact trace influences from the Mesopotamian region and elsewhere on the new states of the Central Gangetic and Bengal Delta regions
80
Origins of yoga and Tantra
and the northern Deccan? Cyrus, the founder of the Achaemenid state
(559–530 BCE), appears to have originated from a culturally Elamite rather
than Iranian area (Potts 2005: 16–17). His son Cambyses was installed during Cyrus’s lifetime as ruler in Mesopotamia in 538 BCE in the ‘traditional
Babylonian manner’, in a ritual involving Cambyses’ approval and adoption by the Babylonian god Marduk, and later went through a similar ritual
in Egypt in relation to an Egyptian deity (Hinnells 1985: 99).
By the time of Darius (522–486 BCE), the Achaemenids clearly had
a developed idea of divinely sanctioned kingship, derived from Avestan
(Zoroastrian) sources, and linked to the idea of the king as being engaged
in a cosmic struggle for the forces of good against those of evil (Skjærvø
2005: 52–81; Hinnells 1985: 98–109, note seal on p. 100). Whether or not
the Achaemenid rulers could be described as wisdom kings in the sense of
Janaka, they clearly saw themselves as kings who ruled by the authority of
Ahura Mazdā and who were responsible for securing peace and well-being
on earth in his name, for combatting the forces of the ‘Lie’, and for educating
their people in the ways of wisdom (Skjærvø 2005: 61–5; Lincoln 2003).
In passing, it seems at least possible that the Vedic devas were regarded
for the most part as belonging to the forces of the Lie in Achaemenid
terms. How far this was an issue for Achaemenid rulers encountering Vedic
Indians it is hard to say. They were probably still conscious of their Indian
subjects as being ‘Aryan’ (see Grenet 2005 on the famous list of Aryan
peoples in the Videvdad), and in any case the Cambyses episode suggests
that the Achaemenids were relatively pragmatic about such matters.
It is conceivable though that the Achaemenid model may itself have been
a significant influence on the evolution of the wisdom king model in the
Central Gangetic region, especially if, as Wright suggests, there was a kind of
polarisation taking place between Brahmanical and counter-Brahmanical
models in the region (Wright 1966: 16–17). If the local rulers were looking for
alternative models for a state to those offered by the Brahmins, they might
well have looked to the Achaemenids.20 Yet one might expect contacts with
the Achaemenid state to have left deeper traces. There have been suggestions
that writing (the Kharos.t.hı̄ script, which may have served as a model for
the Brāhmı̄ script, the ancestor of all modern Indian scripts) came through
Iran, and also the Aśokan practice of public stone inscriptions (Thapar
2003). Punch-marked coinage, the first known form of Indian currency,
may also have come from the Achaemenids, though the evidence seems to
20
Cf. e.g. Kulke and Rothermund 1990: 57. At a later stage, the Indo-Greek kingdoms might have
offered a similar resource.
Two worlds and their interactions
81
point to the Indo-Greek period (Bailey and Mabbett 2003: 57 n.2, referring
to Cribb 1985).
The Northwest was also a channel to Greek knowledge: we know that
there were substantial transfers of knowledge along this route, for example
in the case of Indian astrology, which derives from the Greek system. The
influence of Greek sculpture on Indian sculpture is also obvious and has
long been recognised. We will see some further examples of this kind later,
some relatively secure, others more speculative.21
Magadha and the Bengal Delta
Moving east from the Central Gangetic region, we come to Magadha and
then to the Bengal Delta. Magadha has already been mentioned as one
of the mahājanapadas. It was in the end the most successful of them as
an expansionist state, becoming the basis of the first major north Indian
empire, that of the Nandas. The kings of Magadha in the Buddha’s time
seem to have no real claims to Vedic connections, and in fact the then
ruling dynasty does not seem to have gone back very far, though Magadha
was included in the Mahābharata under the rule of a doubtless fictional
earlier dynasty. Magadha probably owed much of its initial significance to
its natural resources; this was an early area for iron mining (Kosambi 1965:
123). From the Vedic perspective, however, this was a transition region, only
partially Brahminised and still of suspect nature.
The areas further to the east were presumably even more so. Yet there
are mentions of these regions in fairly early Vedic material and the coastal
trading states that evolved in the nearby Ganges-Brahmaputra delta, along
with other coastal settlements such as the Mahanadi delta in Orissa, may
have been of importance from quite early times. The unstable nature of
the Ganges-Brahmaputra delta means that many early port sites may have
disappeared, but Tamluk, the ancient port of Tamralipti, on the western
edge of the delta, has been excavated and dates at least as far back as the
second century BCE, while Chandraketugarh, in the 24 Parganas district,
has an occupational sequence going back to a period before the Northern
Black Polished Ware (NBPW) which is generally associated with the developed kingdoms of the Central Gangetic region (Chakrabarti 1995: 218). We
21
I should note that I have no interest here in trying to claim that everything valuable in South Asia
came from outside. However it is pointless to exclude the possibility of external borrowings on a
priori grounds and as noted there will be further examples in later chapters. I do not see that it
lessens the achievements of the people of South Asia throughout history to recognise that there was
sharing and interaction between cultures.
82
Origins of yoga and Tantra
know from Greek and Roman sources that there were important trading
communities here at a somewhat later stage.
The extent to which the Bengal delta area shared a cultural identity with
present day Eastern UP and Bihar at the start of the period is unclear; I
would tend to assume that as trading settlements grew in the small towns
of the Eastern UP-Bihar area, there was increasing cultural importation by
these urban centres along the trade routes from the delta. The Bengal region
in more recent times developed a caste system with distinct differences from
that of most of North India, with little or no ks.atriya presence, and a basic
division between Brahmin landowners, lower castes and untouchables, but
given that substantial Brahmin settlement in the region probably did not
get underway until well into the first millennium CE this can probably not
be read back very far.
As elsewhere, for example in pre-colonial West Africa or with the cities
of the Hanseatic League in northern Europe, one might expect a coastal
community based on trade to have political models based on the maintenance of the law of the market and of peace between the various trading
communities. It is uncertain how strong a contrast would have existed at
this period between the social and political style of the delta towns and
inland settlements, but there were clearly connections between them, and
a trading state might provide a natural environment in which a wisdom
king ideology might develop. Traders would have had little interest in war,
since their urban bases were natural victims in times of military conquest,
and they would have preferred to see their kings as relaxed and somewhat
hands-off administrators of justice rather than as aggressive conquerors.
It may be worth noting the significance of cult-associations in the West
African region in recent times. Such associations, centred around initiation into the cult of the various spirits of the local pantheon, have offered
in recent times some of the same advantages that early ascetic Buddhist,
Jain or Ājı̄vika ascetic orders may have offered to the growing merchant
community: a trans-local association based on ethical principles, with a
network of cult-centres and fellow practitioners scattered throughout the
cities and settlements of the trading region.22 The initiatory cults of the
Hellenistic world (Demeter at Eleusis, Isis, Mithras) may have offered somewhat similar advantages to their members. I shall suggest in Chapter 6 that
22
For example the ekpe or ngbe societies of the Cross River region of south-eastern Nigeria, which ‘were
instrumental in the transformation of an aggregate of agnatically organized fishing communities [. . .]
into the hub of a vast transethnic exchange system based on the circulation not just of slaves, palm
oil and European commodities, but of sacred knowledge as well’ (Palmie 2006: 100), or the Poro and
Sande societies in Liberia, Sierra Leone and neighbouring regions (Little 1965, 1966; Fulton 1972;
Leopold 1983).
Two worlds and their interactions
83
cult-associations of this kind may have been significant predecessors to the
Buddhist and Jain orders.
south ind ia and the caste system
As I mentioned, South India probably did not come into direct contact with
North India until a considerably later period. As a consequence, we have
information about South Indian society which arguably predates extensive
influence from the ‘Indo-Aryan-speaking’ component in the North, and
this may provide some important clues about the nature of Indian society
before the arrival of these elements. In particular, the southern material
introduces a theme which will be significant throughout this book; the
low-caste or outcaste group with an important spiritual or ritual function. Indeed, George Hart has argued that this non-Indo-European theme
is at least as important for the development of the caste system as the
Indo-European structure represented by the familiar varn.a-system, with its
fourfold division into Brahmin, ks.atriya, vaiśya and śūdra.
The population of the four southern states of modern India, as is well
known, mostly speak Dravidian languages, unrelated to the Indo-Aryan
languages dominant in the north. These areas were only slowly and progressively incorporated into the new order developing in the north, and
they retained their linguistic distinctiveness and a considerable degree of
cultural distinctiveness as well. This is particularly true of the regions that
make up what are now the two southernmost states, Tamilnadu and Kerala.
The two states to the north, Karnataka and Andhra Pradesh, were substantially incorporated into the Sātavāhana and Meghavāhana states from the
first century BCE or so, and presumably came under significant Indic and
Brahmanical influences at this time. The Meghavāhanas also made a substantial incursion into Tamilnadu in the first century BCE, but on the
whole the politics of the south centred around conflicts between relatively
small locally-based regimes, and even when larger-scale states such as the
Pallavas developed, they were themselves southern in origin and reflected
southern emphases.
As these larger states developed, South Indian kings increasingly looked
to the north and to the import of Indic models of religion and of kingship.
These were, however, not the only influences at hand. The Malabar coast,
in what is now Kerala, had strong trading connections to Egypt and the
Mediterranean, as the Periplus of the Erythrean Sea, a guide to the sea-trade
to India written in Egypt in the early second century CE, demonstrates
(Huntingford 1980). Richard Fynes demonstrated a few years ago that the
84
Origins of yoga and Tantra
cult of the Goddess Pattinı̄, important today in Sri Lanka and earlier also
significant in much of South India, is probably derived from the Hellenistic
cult of Isis in Egypt, and was transmitted along this sea route (Fynes 1991,
1993). So, no doubt, was much else. There was a Roman settlement with a
Temple of Augustus at Mouziris on the Malabar coast (Huntingford 1980:
116), so there was at least some presence of European ideas of divine kings
as well as Brahmanical ones.23
Indic influences in the south included not only Brahmanical religion,
but also Buddhism and Jainism. According to Jain tradition, there was a
strong Jain presence in the south from the third century BCE onwards,
and Śravan.a Bel.gol.a in Karnataka remains one of the most important of
Jain pilgrimage sites. Sri Lanka was traditionally converted to Buddhism at
the time of Aśoka and there was a Buddhist presence in the Andhra region
by around the same time. However, substantial evidence for Buddhism
in Tamil-speaking regions only dates from the fourth century CE, with
Brahmanism and Jainism perhaps a couple of centuries earlier (Schalk
1994).
Here, however, we are interested in what might have preceded the arrival
of Brahmanism, Buddhism and Jainism. Our main source is early Tamil
literature. Some of this is very informative but unfortunately, the dating
of early Tamil literature is subject to the usual controversies.24 In addition,
much of what we have has clearly been formalised for literary purposes.
However the Tamil Sangam literature, generally regarded as the earliest
layer of the Tamil literary heritage, can perhaps be dated to the third to
sixth centuries CE onwards.25
George Hart, who works primarily on early Tamil poetry, has developed
an interesting account of Tamil ritual kingship, which he suggests was an
important source for the Indian caste system as we know it today (Hart
1987). Essentially, he argues that the indigenous South Indian political
model was based around the ritual power of the king, supported by a group
of ritual and magical specialists of low social status. Hart has also suggested
that this provides an indigenous model for the caste system, and that we
23
24
25
A recent novel by the Indological scholar Kamil Zvelebil gives a fictional, though historically based,
account of the Roman trade with India (Zvelebil 2001).
All this is today rather political. Given the anti-Brahmanical and anti-Sanskritic tendency of much
Tamil politics in the twentieth century, there is a tendency for some modern Tamil historical writing
to play down Brahmanical influences at the expense of Buddhist and indigenous material, just as
there is a tendency among North Indian writers influenced by Hindu nationalism to emphasise
Vedic and Brahmanical material and push it back into earlier periods.
Datings for this literature vary but the scholarly consensus seems to be settling on this period (e.g.
Zvelebil 1992).
Two worlds and their interactions
85
should look here rather than to the Indo-Aryan varn.a model for central
aspects of the Indian caste system as it later developed.
The question of the origins of caste (here referring to what in Sanskrit
and North Indian vernaculars is referred to as jāti, not varn.a, see below) is
a large and important one, and any substantial treatment of it would take
me well beyond the scope of this book. Controversies over caste in recent
years have centred around a number of related issues: how far one can talk
about a single caste system for India; how far back we can trace caste as
we know it in modern times into earlier Indian society; how dominant the
values associated with caste were in Indian society; how far caste represents
an autonomous domain of hierarchy which can, conceptually at least, be
seen as distinct from the authority of the king.
On all these issues, Louis Dumont’s classic work on the caste system,
Homo Hierarchicus, took a clear position: he held that there was an underlying caste system, of which local caste hierarchies are individual expressions;
that this was the basic structure of pre-modern Brahmanical society; that
its values totally dominated all levels of Indian society; and that the Brahmanical domain of caste hierarchy should, conceptually at least, be kept
separate from the power of the king (L. Dumont 1972).
There has been a great deal of rethinking and further study since the time
of Dumont, and all of these contentions have been variously contested.
Dumont’s work was a major contribution, if only in the clarity with which
it stated the Brahmanical perspective on Indian society (Berreman 1971;
Mencher 1974). More recent re-analyses, however, have surely been right
to emphasise the extent to which the caste system is always about power,
politics and practical success. Caste cannot be treated as primarily derivative
from a hierarchy of relative purity, with other factors as a simple add-on
(Quigley 1993; Searle-Chatterjee and Sharma 1994). If only because of its
close relation to the political domain, caste is constantly open to change and
transformation, and it is hard to believe that Dumont’s idealised picture
of the caste system was ever an empirical reality anywhere, though it has
undoubtedly been a powerful ideological conception in many places and
times.
Susan Bayly’s recent demonstration of the relative modernity of much
of what we now recognise as the caste system, in her recent book Caste,
Society and Politics in India, may itself be open to criticism in details, but
the general picture it paints is persuasive (Bayly 2001). As for values, I have
argued elsewhere that at the very least caste needs to be supplemented within
the value-systems of everyday Indian life by additional systems of values
more oriented around the practicalities of everyday life (Samuel 1997).
86
Origins of yoga and Tantra
South Indian model of caste
King
Society
(except
polluted
groups)
Various low-status and ‘polluted’ groups who support king via ritual services
Figure 4.4. The Early South Indian Caste System according to George Hart
The relevance of Hart’s work on early South India to all this is that he
argues that those aspects of caste which are about polluted and low status
can be traced back to a period of Tamil society which, arguably, had not
yet been substantially restructured in accordance with Brahmanical norms.
The ritual kingship of the early Tamil king was dependent on a variety
of court ritualists, including drummers and other musicians, who were
essential to the maintenance of the king’s power, but who were themselves
regarded as of low and ‘polluted’ status, like the so-called untouchables or
dalit groups of modern Indian society (Hart 1987).26
In other words, the major division in society was between a majority
without internal caste-like divisions and a minority consisting of a number
of small occupationally-polluted groups (Fig. 4.4).
This is a completely different structure from that familiar to us from
early Vedic sources. Here, by contrast, we find the well-known division
into four groups, the three upper varn.as (Brahmin, ks.atriyas and vaiśyas),
and the remainder, the śūdras (Fig. 4.5).
Within this system, we have a series of distinctions, between the ks.atriya
and the Brahmins, between these two and the vaiśyas, and between these
three ‘twice-born’ castes and the śūdras. The first three categories here
26
Dalit and low-status groups have continued to have an important ritual role in Indian society into
modern times, particularly in the South, and there is an extensive body of anthropological material
bearing on this theme. See for example Brubaker 1979.
Two worlds and their interactions
87
Brahmanical Model of Caste (varna)
Three twice-born castes
Ksatriya
Brahmin
´
Vaisya
´
Sudra
Figure 4.5. The Brahmanical Varn.a System
probably derive, like the parallel formulations which Georges Dumézil
pointed out in other Indo-European societies such as Rome, Greece or
Scandinavia, from a common Indo-European basis, while its extension to
a fourfold structure through the addition of the śūdras is presumably a
Brahmanical formulation from North India.27 If Witzel is right, in fact,
we can locate this development in the early ideological restructuring of the
Kuru state, which also insisted on the dominant position of the Brahmins
as the most pure of the varn.as.28
However, as is well-known, this does not constitute the caste system in
any of the various forms in which we find it in present-day South Asia.
As has been shown by a wide variety of anthropological studies, the most
‘typical’ form of the South Asian caste system actually consists of a large
number of in-marrying sub-groups (jāti) (e.g. L. Dumont 1972). Typically
some of these groups are generally accepted by members of the community
as being at the pure end of a spectrum of purity and pollution, others at
being the impure end. Disagreements about status are however endemic,
often phrased in terms of competing claims to Brahmin, ks.atriya or vaiśya
status, and expressed through refusal to accept various kinds of food or drink
27
28
A convenient introduction to Dumézil’s theories is provided by Littleton 1982. Discussion in recent
years has focused on the question of whether a ‘fourth function’ is implied in the general IndoEuropean context, see e.g. N. Allen 1987, 1996.
It may be noted that this Brahmanical dominance was contested outside the Kuru region. We know
from Buddhist texts that the priority of the Brahmins was not accepted by the compilers of the Pali
Canon, who repeatedly represent Śakyamuni as informing Brahmins that they are inferior by caste
status to ks.atriyas. Thus it would seem that while the people of the Central Gangetic region accepted
the general idea of the division into Brahmin, ks.atriyas, and vaiśyas, they rejected the primacy of the
Brahmins that was a key feature of the Kuru-Pañcāla system.
88
Origins of yoga and Tantra
from other groups (implying that the members of the other groups are less
pure). It is also clear that the ultimate authority in terms of determining
status lay in the local ruler. Many agricultural communities also include
a large majority of a single caste, usually peasant farmers, with so-called
untouchable groups owning little or no land and providing agricultural
labour for this ‘dominant’ caste.
This is a very different picture from the varn.a system. At most the
varn.a system provides one component of the caste system, the idea of a
hierarchy of status in which the Brahmins occupy the highest position and
claims to membership of other twice-born varn.a provide the basis for status
competition. The other major component of the caste system is provided,
Hart suggests, by the ritual kingship of South India, with its low-status
polluted ritual practitioners.
These low-status polluted ritual groups are an important issue in relation
to the development of the ascetic orders, and later of Tantra, and I will
return to them in Chapter 6 in this connection. For the present, I want
to note, however, that there are indications that this scheme was not just
a local South Indian issue but may have had a wider distribution. It may
well be that the kind of situation described by Hart, in which the ruler was
surrounded by a number of groups of occupational specialists with ritual
functions, but low or compromised social status, was more widespread
in early times.29 We still find something very much like it today among
Tibetan communities in Ladakh, Nepal and elsewhere in the Himalayas,
where low-status drummers (usually referred to as Mon), blacksmiths and
other specialists retain a significant religious role (e.g. Jest 1976; Kaplanian
1981: 176–89; Dollfus 1989).
In fact, the North Indian system of social stratification itself, in social
reality rather than Brahmanical theory, may have been more like this even
before the encounter with the south got under way. A common pattern
in many areas in recent times, including large parts of Bengal and even
the Punjab, is one where the majority of the population belonged to a
single cultivator-type caste, with small numbers of attached low-caste ritual
specialists of various kinds. This again seems to hark back more to the kind
of model suggested by Hart, especially when it is borne in mind that many
priests, whether or not they claim Brahmin identity, are in various ways still
regarded as being of low or compromised status. The preoccupation with
Brahmanical texts such as the Manusmr.ti (‘Laws of Manu’) with explaining
29
For an example in a pastoralist society, see Galaty 1979. It is interesting though that of the five lowcaste occupations regularly mentioned in the Pali Canon (can.d.āla, nesāda, ven.a, rathakāra, pukkusa,
e.g. Masefield 1986: 148) none appears to have a ritual role.
Two worlds and their interactions
89
how low-status and outcaste groups originated from improper marriages
of various kinds may be an early attempt to give an acceptable ideological
spin to this situation (cf. Tambiah 1973).
ge nder at titudes across south and southeast asia
A further issue needs to be introduced at this point; the variations in gender
attitudes and specifically in the status of women across the entire region of
South and Southeast Asia. Southeast Asia as a region is characterised by a
relatively high status for women and by relatively equal gender relations,
and it seems clear that this pattern has a considerable history in the region:
Relations between the sexes are one of the areas in which a distinctive Southeast
Asian pattern exists. Even the gradual strengthening of the influence of Islam,
Christianity, Buddhism and Confucianism in their respective spheres over the last
four centuries has by no means eliminated this common pattern of relatively high
female autonomy and economic importance. In the sixteenth and seventeenth
centuries the region probably represented one extreme of human experience on
these issues. It could not be said that women were equal to men, since there were
very few areas in which they competed directly. Women had different functions
from men, but these included transplanting and harvesting rice, weaving and
marketing. Their reproductive role gave them magical and ritual powers which
it was difficult for men to match. These factors may explain why the value of
daughters was never questioned in Southeast Asia as it was in China, India, and
the Middle East; on the contrary, ‘the more daughters a man has, the richer he is’.30
All this could certainly not be said of contemporary South Asia. Despite
the significance of female forms of divinity in modern Hinduism, and a
marked role of women in ritual, particularly at the domestic and village
level, the status of women in most of South Asia is undoubtedly lower than
that of men (cf. Samuel 1997). As for the value of daughters, the economic
costs associated with daughters, specifically in relation to dowry payments,
is now so high that the South Asian population has a large deficit of women,
much of it undoubtedly resulting from abortion, infanticide and selective
directing of health and food resources to male children. While the inflation
of dowry in recent years and the vulnerability of the poorer sections of
the South Asian population to international market forces have made this
situation much worse, it also reflects a preference for sons over daughters
that goes back for many centuries.
30
Reid 1988: 629; the quote at the end is from the sixteenth-century Portuguese author Antonio Galvão
(Galvão 1971: 89).
90
Origins of yoga and Tantra
A widespread North Indian, Pakistani and Bangladeshi village marriage
pattern exacerbates women’s low status. Women typically move at marriage
into their in-law’s household, generally in another village at some distance
from their natal household. This creates a structural vulnerability for young
married women, who are isolated in a household where they have no close
ties except to their husband, who is himself typically in a junior position
in an extended family household.
Ideas of female bodily functions, above all menstruation and childbirth,
as intrinsically polluting appear to be deeply rooted in much of South Asian
culture, and clearly provide at least one level of ideological justification for
women being viewed as inferior (Rozario and Samuel 2002a). Southeast
Asia is again different in this respect, and the differences are evident in
terms of the attitudes to and treatment of childbirth practitioners in South
and Southeast Asia. In much of village India, Pakistan and Bangladesh,
childbirth practitioners (dai) are of low status, barely remunerated for their
services and mostly from dalit (‘untouchable’) caste backgrounds. Educated
women are reluctant to take on this work, and even where doctors and
trained nurses or midwives are present they are frequently unwilling to
come into physical contact with women in childbirth. The consequences
for the mortality of mothers and children in childbirth are appalling, and
there is little sign of the situation changing (cf Rozario and Samuel 2002b).
At the same time, South Asia is not uniform in this respect. South India
and Sri Lanka have a markedly more positive attitude to female bodily
functions. There is a widespread practice of ritual celebration of a girl’s
first menstruation in these regions. Traditional childbirth practitioners also
appear to be more respected, and childbirth practices more focused on
the welfare of mother and child, and less on issues of pollution (Samuel
2002b: 10). It has generally been argued by anthropologists that this is
linked to the South Indian marriage and kinship system, which encourages
marriage between related families, so that brides are rarely moving into a
completely isolated situation in their new household, and also operates so
as to maintain status equality between households.31
Any attempt to deal with such complex matters in such a limited compass
is inevitably somewhat of a caricature, and I am aware of the limitations of
the above account. It nevertheless reflects something real about the nature
of South and Southeast Asian societies in recent times, and it raises the
question of the historical dimension of these differences.
31
It has been suggested that the kinship system in the Central Gangetic region may have been more
‘Dravidian’ in form at the time of the Buddha. The evidence for this is weak, see Trautmann 1979.
Two worlds and their interactions
91
Here there are a number of points to be made in brief, some of which I
shall return to in later chapters. An initial point is that the negative valuation
of female biological functions is closely tied up with aspects of VedicBrahmanical religion. Elsewhere I have cited Julia Leslie’s discussion of the
widespread myth of Indra’s guilt, and its transfer to women, where it takes
the form of menstruation. The myth occurs as early as the Taittirı̄yasam
. hitā;
other versions are found in the Mahābhārata, Rāmāyan.a and the Skanda and
Bhāgavata Purān.as (Leslie 1996a). The earth is also a frequent recipient of
Indra’s guilt, and its seasonal changes are treated as a form of menstruation.
Menstruation is thus the sign of a woman’s participation in Brahmin-murder. It
marks her innate impurity, her cyclical fecundity, her uncontrollable sexuality, and,
by extension, the inescapable wickedness of her female nature. (Leslie 1996a: 91)
What is clear, though, is that the Brahmanical set of values, with its basically
soteriological orientation centred around concepts of purity and casteduty, is not the only set of values found within Hindu villages today. A
countervailing set of values, generally referred to by anthropologists as
‘auspiciousness’, and centred around fertility, productivity, and this-worldly
success, is also very much present. Auspiciousness has a close linkage both
with women and with goddesses (Samuel 1997).
I discuss the relationship between these two value systems further in
Chapter 7. There is, I think, a historical as well as a logical relationship
between the two sets of values.
The presence of significant festivals which celebrate the menstruation of
a female deity in parts of East India (Orissa, Bengal and Assam, cf. Marglin
1994, 1995; Samanta 1992: 59–60; Chawla 2006) is one significant indication of an earlier situation which gave more positive recognition to female
values. So also perhaps is the presence of more gender equality and a
more positive orientation to sexuality in various Indian ‘tribal’ populations, though I think we need to approach this material with considerable
caution. Contemporary ‘tribal’ populations in India and South Asia cannot
reliably be treated as a relic of earlier times. While there may be and probably are cultural features which derive from earlier patterns, these cannot
be easily disentangled from more recent developments.
Here it is important to appreciate that ‘tribal’ populations in South and
Southeast Asia today cover a variety of different situations. Those in much
of South Asia have been involved over more than two millennia with mainstream forms of Indian society. Thus there is also a long-term relationship
of mutual influence and of systematic mutual differentiation. The tribes
are what they are today in large part because of their dialectical relationship
92
Origins of yoga and Tantra
with the ‘mainstream’ population (cf. e.g. Gell 1982). In addition, there is
a tendency to see the tribes from the mainstream Hindu perspective precisely in terms of stereotypes of sexual freedom. At the same time, there are
indications of preserved values that are indeed more gender-positive and
more celebratory of female sexuality and fertility (e.g. Elwin 1947; Archer
1974; Zvelebil 1988).
The highland tribal populations of areas such as Arunachal Pradesh,
Assam, Mizoram, Tripura, and Manipur are in a rather different situation, and one can class these in general terms along with many of the
Southeast Asian highland populations. Here, there is also a recognisable
highland-lowland dialectic, but tribal populations have generally had more
autonomy and more ability to conduct life on their own terms, at least until
relatively recent times. There is an interesting body of literature focusing
on these tribal populations and their relationship to mainland Southeast
Asian populations. Durrenberger and Tannenbaum have noted the lack of
penetration of mainstream (i.e. Buddhist) religious discourse into highland
populations, which continue to emphasise values of fertility, productivity
and success values over the soteriological goals of mainstream religious traditions (Durrenberger and Tannenbaum 1989). All this raises a number of
interesting and important issues (cf. Kirsch 1973; Russell 1989; Kammerer
and Tannenbaum 1996), but they are not ones that I can deal with in detail
in this book, which is essentially concerned with lowland and ‘mainstream’
populations.
One question that should perhaps be considered is the possible historical
impact of the Muslim invasions on gender relations in South Asia. There
undoubtedly was an impact here: one can see it, for example, in the dramatic
shift in modes of dress in South Asia around the time of the Muslim
conquests (Fabri 1994). However, the issue is less one of Islam as such than
of the specific Muslim populations that were involved in the conquest, and
the general feudalisation and militarisation of Indian society on a patriarchal
lineage model that took place before and during the Muslim invasions.32
Elsewhere, as in Southeast Asia (Malaysia and Indonesia), Islam itself has
proved quite capable of coexisting with relatively gender-equal societies.
More specifically, the Afghan and similar Central Asian Muslim populations that took a leading role in the Muslim conquests were themselves
based around a patrilineal clan system with strong patriarchal elements,
and the general militarisation of Indian society over the seventh to twelfth
32
This is the process which Davidson 2002a refers to as sāmantization, and which I shall discuss in
more detail in relation to the growth of Tantra and the fierce goddess cults in Chapter 12.
Two worlds and their interactions
93
centuries tended to reinforce the hierarchical and patriarchal elements of the
Indian value system as against the feminine and auspiciousness elements.
All this coincided with the initial impact of Islamic cultures in northwest
India and with the eventual Muslim conquest of most of India, but it was
as much an outgrowth of endogenous factors as of external ones. This issue
will be considered further in Chapter 12.
This concludes our survey of the various social contexts within which
Indic religions were developing in around 500–400 BCE, or which may
have influenced those developments. Much of the evidence I have presented
is sketchy or indirect, or pertains more to the realm of ideology than of solid
empirical fact, but there is perhaps enough to point to some of the variety
of resources available to the developing religious cults of North India at this
time. In the following chapter we turn to look at these cults and practices.
chap t e r 5
Religion in the early states
I begin this chapter with an overview of evidence about South Asian religion before 500 BCE. This discusses both archaeological evidence, mainly
relating to the Indus Valley urban tradition and textual sources relating to
the early religion of Indo-Aryan speaking peoples. In subsequent sections, I
look at the development of Vedic-Brahmanical religion in the Kuru-Pañcāla
region, and at other aspects of early Indian religion, particularly the religion
of local gods and spirits
a rchaeological sources f or early indic religion
I have already provided some discussion of religion in the Indus Valley cultural tradition in Chapter 1. I suggested there that while there are certainly
features of the Indus Valley material that might be interpreted in terms of
continuities with later periods, we know very little for certain. The evidence
is capable of many interpretations, and analyses are so heavily dependent
on reading later practices and concepts into the material that they are of
little help in evaluating whether there really were continuities.
One intriguing indication of possible continuities in religious practices
is provided by a striking group of terracotta figurines from Mehrgarh, dated
to about 2800–2600 BCE, so two or three centuries before the Indus Valley
Integration era. Some of the female figurines from Mehrgarh have a hairparting with a streak of red pigment, and it has been suggested that this can
be related to the modern Hindu practice of women placing sindhur in their
hair-parting as a sign of their married status (e.g. Kenoyer 1998: 44–5). In
fact, it is not clear whether the red pigment on these figurines is confined
to the hair-parting, or found all over the bodies, and it appears that some
at least of the figurines are modern fakes,1 so it may be unwise to put too
much emphasis on the similarity to the modern use of sindhur. The use of
1
Theresa McCullough (personal communication, September 2006).
94
Religion in the early states
95
red ochre in sacred contexts remains common, however, in modern Indian
religious practice more generally. Deity images, especially rocks or natural
features held to represent female deities, are often coated in red pigment,
and the association of red with fertility is also fairly explicit in many folk
ritual contexts. Here again, though, we have nothing at all specific about
how these figurines were used or regarded by the people who made them.
As far as the Indus Valley seal images are concerned, we can perhaps
accept the likelihood of the cult of some kind of goddess, and of sacrificial
offerings to her, on the basis of the ‘fig tree deity’ seals discussed in Chapter 1
(e.g. Fig. 1.2). The existence of theriomorphic deities (gods in animal form)
in the Indus Valley Integration Era also seems fairly likely, on the basis of
the many seals that show animals, most often cattle of some kind, with what
may be offerings in front of them. This might link with the cult of wild yaks,
rams and other animals in Tibet, and I have suggested elsewhere that there
may be some linkages between the Indus Valley culture and early Tibetan
religion (Samuel 2000). But it would be unwise to build too much on
such connections. The linkages between any Indus Valley cult of goddesses
or theriomorphic deities and the Vedic religion, or any other subsequent
Indic religion, are far from obvious. Nor do the horned headdresses visible
in the Indus Valley seals (cf. Figs. 1.1, 1.2) link up with anything obvious
in later Indic cultures.2 The Indo-Aryan immigrants clearly had their own
cattle-centred culture, as is evident both from the R.gveda and from the
comparable Avestan materials, so we hardly need an Indus valley bull cult
to explain the significance of cattle in later times.
My comments in Chapter 1 referred primarily to the so-called ‘Integration Era’ of the Indus Valley cultural tradition, in other words the period
of the large-scale urban society that we know from Mohenjo-daro and
Harappa in particular. This period is now generally dated from approximately 2600 to 1900 BCE. In relation to the period from the end of the Integration Era up to around 500 BCE (in fact, until the start of the Mauryan
state) there is again little direct archaeological evidence relating to religion.3
A number of figurines dating from this period have been regarded by some
authors as representing goddesses. These include, for example, some small
nude figurines from early farming settlements in Maharashtra, which appear
2
3
The only parallel known to me is with the bison-horn headdresses used by some Gond tribes, but
again the connection is highly speculative.
One study from well outside the Indus region is worth mentioning: F. R. Allchin’s work on prehistoric
ash-mounds from around 2000 to 750 BCE in Mysore and Andhra Pradesh. Allchin has suggested
that these mounds may result from annual festivals involving fire and the worship of cattle, similar
to the modern festivals of Holı̄, Dı̄valı̄ and Pongal (Allchin 1963; Sullivan 1971).
96
Origins of yoga and Tantra
to have been associated with storage jars or grain silos (Poster 1986, cat.
Nos. 8–10, pp. 80–2; see also M. Joshi 2002: 40). These figurines may date
from 1300 to 1200 BCE. A group of somewhat later terracotta figures from
the Swat valley in Pakistan have also been interpreted as mother goddess figures (Poster 1986, cat. No.12, p. 84). There is need for caution when it comes
to identifying female figurines as representing a goddess cult (see Haaland
and Haaland 1995; Goodison and Morris 1998), but there may be some justification for these interpretations. A much more substantial iconography of
figures representing female deities can be found from the Mauryan period
onwards, so it would not be surprising if there were earlier examples.
It is interesting in any case to note that while the Vedic tradition seems
not to have employed images of deities until late Mauryan or Śuṅga times
(third to second century BCE) at the earliest, sculptural images of deities
may go back much further in these folk contexts. Given the fragile nature
of the evidence, and the fact that many images may have been made of
substances that would not leave any trace in the archaeological record (e.g.
cow-dung, regularly used for images of deities in the folk tradition in
South Asia in modern times, cf. Samuel 2005a: 259, 280 n.3), such practices
could have been quite common and widely distributed. The same is true
of ritual wall-painting and floor-painting (kolam, alpana, aripan, muggu,
etc.), again very widespread in modern times but unlikely to survive in the
archaeological record (E. Gupta 1983; Nagarajan 1997; Thakur n.d.: 36–50;
Kilambi 1985; Jayakar 1989: 118–19).
t extual sources f or the early religion of the
in do-aryan-speaking peoples
The oldest datable evidence for the early religion of the Indo-Aryanspeaking peoples is the Mitanni inscription, a peace treaty between the
Hurrian kingdom of the Mitanni and the Hittites, and it dates from about
1360–1380 BCE (Thieme 1960). Five male deities are called as witnesses and
keepers of the treaty, and their names are recognisable Vedic Sanskrit for
Mitra, Varun.a, Indra and the twin divine horsemen or Nāsatyas (Aśvins).
All these are well-known Vedic deities, though only Indra is a central figure
in the Vedas, suggesting that the cult of the other deities (who belong to a
group known as the Ādityas) had become less significant in the meantime.4
4
Cf. Brereton on the Ādityas (Brereton 1981). Insler notes that the predominance of Indra is primarily
in books Three, Four, Six and Eight of the R.gveda; in books Five and Seven, there are almost as many
hymns to Varun.a and the Ādityas as to Indra (Insler 2003).
Religion in the early states
97
The early Iranian material, as noted in Chapter 2, is linguistically and
stylistically close to the hymns of the R.gveda. However, it post-dates the
reconstruction of Iranian religion associated with the figure of Zarathustra
and incorporates his rejection of the cult of the daevas (corresponding to
the Vedic devas or gods) so that the Vedic deities do not appear directly
(Insler 2003).5
Stanley Insler suggests that the Vedic and Iranian material together points
to a primarily pastoralist culture, living in small villages and settlements
under chieftains or minor sovereigns who were expected to protect their
subjects and also to judge in disputes. The deities Mitra and Varun.a are
associated with the role of these minor sovereigns. There was also a warlord
chosen temporarily in time of attack, and associated with the deity Indra.
These people seem to have called themselves ārya (perhaps originally meaning ‘hospitable’, ‘cultivated’).
Their primary ritual was a sacrifice or offering performed outdoors under
an open sky and involving the delimitation of a sacred space, its covering
with grass to provide a comfortable seat for the deities, and the lighting
of a fire to provide warmth and represent truth. Food was then offered
to the deities; on big occasions, animals were sacrificed and soma juice6
prepared. It was important to have a priest (Skt. hotr., Avestan haotar)
present to formulate the praises that accompanied the sacred meal, and
which correspond to the R.gvedic hymns or the Avestan gāthās.
th e vedic religion and the wid er religious scene
If Insler’s reconstruction is a reasonable picture of Indo-Aryan religion
at around the middle of the second millennium BCE, before the IndoAryan speakers arrived in North India, then the religion of the R.gveda,
the Brahman.as and associated texts already represent a substantial reformulation. While a relatively simple set of domestic or gr.hya rituals centred
5
6
Of the five Vedic deities in the Mitanni treaty, only Mitra appears unambiguously in the early Iranian
material, though Stanley Insler has suggested that Ahura Mazdā, the new deity associated with the
reforms of Zarathustra, takes over many features of Varun.a, and that indications of Indra, the Nāsatyas
and another Vedic deity, Aryaman, can also be found in Zarathustra’s hymns (the Gāthās). Indra,
Saurva and Nanhaithya, corresponding to Vedic Indra, Sarva and the Nāsatyas, occur in the later
Avesta as members of the rejected daeva category (Insler 2003). Parpola has developed a complex
argument according to which the presence of Mitra-Varun.a alongside Indra in the Mitanni material
indicates a temporary compromise reached in Bactria between asura and deva worshippers (Parpola
1999a). See also Parpola 2005.
The precise nature of soma is still uncertain, though an extract of Ephedra seems the most plausible
current opinion (cf. Houben 2003a, 2003b). See also Spess 2000, who argues for species of Nymphaea
and Nelumbo.
98
Origins of yoga and Tantra
around the domestic sacred fire, which is set up at marriage, was practised at the household level, and is in fact central to the Vedic ideology of
marriage and the household (B. Smith 1989: 146–68), a much more elaborate series of rituals performed by the brahmin on behalf of others had
developed alongside it. This śrauta ritual is a ‘highly complex and very
expensive set of sacrifices requiring the services of an array of ritual specialists’ (Olivelle 1998: xli). It involves several ritual fires and a number of
priests. Several major rituals are explicitly linked to kingship (the rājasūya
or royal consecration and the aśvamedha or horse sacrifice) and the whole
body of material suggests a development and reconstruction in the context
of a much-expanded royal role. It seems clear that much of this would
have taken place some time before the Second Urbanisation got properly
under way. This process can doubtless be associated with the transformation of the arriving Indo-Aryans, along with elements of the existing local
populations, into an increasingly settled and largely agricultural society,
although one that was still largely committed to pastoralist values at an
ideological level.7 The soma ritual continued as a key part of the complex, but it appears that actual knowledge of the soma plant was lost at
an early stage, perhaps before the settlement in Northwest India, so that
soma became a focus for symbolic elaboration rather than the ingestion of
a pharmocologically-active substance.8
I have already referred to Witzel’s arguments regarding this process of
reformulation, which he associates with the development of the Kuru state,
initially a ‘super-tribe’ or tribal federation, and will return to them again
later. At this point, it is important to remind ourselves that these developments were, initially at least, regional developments, confined to the Kuru
state and its immediate neighbours.
As I noted earlier, the development of Indian religions is most often
presented in ‘orthogenetic’ style, as something that grows out of the Vedic
material, with other influences coming in as external influences if at all.
This reflects the Brahmanical textual tradition’s view of itself, but it does not
necessarily represent the way things happened at the time. The Brahmanical
7
8
In this it might resemble a range of East and Southern African populations in more recent times,
which combined a mostly agricultural economy with ritual and ideological complexes that were
still largely based around pastoralism. The Ndembu, studied by Victor Turner, and related Lunda
populations, are classic examples (cf. Turner 1957, 1968, 1970).
A brief description of the soma sacrifice is provided by Olivelle (1998: xlv). One of the problems with
identifying soma with ephedra (see above) is that ephedra has little in the way of psychoactive effect.
However, it is by no means certain that the original soma was psychoactive. The extensive use of
cannabis by modern Hindu and Muslim ascetics in India is nevertheless suggestive of a long-standing
tradition of employment of psychoactive substances (cf. Movik 2000).
Religion in the early states
99
texts are written from within the Vedic-Brahmanical setting, and often at
a much later date than the events to which they refer. The earliest layers of
Buddhist and Jaina material, to the extent that we can reconstruct them,
depict a rather different model, but they too tend to have been read through
assumptions of the kind of Brahmanical presence characteristic of modern
India (see Chapter 6).
The material presented in Chapter 4 suggests a more complex picture. To
make sense of it we need to keep a number of issues distinct that can easily
be confused and have often been conflated in the past. The first distinction
that needs to be held clear is between the immigrant Indo-Aryan-speaking
populations which (presumably) entered South Asia in the course of the
second millennium BCE (possibly continuing up to about 800 BCE or
later in some parts on Witzel’s model), and the settled Indo-Aryan-speaking
populations of the mid to late first millennium BCE. We can assume that
the latter were all by this time mixed populations of immigrants and earlier
residents, though the proportions of the mix doubtless differed from place
to place, with the ratio of immigrants to prior population probably highest
in the Northwest and decreasing eastwards towards Bihar and Bengal and
southwards towards the Deccan plateau.
These groups were all speaking Indo-Aryan languages by 500 BCE or
so, and we can presume that they had all accepted some aspects of the
cultural heritage that came with the immigrant Indo-Aryan speakers. These
would have included the tripartite Indo-European distinction between a
ruling warrior group (ks.atriya), a priesthood carrying out sacrificial ritual
(Brahmins) and a wider population, now largely farmers (vaiśya). We can
probably also assume some presence of Indo-Aryan deities and rituals,
including the deities Indra and Brahma, both prominent in the Buddhist
texts, and perhaps some of the widespread seasonal rituals. I refer to this kind
of material for convenience as the generic Indo-Aryan cultural tradition,
though we should not assume that it was entirely uniform or equally strong
throughout this whole large area.
A second important distinction is between this generic Indo-Aryan cultural tradition and the specific body of rituals and practices which were
developing in the Kuru state and the Kuru-Pañcāla region. I refer to
the latter as the early Vedic-Brahmanical ritual complex or early VedicBrahmanical religion. The use of ‘early’ is meant to distinguish it from
the modified and transformed version of this complex which was actively
propagated throughout the wider Indo-Aryan-speaking region, perhaps
from around 400 BCE onwards. We will consider this development later
in this chapter.
100
Origins of yoga and Tantra
There appears to be evidence of shared material, and also evidence of the
formation of a distinctive self-conscious group in the Kuruks.etra region
which saw itself as ‘Aryan’ as contrasted with other Indo-Aryan-speaking
populations. Aryan in the Iranian context appears to be a term applied to
Indo-Aryan populations as a whole (Shahbazi 2005) so this is presumably
a development in the Indian context. I assume that it was consequent
upon the differentiation of the Indo-Aryan-speaking population in India,
whether before or after their arrival, and perhaps also on differing degrees
of accommodation to and adoption of practices of previous populations in
South Asia. At the same time the Kuru-Pañcāla state was associated with an
evident ‘drive to the East’, a desire or need to convert the populations of the
Central Gangetic region to the new Brahmanical orthodoxy (Heesterman
1962).
It would seem that by the time of the historical Buddha and of the
Jaina teacher Mahāvı̄ra, the generic Indo-Aryan cultural tradition was an
accepted part of society through much of the Central Gangetic region.
There were also Brahmins and a degree of movement between the Brahmins
of this region and those of Kuru-Pañcala. It seems clear, however, that the
nature of Vedic and Brahmanical religion in this region was different and
considerably less dominant than in the Kuru-Pañcāla region. Even in texts
such as the Upanis.ads, which appear to originate in some cases from the
North, in others from the Central Gangetic region, we appear to have an
ongoing dialogue between models of what Brahmins are or should be doing,
about whether Brahmanical knowledge represents the ultimate truth, etc.
(Olivelle 1998: xxxvii–xl). This is even more obviously true, of course, in
relation to the Buddhist and Jaina material.
As for the suggestion, still prevalent in popular literature, that Buddhism
represented a protest against the pre-existing Brahmanical caste system,
there seems little truth in this. The Buddha’s comments on Brahmin claims
of high caste suggest less an opposition to an already imposed caste system
than a refusal by a spiritual leader belonging to an established group of
high status to accept a new imported Vedic-Brahmanical model in which
the Brahmins are supreme.9 References to Vedic material in the earliest
Buddhist literature are limited although there is evident knowledge of the
existence of the R.g, Sāma and Yajur Vedas and suggestions of detailed
engagement with Vedic ideas (e.g. Jurewicz 2000). There are also fairly
clear references to material that is found in the Upanis.ads (Gombrich
9
See Dı̄ghanikāya 3 (Ambat.t.ha Sutta), Dı̄ghanikāya 27 (Aggañña Sutta), Majjhimanikāya 93 (Assalāyana
Sutta) etc.
Religion in the early states
101
1990, 1992), indicating that this material was in circulation in some form
or another, if not necessarily that of even the oldest Upanis.ads as we know
them today. I return to some of these issues in the next chapter, in relation
to the origins of the ascetic tradition.
There are indications, however, as we have seen, of royal patronage for
Kuru-Pañcāla Brahmins in the evolving states and we may suppose that
one of the main things which they could offer was the body of ritual
practices regarding state power which had been developed in the KuruPañcāla context. The model of kingship that developed in South and much
of Southeast Asia, in other words, largely adopted a Vedic-Brahmanical
ritual idiom, as indeed is the case in Buddhist states in modern times.
As we will see, there were extensive efforts on the point of the Buddhist
tradition in later times to provide its own rituals for royalty, but it seems that
the Vedic-Brahmanical tradition established itself as the prevailing model
in this area quite early on.10
The Buddhist and Jaina texts also provide evidence of more local and
specific levels of religious practice, of cults of what were called the laukika
or ‘worldly’ deities.11 They included the cults of guardian deities (yaks.as,
nāgas, etc.) linked to various early states, as well as no doubt more local
cults at the village level such as are found in modern India. In origins, this
seems to have been essentially separate from both the generic Indo-Aryan
cultural tradition and the specific Vedic-Brahmanical traditions from the
Kuru-Pañcāla region. We might also consider here the cults of warrior
deities (Balarāma/ Sam
. kars.an.a, Vāsudeva, Skanda, etc.) and the other early
religious forms out of which Vais.n.ava and Śaiva cults later developed.
These were eventually incorporated into the developing Vedic-Brahmanical
corpus, but may have started out quite separately.
By the first century BCE it is clear that there are cults of protective deities
of towns, states and families,12 often associated with shrines in groves or on
hills outside the city proper, and frequently centred around a tree where
the deity lived. There seems no reason to assume that these do not go back
to before the time of the historical Buddha, especially given their strong
presence in the sūtra narratives. Cults of deities associated with rivers and
lakes are probably also early.
10
11
12
See Klimburg-Salter 1989 and Walter 2000 for Buddhist developments. Also Chapter 12, below.
The term is used by Brahmanical (Patañjali), Buddhist and Jain authors with approximately the
same meaning, in each case opposed to higher levels of deities who include the Vedic deities Indra,
Brahma etc. (deCaroli 2004: 13 for Brahmanical and Jain authors).
E.g. the Buddha’s clan (deCaroli 2004: 15).
102
Origins of yoga and Tantra
These deities have most usually been referred to in modern times, following Coomaraswamy (1928–31), as yaks.as, and I shall follow this terminology
here for convenience, while recognising that I am using yaks.a and yaks.i as
generic terms for deities which would have been referred to in a variety
of ways at the time, including some that would have been called by other
terms, such as nāgas and devas.13 In later times, the yaks.as along with the
nāgas dwindled to beings of minor importance, guardians of treasures, or
tree-dwelling spirits that may frighten an unwary traveller, a status they
retained into modern times in parts of rural South Asia.
In the late first millennium BCE, and for many subsequent centuries,
yaks.as, male and female, were an important part of the religious landscape.
Male deities of this kind were gods of prosperity and protection (Fig. 5.1).
While they have a kind of warrior role as generals over lesser yaks.as – we
will see some of this below, and more in Chapter 12 – they are not portrayed
in a particularly militaristic fashion. Female deities are generally associated
with prosperity and fertility, both human and agricultural, and are often
associated with tree or plant imagery. It is clear from the Purān.as and
other later sources that these deities were worshipped all over North India,
including the Kuru-Pañcāla region; even Kuruks.etra had a set of four yaks.a
deities who acted as dvārapālas for the sacred territory (Bharadwaj 1989).14
While representations of some of the major figures, such as Śrı̄, Kubera
and Sūrya, are found in many different locations, for the most part yaks.as
and nāgas were associated with specific regions and localities. This is clear
in the later discussion of these deities in the Epics and Purān.as, as well
as in Buddhist and Jain material. The yaks.as we know of were generally
linked to sizeable towns or regions, though one can imagine that there was
another layer of more local deities again who have left little or no trace in
the literature.
A list of major regional yaks.a deities from a somewhat later period is
included in a Tantric ritual text, the Mahāmāyūrı̄, where a long series of
yaks.as is invoked by name and location, alongside a whole series of other
deities. The Mahāmāyūrı̄ list, which might date from around the fourth or
fifth century CE,15 was studied and translated by Sylvain Lévi (1915), and
again by D. C. Sircar (1971–72).
13
14
15
DeCaroli suggests ‘spirit-deities’ as a generic term (deCaroli 2004). Other recent treatments include
Sutherland 1991 and Misra 1981.
On other local Yaks.a cults see Agrawala and Motichandra 1960; von Mitterwallner 1989.
It is not included in the two early Chinese translations of the text, made in the early fourth century,
but forms part of a translation made in 516 CE (Sircar 1971–72: 262). I am indebted to Will TuladharDouglas for bringing the Mahāmāyūrı̄ yaks.a list to my attention.
Religion in the early states
Figure 5.1. Unidentified Yaks.a. Vidisha Museum
103
104
Origins of yoga and Tantra
Krakucchanda stays at Pāt.alı̄putra, Aparājita at Sthūn.ā, the Yaks.a Śaila at Bhadrapura and Mānava in the northern quarter.
Vajrapān.i stays at Rājagr.ha; he has his abode at Gr.dhrakūt.a; three times he
traverses the earth as far as the ocean; he has great strength and great might and
his valour spreads over (or, step covers) a hundred yojanas.
The Yaks.a Garud.a stays at Vipula, Citragupta at Sthitimukha and the Yaks.a
Vakula possessing a great army and great strength stays at Rājagr.ha.
The two Yaks.as, Kāla and Upakālaka, live at Kapilavāstu where the sage Buddha
called Śākyaketu and Mahāmuni was born. (Sircar 1971: 265–8)
These are the first few verses, which deal with the area around Pāt.alı̄putra
and the earlier Magadhan capital of Rājagr.ha. The next verses deal with
other towns in the Central Gangetic region associated with the life of
the Buddha, before gradually moving out into the surrounding regions.
The text understandably emphasises those deities who are important for
the Buddhists, such as Vajrapān.i (Fig. 5.2). Vajrapān.i, who appears to have
begun life as the guardian deity of Rājagr.ha, was to become a very important
figure in the evolution of Tantric Buddhism. He was regarded as having
become a devotee of the Buddha and already appears in this role in the
early sutras.
By the first and second centuries CE, Vajrapān.i is frequently depicted in
Buddhist iconography, as in this Gandhāran image, where he is standing
behind the Buddha and holding his vajra, as always. The set of the Four
Great Kings, consisting of Vaiśravan.a (Kubera, the king of the yaks.as),
Dhr.tarās.t.ra, Virūd.haka and, Virūpāks.a, also developed an important role
as guardian deities, and are still often painted on the walls of Buddhist
temples in the Tibetan and East Asian traditions (Fig. 5.3). Other important
yaks.a deities include Man.ibhadra, who seems to have become a guardian
deity for merchants and travellers (Thapan 1997) and Pūrn.abhadra, who are
both included among Kubera’s eight attendant yaks.as in Buddhist material.
One can perhaps get some of the character of these deities, if in a literary
mode, from the account of Vaiśravan.a and Pāñcika helping out prince
Sudhana in the Divyāvadāna:
At that very time, the great king Vaiśravan.a, attended by many yaks.as, many
hundreds, many thousands, many hundreds of thousands of yaks.as, was travelling
that way to a meeting of yaks.as. [. . .] [H]e caught sight of Prince Sudhana and
it occurred to him, ‘This is the Bodhisattva of the present Auspicious Aeon; he is
headed for disaster, setting out for battle! I should help him out. That hill-tribe
chieftain must be made to submit, but without harm being inflicted on any living
being.’ Knowing this, he summoned Pāñcika, the great field marshal of the yaks.as:
Religion in the early states
105
Figure 5.2. Indra disguised as a woodcutter offers grass to Śākyamuni (accompanied by
Vajrapān.i), Gandhara, 1st cent CE, Peshawar Museum
‘Come, Pāñcika! Make the hill-tribe chieftain submit to Prince Sudhana without
a fight and without harm being inflicted on any living being!
‘Very well,’ Pāñcika, the great field marshal of the yaks.as, replied to the great king
Vaiśravan.a, and he created the four divisions of a divine army: men as tall as palm
trees, elephants the size of mountains and horses the size of elephants. Then, using
both all manner of weapons, such as swords, clubs, lances, javelins, discuses, pikes
and axes, and the cacophony of massed musical instruments to inspire great fear,
Pāñcika and that mighty host reached the hill-tribe village. (Tatelman 2005: 261)
Needless to say, the hill-tribe village decide that resistance is pointless and
submit without a fight.
The female yaks.a (yaks.i, sometimes yaks.in.ı̄) imagery raises the wider
question of the role of female deities in popular religion. I have already
alluded to the presence of what appears to be goddess imagery from the
106
Origins of yoga and Tantra
Figure 5.3. Vaiśravan.a. Modern Tibetan wall-painting, Tongsa Gompa, Kalimpong
Religion in the early states
107
late second millennium BCE onwards. The early material is difficult to
interpret conclusively, but a body of archaeological material from Mauryan
times onwards is strongly suggestive of a goddess cult of some kind and/or
a religious attitude to sexuality in the first millennium BCE.
In the first place, we could point to a substantial number of images of
figurines from North and Northeast India, mostly dated to the Maurya
period.16 These appear to develop towards later figures whose iconography
is best known from West Bengal, where the type-site is Chandraketugarh
(see below), but which have evident links to similar figures over much of
North and East India, including present-day Pakistan.17 As noted earlier,
one should not jump to conclusions about goddess cults whenever one finds
a female figure, but in these cases the continuity with later iconography
that can be securely identified with goddesses of fertility and auspiciousness
would seem to justify such an assumption. The so-called ‘ring stones’, also
from the Mauryan period, ornamented with what have been plausibly
interpreted as goddess figures, point in a similar direction, as do a number
of other small pieces from the same period and general context.18 The
prominence of vegetation imagery in the terracotta figures and plaques
from Bengal and elsewhere, and to a lesser extent on the ring stones, is
striking, and make it clear that this is a body of imagery which is associated
with a primarily agricultural society.
In fact, the extent to which early Buddhist stūpas and temples were
also ornamented with imagery concerning fertility and auspiciousness is
quite striking. This varies from the largely floral and animal imagery at the
oldest Sāñcı̄ stūpa (No.2; Karlsson 1999: 88–94) to the so-called Gajalaks.mı̄
imagery (two elephants asperging a goddess, e.g. Fig. 5.4, Karlsson 1999: 92
fig.11; in later iconography at any rate the goddess is identified as Laks.mı̄)
and the male-female couples (the so-called mithuna motif, cf. Agrawala
1983) found extensively on early stūpa and temple railings such as those
16
17
18
For reproductions, see e.g. Poster 1986, cat. nos. 18–19 (Mathura, UP), 20 (UP); Klimburg-Salter
1995, cat. nos. 10 (near Peshawar), 11–14, 18, 20 (Mathura, UP), 15–17, 19, 21–22 (UP). It is tempting
to identify other figures, such as the headless female figurines from Maharashtra (Poster 1986, cat.
nos. 8, 10) as goddess figures, but perhaps safer to be cautious.
See Poster 1986: cat. nos. 23–24 (Kausambi, UP; Poster identifies these as ‘Shri Lakshmi’), 25–26
(Mathura, UP); nos. 33–34 (‘north or eastern India’); also note the couple from Mathura, no. 29.23
(Mathura, UP); Harle and Topsfield 1987, cat. No. 6 (Northwest Province in Pakistan), No. 9 (a
small bronze, Northwest India or Pakistan c.100 CE).
For ring stones, see e.g. Klimburg-Salter 1995, cat. no. 3 (Taxila, 3rd to 2nd century BCE), Joshi
2002 (fig. 2, Bihar, 3rd century BCE; fig. 3, Punjab. 3rd century BCE); Allchin 1995c, 263–9 and fig.
11.31 (Taxila and Bihar). For other related material, see Allchin 1995c, figs. 11.32–11.34.
108
Origins of yoga and Tantra
Figure 5.4. Gajalaks.mı̄ on Stūpa 2, Bharhut Stūpa, Indian Museum, Calcutta
at Bharhut and Bodhgaya and as a recurrent motif at Nāgārjunakon.d.a.19
These issues will however be discussed later in the book. A third body of
significant imagery from a somewhat later period is also worth mentioning
(the earliest examples are dated to the first century CE, Desai 1990). These
are the so-called Lajjāgaurı̄ figures, representing nude female figures with
their legs apart, often with a lotus instead of a head.20
19
20
Much the same is true of the Jaina imagery on the so-called votive tablets or āyāgapat.a (Quintanilla
2000).
Desai 1990: 268–9 and fig. 48.3 (Bhita); Poster cat. no. 55 (Jhusi, UP, 2nd century CE); Agrawala
1983, figs. 6 (Karnataka, c.700 CE) and 9 (Nagarjunakonda, 3rd century CE). Also Brown 1990;
Bolon 1992; Agrawala 1983: pls.6 and 9.
Religion in the early states
109
All this iconographic material points to a religion of fertility and auspiciousness that gives full recognition to the female aspect of the procreative
process and which finds its full expression in the early Buddhist sites that I
have just mentioned.21 This religion is often described as the religion of the
yaks.as (fem.) and regarded as forming a single complex with the cults of local
guardian deities mentioned above. I shall discuss it further in the following chapters in the specific context of its relationship with the developing
śraman.a (Buddhist, Jain etc.) and Vedic-Brahmanical traditions.
In relation to goddess imagery, it is also worth noting the very substantial
iconography from Gandhāra and Mathurā, depicting a female figure, often
with one or more children or other signs of fertility and auspiciousness,
sometimes with a male consort. The female figure has traditionally been
referred to as Hāritı̄ and the male figure as her consort Pāñcika but these are
maybe just labels of convenience; in reality there is as far as I know no solid
evidence to link these figures with the Buddhist yaks.as of those names in
Buddhist textual sources, although the Buddhist Haritı̄ is associated with
children and the seventh-century Chinese pilgrim, Yijing, reports a stūpa
linked to her legend in the Gandhāra region (cf. Samuel 2002b). Here as
elsewhere a group of plausible conjectures has become received opinion.22
What is undoubtedly true however is that this goddess of fertility and
prosperity was an important figure at both Gandhāra and Mathurā, since
hundreds of statues and images of all sizes have been found in both locations.
An interesting question here is the possible linkage with the very similar
Iranian iconography associated with the goddess Ardoxsho (who is depicted
with her consort Pharro).23
In the Buddhist legend of Hāritı̄, she is a dangerous yaks.in.ı̄ who is
responsible for the deaths of many children before the Buddha ‘converts’
21
22
23
The well-known erotic imagery on later Hindu temples is doubtless to a large degree an outgrowth
of this early imagery of sexuality and fertility, as will be discussed in later chapters.
While a few of these images bear inscriptions, none as far as I know mention Hāritı̄. The one
apparent exception I have come across, an inscription mentioned in Mathur 1998: 59 (referring to
an image now in the Central Museum, Lahore) is a false trail, since the part of the inscription that
might have mentioned the name is unreadable (see Archaeological Survey of India Annual Report
1903–4, Calcutta 1906, p. 255). It is possible that the inscription mentions smallpox, but this too
is uncertain. Lerner notes that the identification of the Gandhāran-Mathurān goddess with Hāritı̄
relies primarily on the report by Yijing (Lerner 1984: 145).
Ardoxsho occurs frequently on Kus.an.a coinage. John Huntington comments on a couple of Ardoxsho
images from Gandhara, ‘While I believe that Ardoxsho is conflated with Harı̄tı̄ in symbolic function, I feel that her identification on the coins (Harı̄tı̄ does not occur on Kushan coins that I am
aware of ) and the iconographic component of the Cornucopia as seen in the coins and the two
sculptures mentioned above is enough to keep Ardoxsho as a discrete entity’ (H-Buddhism list,
www.h-net.org/∼buddhism/, 24 Feb. 2005).
110
Origins of yoga and Tantra
her and establishes a monastic cult to her (Samuel 2002b; Strong 1992;
de Caroli 2004). This raises the question of whether these deities were in
fact seen only as benevolent and positive figures. In fact, demonic female
spirits seem often to be seen as responsible for childhood illness in later
sources, including the medical treatises and the Puran.as (Wujastyk 1997,
1999). The principal Hindu goddess of childbirth, S.as.t.hı̄, seems to have
had similar associations (Gadon 1997). One could provisionally imagine a
cult of local deities, many of them female, who are responsible for illness if
offended. I will return to this theme in Chapter 10, since it is closely connected with the origins of some parts of the religious complex referred to as
‘Tantra’.
Thus in both regions, the dominant religious tradition of the emerging
state had to reach some kind of accommodation with the local cults of
the settled agricultural population, what I am calling the yaks.a religion in
this chapter. The yaks.as and yaks.ı̄s were mostly benevolent local gods and
goddesses associated with agriculture, fertility and protection. A number of
deities who were important in the historic development of Indian religion
but have no real connection with the Vedic-Brahmanical pantheon seem to
originate in this general stratum: perhaps the best-known cases are Gan.eśa
and Laks.mı̄. Gan.eśa seems to have come to prominence at a fairly late
stage, perhaps around the fifth to sixth centuries CE,24 but Laks.mı̄, or very
similar goddesses, are clearly present in the iconographical record at a much
earlier time.
One of the most consistent and impressive bodies of yaks.a-type imagery
comes from the Ganges Delta, an area that lay outside the penetration of
Vedic-Brahmanical culture until quite a late stage. Certainly there is little
or no sign of the Vedic deities in the imagery of the Chandraketugarh
terracottas, a wide range of which has been made available in Enamul
Haque’s splendid recent book (Haque 2001).
The first of these images to become known in the West, the well-known
Ashmolean yaks.ı̄ relief, was found at Tamluk, the site of the ancient Bengali
port city of Tāmralipti, in 1883 (Harle and Topsfield 1987: 6–7). It was only
much more recently that a large number of similar figures were uncovered in
the Calcutta area, in particular at the site of Chandraketugarh, with which
they have become associated. It is not certain what the ancient identity
24
Though there are some intriguing elephant figures from Chandraketugarh (e.g. Haque 2001, nos.
C673–677, C681), and elephants, apparently as symbols of prosperity and auspiciousness, are also
common at Sāñcı̄.
Religion in the early states
111
of this city was, but from the richness of the material it was evidently an
important place: Dilip Chakrabarti has suggested that it may have been
the capital of the ancient state of Vaṅga (Chakrabarti n.d.). The date is
also somewhat uncertain, though it is generally assumed that the onset of
urbanisation in this area was about 300 BCE, so that the terracottas must
date from after that. This is obviously a sophisticated urban art. Similar
terracottas have been found in other parts of North India, with a similar
general iconography.
I wrote at the beginning of this chapter about the difficulties of interpreting iconography, and suggested that we cannot establish anything very
definite on the basis of the Indus Valley seals. I think that when we come
to the Chandraketugarh terracottas we are in a different situation. The
iconography here is detailed and consistent over a wide range of material. This is particularly true for the large class of female figures depicted
frontally, often with an elaborate and characteristic headdress involving
a number of large hairpins. While many of the Chandraketugarh images
appear to be representations of ordinary human beings, it is clear in the case
of these images that the figure represented is intended to be divine (Bautze
1995; Haque 2001). Versions of this goddess with the elaborate headdress
are found all over North India, and several hundred have been found at
Chandraketugarh and nearby sites.
In one of the most striking of these images (Bautze Fig. XIII; Haque
no. 501), flowers fall from the sky around the goddess and merge into a
shower of coins from a bag in her hand. The resemblance to the common
modern iconography of Laks.mı̄ as showering money on devotees is remarkable. Perhaps it is not accidental that we find this imagery at what would
presumably have been a major trading centre. In any case, it seems entirely
plausible to identify this figure as a version or more exactly a local equivalent
of Laks.mı̄. Another striking group of presumably divine females consists
of figures represented with wings, often against a background of vegetation that seems to make an interpretation in terms fertility and prosperity
difficult to deny (e.g. Haque 2001, C344).
We should be careful about assimilating all goddesses of this kind into
Laks.mı̄, since as I shall later argue this process of assimilation, which is a
characteristic feature of how Brahmanical religion came to work in later
times, is actually part of what we should be studying rather than something
to be taken for granted. At any rate, at a somewhat later period we find
a variety of female deities of prosperity and fertility in different regions
and contexts, including the Brahmanical Śrı̄-devı̄, the mother-goddess
112
Origins of yoga and Tantra
Am
. bikā, who becomes quite important in Jain contexts, the goddess Hāritı̄,
whom we discussed above, and the snake-goddesses Padmāvatı̄ and Manasā
(cf. Cort 1987; Sen 1953).
Despite the large number of terracottas that have been found at Chandraketugarh, there are relatively few male figures that might be deities. A
few are winged, which makes their identity as deities reasonably certain
(e.g. Haque 2001: C319, C321, C328). These and some other figures bear
some resemblance to the yaks.a figures at Bhārhut. A number of male figurines (mostly small rattles) have been tentatively identified as versions
of the important yaks.a deity Kubera, a classic yaks.a figure. Kubera is also
portrayed as a young man, and the winged figures might tentatively be
connected with him. There are also some elephant figures which might
be seen as early versions of Gan.eśa (e.g. Haque 2002: C673). Nothing in
this pantheon, however, can be unambiguously identified with the VedicBrahmanical pantheon, as distinct from the yaks.a-type deities that later
became associated with it. If anything, the winged figures suggest West
Asian influences, though there are also examples of winged figures from
Bharhut and Sañcı̄ (Haque 2001: 98).
Certainly the emphasis in the terracotta iconography is on well-being
and prosperity. Apart from the goddess plaques, which I think can clearly
be regarded as items for religious devotion in one form or another, and a
few conjectural male deity plaques, much of the other iconography seems
playful and perhaps decorative. It includes a great deal of erotic art, including many representations of mithuna or amorous couples, often in sexually
explicit poses. There has of course been an extensive literature on the symbolism of erotic sculpture in Indic religions and I do not intend to get
involved in it at length here, but I think it is striking that here and at
other relatively early sites such as Mathurā, Bodhgayā, Nāgārjunakon.d.a
and elsewhere the theme is treated naturalistically and seems most easily
interpretable as a generalised representation of good fortune and prosperity.
The Tantric treatment of sexuality is a quite different matter, and its real
development lay several centuries in the future.
While issues of fertility and good fortune are clearly important in
the R.gveda and early Vedic-Brahmanical material, and there are some
indications in both of the presence of female deities, the general VedicBrahmanical picture is quite different. The dominant Vedic deities are
male and this tradition also appears to have been largely aniconic in its
earlier period, in other words there were no images of deities as such. This
was also the case with early Buddhist sites, where the Buddha himself is
Religion in the early states
113
not portrayed directly (Karlsson 1999; Rabe 1999). I turn now to examine
the early Vedic-Brahmanical material in a little more detail.
th e develop m ent of ved ic-brahmanical religion
I have already alluded to the chronological difficulties regarding the dating
of Vedic material (Chapter 2). The current consensus among Western
scholars tends to accept the basic sequencing of the Vedic material developed
by nineteenth-century scholars but it is very difficult to reach any precision with relation to absolute chronology. Witzel’s datings provide a working basis but one should clearly be wary of taking any of these figures
too literally.
A lot has been written about Vedic religion on the basis of the Vedas
themselves and the various Brāhman.as, ritual manuals and so on. There is
little point in presenting this material at length since I have little original
to add. However, it is worth summarising some main points:
There is a basic level of material with close relation to the Indo-Iranian
texts, which can be seen in many of the R.gvedic hymns. This corresponds
to what I have referred to above as a generic Indo-Aryan cultural tradition.
It includes aspects of the cults of the ‘imported’ Indo-Aryan gods such as
Indra, Agni, Varun.a, Mitra and Soma. The ‘shamanic’ and ‘ecstatic’ aspects
of Vedic ritual probably also go back to this layer, including the soma-fuelled
visionary state in which the Vedic .r.sis were supposed to have composed their
hymns (Gonda 1963; Houben 2003b; Thompson 2003), and the famous
description of the keśin or long-haired sage (muni) in Book 10, hymn 136
of the R.gveda (Werner 1989; Deeg 1993). These aspects of the Indo-Aryan
cultural material are significant in terms of the ways in which they may
have contributed to the later role of the Indic ascetic, and I shall return to
them in this context in Chapter 8. The complex figure known as the vrātya
perhaps also goes back to this layer, but will be discussed separately.
A second phase includes the reformulation of the R.gvedic hymns in the
Samaveda and Yajurveda. This is Witzel’s middle period and he associates
it with the creation of the Kuru state ritual and mythology. This formed
the ritual basis (the rājasūya, aśvamedha, etc.) for the warrior king model
discussed in Chapter 4, evidently one of the more successful Kuru-Pañcāla
exports. The aśvamedha or horse-sacrifice, already mentioned in Chapter 4,
makes the point most clearly; it can only be performed when the king has
established his military authority over the region through which the horse
travels. Also in this phase we begin to see an expanded role for a number
114
Origins of yoga and Tantra
of deities who take only a minor part in the R.gvedic hymns. Among the
most significant of these are Vis.n.u and Rudra (Śiva).
Vis.n.u (along with his various avatars) and Śiva were of course to become
in time the two most significant male deities in the Hindu pantheon, each
the focus of a major devotional cult. Both deities were in particular to
become strongly associated with kingship, a process perhaps prefigured in
Vis.n.u’s case by his close relationship with the sacrifice in the Brahman.as, and
moving much further on in the Mahābhārata and Rāmāyan.a.25 Śiva himself,
along with subsidiary forms of Śiva such as Bhairava, was also to become a
major deity associated with yogic and Tantric practice. This development
took many centuries, however, and was preceded by a phase in which we
can see a variety of gods taking on the role of patron deity of kingship
and battle, including Skanda, Sūrya and the Vr.s.ni deities (Sam
. kars.an.a,
Vāsudeva, etc.) among others (see Chapter 9).
The first major text associated with Śiva is the Śatarudrı̄ya, a liturgy which
forms part of the Taittirı̄yasam
. hitā, part of Witzel’s second phase. Śiva here
is invoked mainly under the name Rudra, although the text includes the
first known occurrence of Śiva’s mantra (om
. ) namah. śivāya. Śiva’s role in
this text is as a potential source of disorder, chaos and misfortune who is
invoked to protect and help, though there are also indications of a more
soteriological role (e.g. Sivaramamurti 1976: 24–5).
This combination of a position outside the ordered world of civilised
society and a soteriological role is an important one, which we will find
recurring in a number of forms later in this book. As I have noted above,
there is probably not very much to be gained by seeing Śiva as an inheritance from the Indus Valley. This may or may not be the case, and his
‘outsider’ role may or may not derive from his earlier non-Vedic history.
What is undoubtedly true, however, is that he became the first and supreme
example of a fundamental principle of Indian religious life; the power of
disorder, destruction and transgression, the positive results to be gained
from breaking through the patterns of normal behaviour. Seen in a somewhat different light, one could express this as the need to come to terms
and worship the things that can go wrong and create misfortune.
This is not as alien a pattern as it might seem. In some ways, it is the
basis of the religious orientation of many small-scale preliterate peoples
25
The story of Vis.n.u’s three steps occurs in the Śatapatha Brāhman.a. Vis.n.u plays only a minor role
in the Upanisads, perhaps reflecting their ascetic ambience as much as their date: he occurs in the
Br.hadāran.yaka Upanis.ad 6.4.21 as part of a ritual to have a child; he occurs in the opening invocation
of the Taittirı̄ya Upanis.ad (and its recapitulation at end of Chapter 1); the highest step of Vis.n.u is
mentioned in the Kat.ha Upanis.ad 3.7, 8–9.
Religion in the early states
115
and of village communities within pre-modern states in many parts of the
world.26 Misfortune is seen as the action of the spirits and as a consequence
of their being offended or provoked. It can be countered, if one is lucky,
by reconciling oneself with the offended powers.
Western religion, while accepting the idea of divine punishment, has
tended to associate it with ideas of divine love and divine justice. This goes
with the tendency, strongest in Christianity and marked in Islam as well, to
dichotomise good and evil. Indic religions rarely do this as systematically or
completely as do Western religions. In psychoanalytic terms, Indic deities
encompass the bad father or mother as well as the good father or mother.
Śiva and Kālı̄ are supreme examples of this principle.
This question of the positive value attached to the forces of disorder and
transgression also seems to be a key issue in relation to a group of people
in Vedic society who have often referred to in relation to the growth of
yogic and Tantric practice, the vrātyas. They will turn up in a couple of
later chapters, so I introduce them in some detail here.
There has been a lot of speculation about the vrātyas, much of it perhaps
bordering on the realms of fantasy.27 One reason for confusion is that (as
with a number of Vedic terms) the term vrātya appears to have shifted
its meaning considerably over time.28 A reasonably consistent picture has
however emerged in recent years through the work of Falk, Witzel and
others (Heesterman 1962; Bollée 1981; Falk 1986; White 1991; Witzel 1995c).
At the time of the Kuru-Pañcāla state or its historical equivalent, they seem
to have been associations of young men who have not yet married and
achieved full adult status. These groups of young men went out on raiding
expeditions into neighbouring territories, these expeditions perhaps acting
as a way of canalising the ‘traditional aggression resulting in cattle rustling,
fighting and small scale warfare existing with one’s neighbours’ (Witzel
1995c: 18). They also had a specific ritual function, which Witzel regards
as preserving elements of the old Vedic ritual before the Kuru reformation.
It particularly involved the performance of extended midwinter sacrificial
rituals out in the forest, away from the village community, on behalf of the
community as a whole (White 1991; 95ff ). The vrātyas have been linked to
26
27
28
One should really say, it is one basis of their religious orientation. Other idioms of misfortune, such
as witchcraft, sorcery and soul loss, work in different ways. Most small-scale pre-modern societies
seem to give primary emphasis to one or another of these.
The vrātyas have also regularly been read as proto-shamans, linked to the supposed Vedic shamans
of the R.gveda.
Thus by the time of the Laws of Manu, perhaps around the first or second century CE, the vrātyas
(2.39; 10.20) are sons of ‘twice-born’ groups who fail to undergo the sacred thread ceremony, and
their children in turn become the origin of various non-Aryan or other disvalued groups (10.21–23).
116
Origins of yoga and Tantra
evidence for sodalities or associations of young men in Iranian and other
Indo-European material, and Falk and Witzel appear to imagine them
rather after the model of the age-set system of organisation, known in
particular from the studies of Oxford anthropologists, among others, in
East Africa.
If this is a plausible model, we might suppose that all men probably went
through a period as members of the vrātyas, and that they originally at least
formed the basis of the fighting force of the tribal group in times of war.
From Falk’s material, it seems that some men who for one reason or another
were not able to proceed to the subsequent stage of adult householders
may have remained as vrātyas and become the leaders of the vrātya groups.
Some of these may have been younger sons of families where the oldest son
had taken over as head of the household. As time went on, the shift from a
pastoralist to an increasingly agricultural economy may have meant that the
standard solution to inheritance in a pastoralist society, of redistributing the
herds through marriage or other mechanisms so that each family has enough
to live on, was breaking down. Such vrātyas may have played a part in the
outward expansion of Vedic-Brahmanical culture from the Kuru-Pañcāla
region (see below). We have little direct evidence, but the association of
various eastern groups with vrātyas in Manu is perhaps suggestive that they
may have been involved in settlement in this area, or at least that there were
groups out to the east that could be compared with them.
As for the ritual function of the vrātyas, Falk regards their midwinter
twelve-day sacrifices as related to such Indo-European phenomena as the
Roman Lupercalia and the twelve nights of Christmas. This was the time
when the wild hunter Odin rode through the forests of northern Europe.
Odin’s equivalent in the Vedic context was Rudra, the Vedic prototype of
Śiva. As Rudra’s ‘dogs’ or ‘wolves’ (White 1991: 101) they slay the sacrificial
cow in the food shortages and drought of mid-winter and so help to bring
about the return of prosperity in the following year. The most notorious
vrātya ritual was the mahāvrata or ‘great observance’, which involved ritual
sex between a brahmacārin (presumably, in this context, meaning a young
man otherwise vowed to celibacy) and a prostitute (Gonda 1961).
I would assume that if the kind of picture sketched by Falk and Witzel
has some plausibility, these midwinter rituals might be seen as gradually
disappearing during the second phase of the development of Vedic religion, to be replaced by the seasonal sacrifices and rituals described in the
Brahmanas and later texts.29 If so, the vrātyas might have ceased to exist in
29
See for example the Śatapatha Brāhman.a (2.5.1).
Religion in the early states
117
the sense described here some time before the time of the historical Buddha,
with the term itself coming to be used as a label for various low-caste and
devalued populations, including peoples geographically outside of the areas
of proper Aryan society, which by that time comprised the Middle Ganges
valley as well as the Kuru-Pañcāla region. The vrātyas in the earlier period
though represent an important prototype for the situation of an ‘unorthodox’ group whose activities, though dark and associated with death and
transgression, are nevertheless somehow essential to the wellbeing of society. As we will see in the following chapter, this has led to their being seen,
perhaps surprisingly, as predecessors of the śraman.a movements.30
We have seen two examples above of the presumably ‘transgressive’ use of
sexuality, in the mahāvrata rituals of the vrātyas and also in the aśvamedha
which includes a notorious sequence where the chief queen simulates intercourse with the sacrificed horse. It would be wrong to give the impression
that such practices are characteristic of Vedic ritual as a whole; they are not.
They are however significant as indicative of an attitude towards the ritual
power of ‘transgressive’ sexuality; there would be no point in performing
these rather bizarre sequences if they were not felt to be ritually very powerful (one can hardly imagine that the chief queen enjoyed the experience).
Mainstream Vedic texts are however surprisingly explicit about sexuality in
its proper marital context: in the Upanis.ads, sex is homologised to the sacrifice, and ritual formulae are provided to influence conception.31 The idea
of the brahmacārin, the young celibate male as a source of spiritual power,
is present, but it has not taken on the sense it has in later Brahmanical
religion, as a specific stage in the life-cycle.32
Various processes have been suggested for the development and transformation of Vedic religion and it seems to me that most or all of these have
some explanatory content. For the early phases, these include Heesterman’s
model of the move from the agonistic model of the sacrifice to a ritual concerned primarily with status-affirmation, a development which can be correlated with the move to permanent settlement (Heesterman 1985, 1993) and
Witzel’s argument about the needs of the Kuru state (Witzel 1995c). Later
phases are driven perhaps by the growth of the model of the Brahmanical
30
31
32
Heesterman regards the vrātya practices as representing an earlier stage of Indo-Aryan ritual, before
the Vedic reforms, and as being characterised by violent and conflictual themes that had been
cleaned out of the orthodox forms of the ritual (Heesterman 1962). I am not entirely convinced by
Heesterman’s picture of the early form of the sacrifice (see Heesterman 1985, 1993; B. Smith 1989:
40–6), but would in any case suggest that it is useful to see the vrātyas structurally, in relation to the
mainstream and orthodox ritual of their time, rather than in terms of presumed origins.
E.g. Br.hadāran.yaka Upanis.ad 6.4 (= Olivelle 1998: 88–93).
On the early role of the brahmacārin see Crangle 1994; N. Bhattacharyya 1996: 137–9.
118
Origins of yoga and Tantra
ascetic, which will be discussed in Chapters 6 and 7, and by the development of forms of Vedic ritual suitable for widespread domestic adoption
in the new urban culture. I shall return to the later development of Vedic
religion in Chapter 7. Chapter 6, however, is primarily concerned with the
non-Vedic śraman.a renunciate orders, particularly the early Buddhists and
Jains, and I now turn to consider these orders.
chap t e r 6
The origins of Buddhist and Jaina orders
The origin of the renunciate orders in India, and particularly of monasticism, is one of the most intriguing and significant questions that arises in
the study of Indic religions. The idea of asceticism leading to magical power
is not uncommon in religions of tribal and pre-literate peoples. One might
think here of the idea of vision quests among the Plains Indians, and of the
ascetic practices associated with initiation into spirit-cults in many parts of
the world. We might imagine, though we do not know for certain, that similar practices were common among preliterate societies in earlier periods.
However, the idea of someone permanently committed to an ascetic state is
much less common, and indeed since religious specialists in most preliterate
societies are only part-time practitioners, it would scarcely be practicable.
While there are plenty of examples in the ethnographic record of individuals who are regarded as having special powers and who may be set somewhat
aside from the community as a whole and regarded with some suspicion as
a consequence, these people normally make a living through hunting and
gathering, pastoralism or farming like other members of their society.
As for monasticism, we know nothing much like the Buddhist and Jaina
monastic traditions from elsewhere in the world before this date. Monasticism was in time to become an extremely important and influential institution within world culture, particularly in the Buddhist version, which
spread through most of South, Southeast, Central and East Asia, and in
the later Christian version. Thus the question of how the ascetic orders got
going is well worth some attention.
The renunciate traditions of North India in the fifth and fourth centuries
BCE appear to mark the appearance of a new kind of goal or purpose for
ascetic practice, variously known by terms such as moks.a, nirvān.a, kaivalya
or bodhi (below I shall generally speak of ‘liberation from rebirth’, since
the subtle distinctions of later traditions presumably developed a good deal
later on). In the Vedic world, asceticism is for this-worldly purposes or for
a good afterlife (rebirth in heaven) rather than for the attainment of some
119
120
Origins of yoga and Tantra
radically other state. The new goals appear to be premised on the idea of
a self-conscious individual who finds his or her everyday life as a member
of society radically unsatisfactory. In their mature forms at any rate, they
came to terms to some degree with the everyday world, since those who
were striving to attain them were also seen as acquiring skills and powers
that could be used in the service of the community, but ideologically they
are committed to withdrawal from not only everyday society but the entire
cycle of rebirth (sam
. sāra). This too is, as far as we know, a new development,
and we may ask how and why it arose.
Thus we have several related questions: how the goal of liberation from
rebirth came about among the Jainas, Buddhists and other śraman.a (nonBrahmanical ascetics), how it came about among the Brahmanical ascetics,
and how the idea of collective and organised practice in the context of celibate communities arose. This chapter deals primarily with the Buddhists
and Jainas. In addition to discussing the development of the renunciate
traditions themselves, I will also examine the ways in which Buddhist and
Jaina renunciates engaged with the religious life of the larger communities
who provided the material support for their activities. My discussion will
necessarily also deal to some degree with issues relating to Brahmanical
ascetics, but the full development of the Brahmanical model, and the different mode of engagement of the emergent Brahmanism with the larger
community, is the subject of Chapter 7. Chapter 8 considers how we might
explain and understand all these developments.
the contex t of the early ś r a m a n. a orders
The Sanskrit term śraman.a is a generic term that was used by members of several ascetic orders, including the Jainas, Buddhists and Ājı̄vikas.
We assume that these orders came into existence in the time of the
historical Buddha (i.e. fifth century BCE on current datings). Each of
these renunciate orders had a named founder, Mahavı̄ra (also known as
Nirgrantha Nātaputta or Vardhamāna) for the Jainas, the Buddha
Śākyamuni (Siddhārtha or Gautama) for the Buddhists, Makkhali Gosāla
for the Ājı̄vikas.1 Each order, however, had a tradition of previous teachers, and it seems clear that these founders of renunciate orders operated in
a context in which a variety of such teachers were active.2 Although the
1
2
Our information on the Ājı̄vikas derives from Jaina and Buddhist sources, since their own texts have
not survived (see Basham 1951, Bronkhorst 2000).
In fact the term Ājı̄vikas may have been used as a general term, especially for naked ascetics. Bronkhorst
has recently raised the possibility that the followers of Mahāvı̄ra may have been referred to as Ājı̄vikas
in the Pali scriptures, and the followers of Nirgrantha Nātaputta in the Pali scriptures, who are
The origins of Buddhist and Jaina orders
121
Figure 6.1. Śākyamuni prior to his enlightenment (accompanied by Vajrapān.i) consults a
jat.ila Brahmin. Gandhara, late first century CE, Peshawar Museum
Jaina and Buddhist accounts are from several centuries later, and contain a
large admixture of legendary material, it has generally been accepted that
Mahāvı̄ra, Śākyamuni, and the other main characters mentioned in these
texts were real historical characters, and that the narratives contain a core
of genuine historical events. While I assume myself that this is the case, it is
important to be aware of the distance in time between the sources and the
events being described. There is still a strong tendency in Buddhist studies
to say ‘the Buddha said so-and-so’ when what is meant is that ‘such-andsuch a sutta in the Pali Canon describes the Buddha as saying so-and-so’.3
These three founding figures, Mahāvı̄ra, Śākyamuni and Makkhali
Gosāla, are described as more or less contemporary, with Mahāvı̄ra probably
the earliest. As noted in an earlier chapter, the dating of the historical
3
generally identified with the Jainas, may have been primarily the followers of Mahāvı̄ra’s predecessor
Pārśva (Bronkhorst 2000).
There is nevertheless reason to assume that the sūtras, once compiled, were passed through some
generations with reasonable accuracy – at least enough accuracy to preserve jokes and references that
the Pali commentators no longer understood (Gombrich 1990, 1996).
122
Origins of yoga and Tantra
Buddha is uncertain, but if we accept the present consensus which places
his death in around 400 BCE or a little earlier (see Chapter 2), Mahāvı̄ra’s
death might have taken place in around 425 BCE or a little after (Dundas
2002: 24).
The emergence of Brahmanical ascetics and renunciates is a separate
phenomenon that also needs some explanation. The existence of ascetic
semi-renunciate Brahmins is evident in the Buddhist and Jaina sources
(see e.g. Tsuchida 1991). This is also confirmed from Brahmanical sources,
in particular the Dharma Sūtras, where we find a number of terms, particularly vānaprastha and vaikhānasa, for these people (Bronkhorst 1998a:
18–26). These refer to unmarried or married practitioners who live in the
forest, subsist on forest produce, maintain a ritual fire (like a normal Vedic
household) and pursue ascetic practices. The men are sometimes described
as having matted hair (e.g. Bronkhorst 1998a: 33, citing the Mahābhārata).
These seem to be the same people as the jat.ila or matted-hair Brahmins
described in the Buddhist texts as being relatively highly regarded by the
Buddha (Fig. 6.1), especially by comparison with other, wealthy Brahmins
who are normal householders (gr.hastha) and had been given land and wealth
by local rulers in exchange for their services as Vedic teachers and ritualists
(Tsuchida 1991: 54–7). The stated aim of the way of life of these ascetic
Brahmins is to attain rebirth in heaven. These Brahmins co-exist with
other, fully renunciate practitioners who by contrast to them are seeking
liberation from rebirth, either through withdrawal from action or through
the pursuit of wisdom.4 Presumably these renunciates would have included
the śraman.as as well as Brahmanical renunciates pursuing similar goals.
For the early Dharma Sūtras, which presumably originated in circles
close to the urban Brahmins with their royal endowments, being a gr.hastha
or householder is the only proper way of life, and the alternatives (which
also include being a brahmacārin or unmarried man living at one’s teacher’s
household and after his death with his son) are all to be rejected. The
Chāndogya Upanis.ad, evidently closer to the ascetic circles, presents the
reverse picture (Bronkhorst 1998a: 17–18). What is clear, though, is that
both existed at this time, the vānaprastha brahmins and other ascetics
aimed at liberation from rebirth. It is only at a considerably later point
(in the Manusmr.ti or ‘Laws of Manu’, perhaps first century CE) that
4
E.g. in the Āpastamba Dharmasūtra which has two varieties of these, one again called vānaprastha,
the other parivrāja. Bronkhorst describes these as ‘non-Vedic’ in contrast to the married vānaprastha
with a domestic fire (1998a: 13–20). They do not have a sacred fire. Bronkhorst assumes that the
standard pattern for a Hindu sam
. nyāsin in modern times, where the sacred fire is ‘internalised’ within
the sam
. nyāsin’s body, is a somewhat later development.
The origins of Buddhist and Jaina orders
123
the Brahmanical list of āśramas developed into the now-familiar series of
four (brahmacārin, gr.hastha, vānaprastha and sam
. nyāsin), seen as successive
stages rather than as alternatives (see also Olivelle 1993).
In the later tradition of the sequence of four āśramas, becoming a forestdwelling or vānaprastha Brahmin tends to be seen as something done late
in life, as a precursor to the sam
. nyāsin stage, but this is not necessarily what
is at issue here. These early forest-dwelling Brahmins represent part of the
growth of an ascetic tradition among the Brahmins, a tradition that was in
time justified through the development of a body of stories regarding the
sages (r..si) who wrote the R.gveda, who are themselves generally portrayed
as living the life of a vānaprastha. Thus the Buddhist texts give a standard
list of ten .r.si or Brahmin sages, making a point of their ascetic lifestyle.5
By the time of the Mahābhārata, it is taken for granted that the Vedic .r.sis
lived austere lives in forest hermitages.6
Thus the evidence here points to two types of ascetics, a ‘Vedic’ semirenunciate type maintaining a ritual fire and aiming at rebirth in heaven
and a ‘non-Vedic’ renunciate type, without a fire, who is aiming at liberation from rebirth.7 If, as Bronkhorst suggests, the other types of ascetics,
the renunciates who sought liberation from rebirth, were initially distinct
from the vānaprastha Brahmins, there was evidently interchange between
the two. The early Upanis.ads, such as the Chāndogya and Br.hadāran.yaka,
which can be taken as compilations of materials produced by and for these
various ascetic Brahmin groups, indicate that the goal of liberation from
rebirth, and the idea of achieving it through the pursuit of wisdom, became
widely distributed. The early Buddhist sūtras suggest knowledge of this early
Upanis.adic material (Gombrich 1990), though not necessarily of the texts
as later compiled,8 suggesting that the sayings that they knew were part of
a floating body of wisdom sayings, circulated primarily in oral form.
5
6
7
8
Most of these are figures listed in the R.gveda Sam
. hitā as authors of Vedic hymns, and many of them
are important in later Brahmanical legend: At.t.haka, Vāmaka, Vāmadeva, Vessāmitta, Yamataggi,
Angirasa, Bhāradvāja, Vāset.t.ha, Bhagu (Tsuchida 1991 and Walshe 1995: 121).
Here, as always, ‘forest’ refers more to wild and uncultivated regions than to any particular kind of
vegetation.
This discussion is in part suggested by Olivelle, who makes the important distinction here between
the ‘hermit’ or ‘anchorite’ who has physically withdrawn from society (i.e. the vānaprastha), and the
‘renouncer’ who ‘lives in proximity to civilized society and in close interaction with it’ as a ‘religious
beggar’. The former type became obsolete early on, although it remained as a literary ideal (Olivelle
2003: 272).
Other authors (e.g. Horsch 1968) have argued that the Buddhist sūtras do not show clear evidence
of familiarity with any of the Upanis.ads as complete texts. This is perhaps not surprising. The
Br.hadāran.yaka and Chāndogya Upanis.ads have passages in common, and are evidently compilations
of shorter texts which circulated independently before being assembled in the forms we now know
(cf. Olivelle 1998: 95).
124
Origins of yoga and Tantra
There are other indications of such a generally-shared body of wisdom
sayings and stories, such as that remarkable Jaina text, the Isibhāsāiyam
(Schubring 1942–52). Certainly this text provides a view of early Indian
spirituality that is markedly at contrast with any assumptions about discrete
and opposed traditions. The title of the Isibhāsāiyam, which was included in
the Śvetāmbara Jaina canon, means the Sayings of the R.s.is, and the book
consists of a series of verses attributed to particular named .r.sis. These verses
are mostly of the nature of moral maxims, and have short commentaries.
What is most striking is the identity of the .r.sis who are included in this
text, for while they include Pārśva, Mahāvı̄ra (here called Vaddhamān.a)
and other Jaina figures, they also include Vedic and Brahmanical sages
such as Nārada, Yājñavalkya and others, as well as well-known Buddhist
figures such as Śāriputra and Mahākāśyapa, and Makkhali Gosāla, the
founder of the Ājı̄vikas. We seem here to be in a pluralist society of ascetic
practitioners whose members are quite willing to regard members of the
other semi-renunciate and renunciate religious traditions as ‘sages’ of a
standing comparable to their own.
Similarly, we have seen how the narratives of the renouncer-king Nami
or Nimi appear to have been common property between Buddhist, Jaina
and Brahmanical writers, for all of whom he served as an early model of the
renunciate life (Chapter 4). Another intriguing indication of the extent to
which narratives were shared between traditions is the story of the encounter
between the materialist king Paesi or Pāyāsi and a śraman.a teacher. This
occurs both in a Jaina version, where the śraman.a teacher is a Jaina monk,
and in a Buddhist version (Dı̂ghanikāya, no. 23), where he is a Buddhist
(Bollée 2002; Balcerowicz 2005; Walshe 1995: 351–68). The verbal details
of the dialogue vary somewhat, but the story is recognisably the same. Yet
another example of the same kind is given by a dialogue between a son who
wishes to become an ascetic and rejects the arguments of his father who
tells him first to study the Veda, establish a sacred fire, and have children.
This appears in Jaina, Buddhist and Brahmanical versions (Olivelle 2003:
279–80).
Such cases suggest the development of a shared ‘wisdom’ type literature,
doubtless mostly circulated orally in the first place as the tradition suggests,
and focusing on the attainment of liberation from rebirth. Some of this
would have been the preserve of particular Brahmin groups, other parts of
the various emerging śraman.a orders, Buddhist, Jaina and Ājı̄vikas. There
is nevertheless clearly a considerable body of shared intellectual content
among all of this material, with ideas such as the understanding of karma
The origins of Buddhist and Jaina orders
125
and the analysis of perceived experience being issues for all participants.
The Buddhist and Jaina9 orders then can be seen to have emerged from the
circles in which material such as this was circulating.
While the stories regarding Buddha’s predecessors are generally treated as
legendary, those regarding Mahavı̄ra’s immediate predecessor, whose name
was Pārśva, have often been considered to have some historical content, in
part because some traditions have been preserved regarding the difference
between his teachings and those of Mahāvı̄ra.10 The traditional period of
250 years between Pārśva and Mahāvı̄ra is as insecure as the 218 years between
the Buddha’s death or parinirvān.a and the accession of Aśoka on which
the older Buddhist chronology was dated, so if we want to give Pārśva a
historical date we could probably imagine him as active at any time up to
say the late sixth or early fifth century BCE.11
The stories of Pārśva, and for that matter of his predecessors and of
the previous Buddhas, give some support to an idea that religious orders
or traditions of the śraman.a kind might predate the time of the Buddha
and Mahāvı̄ra (I shall refer to these as ‘proto-śraman.a’ movements). Perhaps significantly, there is some ambiguity about whether Pārśva and his
predecessors required celibacy of their followers, and these pre-Mahāvı̄ra
‘Jaina’ movements appear to have been more open to the participation of
women than Jainas in later times (cf. Williams 1966). This suggests that
Pārśva’s movement, if we accept it as historical, may not have been a fully
renunciate movement. At the same time, it seems to have been an organised
movement of some kind, and the Jaina scriptures describe its followers as
still around in the time of Mahāvı̄ra.12 Thus we seem to be dealing here
with an organised urban movement rather than with an isolated forest
householder-ascetic model such as that of the jat.ila Brahmins. It has been
9
10
11
12
Strictly speaking, the term Jaina is anachronistic at this period. As Peter Fluegel has recently noted,
the earliest known usage of the term dates from the seventeenth and eighteenth centuries CE; its
use as a self-designation may not be much earlier (Fluegel 2005). The issue is not entirely trivial,
since the relationship between the nigganthas referred to in early Buddhist sources and the modern
Jainas is not clear (see main text, and Bronkhorst 2000). However, I shall continue to use the term
for convenience; where not otherwise explicit, it refers to the followers of Mahāvı̄ra.
Possibly, as Williams has suggested (R. Williams 1966), Pārśva’s predecessor Nemi also has some
historical reality. This is a different person from the Nimi or Nami of Videha whom I discussed in
Chapter 4; he is placed in Gujarat and Williams speculates on the possibility of a pre-Pārśva tradition
from Gujarat.
See also Dundas 2002: 32.
The Uttarādhyayana Sūtra, Chapter 23, consists of a dialogue between Keśi, a follower of Pārśva,
and Gautama, a follower of Vardhamāna (Mahāvı̄ra), which results in Keśi adopting the teachings
of Mahāvı̄ra (Jacobi 1895: 119–29). See also Dundas 2002: 30–3.
126
Origins of yoga and Tantra
suggested recently that early Buddhist references to nigan..thas, which have
generally been taken to refer to followers of Mahāvı̄ra, may in fact have
been primarily to members of Pārśva’s movement (Bronkhorst 2000).
It is evident that the early Jaina and Buddhist materials had a great deal in
common, including much of their terminology. The traditional temporal
sequence, in which the Jaina order or orders took form somewhat earlier
than the Buddhist saṅgha, makes some sense. In areas such as its attitude
to asceticism, or its modes of meditation, the Buddhist teachings give the
impression of a modified and developed version of the Jaina tradition. Thus
early Jaina yoga seems primarily to consist of extreme ascetic practices on
the Brahmanical model, while the Buddhists added to this both what was
a novel systematisation of the dhyāna practices, and the new vipaśyanā or
insight practices (see below).13 At the same time, the Buddhist suggestion
that extreme asceticism is inappropriate can be seen as a criticism of the
Jainas as well as of Brahmanical ascetics.
All this suggests, in terms of the new chronology, that we might be
looking at a developing semi-renunciate householder movement in the sixth
and fifth centuries BCE, culminating in the establishment of a full-scale
renunciate movement in the mid to late fifth century. As noted in Chapter 3,
the dating of the Second Urbanisation in North India is uncertain because
of the C14 calibration problem for this period. It would seem, however, that
the emergence of fortified cities at places such as Rājagr.ha, Campā, Ujjayinı̄
and Vārān.ası̄, the capitals of four of the major states described in the Pali
canon and Jaina sources, would have taken place by around 550 BCE at the
latest, perhaps one or two centuries earlier (the city of Śrāvastı̄ is somewhat
later) (Erdosy 1995c: 109–10; Chakrabarti 1995). Thus, as has often been
suggested, we are probably dealing with a movement that got going either
around or shortly after the time of the initial growth of these cities.
While the historical Buddha and Mahāvı̄ra can probably be dated at least
a couple of centuries after the establishment of urban settlements (cf. Bailey
and Mabbett 2003), ‘proto-śraman.a’ movements such as those represented
by the stories of Pārśva and his predecessors may have been closer to the
initial process of the Second Urbanisation. If the cult-association model
discussed in Chapter 4 has any value, there may have been a connection
between these movements and the growth of trading networks and market
centres that led to the early urban centres. In other words, membership in a
cult-association with branches in various urban trading centres would have
been helpful for merchants travelling from one centre to another, since it
13
Bronkhorst 1993; Cousins (personal communication, 2002).
The origins of Buddhist and Jaina orders
127
would have provided them with a network of fellow cult-members in these
places who could be trusted and who shared their values.
I return to this suggestion again in Chapter 8. Here, however, I move to
another proposed influence on the early ascetic orders: the vrātyas.
the v r ā t ya s as predecessors?
Dundas suggested some years ago that the vrātyas, the somewhat mysterious
Vedic ritualists mentioned in Chapter 5, might have been a significant
precedent for the śraman.a traditions (Dundas 1991, 2002). Dundas in fact
did not refer directly to the vrātyas but to ‘ancient Indo-European warrior
brotherhoods’ (1991: 174; 2002: 17). W. B. Bollée, whom he cited as his
source, however, was concerned primarily with the vrātyas, whom he traced
back to earlier Indo-European groups of male warriors (Bollée 1981).
At first sight, it is hard to see what the vrātyas might have to do with the
notoriously pacifist Jainas, or for that matter the Buddhists. The vrātyas
were after all, on the Falk-Witzel model at any rate, gangs of aggressive
young men who went out raiding cattle from neighbouring tribes when
they were not involved in transgressive rituals out in the forests.
There is more however to the argument than perhaps appears at first
sight. Dundas noted the prevalence of martial imagery in Jaina sources –
Jina, after all, means ‘conqueror’, so that, as Dundas put it ‘martial conquest
is the central image and metaphor of Jainism, giving the religion its very
name’ (Dundas 1991: 173). In fact military imagery is quite widespread in
both Jaina and Buddhist traditions. Of course, both the Jainas and the
Buddhists had founders from the ks.atriya or warrior caste. Dundas noted
that Jainism has in fact historically been ambivalent about war, and Jaina
communities have certainly been supported by violent and aggressive rulers.
The same would have to be said about Buddhists, for all of the splendid
example of the Emperor Aśoka’s conversion to Buddhism as a result of
his revulsion at the slaughter involved in the war against Kaliṅga. The
Sinhalese Buddhist willingness to justify King Dut.t.hagaman.i’s slaughter of
the Tamils on religious grounds is only one well-known example, if one
that has become particularly notorious because of its lethal consequences
in modern times.
In fact, ascetics have often been linked with militarism in Indian history
and have frequently even served as warriors.14 Dundas’s real point though
14
One can think of the armed and militaristic sam
. nyāsins known as the nāgas, or of the sam
. nyāsin-led
revolt against the British immortalised in Bankim Chandra Chatterji’s novel Anandamath (Chatterji
128
Origins of yoga and Tantra
was about formal structural models.15 How did the idea come about of
having a formal organisation of ascetics living together and pursuing a
common religious goal?
Terms employed in Jainism and Buddhism to employ groups of ascetics such
as gan.a, ‘troop,’ and saṅgha, ‘assembly,’ are used in early Vedic texts to refer to
the warrior brotherhoods, the young men’s bands which were a feature of Āryan
nomadic life, and the stress found in the old codes of monastic law on requirements
of youth, physical fitness and good birth for Jain and Buddhist monks, along with
the frequent martial imagery of Jainism and its repeated stress on the crushing of
spiritual enemies, may point to a degree of continuity with these earlier types of
warrior. (Dundas 2002: 17)
Certainly this warrior-brotherhood emphasis might help to explain the
ambivalence that these movements have shown to women members
(Sponberg 1992). However, while Dundas’s point about martial imagery
is true enough, it is also true that the central emphasis of the early Buddhist
and Jaina teachings are anything but martial. The vrātyas may have provided an organisational model and some imagery, but it is hard to see them
as the major predecessor of the śraman.a orders.16
the ś r a m a n. a s and the dead
In another sense, however, the vrātyas may have more to offer us. Bollée
suggests that the śraman.a orders may have derived their close association
with the cult of the dead from the vrātyas. Bollée suggests that an association
with the cult of dead was a significant theme for the early śraman.a communities. They frequently settled at sites associated with the dead and seem to
have taken over a significant role in relation to the spirits of the dead. Here,
I think, Bollée was onto something important, and often neglected, though
again it is unclear how much weight to give to the vrātya connection.
15
16
1992). (See also Boullier 1993). One can also consider such ‘export versions’ of Indian asceticism as
the military cult in early Korea, linked to the future Buddha Maitreya (Mohan 2001), or, at a later
date, the Samurai in Japan (Victoria 2005). It is also a matter of some interest that the Iranian branch
of the Indo-Aryan tradition seems to have led to a cult-organisation with an explicitly militaristic
orientation, that of Mithras, at around the same time or somewhat later, though the extent of the
connection is admittedly problematic (Sick 2004).
Witzel has argued similarly, apparently independently of Dundas: ‘the Buddhist saṅgha has, unobserved so far, some vrātya features as well: a single leader of a larger group of equals who wander
about in the countryside and live on extortion (or by begging), stay away from settlements, have
special dress and speech, etc.’ (Witzel 2003: 90).
At the same time, the early use of the term yoga in the context of the warrior’s transfer to heaven
(see White 2006, and Chapters 9 and 14) may provide some support for this connection. This also
connects with the issue I discuss below of the relationship between the śraman.a orders and the cult
of the dead.
The origins of Buddhist and Jaina orders
129
To begin with, it is worth noting that the most salient linkage between
Buddhist monks and lay communities in modern times in both Southeast
Asian and East Asian Buddhist societies has been their role in relation to
death. That Buddhist monks are specialists in death is something of a cliché
of modern ethnography on Buddhist societies (e.g. Tambiah 1970), and it
seems to be a theme that goes back a very long way in Buddhist history
(see deCaroli 2004; White 1986).
As we will see below, this appears to have been part of how the early
śraman.a communities, at any rate the Buddhists, built up links with local
communities. They settled at places associated with the cult of the dead,
and they took over responsibility for dealing with the process of dying and
of looking after the spirits of the dead. That they were prepared and able
to do this is doubtless connected to the idea that they were in a sense
already part of that world. The world of the śraman.as and the Brahmanical
renunciates seeking liberation from rebirth, in other words, like that of the
dead, was opposed to that of the everyday community of the living. In some
ways, it was even more radically opposed than that of the dead themselves,
since these ascetics were structurally outside the cycle of rebirth itself.
This brings something else along with it. The world of death is also a
world of misfortune and of potential threat. To become a specialist in this
area through renouncing secular life is to step into an unsavoury world of
dangerous and problematical supernatural powers that most people are in
fact doing their best to stay away from. Given the high prestige of monasticism in modern Buddhist societies, the splendour of many Buddhist and
Jaina monasteries and temples and, for that matter, the relatively comfortable and worldly lifestyle of many ordinary members of these orders in
modern times, this aspect of the śraman.a traditions is not always conspicuous. There are reasons why we should take it seriously, however, since it
brings us back to a central theme of the book, the ongoing dialogue between
the everyday world and the world of misfortune, with its associations of
death, transgression and the unsavoury.17
We might start with the very deliberate structural stepping outside of
secular society that is involved in becoming a Buddhist or Jaina śraman.a
or a Brahmanical sam
. nyāsin. In the case of the sam
. nyāsin, the developed
form of this transition involves the actual performance of one’s own funeral
ceremony and the ‘internalisation’ of the sacred fire that is a critical part
of the everyday life of the householder. In all cases, including that of the
17
I would like to acknowledge the work of my former student Kim Chongho in helping me to think
through these questions (cf. Kim 2003).
130
Origins of yoga and Tantra
Buddhist and Jaina śraman.a, it involves a thorough rejection of secular
identity, including one’s previous name. The life of the bhiks.u or beggarascetic, living off food-offerings which are, again ideally, being made to
the monk as representative of the community rather than as individual,
points in the same direction (see below). Of course, in practice, family
connections and caste identity may remain important for Buddhist monks
or other ascetic practitioners, who are not necessarily outside the world of
power and politics, and individual monks may become significant actors
in the secular world in their own right, but ideologically the transition is a
significant one.
We might also note the ‘extreme’ behaviour associated with many renunciate ascetics. In the Buddhist tradition we could point to practitioners such
as the Tibetan chödpa (gcod pa) who, ideally at least, spend their lives wandering from place to place, practising in places associated with frightening
and dangerous spirits, where they visualise the offering of their bodies to
the Buddhist deities, local gods, ghosts and demons. Most Tibetan families
would have some ambivalence about a son or daughter becoming a wandering chödpa, but chödpa are also believed to be among the most powerful of
ritual performers and are the people who are expected to deal with the dead
in cases of epidemic disease and generally to cope with serious episodes of
misfortune (Tucci 1980: 92, Samuel 1993: 211–12). They are expected also
to have a certain invulnerability to epidemic diseases and other supernatural hazards, which is not really surprising when one considers that they
are already themselves living within this world of evil and misfortune to a
significant degree.
Much the same is true of the aghori renunciates in later Hindu tradition
(Parry 1994; Svoboda 1986, 1993). The detailed practices of these ascetics
come out of the ‘Tantric’ traditions which I discuss in later chapters, but
the idea that at least one form of ascetic involves a radical rejection of
conventional lifestyle, associated with such unsavoury practices as living
naked with one’s body covered with ashes, eating out of a human skull-bowl,
and the like, is as we will see an old one. Again, there is very considerable
ambivalence about this lifestyle in modern Indian society, as to other more
‘extreme’ forms of yogin, faqir, etc., but the idea that they are such people
and that they may have real powers has by no means disappeared even
from modern secularised South Asia. Nakedness itself was one of the key
indicators of early ascetics, being both standard Ājı̄vika practice and part
of the rule taught by Mahavı̄ra (e.g. Bronkhorst 2000). Symbolically, it is
another very clear rejection of secular identity.
The Buddhist tradition aimed at a ‘middle way’ which opposed more
extreme forms of asceticism, though in practice it was engaged in a constant
The origins of Buddhist and Jaina orders
131
dialectic with them, as with the disputes over the so-called dhutanga practices, a set of ‘extreme’ ascetic practices that were eventually judged to be
allowable but not required of Buddhist monks (e.g. Gombrich 1988: 94–5;
Tambiah 1984). Buddhist monks wore robes, though in principle these were
made of salvaged rags, and although they are described in the early sūtras
as spending most of the year wandering from place to place it is clear that
from an early stage most of the monastic population lived in settled communities. Various other accommodations to secular life have been part of
the history of the Jaina and Brahmanical ascetic traditions though in each
case the ‘extreme’ options have retained a presence. However, the idea of
the Buddhist monk as being protected against and as offering protection
against the malevolent attacks of the spirit world and the realm of misfortune is a key one in the Buddhist tradition as well from relatively early
times, in the form of the paritta suttas, texts which are recited for protection and the generation of ‘auspiciousness’ and which generally derive from
narratives in which the Buddha states that those of his disciples who recite
these texts will be protected against supernatural powers of various kinds
and other afflictions (Tambiah 1970; Gombrich 1971: 201–9; de Silva 1981).
They are also used in other contexts that for pre-modern South Asians were
heavily associated with supernatural danger, such as childbirth.18
The point I am making here is that there is a structural opposition
between the realm of happiness, prosperity, good fortune and worldly
life, on the one hand, and that of the ascetic world-renouncer, whether
a Buddhist or Jaina śraman.a or a Brahmanical ascetic, on the other. This
opposition does not imply a lack of contact: the inhabitants of the worldly
realm need access to the powers of the world-renouncer, above all when
things are not going as they should. In addition, the opposition is not
equally strong or equally significant at all times and in all places within the
sphere of Indic religions. However, we will see it recurring in virtually all
times and places in one form or another, and it will be one of the basic
building blocks of our analysis.19
the v r ata o r vow
The concept of the vrata or vow (not, apparently, related to that of the
vrātya) is also worth some discussion. This term has a long history in Indic
18
19
The Aṅgulimāla Sutta is recited for protection in childbirth, which is conceived of in South Asia as
a time of very high susceptibility to spirit attack (de Silva 1981: 13; cf. Rozario and Samuel 2002b).
Again, Louis Dumont offers a classic treatment (1960, cf. Madan 1988), but I would prefer to provide
a greater degree of historicisation and to treat the question of renunciation less as a given and more
as a variable over time.
132
Origins of yoga and Tantra
religions, and in its modern meaning it consists essentially of a vow or ritual
observance addressed to a particular deity and aimed at influencing that
deity to help oneself or to help another on one’s behalf.
In modern times, vrata are most often performed by women for the benefit of their families (husbands and children) and typically involve fasting
along with the recitation of a story associated with the vrata (Wadley 1980;
S. Robinson 1985; McGee 1992; Gupta 1999). The concept in the R.gveda
and early Brahmanical texts is somewhat different and has been discussed
recently by Lubin (2001). Lubin, following Joel Brereton, interprets the
vrata as initially having the meaning of a kind of commandment expressing the nature of a particular deity and so constitutive of the order of the
world as we know it. ‘Each deity’s vratá defines its nature and role in the
world’ (Lubin 2001: 568). These commandments set up a moral obligation
on human beings who have relationships with the gods to adhere to the
divine patterns and so to act in ways that follow them. In some places in
the R.gveda and in later texts however the term vrata is used to designate
‘rule[s] of ritual actions’ or of ‘specific, initiatory regimens required for
worship and Vedic study’ (Lubin 2001: 566). The vrata of brahmacarya is of
this kind (Atharvaveda 11.5, Śatapatha Brāhman.a 11.5.4). Dundas suggests
that the term in its earlier meaning ‘conveyed the notion of a “calling” in
the sense of a solemn dedication of oneself on a permanent basis to one
single purpose’ (Dundas 2002: 157–8). Certainly it is this sense that appears
to be intended when the five vows of the Jaina follower are referred to as
the mahāvrata (‘great vrata’).20 These five Jaina Great Vows are abstention
from killing and violence of any kind, not lying, not taking what has not
been given, the renunciation of sexual activity, and the renunciation of
attachment (Dundas 2002: 158–9).21
These are very similar to the five lay precepts of the Buddhist tradition,
which are against killing, stealing, sexual misconduct, false speech and the
taking of intoxicants. The term vrata is not used in the same way in the
Buddhist context, perhaps reflecting the early Buddhist critique of Jaina
‘extremism’, but the idea of the various levels of monastic initiation as
involving a set of vows or commitments that define a pattern of life with a
20
21
The same term is used for the major ritual of the vrātyas, referred to in Chapter 5, where it clearly
has a very different meaning.
The Jaina mahāvrata is also very close to the mahāvrata described in Patañjali’s Yogasūtra ii.30–1,
which consists of the practice in all circumstances of non-injury, truthfulness, non-theft, brahmacarya
and aparigraha (not accepting more than is necessary for bodily subsistence). Lorenzen regards this
as the mahāvrata of the Kalamukha ascetics (Lorenzen 1991: 81). The close similarity of the two sets
of vows is unlikely to be coincidence, but perhaps results less from borrowing in one direction or
the other than from both going back to a shared ascetic background.
The origins of Buddhist and Jaina orders
133
particular expected outcome in terms of divine favour or spiritual success
remains.
But what was this expected outcome? What were the teachings of the
Buddha, Mahāvı̄ra and their contemporaries about?
the content of the ś r a m a n. a teachings
There are two main ways in which one can approach this question. One
is to take one or another modern form of the tradition as providing an
authoritative view of the tradition as it existed twenty-five centuries ago, in
the mid-first millennium BCE. In this case the self-understandings of one
or another presentation of the modern Jaina tradition or of one or another
representative of the often substantially different Buddhist schools becomes
primary. The other is to concentrate on the historical and textual evidence,
such as it is, and come up with our own guesses regarding ‘the original
teachings’. Many people have walked along this road with variable and
contradictory results (e.g. Ling 1973; Vetter 1990; Gombrich 1990, 1992,
1996, 1997, 2003; Bronkhorst 1993, 1998a, 1998b, 2000; Lindtner 1991–93,
1999).
What I write here necessarily falls into the second category, and I have
no real confidence of coming up with anything more secure or final than
my predecessors, many of them scholars with competences far greater than
mine in relation to the textual material. Nor does what follows have any
real claim to originality. Something, however, does need to be said at this
point in the book.
I think it helps to begin, at least provisionally, with one key element from
the traditional approaches. This is the awareness that whatever it was that
was involved in the śraman.a teachings required a radical reshaping of the
individual and his or her relationship to everyday life, that transforms the
initiate in such a way that the ordinary, everyday life of the householder
is either left behind or transvalued, so that if one continues to live it, it
is in a radically altered form, as a stage on a progress towards a goal that
lies beyond or behind the appearances and assumptions of everyday life.
The śraman.a traditions in their modern form of the organised religious
traditions of Buddhism and Jainism continue to hold that what they are
transmitting is not a simple set of moral or ethical teachings, or merely a new
intellectual understanding, but something that is intended to transform the
individual at a profound level. This might arise from or be supported by
moral and ethical teachings, or new intellectual understandings, but it is
the transformation itself that is at the centre of the process. I assume that
134
Origins of yoga and Tantra
this at least goes back to early times. If it does not, then the whole business
becomes of much less interest.
Johannes Bronkhorst’s Two Traditions of Meditation in Ancient India
(1993) is a useful starting point for a discussion of the texts themselves.
Bronkhorst re-examined the Buddhist and Jaina texts, along with the early
Brahmanical sources, in some detail to see what they had to say about early
meditation techniques. He concludes that both early Jaina and Brahmanical
practices were essentially attempts to bring the body and/or mind to a halt
through ascetic practice. ‘Among the non-Buddhists (Jainas), meditation
was a forceful effort to restrain the mind and bring it to a standstill’ (1993:
22). In a more ‘advanced’ form of the process, one also attempts to bring
breathing to a stop. Here Bronkhorst is referring to descriptions in Buddhist
sources, but the Jaina sources themselves are essentially in agreement (1993:
31–44), and much the same is true of early Brahmanical sources. In addition,
the idea of fasting to death is also found in both contexts. Bronkhorst
comments in relation to the Jainas that ‘[e]arly Jaina meditation was only
one aspect of a more general attempt to stop all activities of body and mind,
including even breathing’ (1993: 44).
If all this is correct, we might wonder how far we really should be speaking
of ‘meditation techniques’ in the ordinary sense at all in this context. What
seems to be at issue is more a way of dying which leads to escape from
the cycle of rebirth, as opposed to the normal way that ties one back into
the cycle.22 This has remained an important Jaina theme up to modern
times, and the religious ideal of fasting to death has been maintained to the
present day, though not often practised (Dundas 2002: 179–81). Dundas
notes that in this process (sallekhanā) ‘the central austerity of cutting down
the consumption of food is taken to its logical conclusion so that the body
is “scoured out” (sallikhita) of its negative factors and the mind can focus
solely upon spiritual matters as death approaches’ (179).
However, while a ‘holy death’ may be the eventual goal of early Jaina
and Brahmanical practice, these practices are not described in these terms,
either by themselves or by the Buddhists, nor does Bronkhorst imply that
they necessarily should be. In any case, I would rather read this material in
a slightly different sense, which is that the goal of ‘liberation from rebirth’
existed, and was linked to a ‘stopping’ in some sense of the activities of
body and mind, but that a developed technology of how to attain that goal
22
It may be significant that the theme of fasting to death occurs also in the early South Indian material
as a way in which a defeated king might die (Hart 1975: 88–91).
The origins of Buddhist and Jaina orders
135
did not exist. In other words, it was largely up to individual practitioners
how they might pursue that goal.
In fact, other sets of practices existed that are closer to what we might
refer to as meditation, including the four Brahmic States (brahmavihāra).
This well-known set of practices involves the cultivation of maitrı̄ (P. mettā,
love/friendliness towards other beings), karun.ā (compassion), muditā (sympathy towards the joy of others), and upeks.ā (equanimity). They are
described in the early Buddhist texts as having been practised by both
Buddhists and by some non-Buddhists, although the Buddha is reported
as saying that his followers take them further (Bronkhorst 1993: 93–4), and
they are presumably an old component of the shared culture of ascetic
practitioners I referred to earlier in this chapter.23 Bronkhorst suggests that
members of this culture were working on the basic assumption that action
leads to misery and rebirth. ‘In this tradition some attempted to abstain
from action, literally, while others tried to obtain an insight that their real
self, their soul, never partakes of any action anyhow’ (1993: 128).
Thus the key process within this ‘mainstream’ tradition, which for
Bronkhorst includes both Jaina and Brahmanical practitioners, was aimed
at stopping activities of the body and mind. The Buddhists, by comparison, describe their own practice as that of attaining equanimity towards
the senses, rather than making the senses cease to function (Bronkhorst
1993: 30). The key to doing this for the early Buddhist textual tradition is
through the four dhyāna, a standard series of meditative states. The classical
account of this is given in a number of places in the early sūtras, including
the Mahāsaccaka Sutta, where it follows on a description of the Buddha’s
initial attempts at Bodhgayā to practise the forceful stopping of breathing and other such meditations. These practices are here associated with
three Ājı̄vika teachers rather than with the Jainas, but this is probably, as
Bronkhorst suggests, because Saccaka, the Buddha’s interlocutor on this
occasion is himself a Jaina (Bronkhorst 1993: 22–5; cf. Ñān.amoli and Bodhi
1995: 332–43).24
I considered: ‘I recall that when my father the Sakyan was occupied, while I was
sitting in the cool shade of a rose-apple tree, quite secluded from sensual pleasures,
secluded from unwholesome states, I entered upon and abided in the first jhāna,
which is accompanied by applied and sustained thought, with rapture and pleasure
born of seclusion. Could that be the path to enlightenment?’ Then, following on
23
24
On maitrı̄/mettā, see also Schmithausen 1997.
Note that Bronkhorst regards parts of the Mahāsaccaka Sutta as later additions to an original core
(1993: 15–18).
136
Origins of yoga and Tantra
that memory, came the realisation: ‘That is the path to enlightenment.’ (Ñān.amoli
and Bodhi 1995: 340)
The four jhāna (Skt dhyāna) are meditative states characterised by the presence of various components (applied thought, rapture, etc.). The Buddha
explains how he then decided to end his extreme ascetic practices, ate some
food and regained his strength. He then entered the four jhānas in succession in order to purify his mind, and was then able to attain to the liberating
insight that leads to liberation from rebirth (and, in his case, Buddhahood).
In the present text of the Mahāsaccaka Sutta, this consists of three stages,
(1) the Buddha recollected his own previous lives, (2) he understood the
passing away and rebirth of beings according to their actions (karma), and
finally (3) he knew directly the Four Noble Truths and the way leading to
the cessation of the āsava (Skt. āsrava, taints or intoxicants).
This is a standard passage, repeated in many places in the early suttas:
When my concentrated mind was thus purified, bright, unblemished, rid of imperfection, malleable, wieldy, steady, and attained to imperturbability, I directed it
to knowledge of the destruction of the taints. I directly knew as it actually is:
‘This is suffering’; . . . ‘This is the origin of suffering’; . . . ‘This is the cessation of
suffering’; . . . ‘This is the way leading to the cessation of suffering’; . . . ‘These are
the taints’; . . . ‘This is the origin of the taints’; . . . ‘This is the cessation of the
taints’; . . . ‘This is the way leading to the cessation of the taints.’ (Ñān.amoli and
Bodhi 1995: 341–2)
Bronkhorst suggests that neither the Four Noble Truths (suffering, the origin of suffering, the cessation of suffering, the way to the cessation of suffering), nor the account of the recognition, origin etc. of the taints or āsavas,
which is clearly modelled on the Four Noble Truths, originally belonged
in this context. The Four Noble Truths, after all, describe the knowledge
needed to set out on the path to liberation from rebirth, but the Buddha is
now near the end of this path, and he has already recognised the path to liberation in the earlier passage about his youthful experience of the first jhāna
under the rose-apple tree. Instead, the earliest accounts of the Buddha’s liberation from rebirth probably did not specify the nature of the ‘liberating
insight’ (Skt prajñā, Pali paññā) beyond saying that it consisted in the
destruction of the āsrava (Pali āsava), the ‘taints’ or ‘intoxicants’. This is
an inner psychic process, Bronkhorst notes, and one that could hardly be
formulated discursively in a general way. Instead, the Buddha’s teaching
to his advanced disciples was likely to have consisted of personal advice
appropriate for the specific needs of each person (1993: 108–9). The Four
Noble Truths were, Bronkhorst suggests, incorporated into the account of
The origins of Buddhist and Jaina orders
137
the Buddha’s liberation from rebirth at a later stage, in response to pressure
from the ascetic community generally to provide a discursive account of
the nature of the Buddha’s liberating insight (1993: 110–11).25
Thus what was seen as original and new in the Buddhist tradition may
have been not so much the discursive content as the practical method,
which consisted in the application of the dhyāna or meditational states as
a way of bringing about the elimination of what we would see in modern language as subliminal or subconscious impressions in the mind or
body-mind. These are described in a variety of ways in early Buddhist texts
(āsrava, vāsanā, sam
. skāra). Both the use of the dhyāna and the formulation
in terms of the elimination of subconscious impressions were eventually
taken over within the Brahmanical tradition, where they form an important part of the Yogasūtra26 (Bronkhorst 1993: 68–77). Within the Buddhist
tradition, the dhyāna-techniques (śamatha in Skt, samatha in Pali) were progressively developed alongside practices that involved direct achievement
of liberating insight (vipaśyanā in SKt, vipassanā in Pali).27 The question
of the relationship between these approaches and their relative importance
has continued to be a central issue for the Buddhist tradition through to
modern times (cf. Griffiths 1981; Cousins 1996b).
Bronkhorst’s account seems to me to have considerable plausibility, given
the historical distance between even the earliest sources as we have them
and the lifetime of the Buddha. The rose-apple tree episode also strikes me
as a significant element, since it suggests the kind of spontaneous ‘visionary experience’ that occurs quite often in childhood and early adolescence
(cf. Hoffman 1992; E. Robinson 1996; Maxwell and Tschudin 1996). If we
accept this as historical, then Bronkhorst’s conclusion seems reasonable:
‘There seems little reason to doubt that Buddhist meditation was introduced by the founder of Buddhism, i.e. by the historical Buddha’ (1993: 123).
Bronkhorst’s suggestion that the nature of the Buddha’s liberating insight
(prajñā, Pali paññā) was not initially specified fits in some respects with
another intriguing study of early Buddhism, Peter Masefield’s Divine Revelation in Early Buddhism (Masefield 1986; cf. also Harrison 1987). Masefield
points to evidence in the Pali nikāyas, generally assumed to be the earliest
layer of texts in the Pali Canon, that the important division for its compilers
25
26
27
There is some evidence that the first two dhyāna meditations were also practised by Jainas, but this
Jaina meditation tradition seems to have died out at a fairly early stage (Ferreira-Jardim 2005, 2006).
This is the classic text on yoga attributed to Patañjali. This Patañjali is generally regarded as a different
person to the grammarian Patañjali discussed in Chapter 2, and the dating of the Yogasūtra remains
controversial. See Chapter 9.
On Buddhist dhyāna/jhāna, see also Cousins 1973, 1992.
138
Origins of yoga and Tantra
was not between the monk and the layman, but between the ariyasāvaka,
the noble (ariya = Skt. ārya) person who has heard the teachings, acquired
the ‘right view’ and so is on the path to nibbāna (Skt. nirvān.a), and the
puthujjana who has not yet understood the impermanence of the phenomenal world. This division cross-cuts that between laymen and monks.
Laymen (and also the devas or gods) might be ariyasāvaka, and monks
might be puthujjana (Masefield 1986: xvii-xviii). What is also significant
is that becoming a member of the ariyasāvaka (also described in terms of
paññā, insight, or attaining the dhammacakkhu or dharma-vision) is almost
always described as taking place through the actual teaching of the Buddha
(in a few unclear cases, it may be a result of the teaching of one of his disciples). As a result of this encounter with the Buddha, a person attains the
dhammacakkhu and is on the path to a specific level of attainment. Without
it, it would seem, no amount of practice will bring about the transition to
ariyasāvaka status and the path to liberation.
This is, as Harrison notes, in accordance with the way that many contemporary Buddhists, in Theravāda Buddhism in particular, see the situation.
It is often held that nobody has been able to attain nibbāna or any of
the other exalted spiritual goals laid down in the texts since shortly after
the time of the Buddha, and that the best that one can hope for today
is to accumulate enough good karma over successive rebirths to be born
in the time of the next Buddha, Maitreya, and become a member of the
ariyasāvaka in his presence. This is by no means the only understanding
of the situation in modern Buddhist societies, however, even those where
Theravāda Buddhism is practised; as Harrison again notes, it is widely
believed that exceptional individuals do continue to attain these high levels
of spiritual realisation and in fact can act as conduits of magical power for
ordinary people, empowering amulets and other ritually-powerful devices
(cf. Tambiah 1984).
One response to Masefield’s argument is to regard the Buddha and his
contemporaries as ‘charismatic spiritual leaders’ or something of the kind.
In a sense, this is no doubt true. The tradition makes the personal impact
of the Buddha and Mahāvı̄ra on their followers clear. However, it fails
to explain very much about what was going on at the time, and about
why it was that these men were able to have the impact that they did on
the society of their time. I think that the kind of arguments advanced
by Bronkhorst take us considerably further, and they present a view of
the Buddha as someone who had attained, in contemporary terms, deep
psychological insight into the workings of his own body-mind complex,
and so was able to understand how to bring about transformations in the
The origins of Buddhist and Jaina orders
139
body-mind complexes of the people whom he encountered. This image
of the Buddha as someone who had achieved a deep level of insight into
his own body-mind complex was central to the early Buddhist tradition,
and it was associated with a crucial innovation: the idea of a programme of
specific practices through which others might replicate his achievement.28
I suggested many years ago that the Buddha’s teaching was a reformulation of a kind of shamanic training (Samuel 1993: 370–3), and while I
would be wary of using the term ‘shamanic’ now, after having to deal with
the multiple misunderstandings which it has led to in the context of that
book (cf. Samuel 2005a: 8–13), I would still suggest that the historical core
of the Buddha’s teachings was likely to have been a set of practical and
experiential techniques rather than a body of discursive knowledge.
This is in line too with some other recent approaches to the Buddha’s
teaching, for example Sue Hamilton’s emphasis on the role of experience in
the Buddhist analysis of the khandhas (Hamilton 1996, 2000). Hamilton
argues that the khandhas (Skt. khandhas), the five constituents which are
presented in the early teachings, are ‘not a comprehensive analysis of what
a human being is comprised of [. . .] Rather, they are the factors of human
experience [. . .] that one needs to understand in order to achieve the goal of
Buddhist teachings, which is liberation from the cycle of lives’ (Hamilton
2000: 8). In other words, the khandhas began less as a body of discursive
knowledge than an aid to practice, which was gradually transformed over
time into a set of philosophical assertions.
It is easy enough to imagine how the discursive knowledge arose, and how
the hints and suggestions in the discourses attributed to the Buddha became
gradually expanded into a body of doctrine and theory. As for methodology,
the instruction imparted by the Buddha himself seems to have been both
personal and closely tailored to the needs of the individual student. The
Buddhist sūtras themselves preserve suggestions of this, and their formal
structure (‘Thus have I heard at one time . . .’) with its enumeration of
place, audience and situation retains a sense that the teachings need to
be understood in relation to their specific recipients. Yet in the longer
term there was perhaps no alternative but to produce a series of discursive
accounts of methodology, based on texts such as those we saw, in which the
Buddha describes what he had done himself. As we will see, this was not an
28
Precisely what was in the original programme is difficult to recover, since this depends on how one
dates various parts of the early material. However, some kind of mindfulness of breathing would
seem a strong contender as the basic technique for entering the dhyāna states. Cf. Satipat..thāna
Sutta (Dı̄ghanikāya 22, Majjhimanikāya 10; cf. also Nyanaponika Thera 1969); Ānāpānasati Sutta
(Majjhima Nikāya 118; cf. also Ñān.amoli Thera 1964); Cousins 1996b.
140
Origins of yoga and Tantra
entirely satisfactory situation, and the need to produce new methods that
might be effective in achieving the Buddha’s original insight continued over
the succeeding centuries and has indeed continued up to modern times.
How this process of the gradual accumulation of discursive material
regarding both insight and methods worked in relation to the development
of the Buddhist schools as we know them from their literature is a complex
question which we are only beginning to be able to trace in detail (e.g.
Cousins 1992, 1996b; Lindtner 1991–93, 1999; Choong 2005; Kuan 2005;
Ferreira-Jordim 2005), but the overall process is not too hard to imagine.
Since, in time, the sūtras presenting the teaching of the Buddha came to
be seen as authoritative texts, reformulations of theory and of method
(including monastic discipline) could only be achieved by the ongoing
revision of existing sūtras or the production of new sūtras.
Having said this, we need to recall that all discursive formulations of
the insights achieved by the Buddha, Mahāvı̄ra, or the Upanis.adic sages
are just that: attempts to put things down into words. The temptation for
scholars is to provide a rational and discursive model, but the material we
are talking about may simply not be amenable to such approaches.
In the remainder of this chapter I turn to look at an important aspect of
the accommodation which Buddhism and Jainism made to the pre-existing
cultural context they encountered, their relationship to the local gods of
the folk and civic religion.
loc al deities and everyd ay concerns: the buddhist
and jaina pat tern
A presentation of the śraman.a orders which focuses on their internal organisation and on their meditational practices leaves out significant dimensions
of their place in society and religion of the time. The question of the relationship between these traditions and the presumably pre-existing cults of
gods and spirits provides an important additional perspective.
It is well-known that the Buddhist and Jaina sūtras contain many accounts
of encounters with gods and spirits of various kinds, including the major
Brahmanical gods, Sakka (Indra) and Brahmā, and a host of lesser deities,
male and female.29 These figures appear in the narrative on the same level
29
For the Buddhists, see Jootla 1997; DeCaroli 2004; for parallel early Jaina examples, see e.g. the
Uttarādhyayana Sūtra ch. 9 (Jacobi 1895: 35–41, Indra appears to King Nami in the form of a
Brahmin), ch. 12 (Jacobi 1895: 50–4, a yaks.a in a tree), ch. 23 (Jacobi 1895:121–2, various deities
assembling to hear Jaina teachings). Padmanabh Jaini suggests however that Jaina sutta do not
advocate offerings to yaks.a deities and that the Jaina yaks.a cult (on which see e.g. J. Sharma 1989;
Zydenbos 1992; Pal 1994; Cort 1987) was a later development (Jaini 1991: 193).
The origins of Buddhist and Jaina orders
141
Figure 6.2. Stūpa as depicted on Bharhut Stūpa, Indian Museum, Calcutta
and in much the same terms as the other, human characters; they speak to
the Buddha, ask for and are given teachings, etc. (Jootla 1997).30 Spirits of
the dead also play a significant part in the story. The gods are also portrayed
at early Buddhist stūpa sites such as Bhārhut and Sāñcı̄ (Fig. 6.2).
It is also reasonably well known that many sites of significance for early
Buddhism are also sites with important associations with local and regional
deities. The term caitya, nowadays synonymous with stūpa as indicating
the relic-mounds which are primary sites of Buddhist worship and which
enclose the relics of Buddhist saints, including those of the Buddha himself,
originally appears to have meant a spirit-shrine (cf. Trainor 1997; DeCaroli
2004). The trees which were so strongly associated with the yaks.as became
30
All of which again suggests that we would be better off being careful about taking anything in these
accounts too literally, however accurately they may have been preserved.
142
Origins of yoga and Tantra
associated with the Buddha as well; the tree under which Śākyamuni
achieved Buddhahood was the central location for Buddhist devotional
activities, and earlier Buddhas were associated with specific trees under
which they had achieved Buddhahood.
In a recent book (DeCaroli 2004), Robert DeCaroli has analysed the
relationship between Buddhism and the ‘spirit-deities’ (to use his term) in
detail, arguing that it needs to be taken more seriously than has generally
been the case in the past.31 Specifically, he suggests, following in part from
the work of Gregory Schopen about the association between early Buddhist
sites and places of the cult of the dead (Schopen 1996), that a major initial
function of the monks was to deal with the spirit-world in general and
the spirits of the dead in particular (e.g. DeCaroli 2004: 87–103).32 I have
referred earlier in this chapter to this question of the association with the
dead, in particular in relation to the vrātyas.33
As I noted, DeCaroli uses the term ‘spirit-deity’ as a general term for both
the spirits of the dead and the local and regional deities. More customarily,
the local and regional deities have been referred to by the generic term
yaks.a, a term which seems to have had a variety of applications in the
period, being used both generically (e.g. in the Āt.ānāt.iya Sutta, see below)
and more specifically to refer to a sub-category of such beings associated
with nature, fertility etc. Neither terminology is entirely appropriate, but
deities of these kinds were clearly significant for the Buddhists and for the
general population of the time. We have already seen something of them
in Chapter 5.
yak sa religion at the bh ārhut st ūpa
The well-known remains of the Bhārhut stūpa34 that today form a highlight
of the Indian Museum in Calcutta provide the single most substantial body
of sculptural material regarding these regional deities and the relationship
31
32
33
34
Richard Cohen has argued a similar case for a later period in an important article on Ajanta, where
he also emphasises the political dimensions of the spirit-deity cults (Cohen 1998).
DeCaroli also notes that the offerings to monks echo the offerings made to the dead. He suggests that
a relationship between the two is strongly implied by the Buddhist narratives of the food offered to
Śākyamuni immediately before his enlightenment, which takes place close to Gayā, closely associated
with Brahmanical rites for the dead (DeCaroli 2004: 106–14).
Note that the Wheel of Rebirth incorporates everyone into the family – non-relatives, gods, demons,
spirits of the dead, even animals.
A stūpa contains relics of the Buddha or of someone who has attained a high level of Buddhist
realisation. In later Buddhist thought the stūpa became a key symbolic representation of the central
goals of Buddhism.
The origins of Buddhist and Jaina orders
143
Figure 6.3. Two Deities. Bharhut Stūpa, Indian Museum, Calcutta
that developed between them and Buddhism. The remains consist primarily
of parts of a stone railing that encircled the stūpa.
Much of the railing consists of decorative motifs or scenes from Jātaka
stories, but at intervals there are large figures carved in relief from the
stone. Most of these have inscriptions, which identify them as yaks.as, nāgas
or devas. Similar individual figures are known from other sites, but Bhārhut
is the only place where we have a large number of them from a single context, or where they are systematically identified. Clearly, their positioning
presents them as in some sense accessory to or attendant upon the central
message of Buddhism as figured by the shrine. Since a stūpa such as that at
Bhārhut is a large state edifice, the yaks.as could also have a political message.
Unfortunately, we only have part of the Bhārhut railing, and perhaps
one-third of the total number of yaks.a figures. It is not really possible to
reconstruct the complete set of deities on the railings. In addition, the
railing appears to have been rebuilt at some point, and the surviving figures
belong to at least two different phases, so what we now have does not
necessarily derive from an original and consistent scheme.35 However, it
35
See Barua 1934–37; Kala 1951; Coomaraswamy 1956.
144
Origins of yoga and Tantra
appears that the scheme included the four Great Kings or Lokapālas, a
set of yaks.a-type deities, already mentioned in Chapter 5, who have the
role of protectors of the four directions within Buddhism. Two of these
figures survive (Virūd.haka, the deity of the South, and Kubera, the deity
of the North). The scheme may also have included a set of four images
of Laks.mı̄, or rather Sirima (Śrı̄-ma) devatā, as the goddess is called in
the railing inscriptions. Kubera is depicted as a young man, rather than
the corpulent middle-aged figure more familiar from later iconography.
Sirima has an upright posture and is rather more formal in style than the
remaining deities, male and female, who are variously identified as yaks.as,
devas or nāgas.
Laks.mı̄ is incidentally also depicted in several other places on the railing
in one of her classic iconographical forms, that in which she is asperged
by two elephants to right and left, a form that has continued in use right
up to the present day (see the Sāñcı̄ image, Fig. 5.4). These can perhaps be
interpreted as generalised symbols of prosperity and good fortune. The large
figures however are clearly something rather more than that, both because
of their size, but also because they are carefully identified by inscriptions
and individualised in their portrayals.
We do not really have anything comparable to the Bhārhut iconographical scheme from any other site of the period, though we have quite a
few individual figures from elsewhere. However Benimadhab Barua, who
published the most detailed study of the stūpa in the 1930s, compares the
iconographical scheme to the description of the construction of the Great
Stūpa in Chapters 29 and 30 of the Mahāvam
. sa, the Sinhalese religious
chronicle. This stūpa included images of the Four Great Kings, the ThirtyThree Gods, Thirty-Two Celestial Maidens and Twenty-Eight Chiefs of
the Yaks.as, raising their folded hands.36
At Chandraketugarh we saw the yaks.a-type deities in isolation. Here
they have been brought into connection with a new sacred order, that of
the Buddha. As we learn from the Buddhist sūtras, some of these deities
are thought of as obedient to the Buddha, others as disliking or resisting
his teachings. Chief among those who are followers of the Buddha are the
Four Great Kings of the four directions, and it is their leader, Kubera,
36
‘At the four quarters of the heaven stood the (figures of ) the four Great kings, and the thirty-three
gods and the thirty-two (celestial) maidens and the twenty-eight chiefs of the yakkhas; but above
these devas raising their folded hands, vases filled with flowers likewise, dancing devatas and devatas
playing instruments of music, devas with mirrors in their hands, and devas also bearing flowers and
branches, devas with lotus-blossoms and so forth in their hands and other devas of many kinds, rows
of arches made of gems and (rows) of dhammacakkas; rows of sword-bearing devas and also devas
bearing pitchers.’ (Geiger 1964: 207).
The origins of Buddhist and Jaina orders
145
who comes to the Buddha in the Āt.ānāt.iya Sutta to provide him (or rather
his followers) with spells that can be used to keep the other yaks.as under
control should they cause trouble for the Buddha’s followers. The Four
Kings became major figures in Buddhist iconography in Tibet and East
Asia.37
In general, though, the yaks.a deities can be incorporated into the
Buddhist scheme as lesser beings that rule over the everyday or laukika
world. This, I think, is the message conveyed by the Bhārhut images. The
contrast with the Buddha, who rules over the higher, lokottara, realms, is
explicit not only in the folded hands and respectful postures of the Bhārhut
yaks.a deities but also in the fact that they are represented at all. The Buddha
at Bhārhut is not depicted directly in iconographic form, except in the sense
that the central stūpa itself is a representation of him.
As I noted above, we should see a construction like the Bharhut stūpa as
a political statement as much as a religious one. In signalling his support for
the universal order of the Buddha, and portraying the various regional spiritdeities as submitting respectfully to that order, the ruler who sponsored the
stūpa construction was surely making an implicit claim to a similar status.
The moral and ethical dimension of the Buddhist teachings were also, as
the Aśokan inscriptions suggest, an intrinsic part of the package.
Note that we do not see anything in either body of material, or in
other comparable material, to suggest the presence of fierce or aggressive
goddesses. There have been suggestions that they can be found in the
Vedas, but our evidence for them is really from a later period altogether,
and I will leave it to a subsequent chapter (Chapter 10).
The evidence from these early sites and from the early textual material
suggests that Buddhism from its initial stages took on a significant role in
relation to the existing local and regional spirit-cults. Much the same was
doubtless true of the Jaina tradition in the areas where it became established.
These spirit-cults, rather as in countries such as Thailand today, would
have existed at many different levels, from protective spirits of households
to protective spirits of wider communities, probably (from later evidence)
mainly at trees or sacred groves outside towns (caityas), to regional cults
associated with emergent states. There would also have been concern with
the potentially malevolent role both of these spirits and of the spirits of the
dead.
37
The Bharhut Stupa, and the description in the Mahāvam
. sa, are also reminiscent of the description
in Mahāsamaya Sutta, where a huge array of deities comes to attend on the Buddha (Dı̄ghanikāya,
20 = Walshe 1995: 315–20).
146
Origins of yoga and Tantra
The role of the śraman.a traditions as far as the lay population was concerned was to bring the spirits under control, to civilise them and incorporate them within the community, and to provide an overall ideological
framework within which their cult could be maintained on an ongoing
basis. The cult itself, as again in Southeast Asia today, would probably have
been mostly carried out by lay priests, not by monks, though the monastery
provided a centre of spiritual power and expertise that could be called upon
in cases of emergency.
buddhist and j aina at titud es to the yaks. as
Thus in regions where Buddhism had become a state or official religion,
the yaks.as were subordinated to Buddhism but their cult was allowed to
continue. They were expected to be polite and respectful to the Dharma
and its representatives, and could be threatened by ritual means if they
caused trouble, but the relationship was one of a harmonious coexistence, in
which the yaks.a deities represented the laukika (this-worldly) values while
the Buddha, and in later times the developing cult of bodhisattvas and
other Buddhist deities, represented the lokottara or trans-worldly values of
rebirth and enlightenment. A number of the yaks.a deities were regarded as
particularly close to the Buddhist teachings and as acting as their protectors.
We have seen examples in the case of Vajrapān.i and the Four Great Kings.
Much the same is true in the Jaina tradition, where the tı̄rthaṅkaras are
portrayed with attendant yaks.a and yaks.ı̄ deities, who can be appealed to
to serve the mundane needs of their followers (Cort 1987; J. Sharma 1989).
This is a picture that is familiar from the modern ethnography of the
Theravada societies of Southeast Asia. While there are obviously local variations and issues in each of the major Southeast Asian cultures, we can,
I think, see in the relationship between Buddhism and the spirit cults in
places such as rural Thailand or Burma, as it was in fairly recent times, a
demonstration of how this kind of situation worked. For an example of a
major regional centre, we could look at the city of Chiang Mai in Northern
Thailand, formerly the capital of the state of Lan Na, which goes back to
the thirteenth century and was through much of its history more or less
independent of the Thai state at Sukhothai or Ayutthaya (Wijeyewardene
1986; Davis 1984; Rhum 1987; Aasen 1998: 65–7).
Chiang Mai is a walled city with a series of important monasteries,
one of them containing a crystal Buddha which had much the same stateprotective function for the Lan Na state as the well-known Emerald Buddha
at Bangkok (Aasen 1998: 85–6). Close to the city is the sacred hill of Doi
The origins of Buddhist and Jaina orders
147
Suthep, with a famous Buddhist stūpa containing a relic of the Buddha
and supposedly visited by him in his historical lifetime (Swearer 1976;
Wijeyewardene 1986; Rhum 1987; Samuel 2001a: 399–400).
There is no doubt that at one level this is a prototypically Theravāda
Buddhist community. Yet the equivalent of the yaks.a cult is still very much
alive. As at Bangkok, there is a City Pillar shrine that is the place where
the guardian spirit of the city is worshipped. The hill of Doi Suthep also
has guardian spirits, who receive annual buffalo sacrifice and possess local
spirit-mediums. The myth of these guardian spirits regards them as the
parents of a local Buddhist .r.si, Vāsudeva (the origin of the name Suthep, in
fact) whose story is part of the origin myth of the first Buddhist kingdom
of the region, the Mon kingdom of Haripuñjaya.38
Chiang Mai was a sophisticated urban centre, as of course Bhārhut must
have been. Folk and village religion leaves little direct trace in the archaeological record but we can imagine that it existed and that it dealt with
smaller-scale deities of a basically similar type, as we can again see in the
village ethnography of contemporary Thailand and Burma. In the religious
life of the Northeast Thai village studied by Stanley Tambiah, for example,
the village’s annual ritual cycle is a complex interweaving of Buddhist festivals linked to the universal Theravada Buddhist festival cycle (and to the
Thai state) with local celebrations tied in with the village’s agricultural cycle
and directed towards the founding ancestor spirits of the village and the
regional protector spirit residing in the local marsh. This marsh-dwelling
deity is identified, as is apparently quite common in rural Thailand, with
the Buddha’s miracle-working disciple Upagupta (Tambiah 1970; Strong
1992; cf. also Samuel 2001a).
The spirit-cult is about this-worldly success and prosperity, while the
values of Buddhism are directed towards karma, rebirth, and, if only indirectly, the attainment of nirvān.ā. Yet the two are closely interwoven. The
young men of the village spend a period as monks in the monastery, but
this is not expected to be the beginning of a monastic career, and only a tiny
minority remain as monks. For the majority, this temporary ordination is
in fact a rite of passage to adulthood and marriage. The central transactions between villagers and monks relate to the making of merit, or positive
karma leading to a good rebirth, by donations of food, clothing and other
items to the monks, and through the temporary ordination of young men
38
I have written elsewhere about the role of .r.sis and sages in both Brahmanical and Buddhist contexts
as providing local associations and connections for religious concepts and traditions that in fact
originated outside a particular region (Samuel 2001a).
148
Origins of yoga and Tantra
as monks. Both of these are meritorious actions helping to ensure a good
life in one’s next rebirth.
In his 1970 book, Tambiah summarised the relationships between rituals,
concepts and values in the following diagram. The diagram belongs in
some respects to an era of structuralist over-tidiness which Tambiah himself
moved away from in his later work, but gives a useful sense of the range of
Figure 6.4. S.J. Tambiah’s Diagram of the Value System of Northeast Thai Village Society
The origins of Buddhist and Jaina orders
149
concerns of village religion, and of the relatively limited place of Buddhism
within it:
Here the sukhwan rituals are rituals relating to the idea of a life-force,
soul or inner vitality (khwan) which can be lost from the body and which is
summoned back at times of danger and also of rites of passage; marriage in
the village is a sukhwan rite, not a Buddhist ritual. The cult of guardian spirits and rites of malevolent spirits refer to positive and negative relationships
with local spirits, as guardians and as causes of illness. The relationship
of Buddhist monks to these three spheres is tangential, although some
Buddhist monks serve as exorcists, and paritta rituals (the recitation of specific Buddhist sutta with magical-protective functions) may be integrated
with spirit-cult practices. Overall, however, in this village, as in other village
ethnographies from Thailand, Burma and Sri Lanka (Brohm 1963; Ames
1964; Spiro 1967, 1971; Davis 1985; Rhum 1994), the opposition between
‘worldly’ and Buddhist religion (laukika and lokottara) is ideologically quite
clear. They are both seen as legitimate, and there is a structural relationship between their underlying values (as Tambiah’s diagram is intended to
show), but they occupy separate spheres.
In fact, some analysts of village Buddhism in Southeast Asia have spoken of separate religions (e.g. Spiro 1967). This, I think, is to go too far.
Buddhism and the spirit cults, sukhwan rituals, etc., are clearly related in
all kinds of ways in village thought and practice. The idea, however, of a
separate set of village concerns dealt with by a separate body of people,
who are specifically not Buddhist monks, is clearly present. The Buddhist
monk does not interfere in rituals of everyday life, except to be present
as a way of generating merit or auspiciousness. Major village festivals may
be Buddhist dominated or lay-dominated or involve elements from both,
but the two spheres remain conceptually separate, and the monastic-lay
opposition (which corresponds clearly to celibate versus non-celibate, to
other-worldly versus this-worldly) runs throughout the society.
The details vary in other Theravādin countries (temporary ordination is
a characteristically Thai and Burmese feature, and is not found in Sri Lanka,
for example) but we find much the same overall pattern. Western observers
often used to interpret this as a situation in which Buddhism is a superficial
veneer over the spirit-cults, but the anthropological studies of the 1960s
onwards made it clear that it is a much more complex situation in which
both are an integral part of the society. Buddhism provides a specific ethical
dimension to village life, linked to the Buddhist teachings on virtuous and
non-virtuous action and their consequences. People’s respect for Buddhist
values may be real, and may be a real part of their understanding of the
150
Origins of yoga and Tantra
world, even if they have no expectation or intention of seriously pursuing
the Buddhist goal of nirvān.a in their own lifetime. At the same time, those
values come, in a sense, from outside the village community, and are aligned
with the official values of the wider state of which the village forms part.
Buddhism, in other words, is a state ideology as well as an integral part of
village religion.
Both Buddhists and Jainas were prepared to include the Brahmanical
deities in their scheme on essentially the same basis as the yaks.as, as lesser
spirits who could be invoked for this-worldly and pragmatic purposes as
long as they accepted their inferior position. There are no specifically Brahmanical figures included at Bhārhut, but we can find them elsewhere, for
example in the early Buddha images from Mathurā, Gandhāra and elsewhere in which Indra and Brahmā are shown as smaller figures on either side
of the Buddha, often asking him to teach (e.g. Klimburg-Salter 1995: 188
no.168, 189 no.169). The Mahāmāyūrı̄ list mentioned in Chapter 5 includes
many of the well-known Brahmanical deities, each treated as a yaks.a with a
specific location. Unfortunately, not all of these locations are identifiable,
but they include Vis.n.u at Dvaraka, Śiva at Śivapur (Shorkot in the Punjab),
Siva’s son Kārttikeya at Rohitaka or Rohtak, Indra at Indrapura (Indore in
Bulandshahr Distt., UP or Indore in MP) and so on (Sircar 1971–2: 270,
279, 275, 280). I imagine the Buddhist compiler of this text taking a certain
ironic pleasure in treating what were undoubtedly by this period major
all-India deities as, in effect, local protective spirits. Similarly, the At.ānāt.iya
Sutta closes with a list of yaks.as (this is what the text calls them) who may
be invoked as protection against those disobedient even to the Four Great
Kings. The first members of the list are Brahmanical deities or sages: Inda,
Soma, Varun.a, Bharadvaja and Pajapati.
This inclusion of the early Brahmanical deities, again, is a feature of modern Theravādin communities: the city pillar shrine in Chiang Mai is linked
to Indra as well as to the city’s protective spirit, and rituals are still performed there to ensure the rains, a traditional Vedic function of Indra. The
most popular shrine in downtown Bangkok, outside the Hotel Erevan,
is a shrine to Brahmā, regularly invoked for this-worldly and pragmatic
purposes (Majupuria 1993). In Sri Lanka, Indra, or Sakka as he is usually called in Pali, is regarded the protector of the island under Buddhist
auspices, with a set of four gods (the list varies somewhat) as his immediate assistants, including Vis.n.u, Pattinı̄, and Kārttikeya (Ames 1964, 1966;
Seneviratne 1978; Holt 2004). The specific cult of the deities here reflects the
close historical relationships between the Kandyan state and South India.
The origins of Buddhist and Jaina orders
151
The general conceptual structure, however, including the dominance of
Indra, a deity who has been largely marginalised in India proper, is very
much the same as in other Theravādin countries.
There is much more that could be said about this general situation, in
which the developing Buddhist and Jaina traditions constructed a harmonious if hierarchical relationship between themselves and the cults of local
yaks.a-type deities. For example, many of the sites associated with both the
teachings of the historical Mahāvı̄ra and Śākyamuni, and with their early
religious centres, seem to have had previous yaks.a associations. I have said
enough here for my purposes, however, except that I want to emphasise
the contrast with the pattern we find generally in modern Hindu communities. Here the local deities, even where they clearly have strong local
identities, are (officially at least) regarded not as a separate sphere of worldly
gods, but as local transforms of an all-India or regional deity such as Śiva or
Pārvatı̄. This is so even in major temples – we can think, for example, of the
great goddess temples of South India, such as Minaks.ı̄ at Madurai, which
I will discuss briefly in Chapter 7, of Kāmāks.ı̄ at Kanchi, or of Kumārı̄
at Kanyakumari, all goddesses who have a very strong local identity and
individuality, but are officially regarded as transforms of Pārvatı̄ or Durgā.
In the north, the myth of the Śākta pı̄t.has similarly links many local and
regional religious sites to the Śiva-Pārvatı̄ story. Other myths interpret local
shrines in terms of Kr.s.n.a, Vis.n.u and other all-India deities – Jagannāth at
Puri would be a classic example (Eschmann et al. 1986).
This – to anticipate my later argument – can be seen as typical of ‘Tantric’
religion, using ‘Tantric’ in a fairly wide sense. A key issue here will be the
nature of the man.d.ala as a spiritual and political model (see Chapter 9).
For the present, though, we might look at the Matsya Purān.a’s story of the
yaks.a Harikeśa, son of the great yaks.a king Pūrn.abhadra. Harikeśa was a
devotee of Śiva, and after quarrelling with his father went to Vārān.ası̄ and
engaged in a thousand years of austerities ‘with his mind intent on Śiva’.
Eventually Śiva, pleased with his devotion, allowed him to stay permanently
in Vārān.ası̄ and appointed him as the ks.etrapāla or guardian of the city, with
four other yaks.as as his subordinates (Agrawala and Motichandra 1960).
Today, though, the spiritual guardianship of Vārān.ası̄ is not in the
hands of yaks.as, whether Buddhist or Śaivite, but those of a set of Bhairava
deities, regarded as fierce transforms of Śiva. This was perhaps already
under way when the Mahāmāyūrı̄’s catalogue was written, since it refers to
the yaks.a of Varan.ası̄ as Mahākala, which in Buddhist contexts generally
denotes a fierce transform of Śiva who is acting as a protective deity. The
152
Origins of yoga and Tantra
nature of this transition will be one of the key themes of Part 2 (Chapters 9
to 13) of this book.39 First I turn, in Chapter 7, to the developments in the
Brahmanical tradition from the fifth century BCE onwards that paralleled
those that we have considered in the present chapter among the Buddhists
and Jains.
39
Tibet, by the way, offers a mixture between the two models, early Buddhist coexistence and Tantric
incorporation. In this respect, as in others, Tibet represents a much more radical transformation
of Tantric religion, Śaiva or Buddhist, in the South Asian context than we find among the Newars
or in Bali (Samuel 1982, 1993). Some major mountain deities (Kailash, Tsari) were incorporated via
the Cakrasam
. vara mythos (Macdonald 1990; Huber 1999). Possibly Panden Lhamo, the protective
deity of Lhasa and of the Gelugpa order, and apparently related to a pre-Buddhist deity of Tibet, is
a similar case. The Bhairava or Mahākāla pattern of protector is also important in monastic ritual
contexts. For the most part, though, Tibetan local deities are treated as a separate and lower order
of beings, not as transforms of Buddhist deities. There can be some overlap and confusion here,
and the notorious Shugden controversy, which divided the Tibetan refugee community in recent
years, was in part about whether this deity should be regarded as a (possibly malevolent) worldly
spirit-deity or an enlightened Tantric protector and transform of Mañjuśrı̄ (cf. Dreyfus 1998).
chap t e r 7
The Brahmanical alternative
I referred in Chapter 6 to Brahmanical asceticism, but did not go into much
detail about how it was practised and understood, and how it related to the
‘Vedic’ pattern of religion developing in the Kuru-Pañcāla region. In this
chapter I explore the development of asceticism on the Brahmanical side in
more detail, and discuss how Brahmanical thinkers and priests responded to
the challenge of the śraman.a traditions which we considered in Chapter 6.
At the end of this chapter, I look at the way in which the Brahmanical
alternative related to regional and local deities and cults, contrasting its
approach to the Buddhist and Jain approach sketched in Chapter 6.
varieties of brahmins
I referred in Chapter 6 to the indications of a variety of different Brahmin
roles in the fifth and fourth centuries BCE or thereabouts, and mentioned
Patrick Olivelle’s demonstration in his book The Āśrama System that the socalled āśramas originally represented alternative ways of life rather than
a sequence of stages through which all men of the Vedic elite groups
(Brahmin, ks.atriya, vaiśya) should progress (Olivelle 1993). Olivelle notes
that there ‘appear to have been tensions and rivalries between the traditional Brahmins of the villages, who were the heirs and guardians of the
vedic world, and the newly-urbanized Brahmins’ (1993: 59). Elsewhere
in The Āśrama System, Olivelle points to the references in Buddhist and
Brahmanical sources to married semi-renunciate Brahmins, characterised
by their matted or braided hair (jat.ila) hairstyle, living in uninhabited areas
outside towns with their wives and children. We have already met these
jat.ila Brahmins in Chapter 6. They maintained a sacred fire, and their
life ‘centered around the maintenance of and the offering of oblations
in the sacred fire’ (1993: 24). Their residences were described as āśrama;
Olivelle argues that this is the original meaning of āśrama, preceding
the later usage to refer to a particular period or stage of life. They also
153
154
Origins of yoga and Tantra
performed tapas or austerity and were part way at least to an ascetic
lifestyle.
Interestingly, these jat.ila Brahmins were viewed with respect by the early
Buddhists. Olivelle notes, for example, that the Vinaya code exempts them
from the four months probationary period normally required before admission to the saṅgha (1993: 21). Buddhist sources often describe Śākyamuni
as pointing out that this was how Brahmins used to live and ought to
live, by contrast with the village-dwelling Brahmins. However, the Buddha
argues that they need to abandon the sacred fire and go beyond the life of
a householder.
As we saw in Chapter 6, we seem to have a variety of Brahmins at this
period: traditional village-dwelling Brahmins, urban Brahmins who were
perhaps taking on the role of spiritual teachers of the new Brahmanical
ritual practices for householders (and so were in competition for patronage with the Buddhists and Jains), semi-renunciate Brahmins who had
retreated to settlements outside the town or village, and, apparently a later
development, fully-fledged Brahmanical renunciates who had abandoned
their householder lifestyle and also abandoned (or rather internalised) their
sacred fires. The unpopulated areas with which these last two groups were
associated were also of importance to the non-Brahmanical ascetic traditions. As I pointed out in Chapter 6, these Brahmanical ascetics seem to
have been part of a common ascetic culture with a tradition of wisdom
practices, teachings and sayings that were at least partly shared. They also
shared with them a liking for the ‘forest’, in other words for uncleared and
unpopulated areas outside the cities and their agricultural hinterland.
brahmanical asceticism
What, though, were the sources within the Brahmanical tradition that
might lead to such a development? I referred in Chapter 1 to the ‘shamanic’
and ‘ecstatic’ elements of the early Vedic material. I am somewhat wary
about using the word ‘shaman’ these days (see Samuel 2005a: 9–13), but if
one takes the contrast between shaman and priest to be, at least in part,
that the shaman deals with the gods or spirit as an equal, while the priest
treats them from a position of inferiority, then it is clear that the R.gveda
at least looks back towards a shamanic kind of relationship with the gods.
The .r.sis to whom the Vedic hymns are attributed, whether or not they
actually existed as historical figures, were seen as having power to compel
the gods and so compel the natural order (Gonda 1963; see also Crangle
1994). So, one assumes, were the ritualists who claimed descent from them,
The Brahmanical alternative
155
at least initially. Different groups of these shamanic-type .r.sis were employed
by various chieftains and warriors to further their own interests, and they
might invoke the spirits against each other and try to interfere with each
other’s rituals.
At some point, however, no doubt gradually over time, things changed,
and the Brahmins of a later period, though still tracing their ancestries to
clans or gotra founded by one or another of the original seven .r.sis took on
a much more priestly role, as sacrificial performers. The Vedic hymns were
gradually transformed into ritual formulae to be recited at various stages
of the Brahmanical ritual, which was elaborated into a set of formalised
procedures and the conflictual elements played down.1 Falk, Heesterman
and others have written about this development, as has Witzel, who as we
have seen associated it particularly with the Kuru-Pañcāla state. I assume
that he has the general place and period more or less right, and that this
development can be linked in particular with the shift from a primarily
pastoral-nomadic economy to a situation where the Vedic population was
living among and increasingly merging with agricultural populations in
North India, many of them probably already Indo-Aryan speaking. The
stories of the seven .r.sis, and of the conflicts between the .r.si or their associated
tribes, such as that between Vāsis.t.ha and Viśvāmitra, gradually became
mythic material that was reworked in a variety of contexts (see Chapter 9,
where I discuss the role of these two .r.sis in Book One of the Rāmāyan.a).
To the extent that the sacrificial rituals became about the assertion of
status, they also perhaps increasingly lost their explicitly instrumental character. If Witzel and Falk are correct, however, the process of regularisation
of Vedic ritual left some important functions to be dealt with by other
kinds of ritual specialists. In Chapter 5, we considered the special case of
the vrātyas, but there are other examples. The hymns of the Atharvaveda,
including many pragmatically-oriented rituals, were the preserve of a separate group of Brahmins from those associated with the first three Vedas.
Thus the Vedic material retained the possibility of providing a body of magical techniques for pragmatic purposes, and the idioms that it provided for
doing this are of significance for later developments. It is worth looking
briefly at two complexes of ideas, one relating to the terms tapas and śrama
and the second relating to the term vrata.
The term tapas, which is often translated as ‘austerity’, in the sense of
carrying out demanding or extreme ascetic practices, derives from the root
1
While the high status of the Brahmins in Kuru-Pañcala is evident, it is probably also worth bearing
in mind Richard Lariviere’s point that our view is biased in their favour by the Brahmanical nature
of most of our sources. The king, even in Kuru-Pañcāla, retained the real power (Lariviere 1997).
156
Origins of yoga and Tantra
tap, to give heat, to make hot, to be hot. Walter Kaelber’s Tapta Marga
(1989) traces the associations of this root in Vedic literature with the fire
and with the sun which provides the heat needed to ripen the fields and
which also generates rain. Tapas is a basic creative force, with close links
also to sexual desire and excitation, intercourse and orgasm.
Hints of these associations can already be found in the R.gveda. Their
full development, however, takes place in the Brahman.as and especially
in the Śrauta Sūtras, where ascetic practice has become a key part of the
conception of the sacrificer’s role (Kaelber 1989: 23). Tapas thus becomes
something which can be built up within one’s self, so generating power:
Crucially, man may develop or generate such ‘magical heat’ within himself through
the performance of various self-imposed austerities, such as fasting or chastity. [. . .]
In that such austerities generate heat within man, they too are regarded as a form
of tapas. Tapas refers therefore not only to heated power but also to the heated
effort which produces it. [. . .] The accumulated tapas of asceticism ‘saturates’ the
devotee, making him a reservoir of heated potency. This power may manifest itself
as a sexual and fecundating energy which when released generates rainfall, fertile
fields, and biological offspring. Through an asceticism that ‘overcomes’ nature,
man is able to control nature. (Kaelber 1989: 144)
This generates another series of images; the heated sweat of the sacrificer
corresponds to the heated milk in the cauldron of the pravargya, a hot milk
sacrifice to the Aśvins, and the overflowing of the cauldron corresponds to
ejaculation.
[T]he heat assumed by the devotee in the presence of the sacred fire makes him
susceptible to visions and divine revelation. Agni, himself a seer, provides the
devotee with ‘head-heat,’ turning him into a seer as well. Relatedly, the heated
effort of ascetic practice kindles an ‘inner fire’ of illumination, yielding ecstatic
insight. Like the paradigmatic .r.sis, the earthly ascetic is able to ‘see’ or behold
through his tapas, through his self-imposed austerity. In this context his tapas takes
the form of ‘cognitive brooding’ or ‘intense meditation’.[. . .] Once again, however,
we see a clear transparency to that natural heat generated by a hen as she broods
over her eggs. Relatedly, visions mature in the heated mind just as crops mature
and ripen upon a gently heated earth. (Kaelber 1989: 145–6)
Closely related to tapas is another term; śrāma, a Vedic root meaning ‘to
strive’, ‘toil’ or ‘to exert oneself ’. Here again this can refer to the accumulation of creative power, or its sexual expression: the creator god Prajāpati’s
creative intercourse with the earth, for example, is referred to as śrāma (Olivelle 1993: 9). One can exert oneself through sexual intercourse, so creating
new life, or one can exert oneself in ascetic practice (tapas). A well-known
dialogue hymn from Book One of the R.gveda, in which the speakers are
The Brahmanical alternative
157
identified as Lopāmudrā, the wife of the .r.si Agastya, the .r.si himself, and the
commenting Vedic poet, seems to counterpose the two meanings of ‘toil’,
with Lopāmudrā succeeding in distracting her husband from ascetic ‘toil’
to sexual ‘toil’:
1. [Lopāmudrā:] ‘For many autumns past I have toiled, night and day, and each
dawn has brought old age closer, age that distorts the glory of bodies. Virile
men should go to their wives.
2. ‘For even the men of the past, who acted according to the Law and talked about
the Law with the gods, broke off when they did not find the end. Women
should unite with virile men.’
3. [Agastya:] ‘Not in vain is all this toil, which the gods encourage. We two must
always strive against each other, and by this we will win the race that is won by
a hundred means, when we merge together as a couple.’
4. [Lopāmudrā:] ‘Desire has come upon me for the bull who roars and is held back,
desire engulfing me from this side, that side, all sides.’
[The poet:] Lopāmudrā draws out the virile bull: the foolish woman sucks dry
the panting wise man.
5. [Agastya:] ‘By this Soma which I have drunk, in my innermost heart I say: Let
him forgive us if we have sinned, for a mortal is full of many desires.’
6. Agastya, digging with spades, wishing for children, progeny, and strength, nourished both ways, for he was a powerful sage. He found fulfilment of his real
hopes among the gods. (O’Flaherty 1981: 250–51 = R.gveda 1.179)
This hymn is one of a number of dialogue hymns, in all of which one
character attempts to persuade the other to sexual activity (O’Flaherty
1981: 245). While we can only guess at the original ritual context, it seems
fairly clear that the central theme here is the diversion of the ascetic energy
of the tapasvin, the person who is generating tapas, to productive ends.
However, śrām is the root of śraman.a, the word for renunciate ascetic that
at a later period becomes clearly opposed to the Brahmanical priest, being
typified by the Buddhist and Jaina śraman.as (Olivelle 1993: 9).
Not all śrāman.as, however, were outside the Vedic fold. Even in the
R.gveda, there seems to have been a range of different kinds of practitioners
besides the sacrificial priests. These include the long-haired (keśin) sage or
muni described in R.gveda 10.136, a text that has often been interpreted as
describing a kind of shamanic practice (cf. Werner 1989; Deeg 1993). This
is a translation of the first four verses:
1. Long-hair holds fire, holds the drug, holds sky and earth. Long-hair reveals
everything, so that everyone can see the sun. Long-hair declares the light.
2. These ascetics, swathed in wind, put dirty red rags on. When gods enter them,
they ride with the rush of the wind.
158
Origins of yoga and Tantra
3. ‘Crazy with asceticism, we have mounted the wind. Our bodies are all you mere
mortals can see.’
4. He sails through the air, looking down on all shapes below. The ascetic is friend
to this god and that god, devoted to what is well done. (O’Flaherty 1981: 137–8)
In the final verse, the long-haired sage ‘drinks from the cup, sharing the
drug with Rudra’ (O’Flaherty 1981: 138). He seems to be a naked sage
(‘swathed in wind’) and there are suggestions of control over sexual power
(Olivelle 1993: 12–14). The association with Rudra, the god who provides
the prototype for the later Śiva, and who in the Vedas is an ambivalent
figure who is not fully incorporated into the Vedic pantheon, is intriguing
in view of the later linkage between Śiva and ascetics. But there are other
possible ascetic figures too in the Vedas, including the vrātyas, who have
been discussed in Chapters 5 and 6, and the brahmacārin, a figure who we
will examine briefly below.
For the early Vedic context, Kaelber appears willing to follow Eliade,
Oldenburg and others in comparing the ‘heated passage’ leading to insight
to the heated passage of ‘classical’ shamanic initiation, though these are of
course homologies rather than suggestions of historical linkages (Kaelber
1989: 125–41). Whether or not we use the term ‘shaman’ for the long-haired
muni, the vrātya or for that matter for the .r.sis themselves, it is clear that we
have a tradition of figures who communicate with the gods while in ecstatic
states, and who have visionary powers as a result of their divine contact.
These figures exercise tapas in some sense, and this is seen in terms of the
accumulation of sexual or quasi-sexual ‘heat’. They are also, in some cases
at least, expected to put their power to the service of the community (e.g.
Kaelber 1989: 17–22).
The sexual dimension to Vedic asceticism helps to explain why the brahmacārin, who is defined by his abstinence from sexual intercourse, is a
central figure (Kaelber 1989: 17–22). The logic here is that sexual ‘energy’
is either expended in family life for the production of children or for pleasure, or directed towards the attainment of ascetic power. In later tradition,
the primary reference of the term brahmacārin is to the stage of being a
Vedic student in the four-āśrama scheme, but this scheme is itself a later
development. What is clear in the earlier material is that the brahmacārin
is a repository of power built up through his tapas (in this sense meaning
‘ascetic practice’) and this tapas can be put to the use of the community. For
Kaelber, a key example is the well-known mahavrāta ritual of the vrātyas,
described in the Jaiminı̄ya Brāhman.a and the Śrauta Sūtras, which appears
to have involved ritual intercourse between a brahmacārin and a prostitute
The Brahmanical alternative
159
(see Chapter 5, and Gonda 1961). Here both brahmacārin and prostitute
are sources of power. The creative ritual role of the brahmacārin is already
stated in the Atharvaveda (11, 5):
1. Stirring both worlds the Brahmacārı̄ moveth: in him the deities are all oneminded, He hath established firmly earth and heaven: he satisfies his Master
with his Fervour [=tapas]
2. After the Brahmacārı̄ go the Fathers, the heavenly hosts, all Gods in separate
order, After him too have the Gandharvas followed, thirty and three, three
hundred, and six thousand. He satisfies all Gods with his devotion . . .
6. Lighted by fuel goes the Brahmacārı̄, clad in blackbuck skin, consecrate, longbearded. Swiftly he goes from east to northern ocean, grasping the worlds,
oft bringing them anear him . . .
12. Thundering, shouting, ruddy-hued, and pallid, he bears along the earth great
manly vigour.
Down on the ridge of earth the Brahmacārı̄ pours seed, and this gives life
to heaven’s four regions . . . (Griffith 1916, vol. 2: 68–70 = vv.1, 2, 6, 12; cf
Kaelber 1989: 17)
In the Vedic material, the point of the exercise is to put the power of tapas
at the service of the community, to generate fertility and productivity, in a
word what contemporary anthropologists of India call ‘auspiciousness’ (V.
Das 1987; Marglin 1985a; Madan 1985; cf. Samuel 1997). This is what the
post-Vedic tradition referred to as the karma-kān.d.a. But the forest ascetic
tradition, as witnessed in the early Upanis.ads, moves in another direction.
Here tapas is used to drive the search for knowledge (jñāna) and liberation
(moks.a), a search which involves the rejection of ordinary social identity and
of obligation to the community through identification with the Absolute.
This jñāna-kān.d.a is a direct assault on the traditional view that ascetic
practice is for the good of the community:
The tradition and spirit of the karma-kān.d.a was at first challenged by the innovations of the jñāna-kān.d.a. This took place on many fronts. Traditional ritual
and its rewards were challenged by a saving knowledge. The prestige of priest and
sacrificer was challenged by ascetics of the forest. [. . .] The value of worldly goals
was challenged by a pessimism regarding sam
. sāra. [. . .] The value of karma was
challenged by a salvation beyond action and the value of dharma was challenged
by a salvation beyond conventional morality.
In time, however, the orthodoxy and orthopraxis of the karma-kān.d.a met these
challenges. The result has been termed the Brāhman.ic(al) synthesis. In point of
fact, however, it was not so much a synthesis of the two trends as it was the
absorption of one tradition by the other. The orthodoxy of the karma-kān.d.a found
room for the innovations of the jñāna-kān.d.a and brought them into its fold. [. . .]
160
Origins of yoga and Tantra
Although moks.a is cited as the culmination of Vedic values, dharma invariably
remains primary in the law books. Although knowledge is lauded, ritual action
remains paramount in the Śrauta Sūtras. Despite the elitist goal of moks.a, the
priesthood and its constituency still sought rainfall, fertility, pleasure, success, and
a repertoire of worldly ends. (Kaelber 1989: 79)
What was at stake here was the incorporation of the Vedic version of
the śraman.a option within the Brahmanical fold, and the defusing of the
challenge that it posed to the priestly families. As Kaelber notes, one can
go for an ‘orthogenetic’ model here, following Biardeau and Heesterman
among others, and stress the continuity of these developments with the
original Vedic material, or, following writers such as Eliade, Dumont or
Olivelle, one can see a process of ‘challenge and assimilation’ in which new
models and approaches, not necessarily emerging from within the Vedic
context, are confronted and incorporated more or less successfully. Kaelber
prefers the latter approach, though as he rightly notes, neither can provide
a complete model of processes of such complexity (1989: 107–9). I would
agree, and suggest that this is generally a more productive way to approach
these complex processes than purely in terms of the internal dynamics
of Brahmanical thought. The ‘orthogenetic’ approach offers elegance and
closure, but it lacks a certain degree of realism, especially given that the
logic of Vedic thought and practice is such that new developments are
regularly incorporated through the use of resources already present within
the tradition.
The actual assimilation, in Kaelber’s view, took place through the adaptation of the brahmacārin model to include the rival forms of ascetic practice. These became included within the Brahmanical model through the
four āśrama model (brahmacārin, gr.hastha, vānaprastha, sam
. nyāsin) that
was to become normative in later times. Originally, though, as we saw
in Chapter 6, these four stages (brahmacārin, householder, forest-dwelling
semi-renunciate ascetic and fully renunciate ascetic) were alternatives rather
than sequential. In the new model, brahmacārin gradually became a kind
of period of studentship before married life, while the forest-dweller and
renunciate options were effectively tamed and incorporated. Renunciation, rather than being a rejection of dharma and Brahmanical regulation,
became progressively regulated by Dharmic codes. It now became conceptualised as a period of penance, purification and preparation for death at
the end of one’s life (1989: 121). In reality, many, probably the majority, of
sam
. nyāsins even today renounce earlier in their lives, and their ascetic career
is really an alternative to that of the householder rather than a concluding
The Brahmanical alternative
161
winding-down after secular obligations have been fulfilled. At the ideological level, though, the Brahmanical solution achieved total dominance
within texts such as the Laws of Manu.
We also note that these terms connect up with a view of the brahmacārin
or celibate young man as a reservoir of sexual and magical potency. If we
can judge from the pervasive imagery of beautiful young women, male
and female couples, motherhood and vegetative fertility, the yaks.a religion
tended to celebrate adult sexuality, both male and female. By contrast, it
seems, the Vedic culture preferred to celebrate the unmarried male, a choice
that makes some sense in its earlier incarnation as a pastoralist culture with
a strong emphasis on conflict, raiding and fighting. If so, we can perhaps see
here the origins of the specifically Indic tension over sexuality. While postChristian Europe was concerned about whether sex was sinful, the Indic
problem was whether it represented a loss of power. I consider these issues
further in Chapter 8, which forms an interlude before Part 2 of the book.
In relation to asceticism, the term vrata, which has already received
some discussion in Chapter 6, is also of significance. I noted in Chapter 6
Timothy Lubin’s suggestion, developing ideas from Paul Hacker and Joel
Brereton, that the original meaning of vrata was something like ‘rule’ in the
sense of ‘fixed, characteristic mode of behaviour that manifest one’s will’
(Lubin 2001: 568). Thus one can speak of the vrata of a particular deity, as
defining that deity’s ‘nature and role in the world’ (2001: 568). Collectively,
these ‘pattern[s] of divine agency and authority’ can be seen as ‘defin[ing] a
cosmic and social order that calls for a human response’ (2001: 575). Lubin
suggests that
what the early Veda contributes to the later-attested ascetical rules is the basic
idea that observing a rule (vrata) of ritual service can put a human worshipper
in accordance with divine laws (vrata) and thereby confer divine blessings. The
Atharva Veda and the various brāhman.a texts combine this premise with the notion
that the observance of such rules requires sustained exertion (śrama) and fervid
dedication to fasting and celibacy (i.e. tapas). (Lubin 2001: 578–9)
Asceticism is of course very important in later Indic religion, and the idea
that through tapas or austerity one can compel the gods to behave in certain
ways is a basic trope that runs not only throughout Brahmanical asceticism
but also later Tantric practice. There is a striking similarity between Lubin’s
early Vedic version of vrata and the mature Śaiva or Buddhist Tantric idea
that by doing sādhana of a particular deity one, in effect, compels that
deity to act according to the deity’s own prior promise or will and so brings
about the desired result. An important additional element, however, is that
162
Origins of yoga and Tantra
in many forms of Tantric ritual one also identifies with the deity. We will
explore the development of this idea in later chapters.
Whatever the initial origins of the idea of asceticism as generating magical
power,2 this concept of the vrata as a pattern of behaviour that can compel
divine blessings would seem to be an important early contribution to it. We
have evidence from the Brāhman.as, from early Buddhist sūtras and other
sources of people who follow vrata modelled on the behaviour of various
animals, such as a ox-vrata or dog-vrata.3 Daniel Ingalls, commenting on
the Pāśupatas, an early group of ascetics who were worshippers of Śiva
as Pāśupati, the Lord of the Beasts, suggests that ‘[t]he aspirant hoped to
transform himself first into the Lord’s beast and finally into the Lord of
Beasts himself’ (Ingalls 1962: 295). We will see more of the Pāśupatas in
Chapter 10.
The Buddha is famously described as being dismissive of this kind of
thing, though the Buddhist account also accepts implicitly at least that such
vrata are meant as a serious spiritual path. In the Kukkuravatika Sutta, the
Buddha tells Pun.n.a and Seniya, an ascetic who is pursuing the ox-vrata
and a naked ascetic who has long pursued the dog-vrata, that their vows
will at best lead them to rebirth in the company of oxen and dogs, not
the company of gods. However, when Seniya, the naked dog-vrata ascetic,
asks to join the Buddhist order, he is offered full admission after the usual
probation,4 and soon attains the state of arhat (Majjhima Nikāya 67, e.g.
Ñān.amoli and Bodhi 1995: 493–7). As I noted in Chapter 6, it is not so
difficult to see the rules for order of the Buddhist saṅgha as themselves
constituting a kind of vrata intended to lead to a certain kind of result.
It is worth taking a brief look at the speculative and philosophical dimensions of the Vedic material, since these present another dimension of the
ideological movement which is taking place here. Some of the late R.gvedic
hymns, such as the well-known ‘creation hymns,’ 10.121 and 10.129, already
have a strong speculative element (e.g. O’Flaherty 1981: 25–9). In fact, the
term, ‘already’ is probably not appropriate, since it would seem to suggest that nobody ever speculated about the nature of reality in the tribalshamanic and pastoral context we are supposing for the early Indo-Aryan
2
3
4
Nick Allen has suggested that it goes back to shared Indo-European narrative structures (N. Allen
1998).
For references from Brahmanical sources, see Olivelle 1992: 109–11, which cites the Mahābhārata
(5.97.14) and the Bhāgavata Purān.a (5.5.32); Olivelle also gives further Buddhist references.
Seniya responds by suggesting that he should undertake a four-year probation rather than a fourmonth probation (Ñan.amoli and Bodhi 1995: 497). This might be a way to indicate the disfavour
with which these practices are held, but it seems more likely to be simply an expression of Seniya’s
level of commitment, which is clearly depicted as being higher than that of his ox-vrata companion.
The Brahmanical alternative
163
peoples. There is plenty of ethnographic evidence to demonstrate that
sophisticated philosophical thought can be found among preliterate pastoralist peoples.5 For such peoples to put their reflections into verse and
hand them on to succeeding generations is perhaps less frequent, and these
speculative hymns are generally regarded as a late addition to the R.gvedic
collection.
These hymns nevertheless suggest a move to the detailed exploration
of abstract philosophical notions. Such exploration is characteristic of the
next phase of the Vedic material and has been explored by a variety of
authors (e.g. Miller 1985; B. Smith 1989, 1994). The Vedas were used by
later Brahmanical tradition as a foundational text, and later developments
were constantly referred back to precedents in the Vedas, so there is a systemic tendency in Brahmanical scholarship to read the Vedic material as
more philosophical and less ritually-oriented than it perhaps was. For all
this, there is little doubt that the Kuru-Pañcāla theorists had a sophisticated set of understandings regarding the divine order and the system of
correspondences within experienced reality. Central to this was the multivalent concept of bráhman (B. Smith 1989: 69–72; Witzel 2003: 70–1).6
Witzel, following Thieme and others, suggests that the original Vedic sense
of bráhman is something like ‘the “formulation” or capturing in words of
a significant and non-self-evident truth’, and notes that ‘[t]he formulator
(brahmán) of such truths has special powers, effecting this world and the
cosmos’ (Witzel 2003: 70). Later the term develops a wider and more cosmological meaning, with bráhman as ‘the source and foundation of all that
exists – the nexus of all cosmic connections – and the connective force itself
lying behind all knowledge and action that constructs ontologically viable
forms’ (B. Smith 1989: 72), eventually personified as a creator-deity.
Brahmins clearly have a special relationship with bráhman, but since this
relationship is founded in the ability of the original Brahmin .r.sis to ‘capture’
reality in their inspired verses, the question of whether Brahmins in later
times were valid representatives of the original .r.sis could always be raised.
The Buddhists, as we saw, argued that the Brahmins employed as court
ritualists by the new rulers (the karma-kān.d.a Brahmins, in Kaelber’s terms)
had lost any valid connection with the original Brahmanical tradition. The
5
6
Evans-Pritchard’s Nuer Religion is a classic demonstration (Evans-Pritchard 1956). As for Indo-Iranian
and Indo-Aryan pastoralists, the similarity has long been noted between the Vedic concept of .rta
(divine order, etc.; Witzel prefers ‘active realization of truth’) and the Avestan (Iranian) concept of
as.a (e.g. Witzel 2003: 70). If this implies a common origin, then the concept would go back to the
pastoralist context.
I have generally omitted Vedic accents but in this case they resolve an important ambiguity.
164
Origins of yoga and Tantra
ascetic forest-dwelling Brahmins (the jñāna-kān.d.a Brahmins, in Kaelber’s
terms) were closer, but the true ‘Brahmins’, the true knowers of the ‘nexus
of all cosmic connections’, were the Buddhists themselves.
The jñāna-kān.d.a Brahmins would presumably have rejected such an
assertion, however close they might be in practice to the Buddhist and
Jaina monks. However, they accepted the idea of knowledge of bráhman
as liberating insight. This is the basis of the position taken in the early
Upanis.ads, the body of texts associated with this group.
The status of the karma-kān.d.a Brahmins was not based on their personal attainment to knowledge of bráhman, but on their possession of the
lineages of transmission of the Vedic hymns, which gave them the right to
recite these ritually-powerful verses. However, they too needed to respond
to the Buddhist and Jaina critique, and in practice, as Kaelber notes in the
passages discussed above, they sought to adopt some of the ascetic developments into their own practices. This led to the semi-ascetic style that became
characteristic of later Brahmanical culture. How far this was a simple need
to counter the rhetoric of the Buddhists and other śraman.a groups, and
how far it derived from a desire to incorporate what were seen as important
and valid elements of the ascetic approach, it is hard to say. I imagine that
both factors were operative. The specific intellectual resources of the Vedic
tradition, however, led to a particular construction of asceticism which,
as we have already noted, tended to focus on the ‘taming’ and directing of human sexuality, specifically adolescent male sexuality, for ascetic
ends.
Certainly, the transformations in Brahmanical ritual and practice during
this period were by no means purely defensive. They also laid the foundations for the future integration of the Brahmins within the wider context
of Indic society. The recent work of Timothy Lubin provides a useful analysis of this process (Lubin 2005). Lubin has suggested that in the Gr.hya
Sūtras (texts on household ritual) and elsewhere we can see a deliberate
simplification of the complex śrauta sacrifices to provide rituals that could
be performed by the householder under the instruction and direction of
Brahmin priests.
In the Gr.hyasūtras, the claim is first made that study of the Veda is not merely
available to but incumbent on ks.atriyas and vaiśyas as well as brahmins, with the
corollary that initiation and the daily use of Vedic mantras become the defining
mark of elite, Ārya status in a religiously and ethnically diverse society. The trend
toward identifying initiation and brahmacarya (rather than marriage) as the starting
point for constructing a framework for an orthoprax life of piety, and the multiplication of similar vratas as a framework for personal piety,were developments
The Brahmanical alternative
165
parallel with the rise of ascetical (Śraman.a) movements such as Buddhism. (Lubin
2005: 88)
Elsewhere in the same article he suggests that the effect of this development
may have been to ‘divide the territory’ between the Brahmins and the
śraman.a movements. Thus Lubin suggests that
Buddhism initially won out in the urban zones, where traditional social and cultural
structures were fragmentary and diluted. Meanwhile, Brahmanism reinvented itself
in a form that simultaneously provided a model of domestic piety and personal
sanctification (in a range of degrees of ascetic rigor) that had, especially in rural
society, much of the appeal that Buddhist piety had in the cities and along trade
routes: namely, a code of self-discipline and direct personal access to the presumed
power of sacred mantras. (Lubin 2005: 91–2)
The success of the Brahminical strategy can be seen in the increase in the
use of Sanskrit and the increasing evidence of royal patronage for Brahmins in the post-Mauryan period. Here an important role was played by
foreign rulers (Scythians and Kus.ānas) for whom the use of Sanskrit and
the patronage of Brahmins helped to legitimise their position. The pattern
was then adopted and further popularised by the imperial Guptas (Lubin
2005: 92–7).
local deities and everyd ay concerns:
the brahmanical pat tern
Lubin points to the way in which this model meant that the Brahmins
remained relatively decentralised, with a firm base at the village level. They
never developed the degree of centralisation in large monasteries in or
near cities that was characteristic of the Buddhist and Jaina pattern, and
which was eventually to lead to the vulnerability of those traditions to loss
of state support. Brahmins became court ritualists, often associated with
spectacular and wealthy temples, and they became royal administrators
and bureaucrats, but they also built up a solid base of family land-holdings
within South Asian villages that enabled them to survive the loss of power
and support at the centre during the long periods of Muslim and British
domination. Above all, this was a question of property gifted to specific
Brahmin families and handed down to their descendants.
The downside, if that is the correct term, to the semi-ascetic Brahmin
model was the linkage between Brahmins and purity that became the basic
ideological underpinning of the Hindu caste system (L. Dumont 1972).
Here I am speaking not so much of the cost to the Brahmins themselves
166
Origins of yoga and Tantra
as to the wider society. If Brahmins responded to the śraman.a challenge
by asserting their own purity through a semi-ascetic lifestyle, the necessary
corollary was that others must be less pure, and that there had to be an
opposite ideological pole to the pure Brahmin. As Dumont put it, ‘The
whole is founded on the necessary and hierarchical coexistence of the two
opposites’ (L. Dumont 1972: 81). This other pole was formed by the various
dalit (‘untouchable’) groups, the low-caste groups who were associated with
occupations seen as intrinsically impure.
Thus the caste structure of Indian village society, from the Brahmin point
of view at any rate, became based on a hierarchical opposition between the
pure Brahmin and the impure untouchable. Texts such as the R.gveda’s wellknown ‘Purus.a-Sūkta’ (Hymn 10.90, see e.g. O’Flaherty 1981: 29–32), in
which the four varn.a originate from the dismemberment of the primaeval
cosmic giant, with the Brahmin arising from his mouth, the ks.atriya from
his arms, the vaiśya from his thighs and the śūdra from his feet, are frequently
referred to in this connection.
As I described in Chapter 4, however, the caste hierarchy as it developed in
later times was in fact a very different business from the varn.a classificatory
scheme, and could encompass a large number of separate hereditary groups
( jāti), with varying claims to one or another varn.a status, as well as others
which were seen as outside the varn.a system altogether (cf. L. Dumont
1972; Quigley 1995; Searle-Chatterjee and Sharma 1994; Bayly 2001). Here
the South Indian (and perhaps originally more widespread) conception
of the necessity of low-caste polluted groups to handle the dangerous
magical power of the king (see Chapter 4) may have made an important
contribution.
Purity here seems originally to have been a question of the ritual purity
necessary for the proper performance of Vedic sacrifices, a theme which
doubtless goes back to the Indo-Aryan-speaking immigrant population,
since it is also a key issue for the Zoroastrian tradition. In the new Brahmanical social and religious order, however, it became a more generalised
and intrinsic question, with the lower groups in the caste hierarchy defined
as intrinsically impure, and women defined by their bodily functions as
periodically impure. In both cases, the human consequences were to be
severe, and are still very much active today.
In the 1980s, a number of authors developed a critique of Dumont’s
analysis through pointing to, in effect, the existence of a religion of fertility and everyday life as a counterpoint to Brahmanical religion in the
modern Hindu context. This is the religious complex that has generally
been referred to by anthropologists as ‘auspiciousness’. The writers who
raised the question of auspiciousness, such as Triloki Nath Madan (1985,
The Brahmanical alternative
167
1987), Frédérique Marglin (1982, 1985a, 1985b) and Veena Das (1987), were
involved in a sympathetic critique and extension of Dumont’s work rather
than a radical rejection. These writers agreed with Dumont in taking the
question of values in South Asian societies seriously, but they felt that
Dumont’s account was over-simple and inadequate. Purity was important
in Indian society, but it was not the only central value. So ‘auspiciousness’
was seen as being an independent value, as delimiting a separate sphere, or
(at least with Veena Das and T. N. Madan) as defining a second axis around
which the structural analysis of Indian society might be based. ‘Auspiciousness’ became a general label for concerns with attaining health, success and
prosperity and avoiding death, disease, poverty and misfortune.
There seems to be a close relationship between auspiciousness and gender.
Women’s ritual is largely about generating auspiciousness.7 The converse
is also true: much ritual concerned with auspiciousness is in fact ritual performed by women, as with the auspicious diagrams (kolam, etc.) drawn on
or in front of the threshold of houses by women each morning in parts
of India (Nagarajan 1997), or with the vrat rituals performed by Hindu
women for the welfare of their husbands and children (S. Robinson 1985;
S. Gupta 1999). In addition, the deities invoked (for example, S.as.t.hı̄ or
Laks.mı̄) are frequently female, and women act as key signifiers of auspiciousness and inauspiciousness (as in the concept of a married woman as
sumaṅgalı̄ or intrinsically auspicious, or of a widow as inauspicious). This
new focus on auspiciousness therefore provided a context in which the role
of women in ritual and religious life received considerably more attention
from anthropologists than had previously been the case.
A convenient starting point for a discussion of this literature on auspiciousness might be Veena Das’s diagrammatic representation (V. Das 1987:
143; also reproduced in Marglin 1985a: 294), in which the idea of the two
axes of purity-pollution and auspiciousness appears particularly clearly.
Das’s diagram was modelled on the diagram from Tambiah’s 1970 book
that I reproduced in Chapter 6 (Fig. 6.4). Das’s diagram, like Tambiah’s,
reflects the tendency for abstract formulism of the anthropology of the time,
but, again like Tambiah’s, it presents a useful perspective on village values.8
However, the similarities between the two diagrams are more apparent than
real.
7
8
See, for example, the strı̄ ācār associated with marriage in rural West Bengal and Bangladesh (Fruzzetti
1990). See also Samuel 1997. A related issue here is whether one interprets such rituals in terms of
what they mean to their performers or to male ritual experts within the society. See Brown on the
meanings of Mithilā art (C. Brown 1996).
It should also be noted that Tambiah’s diagram is intended to represent the constellation of values
within a particular Thai (more accurately Lao or Isan) village, whereas Das’s is a generalisation across
Indian society as a whole.
Origins of yoga and Tantra
168
Sas
ht
ben i Mo
ev the
ole
sp
r ie s t
Bar
ber
,W
rth
ta
Pre ta
B hu
n
r
sto n
ce atio
An piti
o
Pr
ns
ma
ah
br
ath
Ma
ha
devis
De
Mot
h
male er G
vo od
l
e
nt des
as
Ba
r
b
er,
Do
Impure,
liminal
states,
associated
with the
disarticulated
states of
the body
n
si
se ects
p
Pure
Br
ah
bi
M
la
Ka
n
ild
an
ria
rm
ge
ar
Ch
rma
a
an
he
as
W m
m
mi
k sh
i
se s
d es
od ects
r G asp
nt
hn
La
u-
he
as
Pure,
bounded
states,
associated
with the
articulated
states of
the body
Shiva
-Pa
rva
ti,
Vi
s
Auspicious activities associated
with life and the right hand
Ance
sto
r
Inauspicious activities associated
with death and the left hand
Figure 7.1. Veena Das’s Diagram of the Value System of Indian Society
There are a number of things which might be noted about this diagram,
but I will limit myself here to one major point: unlike those in the Buddhist
diagram, the two sets of values are closely entangled with each other. In fact,
an ongoing problem with ‘auspiciousness’ is that it does not correspond to
a clear indigenous category; Das, Marglin (1985a) and Madan (1985, 1987)
refer to a variety of indigenous terms (notably śubha and mangala) and, as
Parry (1991) eventually noted, it does not seem possible to derive a single
coherent category of non-Brahmanical concerns here. There is virtually
no village ritual that cannot ideally be done according to Brahmanical
prescriptions, which is not to say that all roles will necessarily be taken
by Brahmins, since some of them may involve inappropriate involvement
with impurity. Brahmanical values of purity pervade the whole field and,
in effect, claim dominance over the world of everyday life.
This is most notable in relation to women’s bodily functions, which
have a close linkage in Brahmanical thought to the whole question of
fertility in nature more generally. In Chapter 4, I discussed Julia Leslie’s
analysis of a widespread myth of menstruation, which sees its origin in a
The Brahmanical alternative
169
process by which the god Indra transfers his guilt at an unavoidable act of
Brahmin-killing to women, to nature and to the earth. Women’s fertility,
and the earth’s fertility, are somehow intrinsically linked to their polluted
and devalued nature.
Of course, there is a kind of built-in irony or contradiction here, since
the birth of children is necessary for the existence of society, and the birth
of male children at least is a highly desired event. The relationship between
auspiciousness and purity is complex and often contradictory. Groups that
are impure may well also be auspicious, as with the hijra (eunuchs) who
have an important role in much of India as bearers of auspiciousness in the
context of marriage and the birth of male children (e.g. Nanda 1990: 1–6).
It is, however, perhaps not surprising that the contradictions are most
conspicuous in relation to women and their role in village society (Samuel
1997). Brahmanical thought, as mentioned above, centres around the ideal
of the celibate young male. Women’s bodily functions, particularly menstruation and childbirth, are seen as polluting and problematic within this
set of values. The ‘auspiciousness’ principle, by contrast, tends to focus
around questions of prosperity, good health, and human and animal fertility. The imagery of the goddess Laks.mı̄, with its close affiliations to the
non-Brahmanical yaks.a religion, is a key representation of ‘auspiciousness’,
and a young fertile married woman is the most immediate human representative of Laks.mı̄. Yet, for the Brahmanical scheme of things, childbirth
is seen as the most polluting of bodily processes, and traditional childbirth
attendants generally belong to the lowest of social categories and are viewed
with little or no respect.
In Chapter 4, I alluded briefly to the problematic consequences of this
entangled relationship between what is desired and what is pure for childbirth in contemporary South Asia (see also Rozario and Samuel 2002a).
Here it is perhaps enough to note that this mutual entanglement creates a
very different relationship to the everyday religion of the village than that
found in the Buddhist context in Southeast Asia (see Chapter 6), or than
that which seems to have been characteristic of Jaina and Buddhist contexts
in South Asia in the past.9
Women clearly have a ‘lesser’ status in some respects in Southeast
Asian Buddhist societies as well, most conspicuously in relation to their
9
To a considerable degree, modern Jains have preserved this relationship in modern India, although
there has been some degree of assimilation to Brahmanical values. Jain communities include hereditary
Brahmins who carry out Brahmanical ritual functions for the community, but they are not regarded
as ideologically dominant in the way that Brahmins in Hindu communities are. Cf. Carrithers and
Humphrey 1991; Cort 2001.
170
Origins of yoga and Tantra
participation in monastic Buddhism (cf. Keyes 1984, 1986; Kirsch 1985).
This negative valuation of women as Buddhist practitioners does not seem,
however, to imply the same devaluation and loss of autonomy characteristic
of women’s situation in many modern Hindu communities. Some authors
(e.g. Reid 1988) have seen this greater autonomy of women in Southeast
Asian societies as a specifically Southeast Asian characteristic, and certainly
it was true in the pre-modern period for Muslims in Malaya and Indonesia,
and Christians in the Philippines, and tribal populations in many areas, as
much as for Buddhists. My point here though is not to construct causal
relationships, and it would be simplistic, to say the least, to claim that
‘Hinduism’ led to less freedom for women, and ‘Buddhism’ to more. The
situation in the pre-modern period was the result of a complex series of
historical developments in which the role of developing religious traditions
to other aspects of society is best seen in terms of what Weber would have
called ‘elective affinities’ (Wahlverwandschaften).
Thus one might suggest that, historically, restrictive forms of Hinduism
and Islam made sense in relation to the specific forms of patrilineal kinship
and marriage found in most South Asian village societies, particularly as the
process of ‘sāmantisation’ in the course of the first millennium CE led to an
accentuation of the patriarchal features of these societies. They made less
sense in Southeast Asia, where matrilineal and bilateral kinship practices
are more common. To the extent that one or another religious tradition
was adopted, the inner logic of each tradition allowed for and encouraged a
further intensification of the particular pattern that became associated with
it. Thus, for example, restriction of women (parda, etc.) became a sign of
high status in Hindu societies throughout much of North India, including
present-day Bangladesh and Pakistan, and was adopted where possible by
groups that wished to claim higher status.10 All this, however, is only part
of the story, and clearly there was also a great deal of historical contingency
at work.
The (pre-Tantric) Buddhist and Jaina construction of things saw everyday life, including its ritual aspects, as a separate (laukika, worldly) sphere
which co-existed with the teachings and practices of Buddhist and Jaina
monastics but was not properly their concern. As in countries such as Thailand and Burma in recent times, there would presumably have been a range
of priests and minor ritual practitioners at the village level, as well as rituals
performed by householders and lay ritual specialists, but none of these had
10
South India, which has a quite different (‘Dravidian’) system of kinship and marriage, had a somewhat
different history, and women’s status today is in some respects significantly better in these regions
(Samuel 2002b: 9–11).
The Brahmanical alternative
171
the kind of ideological dominance claimed by the Brahmins in the Brahmanical model. Thus weddings in Buddhist Southeast Asia, Sri Lanka or
Tibet were not Buddhist ceremonies, although in some cases monks might
be invited to generate auspiciousness and merit (good karma) in association
with them through recitation of Budhist sūtras or other auspicious acts.
By contrast, the Brahmins as village-level ritual practitioners saw their
role as setting the authoritative mode of performance of all of the rituals of everyday life (the sam
. skāras, see e.g. Pandey 1969). In reality, this
‘Sanskritisation’ (perhaps one should rather say ‘Brahminisation’) of village
life is far from complete, even today, with the performance of the sam
. skāras
by Brahmin priests often confined to the wealthier (and so usually higher)
castes who are in a position to employ them, but the ideological dominance
of the Brahmanical perspective in the wider society is well-established, with
village rituals often seen as faulty and incorrect due to a lack of knowledge
of the proper Brahmanical way to do things. This is however part of an
ongoing dialectic between lay people and religious experts, village level and
urban centre which would seem to go back to the early period we are here
discussing (e.g. Marriott 1969b).
If we were to look for a Brahmanical equivalent to Chiang Mai, the
Thai city which I cited as an example of the Buddhist relationship between
power and the state, we might select Madurai in Tamil Nadu, a city of
similar size and like Chiang Mai the capital of a medium-sized kingdom
for hundreds of years, initially during the Pān.d.ya period (seventh to thirteenth centuries) and more recently under the Nayaka kings (sixteenth to
eighteenth centuries). Little survives from the Pān.d.ya period; the city in its
present form dates mostly from the seventeenth century, and is centred not,
like Chiang Mai, on a modest city pillar shrine where Indra and the city
guardian spirits are worshipped, but on the famous and splendid temple of
the goddess Mināks.ı̄ (Hudson 1993; Harman 1989; Mitchell 1989: 449–52;
Volwahsen 1969: 80–6, 88).
Mināks.ı̄ was almost certainly a local deity in origin, but is now regarded
as a transform of Pārvatı̄ and married to Śiva, who under his local name
of Sundareśvara has a shrine next to hers at the centre of the temple.11
She is also regarded as sister of Vis.n.u, and the ritual cycle of the temple
11
Historically, it seems that early Brahmanical temples in South India were mostly dedicated to Śiva
or other male deities. It was only from the thirteenth century onwards that shrines to the local
goddesses, generally treated as consorts of a male Brahmanical deity, were incorporated within them
(Stein 1994: 237, 465). By the time of construction of the Madurai temples as we know it today, this
development had taken place. The temple today contains dual shrines to both Mināks.ı̄ and Śiva,
but it is evident that Mināks.ı̄ is the dominant partner.
172
Origins of yoga and Tantra
focuses on her the three deities of Mināks.ı̄, Śiva and Vis.n.u (Harman 1992:
64–99). What might originally have been a shrine of a local deity has been
transformed into a temple integrally tied into all-India Brahmanical cults
and focusing on Śiva, here functioning as a deity of kingship, and on his
marriage to the local goddess.
The nearby hill-shrine of Tirupparaṅkunram, once a Jain sanctuary,
and now one of the six famous shrines of the Tamil deity Murugan, has
likewise been integrated into the Brahmanical structure through Murugan’s
assimilation to Skanda, son of Śiva; the mythology of the site now centres
around Murugan-Skanda’s marriage to Devasenā (Clothey 1978: 124–7;
cf. also Zvelebil 1991). The deities from this shrine, along with Mināks.ı̄’s
brother Vis.n.u and his entourage from the important local Vais.n.ava temple
of Alakarkovil are also invited to the annual wedding ceremony of Mināks.ı̄
at Madurai (Harman 1989: 98).
Thus we see contrasting orientations towards the local deities and the
religions of everyday life in the Brahmanical and Buddhist contexts; Brahmanical incorporation and mutual entanglement in the one case, parallel
but distinct spheres in the other. Some further discussion of some of these
issues is provided in the following Interlude (Chapter 8), which brings the
first major part of the book to an end. The second major part commences
with the subsequent chapter (Chapter 9), and will focus on the development of the various religious practices that are generally now labelled as
‘Tantra’.
chap t e r 8
Interlude: Asceticism and celibacy in
Indic religions
Chapters 3 to 7 have presented a survey of the evidence relating to the
first stages of the history of Indic religions from the Second Urbanisation
onwards. Key issues here have included the growth of the ascetic orders
and of Brahmanical asceticism and the development of a series of new
techniques of mind-body cultivation in the context of these ascetic groups.
We have also seen something of the wider context of these ascetic traditions
in the religious life of North Indian communities (particularly in relation to
the so-called yaks.a religion) and looked forward to a glimpse of the ‘mature’
relationship between Buddhism or Jainism and village religion on the one
hand, and Brahmanical religion on the other. In this Interlude, I step back
a little to look at some further issues relating to these developments.
I start by asking why the ascetic orders and the new goal of liberation
from rebirth developed at the time when it did. Who and what were the
Buddha, Pārśva, Mahāvı̄ra, Makkhali Gosāla and the other spiritual leaders
of the śraman.a tradition, to the extent that we can take them as historical
figures?
Each generation notoriously remakes major historical figures such as the
Buddha or Mahāvı̄ra in its own image and we can hardly expect our own
time to be exempt from such a rule. However, looking at the evidence presented in the previous two chapters, it would seem that the key component
of the message of these teachers, at least as understood by its early transmitters, was not primarily a rational theory of the nature of the universe. It was
rather the ideal of the renunciation of ordinary life and of the emotions,
feelings, impulses that tied one to it, a renunciation that also included a
rejection of family, descendants and rebirth. At the centre was a conversion experience brought about through contact with a charismatic leader,
and an ideal of a heroic struggle against the emotional entanglements and
deep-seated volitional impulses of ordinary life, aimed at the achievement
of a state of power, self-control and equanimity which is contrasted with
the suffering and confusion of everyday life.
173
174
Origins of yoga and Tantra
The relationship to a specifically renunciate lifestyle is a complex one. All
these teachers had lay followers. If Masefield (1986) is right, lay followers
were as capable of the conversion experience as renunciates, and they might
or might not subsequently proceed to bhiks.u or wandering mendicant
status. The key element was a reversal of everyday life such that the goal
of liberation from rebirth came to be the centre around which their lives
was now constructed. Following this reversal, the emotional entanglements
of everyday life came to be seen as traps and obstacles to the achievement
of that goal. Each tradition had its own specific set of techniques for how
one then proceeded to actualise this vision, with the Jains and Ājı̄vikas,
along with the Brahmanical renunciates who developed at around the same
time, emphasising ascetic practice, with simple contemplative exercises (the
brahmavihāras, perhaps some simple dhyāna practices), while the Buddhists
emphasised the development of the full range of dhyāna practices as a
technique for calming and purifying the mind.
The relationship between practice and insight (prajñā, vipaśyanā) was
an ongoing issue within these traditions, with some later Buddhist meditation traditions developing a variety of dhyāna or samādhi practices
and others treating dhyāna essentially as a prelude to insight, though
arguably the reduction of dhyāna to a brief preliminary found in modern
Theravādin meditation is a recent development (Cousins 1996b). In the
modern Southeast Asian context, the dhyāna (samatha) practices are closely
linked to the ‘pragmatic’ exercise of magical power (e.g. Houtman 1996,
1999). It seems likely that such magical practices were always part of the
stock in trade of those śraman.as who were dependent on individual lay
support, though they might become less important in contexts where large
monasteries were supported by the state or by urban or rural communities.
The development of monasticism, in the sense of substantial communities of renunciates living together, seems to have taken place relatively
quickly, and they were certainly well established by the time when the
sūtras and vinaya texts were written down in the second and first centuries
BCE. The state support for Buddhism and the śraman.a traditions at the
time of Aśoka would doubtless have provided the necessary conditions,
even if this did not exist already.
e xpl ain ing the grow th of the new ascetic traditions
What, however, were the social and cultural traditions that led to these
developments? Strictly speaking, we do not have any evidence that the
goal of liberation from rebirth itself was new, only that it is not present
in the Vedic material. It might have had an earlier presence outside the
Asceticism and celibacy in Indic religions
175
Vedic region, and the tradition of wisdom kings of Mithilā and other early
renunciate rulers might point in this direction. It is hard, though, to think
of it as developing in a small-scale village society. The idea of liberation
from rebirth presupposes that one can see the structure of society ‘from
outside’, in the form of the cycle of rebirth. One cannot rule out that such
an idea and goal might occur to isolated individuals in a small-scale village
society, but it seems unlikely that it would come to be shared by enough
individuals to constitute a movement.1
Yet this is implied by the coming into being of organisations such as the
Buddhist and Jain orders. We have here groups of people who felt it appropriate and meaningful to get together so as to pursue this goal as a group,
along with further larger communities of people for whom this project
somehow meant enough that they were willing to provide it with material
support. This could surely only happen in a context where people’s experience led them to start to see their own particular cultural form of human
life in society from the outside. That in turn probably implies that they had
several such examples of different forms of human society to think with.
In other words, the renunciate orders are unlikely to have developed until
there were urban centres and trade between people from different localities.
In these circumstances, an understanding could come into being that the
way things were in one’s own village or community was not simply the
way things always had been and must be, but one of a number of possible
options, and not necessarily, therefore, the best of those options.
That all of this is tied up with the new urban civilisation growing up at
this time in the Central Gangetic region in North India has been widely
assumed. Indeed, it seems to be almost inescapable, though it is less clear
how long the cities had been in existence when the new ascetic orders got
going or precisely how one might explain the relationship between the new
movements and the urban milieu. On current datings, Greg Bailey and Ian
Mabbett are probably correct in arguing in their recent Sociology of Early
Buddhism that the rise of the ascetic orders followed the initial growth of
cities: ‘early Buddhism developed as a consequence of a changed situation,
rather than of a rapidly changing one’ (Bailey and Mabbett 2003: 260).2
1
2
I suggested in Samuel 1990 that such an experience may be a necessary part of how shamanic
practitioners operate, although the viewing from outside here may also take place, and more typically
does take place, in a mythological language in which social life is constituted through the action of
the spirits. In the case of the early ascetic groups, however, we need to assume a context in which
such an understanding might come to be widely shared.
Their account is not directly concerned with Jainism, but the same comment would apply in the Jaina
case. The ‘proto-śraman.a’ movements discussed in Chapter 6, if they had historical reality, might
have been closer to the time when cities began to develop, and might indeed have been involved with
their initial growth (see below).
176
Origins of yoga and Tantra
The connection is again best seen in terms of what Weber would have
called an ‘elective affinity’ rather than a process of causation. Something
about the new form of life which was emerging meant that the Buddhist
and Jain communities made sense, to all the parties concerned, whether
those who became fully-committed ascetic members of the new orders,
those who were lay followers and providers of food and material support,
or the rulers and elite groups who minimally had to be prepared to tolerate
this new development and who, if the Buddhist and Jain accounts are
reliable, in fact often became active supporters of the new teachings. I
think it is partly this multiplicity of actors, and the fact that we are looking
at an ‘elective affinity’ rather than a causal relationship, that leads to the
multiplicity and (real or apparent) inconsistency of explanations noted by
Bailey and Mabbett’s (Bailey and Mabbett 2003: 13–36).3
It is clear in any case that the new communities were from the beginning engaged with patrons and supporters in the cities, and were typically
located close to them though not within them, in locations that were easily accessible to their populations. At this stage, I return to the question I
posed at the start of this chapter. Why did the ascetic movements and the
institution of celibate monasticism develop in India at this time?
In anthropological terms, what was happening could best be described
as a revitalisation movement, a restructuring of modes of thinking and
feeling, often through visionary processes, such as we see in many times of
abrupt and radical change in human history and society. There was certainly enough change going on at this time, with the growth of new urban
communities and a new lifestyle associated with them, and, presumably, the
move to these new centres of a population which had previously lived in a
much smaller-scale and culturally-enclosed context. However, demonstrating a closer relationship between these changes and the new movements is
complex.
It has of course long been observed that the new traditions had an ethical
component. From the point of view of practice as a renunciate, the ethical
component is in a sense secondary, since it is primarily an aid to reducing
one’s involvement in the affairs of the everyday world. However, from the
point of view of the lay life, the ethical version is more central. Thus it is
clearly quite possible to see a need for ethical guidelines being met by the
3
Attempts to find linear causal relationships are in fact not particularly useful in social analysis, since
even when there is an event which has a massive impact on a community (war, famine, earthquake),
the response is socially and culturally mediated, and may take any of a variety of forms. For similar
reasons, the classical Popperian vocabulary of falsification has limited applicability to the social
sciences, including their historical applications.
Asceticism and celibacy in Indic religions
177
new religions, in a situation where choices have become much wider and
indeed the idea of the individual as having choices to make has become
much more salient. One can also see the new traditions as providing rational
arguments to legitimise and explain the gradual abandonment of ritual
practices that made sense in the village context but had less meaning in the
city. Meanwhile, the Brahmins were reformulating their own practices for
performance by householders under Brahmanical guidance, to some extent
in competition with this new development (cf. Lubin 2005, and Chapter 7).
Doubtless this is one reason why the new śraman.a movements were more
comfortable with the semi-ascetic wilderness-dwelling jat.ila Brahmins, who
were potential allies, than with the new class of urban Brahmans who were
their competitors for support and sponsorship.
One can also see ways in which the rulers of these new states could find the
śraman.a traditions, with their universal ethical principles and also perhaps
their ks.atriya sympathies, a useful ideology for their new states. As Olivelle
has noted, the early Upanis.ads also suggest a need to accommodate to the
needs of kings (1998: xxxv), and if the Central Gangetic region already had,
as I have suggested, a tradition of wisdom-kings with renunciate tendencies,
there would be a natural synergy between the two developments, even if
there were no direct influence from the wisdom-king traditions upon the
evolving śraman.a movement itself.4
The Buddhist tradition’s classic emphasis on dukkha, on the suffering or
unsatisfactory nature of everyday life, and on meditation as an important
aspect of how one deals with it, is also worth taking seriously as a clue to what
people might have been looking for when they turned to the Buddhists and
the other new movements. Suffering in the old village context – the ‘ancient
matrix’, as Stan Mumford once called it, quoting Bakhtin (Mumford 1989:
16ff, cf. Samuel 1993: 6) – was at least part of an ongoing cycle of life in
which the individual was deeply connected with the wider community. The
city was a context in which this connectedness was no longer there. The
urban individual had to deal with a world in which he or she had to make
his or her own way, with little guidance from the past (cf. Gombrich 1988:
72–81). The early Buddhists and the other groupings within the śraman.a
movement provided a guide to the path, techniques to calm and understand
the suffering of the newly self-conscious individual, and a group of likeminded people with whom one could construct a new kind of kinship and
community based upon a common goal.
4
In fact Olivelle does not take the wisdom-king traditions seriously, and implies that the Brahmins
presented their new doctrines as being taught by kings as a deliberate device to encourage acceptance
by their royal patrons. This may be a little too cynical (Olivelle 1998: xxxv).
178
Origins of yoga and Tantra
Was there actually more dukkha around in the new urban environment?
There have been various attempts to link the growth of the ascetic orders to
the negative impact of social change and the new urban way of life. There
is probably at least some truth in these suggestions. Richard Gombrich, for
example, suggests that the new cities provided an environment in which
disease would flourish, so that illness and death were widespread. ‘It is
quite possible that in the Buddha’s environment disease and sudden death
had actually become more frequent. Maybe it is no accident that the early
Buddhists were fond of medical metaphors, describing the Buddha as the
great physician, etc.’ (Gombrich 1988: 57).
There is perhaps a further issue here, which relates to my earlier point
about viewing society from the ‘outside’. The new social environment had
the potential to create a new kind of self-awareness, particularly among
people such as merchants, entrepreneurs, government officials and other
people who were living in an unpredictable and high risk environment
in which individual choices were difficult to make and might have a dramatic impact on their future prospects. The village, by contrast, was an
environment in which people’s lives were more or less pre-given; for most
people, their life-prospects, illness or accident aside, would be much the
same as their father’s or mother’s. Harvests could vary from year to year
with the weather, floods and other natural disasters were always a possibility, but farmers tend to have tried and true techniques for dealing with
them. In such a situation, there is only a limited scope needed for individual
judgement.
All this would have been very much less true in the new urban environment. There would be many examples of different life-choices around, and
for some people at least a constant need to make one’s own choices about
the course of one’s own life. At the same time, individuals were still part of
a complex web of obligations and commitments to spouses, children and
other relatives, with little certainty as to whether they could meet them. It
is not unreasonable to see this as a situation in which people might start
to question the value and meaningfulness of any of the choices they are
making within the ‘round of rebirth’.
If life was risky and unpredictable, with outcomes that no longer seemed
clearly linked to one’s immediate actions, one can see some of the attraction
of the choice of withdrawal from society. To become a renunciate was
to accept one’s own death, in both the physical and (more immediately)
social sense. The renunciate’s previous social personality and role has been
abandoned and has gone for ever. He or she was, in theory at least, no
Asceticism and celibacy in Indic religions
179
longer indebted to anyone in lay society, and no longer obliged to undertake
anything. The renunciate was free, except for the limited rules structuring
the life of the monastic community, and the even more circumscribed rules
regarding support from the laity.
In reality, of course, Buddhist or Jain monks or nuns, or Brahmanical
sam
. nyāsins, were not necessarily totally free from the entanglements of their
previous social connections, but the attraction of the ideal state is perhaps
understandable, especially now that it had been linked to the new goal of
liberation from rebirth and offered the company of a community oriented
towards the same goal.
Why though did the new ascetic orders appeal to the individuals who
became lay followers and supporters? Much of the argument about the
new kind of awareness associated with the urban lifestyle might apply to
lay followers as well, particularly for the relatively prosperous, wealthy and
powerful individuals who made up much of the following of the new orders.
One might also consider some of the practical aspects of involvement with
the Buddhist order. The connection between Buddhism and trade has
long been noticed. As Bailey and Mabbett note, ‘It is possible that trade
was mentioned ubiquitously in Buddhist literature not just because it was
conspicuous in the society reflected in the texts, but because the actual
development and expansion of Buddhism was so closely connected with it.
Buddhist monks often travelled with caravans of merchants [. . .]. It is likely
the extension of Buddhist culture into the Deccan was closely associated
with trade’ (2003: 62; see also Heitzman 1984; O’Connor 1989; Lewis 2000:
49–88).
Weber’s argument for the ‘elective affinity’ between the mercantile culture of north-western Europe in the later Middle Ages and the rise of
Protestantism cannot be transferred as a whole to the different situation of
fifth and fourth century BCE North India, but one wonders if some of the
same elements might hold. Certainly the restrained, moderate lifestyle of
the lay Buddhist would go well with the trading ethos of that time as of
the Netherlands, Germany or Switzerland in the sixteenth and seventeenth
centuries (cf. Gombrich 1988: 76–81).
One can also point to the practical utility for people engaged in longdistance trade, or in travel to distant places on administrative or other
missions, of having a community of people with whom they could claim
fellowship, whom they had some reason to trust, and from whom they could
expect support in distant cities and locations. Trading communities such as
the Buddhist Newars of the Kathmandu Valley are perhaps a good modern
180
Origins of yoga and Tantra
example of this kind of situation, with Newar communities scattered as far
as Lhasa, Kalimpong and Kashmir (Lewis 2000: 53–4). So of course are the
modern Jains. There are parallels here with the mystery cults of the classical
and Hellenistic world (Eleusis, Isis, etc.), with the Sufi brotherhoods that
provided fellowship and a home away from home for Islamic travellers
throughout the long trade routes across Europe, Asia and North Africa,
or with the cult groups (Ekpe, Poro, Sande etc.) that existed into modern
times through the trading states of West Africa, providing linkages between
people over vast areas which were politically various and often, for someone
without local connections, quite dangerous. As I suggested in Chapter 6,
the early proto-śraman.a communities, such as the followers of Pārśva, if
they had historical reality, might be seen as local equivalents of such cult
groups, and as providing a similar mixture of spiritual insight and practical
assistance.
In asking why the new ascetic orders appealed to rulers of the time, it
is of course true that rulers were also human beings dealing with a new
and challenging situation. Some of the arguments about new forms of
self-awareness might apply as strongly to them as others, especially given
the risks and unpredictability of being a ruler of a North Indian state at
that time. But kings also had practical concerns and one can see ways in
which Buddhism and Jainism might be of use to them. Here the ethical
component of the new traditions has already been mentioned. Bailey and
Mabbett have also advanced an interesting argument about the role of the
śraman.as as intermediaries between the expanding centres of urban power
and the various populations they encountered:
[Ś]raman.a teachers were not just rustic medicine men from the wilderness. They
were active everywhere. They could therefore be co-opted to stand for the solidarity
of the kingdom, a solidarity that was cemented by a new message that insisted upon
the universality of values, and subverted the privileged authority of Vedic rituals
and myths which were controlled by a special group. This sort of message was
just what rulers needed when they were trying to bring beneath their dominion
communities too diverse in culture and origins to be accommodated within a
ready-made Sanskrit-brāhmanical image. (Bailey and Mabbett 2003: 175)
I assume that some such arguments give us a reasonable grasp on how
and why the ascetic traditions developed at the time that they did. In the
remainder of this interlude I move to consider the wider question of the
two different kinds of accommodation we have seen between religion and
society, the Buddhist and/or Jaina pattern and the Brahmanical pattern. I
begin with the question of celibacy.
Asceticism and celibacy in Indic religions
181
the logic of male celibacy
I have already pointed to the central relationship in Brahmanical thought
between celibacy, religion and male identity.5 The logic behind this centrality is quite explicit in modern Hindu contexts, and undoubtedly goes
back a very long way; we can see it in the work by Kaelber on Vedic texts
which I discussed in Chapter 7 (Kaelber 1989). The general Indic notions
regarding seminal continence and loss are well known in both textual and
popular sources, and have been highlighted in recent years by a number of
authors, perhaps most notably by Joseph Alter in his studies of contemporary North Indian wrestlers (Alter 1997; see also Khandelwal 2001). These
notions imply that personal power and authority arise from abstention
from and/or control over male ejaculation.
This theme links up closely to that of the renouncer (e.g. L. Dumont
1960). Indic religions, including Buddhists and Jainas as well as modern
Brahmanical thinkers, have seen and still today see the ascetic as someone
who renounces and rejects, above all, the claims and appeal of family life
and of sexuality, and so of the society that is built upon the foundation
of the life of the householder. Here the renouncer is implicitly figured as
male, while women routinely serve as representations of the family life to
be renounced.
Thus spiritual power, which is expressible through various this-worldly
results, is derived from asceticism, while the normative mode of asceticism
is carried out by males, and involves the rejection of, or at least conscious
control over, sexuality. Correlatively, as we learn from the Puran.as and other
sources of Hindu legend, if an ascetic was becoming so powerful that the
gods felt threatened and the order of nature was being interfered with, the
standard solution was to send along an apsarās or celestial dancing-girl to
seduce him and so destroy his power.
It is worth stressing that none of this necessarily has anything to do
with morality. When the ascetic succumbs to the apsarās’s wiles, this is not
a moral fall, but a loss of self-control, leading to a loss of spiritual power.
This is only a moral or ethical issue in so far as self-control is seen as morally
good. In the Buddhist or Jain traditions, one can argue more convincingly
that some degree of positive moral value has become attached to celibacy,
but here again the key issue would appear to be self-control as part of a
process of discipline. The discipline is justified not for its own sake but
5
By celibacy here, I imply the rejection of (hetero-)sexual relations, since the term is also sometimes
used to refer to the rejection of marriage. See Bell and Sobo 2001.
182
Origins of yoga and Tantra
because it is held to lead to a state of enlightenment or liberation.6 Here
there is an explicit contrast with the idea in mainstream Christian traditions
that celibacy is a morally superior state in its own right, as opposed to any
end-result that is supposed to arise from its practice.
In reality, things are more complex. Celibacy is associated with purity, and
purity is undoubtedly seen as a valued state in Indian society.7 Sexuality,
women and family life are all in various ways associated with impurity
and pollution.8 There undoubtedly are senses in which the normative
Brahmanical tradition regards purity as morally superior, and the Brahman
as morally superior to the untouchable, though even here it is difficult to
be too categorical. As always in India, it is easy to find support for both
a statement and its opposite. When Mahatma Gandhi proclaimed that in
an ideal Hindu society, Brahmans and street-sweepers would go on being
Brahmans and street-sweepers, but each would be regarded as of equal value
as a human being, he was speaking out of a valid strain within the Indian
tradition, but it was one that had little correspondence to the social reality
of his time.
Ascetics are prototypically male, and the role of female ascetics in Indic
societies has been and is today marginal and problematic. The goddess
Pārvatı̄’s ascetic practice is only superficially an exception, since it was
aimed at acquiring a husband, in her case the god Śiva. In fact, as is well
known, women’s religious practice regularly involves fasting and asceticism
of various kinds, but its proper aims are generally held to be the establishment and welfare of her family (S. Robinson 1985; M. McGee 1992). This
introduces a certain blurriness into the logic. If spiritual power derives in
some sense from control over the ejaculation of semen, it requires a male
physiology. Women, though, clearly are seen as able to accumulate power
through ascetic practice, even if its utilisation is ideally confined to the
welfare of husband and household.
One solution here, and it is followed in later traditions such as Vajrayāna
(Tantric) Buddhism, is to see both male and female essences as present in
both men and women, so that the internal energy processes are more or
less the same despite the physiological differences between them. But in
practice there is a strong strain of thought that implies that ascetic celibacy
6
7
8
In the Tibetan context, it is not unusual to have a married, non-celibate lama presiding over a
monastery of celibate practitioners. There is really no contradiction in this situation, since the
normative assumption is that the non-celibate lama has transcended the need for celibacy, and is
capable of sexual relations without attachment.
Here again Louis Dumont gives the classic account (L. Dumont 1972).
Cf. Leslie 1996a; Rozario and Samuel 2002a.
Asceticism and celibacy in Indic religions
183
is essentially a male business. It is clear that Hindu, Buddhist and Jain
traditions are all uncomfortable with female ascetic practitioners, and none
treats them as fully equal to men. This reflects the complex relationship
between household and this-worldly-centred ritual and religious life, which
is seen as a proper concern of women, and ritual and religious life that is
aimed at more transcendental and other-worldly goals, which is not so seen.
In any case, the equation between celibacy and male spiritual power
is quite familiar, and it is expressed in various ways across Brahmanical
Hinduism, and across the Jain and Buddhist traditions. No doubt it gains
some of its familiarity and obviousness from the idea of celibacy in the
context of Christian monasticism. Buddhist and Christian monasticism in
particular have spread over vast areas of the planet and acquired a very
substantial presence in many societies. Yet, perhaps as a result, we may be
failing to look closely enough at what is going on when Indic religions take
as perhaps their single most culturally-valued goal the pursuit of spiritual
power by male ascetics.
warrior brotherhoods and wisdom traditions
One approach is to ask how this pattern developed historically. One origin
has been identified, perhaps somewhat speculatively, in the social and religious role of young men in Indo-European pastoralist society. In Chapter 6
I mentioned the suggestion made by the British scholar of Jainism, Paul
Dundas, which in its turn derived in part from work by Willem Bollée, that
the early Buddhist and Jain monastic orders were a kind of transformation
of Indo-European male warrior brotherhoods, and that a transitional stage
could be seen in that shadowy and mysterious group in late Vedic and early
Brahmanical texts, the vrātyas (Bollée 1981; Dundas 1991).
This suggestion has its problems, but it seems clear enough in any case
that a major strain of early Indic religious life valorised male celibacy and
purity. One can, I think, see an early version of this strain as being located
in the Kuru-Pañcāla region, and as linked in many ways to the values of a
pastoralist society, or at least of a pastoralist elite.
As for the vrātyas, scholarship since Bollée’s time seems to support the
idea that they were, in the earlier Vedic material at least, something like a
warrior age-set, a phase through which the young men of the tribe go prior to
marriage (see Chapter 5, and Falk 1986). The vrātyas were unmarried young
men, active as cattle-raiders and warriors, and they also had a significant
ritual role, which was tied up to the ongoing fertility and productivity of the
land. All this came to an end when they married and become incorporated
184
Origins of yoga and Tantra
into ordinary society as householders, though a small minority, unable or
unwilling for one reason or another to move to the householder stage,
remained as permanent vrātyas and acted as leaders of the vrātya troops.
The important point, though, is that we see here an initial context in which
an early Indic society makes use of and valorises the specific age group of
young unmarried men.
As I suggested in Chapter 5, iconographical sources from the Central
Gangetic and Bengal Delta regions, such as the Chandraketugarh terracottas, suggest a religion focusing on something quite different: the cult
of a goddess of fertility and prosperity. This material can be seen as an
early instance of another pervasive theme in Indian thought and imagery,
most familiar in the mithuna (male and female couple) images which run
through Indian sacred architecture from the early Buddhist stūpas to the
spectacular erotic sculpture of Khajuraho and Konarak. I will have some
comments on these later temples and their imagery in Chapter 12, but here
I would point to the more restrained but much more widespread imagery
that we find throughout Buddhist, Hindu and Jaina architecture.
We could consider here, for example, the male and female couples
who separate the narrative scenes from the Buddha’s previous lives on the
Buddhist stupa at Nagarjunakonda, dating from around the third century
CE (Agrawala 1983; figs. 79–83), although there are thousands of similar
scenes that could be chosen. I would suggest that in this imagery we can
see an alternative conception of both kingship and human society to that
which centres around the celibate young man. If the celibate young man as
hero had his natural home in the pastoralist ideology of the north, then the
agricultural societies of the north-east, the homeland of Buddhism and the
ascetic traditions, tended instead to celebrate agricultural prosperity and
fertility.
ascetic practices
It is also to this Central Gangetic region that our early evidence of asceticism
and yogic practices aimed at liberation or enlightenment refers. I would
suggest that it is in the interaction between these two regions that we can see
the origins of Brahmanical asceticism, and of the growth of the idea of the
Brahmin as in part an ascetic. I think that there may well be some truth in
Bollée’s suggestion that the vrātya and similar Vedic social institutions were
reshaped in the process as something rather different, forming the basic
structure of the new ascetic orders, though we can hardly exclude that,
as the Buddhist and Jain traditions both assert, there were earlier teachers
Asceticism and celibacy in Indic religions
185
of a path to enlightenment. While these figures seem largely legendary
or mythical, with the possible exception of Pārśva, the twenty-third Jain
tı̄rthaṅkara, they are mostly also located in the Northeast Indian region.9
Given that the idea of permanent ascetic withdrawal seems itself not to
have been a feature of the earliest Vedic context, this perhaps suggests that
the key stage as far as celibate power is concerned in the Vedic context was
the brahmacārin himself, particularly the brahmacārin who, like the vrātya
leaders, remains in the celibate role rather than marrying. In particular, the
brahmacārin who moved directly into a renunciate lifestyle may be regarded
as the prototypical generator of ascetic power.
In later Indian religious history, we might think here of figures such
as Śankara, who are described as adopting the ascetic lifestyle from youth
without ever becoming a householder. The iconography of Śankara depicts
him as a young man, and his biographies describe him as a child prodigy
who died at the age of thirty-two. We might think too of the iconography
of young ascetic deities such as Ayyappan at Śabarimala, the ascetic form of
Murugan (Skanda) who is worshipped at the great Tamil shrine of Palani, or
Daks.inamūrti, the ascetic teacher form of Śiva. It may be no accident that
these are all celibate forms of deities who elsewhere are strongly associated
both with male sexuality and with warriorship.
Is it reasonable then to see the brahmacārin as a kind of ‘civilised’ and
‘spiritualised’ version of the original role of the vrātya as young male warrior?
If so, we can perhaps begin to sense how the cultural and psychic valorisation
of young men that we already saw in the vrātya role was transformed into
something rather different, as a key symbolisation of the value of purity
that was at the peak of the Brahmanical scale of values.
c e li bacy and m ale identit y in contemporary buddhism
Here it is useful to move to more contemporary material, while remaining
cautious about the extent to which the present can be read into the past.
The question of celibacy and male identity in contemporary Buddhism
nevertheless has some interesting continuities with the material considered above. I shall look specifically at material from Northern Thailand,
although there are comparable features in other parts of Thailand and in
other Southeast Asian societies. As mentioned in Chapter Six, ordination is
usually treated by Thais and other Southeast Asian Buddhists as a temporary
9
If Williams is correct in taking Nemi, the twenty-second tı̄rthaṅkara, as historical (R. Williams 1966),
we would need to include Gujarat as well. Gujarat would have been part of the ‘non-Āryan’ world
at that time, and like the Central Gangetic region and Bengal Delta had strong early links to trade.
186
Origins of yoga and Tantra
period of transition to the marriageable state (Tambiah 1970; Spiro 1971;
Davis 1984; Rhum 1994). Thus what we actually have in relation to most
of the male population is a period when young men withdraw from society
and live in an exclusively male group, followed by their return to society, marriage and incorporation as male householders. Structurally at least,
we can see here a considerable similarity to the old vrātya pattern. If we
recall the suggestions in Northern and Northeast Thai ethnography that
the celibacy of the ordinands is, in some respect, tied up with the coming
of the rains, the fertility and productivity of the land (e.g. Tambiah 1970),
the similarity is even stronger.
What is different in the Buddhist pattern, presumably, is the absence
of the warrior role, and the stress on the ethical dimension of Buddhism,
particularly on moral restraint, represented here above all by the sexual
purity (i.e. celibacy) that is a central element of the definition of a Buddhist
monk in contemporary Southeast Asian states. As Charles Keyes notes, the
high status of monastic celibacy creates certain paradoxes for male identity
in Northern Thai society (Keyes 1986). The ideal male is a Buddhist monk,
a bhikkhu, so where does this leave the male householder?
Keyes sees the practice of temporary ordination as a way in which the
ambiguities of the male Buddhist householder role are partially resolved.
Temporary ordination is a rite of passage to adult male status. It involves a
reformulation of man’s link to his mother in the course of the ordination
ritual, and it is normally followed by marriage after a short interval.
Some features of Keyes’ analysis are specific to Northern Thailand. This is
a society with matrilocal residence, meaning that girls remain in the same
household at marriage, and men move to their future wife’s household.
This is associated with a sense of women and women-centred households
as the stable elements in society, and men as mobile and unrooted, which
is quite different from, for example, the standard North Indian pattern,
where women move to the husband’s household and have a marginal status
within it for many years. However the idea of the ‘morally tempered male
householder’ as a compromise between renunciation (the bhikkhu) and
over-indulgence (the tough aggressive male personality) would seem to
have wider applicability. So does the sense that a woman’s role is defined
by the domestic context (Keyes 1984; Kirsch 1985; Kawanami 2001).10
I would emphasise that this is not just a question of Buddhist monasticism in Thailand serving as a rite of passage for young men and providing a
certain degree of moral tempering. One can also see this in slightly different
10
On the female role in Theravāda Buddhist societies see also Andaya 1994.
Asceticism and celibacy in Indic religions
187
terms, in terms of what one might call the psychic economy of Thai peasant
life. The energy and psychic orientation of young men is being directed in a
particular direction through the practice of temporary ordination. Looking
at the situation in this way makes it considerably clearer why the monastic
role is largely seen as irrelevant and inappropriate for women. It may also
help to explain the continuing relevance of imagery of warfare, warriorship
and conquest within Buddhism, in relation to the conquest of the ordinary
self and its motivations.
In relation to everyday life (and to women in so far as they are seen as
located in the domain of everyday life), this pattern constructs a separation between the two domains: laukika and lokottara, worldly and beyond
the world. Although, in doctrinal terms, the Buddhist teachings can seem
dismissive of the concerns of everyday life, the characteristic orientation
towards everyday life in Buddhist societies is rather to treat it as a separate
domain with its own integrity and logic and its own largely autonomous
ritual life (see Chapters 6 and 7). Thus, as previously noted, marriage
in Buddhist societies was generally not a Buddhist ritual or sacrament,
but a secular contract whose ritual dimensions were linked to the creation of auspiciousness and good fortune, matters essentially disconnected
with the Buddhist goal of enlightenment. The critical point of intersection
between the two domains, everyday and transcendental, was death, the one
point in the life cycle where Buddhist clergy played a critical and central
role.
In many ways, this is all very Indic, if not particularly ‘Indian’ in the sense
of resembling the situation in modern India. The Thais themselves seem
to have arrived in Thailand fairly late, from the tenth century onwards.
Theravada Buddhism, however, was already well established in the region
at that time and the general pattern of male identity here seems similar
enough to that in other parts of Buddhist Southeast Asia to allow one to
assume that it is to a substantial degree of Indic origins.
c e li bacy an d m ale id entit y in brahmanical societ y
Thus, if temporary ordination can be seen as a process that should ideally
result in all men having a kind of spiritually-tempered male identity, somewhere between the extremes of ascetic and warrior, what we could call the
mature Brahmanical model moves in a quite different direction. Here, it
is assumed that different social groups within the village community take
on different roles, with the Brahmins towards the ascetic pole, the ks.atriyas
towards the warrior pole. Of course, the varn.a scheme has only a tenuous
188
Origins of yoga and Tantra
relationship to the social reality of Indian villages in modern times, but it
retains an important ideological role.
Purity in the Brahmanical structure of ideas is seen as something which
varies up and down the social hierarchy, again an idea that has a strong
presence in South Asian communities, if increasingly contested by groups
who find themselves defined as impure or polluted. From the Brahmanical
point of view at least, spirituality is associated less with a stage in the lifecycle than with an elite group within society. Yet in other ways, what we have
seen in Thailand contrasts with what we could call the typical Brahmanical
construction of things.
Thus where the dominant Southeast Asian solution was to see male purity
as a phase in the life-cycle, the mature Brahmanical mode of thinking saw
it as the function or role of a particular social group. Here purity and
spirituality become part of a hierarchical construction of society, in which
the purity of the Brahmin was counterposed to the impurity and thisworldly orientation of lower orders of society. At the same time, everyday
life was systematically structured through a series of sacramental rituals, the
sam
. skāra, performed by Brahmin priests.
This is a more complex pattern, and in fact it involves two ascetic roles:
the ideally semi-ascetic (but married) role of the Brahman within caste
society, and the fully-ascetic role of the sam
. nyāsin outside caste society
proper. It took quite a while to get to this point, historically, and even
longer for village ritual life to be fully incorporated into the Brahmanical
system, a process that in most parts of India had not been fully achieved
even in modern times. However, the pattern proved to be a durable one.
Its greatest strength was perhaps its ability to survive in the absence of state
patronage. The inability of monastic Buddhism to do the same proved to
be a crucial weakness after the Islamic takeover and the collapse, in all but
a few marginal parts of South Asia, notably Sri Lanka and the Nepal Valley,
of the state regimes which had traditionally supported Buddhism.
At the same time, the incorporation of women, the lower castes, and
everyday life into the Brahmanical structure came at a price, which was the
dominance of the ideology of purity, and the definition of female biology
as intrinsically impure. The Brahmanical societies of South Asia put much
more work, culturally, into an ongoing struggle to maintain purity than
did the Buddhist societies of Southeast Asia, and this cultural concern or
obsession had destructive consequences for women in particular, as I noted
in Chapters 4 and 7.
Thus it seems that the role of the renunciate, preserved in a relatively early
form in the role of the bhikkhu among the Buddhist societies of modern
Asceticism and celibacy in Indic religions
189
Southeast Asia, has in a sense bifurcated in the Brahmanical context. On
the one hand, we have the semi-renunciate village Brahman, who at least in
theory plays a central role in the ritual life of the village.11 On the other, there
is the fully-renunciate sam
. nyāsin. Each has acquired some of the psychic
and symbolic value of the brahmacārin. Put otherwise, the brahmacārin acts
as a kind of bridge between the two, with the possibility of movement in
either direction. Both patterns, Buddhist and Brahmanical, nevertheless,
have in-built tensions and complexities: and in both cases, women have
an ambivalent and difficult relationship to a social and ideological order
constructed around the supremacy of certain kinds of men.
conclusion
I have tried to suggest here the two major directions in which the Indic religions were to develop, the Brahmanical and that represented by the śrāman.a
movements, both Buddhist and Jain. I have also sketched something of
the contrasting orientations towards pragmatic and everyday religion, and
towards women and their religious and biological functions, which we find
today associated with them.
Both patterns, I suggest, originated in various ways in the complex
interaction between Vedic culture and the wisdom-oriented traditions and
earth-centred rituals of fertility and prosperity of Northeast India. They
constituted different mixtures of these elements, and we can see these in
transformed versions in the later developments of Brahmanical Hinduism
within India, and of Theravāda Buddhism in particular outside it.
It is important to appreciate here that the role of asceticism and celibacy
within Buddhism and other Indic traditions can only make sense in relation
to the wider assumptions and practices within the society where they are
practised, particularly the assumptions regarding gender and sexuality. It
is worth seeing this issue in the context of the psychic or cultural balance
within society as a whole. Aggression, violence and competitiveness are
clearly modes of human social behaviour which have appeared in one form
or another in most or all human societies, and particularly perhaps among
young men. The formation of groups of young men, whether neighbourhood gangs, football teams, or vrātya bands raiding the neighbouring tribe’s
cattle, provide ways of channelling such behavioural modes in more or less
destructive directions. The idea of young men undergoing a temporary
period as a Buddhist monk or in some other semi-ascetic role could be
11
I say ‘in theory’, because most men from Brahmin families are not priests and perhaps never were.
190
Origins of yoga and Tantra
seen as a way of taking that aggression and competitiveness and directing
it inwards, into the quest of self-mastery. It is perhaps not surprising that
the imagery of warriorship and victory that remain significant within Buddhism has occasionally expressed itself in real-life violence and war (e.g.
Mohan 2001; Victoria 2005). The problem is nevertheless one that every
society has to deal with, our own included, and the ‘temporary ascetic’
approach may well be judged, in relation to the historical record, one of
the more successful.
part t wo
Tantra
The Tantras – there is hardly any other kind of literature that has
met with so much abuse, particularly by those who never read or
seriously studied a single line of it; or that has so much fascinated
those who on the testimony of misinformed and uninformed people
thought the Tantras to be a most powerful, and hence strictly guarded
means for the gratification of purely biological urges. Only very few
people tried to form an opinion of the Tantras by their own. It is
true the Tantras are nothing for those who are so pure in mind and,
alas! so poor-minded that they are unable to see that actual life is
different from the fantastic and mutually contradictory theories and
ideas they have about it; nor are the Tantras meant for those who
consider life to be nothing else but a chronique scandaleuse. But since
it is easier to follow extremes than to weigh the evidence and to decide
upon a middle path, there can be no doubt that these extremists have
done great harm to the study and understanding of what the Tantras
have to tell. For it is by their verdict – unjustified abuse based upon
wilful ignorance and misconceptions about the aim of the Tantras
engendered by this ignorance – that the Tantras are nowadays held in
contempt and considered to be something depraved and mean. Yet
the fact is that the Tantras contain a very sound and healthy view of
life.
Herbert Guenther, Yuganaddha: The Tantric View of Life
(Guenther 1969: 3)
chap t e r 9
The classical synthesis
Part Two of the book is focused on the period from the fourth to twelfth centuries CE, essentially from the foundation of the Gupta Empire in around
320 CE to the final establishment of Muslim rule over North and Northeast
India. I shall be particularly concerned with the growth of ‘Tantric’ forms
of religion. To begin with, some discussion is necessary regarding the years
between Part One and Part Two. This is the period in which what might be
regarded as the classical synthesis of Indian culture took shape, a synthesis
that is most familiar through the great artistic and literary achievements of
the Gupta period in North India.
Part One closed with the initial development of what were to become the
two alternative and parallel cultural patterns that would shape the history
of Indic religions over the centuries to follow. These were the Buddhist/Jain
pattern and the Brahmanical pattern. The Buddhist/Jain pattern worked in
terms of large monastic religious centres and treated the world of popular
and civic religion and of the rituals of everyday life as a parallel sphere that
was largely beyond the concern of Buddhist and Jain religious specialists.
It appears to have been closely linked to urban centres and the trading and
administrative groups within these centres, and initially received a high
level of state support. Buddhism and Jainism also developed pilgrimage
and cult centres where there was a major monastic presence, including
those at the sites associated with the events of the life of the Buddha and
the tı̄rthaṅkaras.
It is not really clear how far Buddhism and Jainism penetrated to the
village level in areas away from big cities.1 In later years, in Sri Lanka,
1
The only substantial regional study of Buddhist temples in South Asia known to me is Julia Shaw’s
work from the Vidisā region. The monasteries in this study were distributed over a fairly large area,
but they are all quite close to a major administrative centre of the period (Vidisa /Besnagar), and
may not be representative of more remote areas (J. Shaw 2000; J. Shaw and Sutcliffe 2001). Most
of the monastery sites uncovered or located in Bangladesh are also close to major administrative
centres (Dhaka-Vikrampur, Mainamati-Comilla, Mahasthan) (Chakrabarti 2001; Samuel 2002c).
193
194
Origins of yoga and Tantra
Southeast Asia and Tibet and in parts of East Asia, Buddhism became a
religion of the countryside and the village as well as of the urban centres,
though it has mostly coexisted with other religious complexes which have
dealt with the affairs of this world: the spirit-cults of Burma, Thailand
and Sri Lanka, and with Confucianism, Taoism, shamanism and Shinto
in East Asia. Here Buddhism and Jainism were following the inner logic
of their ascetic, renunciate and world-transcending origins. The power of
the Buddha in these countries became an important source of magical
ritual technology, which was harnessed for this-worldly purposes (Tambiah
1984; Spiro 1971; Tannenbaum 1987, 1989) but Buddhism maintained some
ideological distance from the affairs of the world and of the gods and
spirits of this world. Only in those countries where Vajrayāna (Tantric)
Buddhism became dominant (Nepal, Bali, Tibet, Mongolia) did Buddhism
itself become so integrally involved in worldly and pragmatic issues that
one can speak of a single religion rather than the co-existence of two or
three distinct religious complexes (Samuel 1982; Samuel 1993).2 We will see
something later of the developments that allowed this integral relationship
to take form.
The Brahmanical pattern, by contrast, saw itself from the beginning as
integrally involved in worldly affairs, and established itself so successfully in
this regard that, even in the modern Buddhist countries of Southeast Asia
and Sri Lanka, some kind of Brahmanical presence exists in both urban and
rural settings.3 Brahmins, as hereditary priestly families, became a presence
at the rural level throughout India, often with substantial land-holdings.
This did not mean that all village ritual was performed by Brahmins, or
that all (or even most) Brahmins were employed primarily as ritual performers, but it established the ideological supremacy of Brahmanical ritual
as the ‘correct’ way in which all kinds of ritual, including the most humble
domestic practice, should be performed. At the other extreme, Brahmins
largely came to monopolise the ritual of state and kingship. Buddhist state
rituals were developed (Klimburg-Salter 1989; Walter 2000), but even in
the modern Buddhist kingdom of Thailand, the king is installed by court
Brahmin priests.
2
3
Keyes suggests that the situation was similar in Southeast Asia prior to the missionising activities of
the thirteenth to sixteenth centuries, in other words there were relatively small numbers of Buddhist
monks living ‘in temple-monasteries in or near the major political centres’ (Keyes 1987: 133).
The Bon religion of Tibet in recent centuries is best regarded as a variant form of Tibetan Vajrayana
Buddhism. In any case Buddhism and Bon are competitors for the same territory, rather than religions
which operate in parallel but separate spheres.
For rural Brahmins in Thailand and Burma see Brohm 1963, Tambiah 1970 ( paahm), etc.
The classical synthesis
195
The establishment of Brahmanical Hinduism as a state religion can be
associated above all with the Gupta dynasty in North and Central India
(c. 320–c. 510 CE). I use the term ‘Brahmanical Hinduism’ here, because I
think that if we want to use the term Hinduism at all before the nineteenth
century then this is the point at which we can reasonably start using it. By
around 500 CE, we have state religions focused on the theistic cult of deities,
Śiva and Vis.n.u, who are still central figures of modern Hinduism, we have
most of the major schools of Hindu philosophy at least in embryonic form,
we have temples and temple rituals which are recognisably continuous with
the versions we know today, and we have much of the social regulation of
later Hindu society. In addition, this religion was the dominant religion
of much of the population of South Asia. At the same time, much of the
development of the caste system lay ahead, and the degree of penetration of
this new Brahmanical culture at the village level was probably considerably
more limited than in modern times.
Some account is needed here of how these developments came about. I
start with a brief summary of political developments over the period. There
are the usual provisos for dates and details, though in comparison with the
periods we have looked at in Chapters 3 to 7, by this period we begin to get
substantial amounts of epigraphic material, and the Chinese translations
and early manuscript material also give us some hold on dating for the
Buddhist material specifically. There are still a lot of open questions, but
at the level at which I am dealing with the material they are for the most
part perhaps not too critical.
After the collapse of the Mauryan empire (c. 185 BCE) there followed a long period during which a variety of large-scale regional powers
existed. These included ruling dynasties originating from within the limits
of present-day South Asia (Śuṅga, Sātavāhana, Meghavāhana, Khāravela,
Iks.vāku,Vākāt.aka,Gupta)andothersoriginatingfromoutside(Indo-Greek/
Bactrian, Śaka, Indo-Parthian, Kus.ān.a).
It is difficult in retrospect to know how stable social life was, or appeared
to be, for those who lived in South Asia at this time. Looking from a vantagepoint two thousand years further on, this looks like a period of political
instability for much of the sub-continent, with many areas moving back and
forth between the control of one or another regional power. The situation is
complicated by historical uncertainties, such as the still not entirely resolved
question of the dating of the Kus.ān.a dynasty (cf. Sims-Williams and Cribb
1996). Some of these dynasties, however, lasted for considerable periods, at
least in their core regions. This is particularly true for the two most important powers, the Sātavāhana and the Kus.ān.a states. The Sātavāhanas lasted
Origins of yoga and Tantra
196
Limit of areas
probably included
in Kus.ān.a Empire
K
U
S.
Purus.apura
Limit of areas definitely
included at some time
within Kus.ān.a Empire
Ā
N
.
A
Limit of areas
probably included
in Kus.ān.a Empire
W.
Ś A
Mathurā
S
K
A
S
S
Ā
Ujjayinı̄
T
Ā
Pratis. t.hāna
H
A
N
AS
W. Śaka (1st cent. C.E.)
V
Core areas
A
Girinagara
Probable maximum
limit of areas at some
time under W. Śaka rule
W. Śaka (early 2nd cent. C.E.)
Sātavāhana
Probable maximum limit of areas at
some time under Sātavāhana rule
Figure 9.1. Map of Kus.ān.a and Sātavāhana States, first to third centuries CE
The classical synthesis
197
for about two and a half centuries (50 BCE to 200 CE), the Kus.ān.as for
perhaps two centuries (50–250 CE). For much of the first to third centuries,
large parts of South Asia were divided between these two major powers. It
may have seemed like a period of prosperity and relative peace, especially
for many of the inhabitants of the Satavahana and Kus.an.a states.
The Kus.ān.a realm extended into Central Asia and this was also a period
of extensive trade through to the Parthian and Roman empires to the West
and increasingly into Southeast Asia and China to the East. Buddhism was
spreading along these trade channels to Central Asia and to China, and
other cultural influences were doubtless coming in the reverse direction.
Some of India’s greatest creative thinkers and artists lived in this period, if we
accept the most likely chronologies: these included, for example, the great
philosopher Nāgārjuna and the poet Aśvaghos.a. It was during this period
too that the great stūpas at Amarāvatı̄ and Nāgārjunakon.d.a were erected,
and their fragmentary remains bear witness to the immense sophistication
of South Asian art and life at this period. We are accustomed to think
of Gupta rule as the ‘golden age’ of Indian artistic and cultural achievement, perhaps because of a later prejudice against the ‘foreign’ rule of the
Indo-Greek, Kus.ān.a and Śaka dynasties and their patronage of Buddhism,
but the achievements of the Gupta period would not have been possible
without the foundations, and the cultural hybridisation, of the preceding
centuries.
A period of dynastic transition followed in the late third and early fourth
centuries, with the situation stabilising again in the fourth century CE,
with Gupta control over most of North India, Vākāt.aka rule in Central
India and Kus.ān.a rule continuing for a while in the Northwest. The Guptas
and Vakat.akas4 were both primarily pro-Brahmanical in their religious
preferences, although Buddhism continued to flourish during this period.
We can assume a pluralist religious situation, much in fact as has existed in
India in more recent times, with individual families, occupational and caste
groups affiliated by and large to one or another local version of Brahmanical, Buddhist, or Jaina religion.5 The Buddhists and other śraman.a groups
were probably strongest in the urban areas, among trading and artisan
communities, while the Brahmanical lineages gradually gained strength in
the countryside as more and more land was gifted to Brahmin families.
Buddhist and Jaina shrines and monasteries and Brahmanical temples were
4
5
The Vākāt.akas should probably be considered as comprising two separate states, the Eastern Vākāt.akas
with their capital at Nandivardana and the Western Vākāt.akas with their capital at Vatsagulma (Kulke
2004). For the Vākāt.akas in general, see Bakker 2004.
Also Ājı̄vika, though they never seem to have received a high level of state patronage (Basham 1951).
Origins of yoga and Tantra
198
Limit of areas annexed
by Samudragupta
Limit of areas made
tributary by Samudragupta
G
U
P
T
A
Prayāga
Nandivardhana
Vatsagulma
A
Ā T.
K
V Ā
K
A
Pātaliputra
S
S
Limit of areas made
tributary by
Samudragupta
Maximum extent of areas
at some time under Vākāt.aka
overlordship
Limit of areas at
some time under
Vākāt.aka rule
PALLAVAS
Kān̄cı̄puram
Core areas
Gupta
Vākāt.aka
Pallava
Figure 9.2. Map of Gupta, Vākāt.aka and Pallava States, fourth to sixth centuries CE
scattered throughout the countryside, as no doubt were places of power
linked to the various local deities of the yaks.a class. The ruling families
probably felt some need to patronise all the major traditions, though over
time the balance shifted increasingly towards the Brahmins and what was
eventually to become Hinduism.
The classical synthesis
199
These dynasties ruled over much the same regions that we have been considering in Chapters 4 to 7, and over which the Mauryan empire had ruled,
but another important feature of this period is that the newly-established
Indic cultural patterns were by now expanding into new regions: Central
Asia, South India and Southeast Asia.
Central Asia has already been mentioned: Kus.ān.a rule, like Indo-Greek
rule before it, led to regions such as Bactria and modern Afghanistan being
confluences of Indic and Central Asian cultural patterns, with explicitly
Buddhist kingdoms being created at locations such as Bāmiyān (KlimburgSalter 1989). Over the succeeding centuries, the Buddhist regions of Central
Asia would arguably make a major contribution to the development of the
Mahāyāna and early Vajrayāna. The importance of Gandhāra, the gateway
to Central Asia and the location of the great early centre of scholarship
at Taxila (Taks.aśı̄lā), has already been noted in Chapters 3 and 4. Asaṅga
and Vasubandhu, two of the greatest of Mahāyāna Buddhist scholars, came
from this region (Davidson 2002b: 159–60). The regions beyond Gandhara
were however also significant. It has been plausibly argued that the cult
of Maitreya derives in part at least from Iran (Nattier 1988: 45 n.51); the
Avatam
. saka Sūtra may have been compiled and partly authored ‘within
the Indic cultural sphere of Central Asia’ (P. Williams 1989: 121), while
the cult of Mañjuśrı̄ also has strong associations with Inner Asia (Gibson
1997: 43–5). There were also non-Buddhist influences, particularly in the
Kus.ān.a period, and a significant cult of Śiva developed in Bactria (Scott
1993).
By the first and second centuries CE, the Dravidian-speaking regions of
the south were also increasingly being incorporated into the general North
and Central Indian cultural pattern, as were parts at least of Southeast Asia.
The Pallava kingdom in South India was largely Brahmanical in orientation
although it included a substantial Jain and Buddhist population, while Indic
states were also beginning to develop in Southeast Asia. Tamilnadu was
however to remain an important centre for Buddhism for many centuries,
with connections to the Buddhists in Sri Lanka.
In South India and Southeast Asia, Indic models of the state were to
have a massive impact, and the same can perhaps also be said for areas
of Eastern India such as East Bengal and Assam which had arguably been
marginal to the previous North Indian cultural complexes. In South India,
as in East India, the records of Brahmin settlement, witnessed by land-grant
inscriptions, are an important indication of the process of ‘Indianisation’.
I avoid the term ‘Sanskritisation’ in this context, since what is involved is
the borrowing by a state of a whole model of social organisation, rather
than the attempt of a group to raise its own status vis-à-vis other groups.
Origins of yoga and Tantra
200
P
y
PEIKTHANOMYO
Peikthanomyo
u
ŚRĪKS.ETRA
Śrı̄ks.etra M
o
A
N
A
s
H
Ā R
DV
M
AV
TUNSUN
P A
AT
I¯
U
N
LA
F
TAK
KO
Oc Eo
TĀ M
Takuapa
er
C
n
Suddhammavatı̄
Khm
(Thaton)
Tāmbralinga?
B
EARLY
R
AL
TRADING Chu-li? LINGA
AN Langkasuka?
STATES Kadāram/ GKAS
M
RA A
DĀ TĀR
KA KA
Chieh-ch’a
UK
A
Figure 9.3. Early states in SE Asia to c. 650 CE
However this undoubtedly was a culture closely associated with the Sanskrit
language.
While there are differences between South India and Southeast Asia,
there is also much to be said for seeing these as parallel developments. The
spread of Indian influence through Southeast Asia is a complex process and
only partly understood (Coedès 1975; Mabbett 1977a, 1977b; Wheatley
1983; Bentley 1986). Rice cultivation in Southeast Asia would seem to go
back at least to the later third millennium BCE, bronze to some date after
1500 BCE, and iron to around 500 BCE. By this stage, Southeast Asia was
becoming ‘a vital component of an exchange network which linked the
Roman and Chinese empires’ (Higham 1999: 82).
A number of states developed in the region from around 200 CE onwards,
initially at Oc Eo (‘Funan’) in the lower Mekong delta, later at Dvāravatı̄
The classical synthesis
201
in central Thailand, at Angkor in Cambodia, and Champa in presentday coastal Vietnam, among others. In Burma, a group of Pyu city-states
belong to the later third century CE onwards, and were linked to the Bengal region through further coastal Mon and Arakanese kingdoms (Stargardt
2000). These states gradually came under Indic cultural influences, incorporating aspects of both Buddhist and Brahmanical religion. As Mabbett
rightly notes, ‘Indianisation’ here was not necessarily a radically different
process from similar processes in South Asia proper, particularly in nonIndo-Aryan speaking areas or those without significant local populations
claiming Brahmin or ks.atriya status:
India is not an integrated social and cultural unit in which a particular system of
ideas controls a system of social behaviour. It is an arena of vastly different groups
jostling together, with their several ethnic affinities, languages and cultures, and
Sanskrit lore is like a great stain that has spread unevenly throughout the whole,
changing its colour repeatedly as it came into contact with different elements.
Southeast Asia is the same, and the differences within the two zones are as great as
those between. There is no reason in principle why the stain of Sanskrit lore should
not spread across into Southeast Asia by the same social processes that advance it
in India, growing fainter or more mutable as it recedes further from its original
source. There need be no reason why there should be a difference in kind. And the
record of ‘Indianized kingdoms is testimony that this is what happened. (Mabbett
1977b: 160)
As the existence of Chinese translations of Buddhist scriptures from the
first century CE onwards indicates, Buddhism was also being exported to
East Asia. Here, however, there were already well-developed local models
of state organisation and Indic models of society and the state had much
less impact. Buddhism was regarded as a foreign religion in China and its
situation was always vulnerable to nationalist attacks, such as the major
persecutions of Buddhists during the later T’ang period (Weinstein 1987,
and see Chapter 12 below).
early theism: bh āgavata and śaiva cults
A major religious development during this period was the growth of a new
kind of conception of the deity. Whatever the differences between Vedic
deities and yaks.a cults, there is no real evidence that either class of deity was
conceived of as a transcendental and all-embracing presence, in the way
that Śiva and Vis.n.u were later to be imagined.6
6
Here I am not particularly using ‘transcendental’ by contrast with ‘immanent’, but simply to indicate
that the deity has taken on a universal and cosmic role. As is well known, Indian philosophical
202
Origins of yoga and Tantra
One can certainly see signs of a movement in this direction in some
of the later Brāhman.a speculations and the earlier Upanis.ads. Thus, references such as that to the deity Skanda in the Chāndogya Upanis.ad (7.26,
cf. Olivelle 1998: 166) suggest that the author or authors are constructing
a relationship with an independent cult of Skanda.7 It becomes explicit
in texts such as the Śvetāśvara Upanis.ad (perhaps second or first century BCE), in which the universal deity is Śiva, or the Bhagavadgı̄tā (perhaps first or second century CE), where he is Vis.n.u as revealed through
Kr.s.n.a. These texts suggest a coming together of the philosophical and
metaphysical tendencies of the Upanis.ads with a commitment to a specific universal deity that derives from other sources. Similarly, I would
assume that the authors of the Śvetāśvara Upanis.ad and the Bhagavadgı̄tā
were providing, as it were, a sophisticated metaphysical understanding for
deities who had already assumed a considerable importance in the external
world.
For all this, our evidence for the cult of these deities from non-textual
sources is initially limited. The first sculptural representations of any of the
new deities dates from around the first century BCE, either at around the
same time as the earliest yaks.a imagery or a little later. These early deity
images include representations of the Bhāgavata deities, who represent the
earliest sculptural phase of the Vais.n.ava cults. A little later (first and second
centuries CE) we also have representations of some of these deities on
Indo-Greek coinage.
The origins of the Bhāgavata deities are connected with the Vr.s.n.i people of North India, an initially fairly marginal group who were coming to
dominance during the period of the composition of the Mahābhārata. The
cult of Kr.s.n.a as Vasudeva seems to have risen to prominence along with
them.8 The main figures are two male warrior deities, Vāsudeva (Kr.s.n.a)
and Sam
. kars.an.a (Balarāma), sometimes expanded through the addition of
a goddess, their sister or consort Ekānam
. śa/Nidrā to make the so-called
Vr.s.n.i Triad (cf. Couture and Schmid 2001; Yokochi 2004).9 We have early
7
8
9
traditions developed a variety of ways of conceptualising the relationship between universal deity and
devotee in later centuries, including monist, dualist and more complex options.
The Chāndogya Upanis.ad is generally regarded as one of the earliest Upanis.ads (fifth to fourth
centuries BCE, if one accepts the ‘late’ dating for the Buddha), but this looks like a later addition.
Skanda is identified in this text with Sanatkumāra (‘Ever-Young’), elsewhere regarded as a form of
or a son of Brahma and apparently a deity associated with the ascetic tradition. Sanatkumara also
occurs in Buddhist texts as a proponent of the Buddha’s teachings (e.g. Dı̄ghanikāya 19).
The grammarian Pan.ini explains the term ‘Vasudevaka’ as meaning ‘a bhakta [i.e. worshipper or
devotee] of Vasudeva’, but as noted in Chapter 2, the dating of Pan.ini is far from certain.
For reproductions see Srinivasan 1997, plates 16.13 to 16.15 (Sam
. kars.an.a/Balarāma), plates 18.10 and
18.11 (Vāsudeva/Kr.s.n.a), plate 16.5 and Couture and Schmid 2001 and p. 181, fig. 1 (Vr.s.n.i Triad).
The classical synthesis
203
representations of both Vāsudeva and Sam
. kars.an.a in a numismatic context, on a coin of Agathocles dated to the first century BCE (Srinivasan
1997, plates 16.6 and 16.7), as well as some early sculptural representations
(Srinivasan 1997, plates 16.13–16.15). These are clearly warrior deities, relatively crude figures holding conspicuous weapons. There are also a number
of inscriptional references, of which the most famous is the Garud.a pillar
inscription of the Greek ambassador Heliodorus, who styles himself as a
Bhāgavata, near Besnagar (Colas 2003: 230–2). The deity Nārāyan.a, mentioned in some late Vedic texts (Brāhman.as and Āran.yakas) is also associated
with Vāsudeva and Sam
. kars.an.a in some of these inscriptions, though the
precise relationship is not clear.
These deities were in time to be brought together with the cult of Vis.n.u,
another deity who appears in several hymns in the R.gveda, often in association with Indra, as well as in later Vedic texts. These deities came to
form a scheme in which Vāsudeva (Kr.s.n.a), Sam
. kars.an.a, and Kr.s.n.a’s son
and grandson. Pradyumna and Aniruddha, are all treated as emanations
(vyūha) of Nārāyan.a/Vis.n.u as supreme godhead. This is the basis of the
later Pañcaratra ritual tradition, as well as a precursor to the eventual concept of the multiple avatars or divine manifestations of Vis.n.u. As is well
known, Kr.s.n.a is also treated as a manifestation of Vis.n.u in the Bhagavadgı̄tā,
which forms part of the Sanskrit Mahābhārata.
Dennis Hudson has suggested in a recent article that the Bhāgavata
movement developed ritual procedures that legitimised the role of rulers
who were technically śūdra or foreign (mleccha), and that much of the rise
to prominence of Bhāgavata religion can be understood in relation to their
association with such rulers (Hudson 2002: 135). Some of his argument
depends on an assumption that we can read Pañcaratra ritual into a considerably earlier period than is usually supposed. If we decide not to follow this
part of Hudson’s analysis, then his model of the Bhāgavatas as providing
ritual services for upwardly-mobile dynasties of questionable origins might
apply to later dynasties such as the Guptas and Pallavas rather than to the
time of their initial growth. In any case, Brahmin theorists were doubtless
working overtime on solutions for problems of this kind from quite an
early date, and Vis.n.u was to become in later Indic tradition the deity most
centrally associated with royal power and rule.
We find early Śiva images and early forms of the liṅga, the phallic representation of Śiva, at around the same time (von Mitterwallner 1984;
Srinivasan 1984; N. Joshi 1984), and both Śiva (Oesho = Īśa, one of Śiva’s
titles) and the warrior-god Skanda figure on Kus.ān.a coinage, in forms
which may reflect some degree of syncretism with Iranian and Bactrian
204
Origins of yoga and Tantra
deities (first and second centuries CE).10 Early depictions of Skanda present
him as a warrior deity (e.g. Czuma 1985, pl.93); as noted above, he is
mentioned in the Chāndogya Upanis.ad and was eventually identified as a son
of Śiva.
As usual, the precise dates for these developments are unclear; the spectacular early liṅga at Gud.imallam in Andhra Pradesh has been variously
dated from third century BCE to fourth century CE (von Mitterwallner 1984: 13 and pl.18; Srinivasan 1997: pl.17.9). However, Śiva had clearly
become an important deity associated with royal power by the third and
fourth centuries CE. He also became a central figure for the whole complex of ideas that we refer to as ‘Tantra’ (see Chapter 10). He is traced back
to the god Rudra, who appears in the R.gveda as a marginal figure, ‘the
embodiment of wildness and unpredictable danger, he is addressed more
with the hope of keeping him at bay than with the wish to bring him near’
(O’Flaherty 1981: 219). Another fairly early Vedic text, the Śatarudriya, is a
litany accompanying offerings to Rudra: Gonda describes it as representing Rudra-Śiva ‘in his ambivalent character, both as a malevolent and as a
benevolent deity’ and notes that Rudra-Śiva in this text is ‘unmistakeably
on the way to become an All-God’ (Gonda 1980: 82).
Perhaps understandably, Śiva is one of the deities in relation to which
the conflict between ‘orthogenetic’ and more complex historical models has
been most intense (see e.g. Srinivasan 1997; Bakker 1999, 2001). I do not
have anything substantial to add to the controversy on this theme, though it
is worth noting that while Śiva is a major figure in the Mahābhārata, liṅgaworship at any rate has a marginal role in Brahmanical texts until about
the fifth and sixth centuries CE (Bakker 2001: 404). This might suggest
that liṅga-worship was a more popular form of worship, and therefore
that a model of multiple origins for Śiva may be more useful than an
orthogenetic account. It is also perhaps significant that village gods and
goddesses tend to be viewed as forms of Śiva and his consort (Durgā, Kālı̄,
Umā, Pārvatı̄), rather than as forms of Vis.n.u and his consort (Laks.mı̄). As
in other areas, part of the genius of Brahmanical thought was to rework
whatever it absorbed within its own vocabulary.
What is more relevant for my own argument is the relationship between
Vis.n.u, as the deity who became identified with the appropriate and legitimate exercise of power, and Śiva, who maintained an edgier, more dangerous
and potentially destructive persona. This is not so much a question of ‘good’
and ‘bad’ as of ‘civilized’ and ‘uncivilized’ forms of power. Vis.n.u, in part at
10
Joe Cribb has, however, argued against the identification of the ‘Oesho’ figure with Śiva (1997).
The classical synthesis
205
least, represents the incorporation of the ideal of the just king or wisdom
king into the Brahmanical model, as his association with Rāma suggests,
although like all major Indian deities he is a complex figure who can take
on a variety of roles, positive and negative. Śiva seems to have become a
deity of royal power at a somewhat later stage, and this development is
closely tied up with the development of dangerous and powerful forms of
ritual to protect royalty and to support its less peaceful projects; in other
words, with the early stages of Tantra. I return to this theme in Chapter 10.
To return to the pre-Kus.ān.a and Kus.ān.a context in which we first see
these deities, it is worth noting that the distinction between the new deities
and the yaks.a cults is not necessarily a sharp one. Gods such as Vāsudeva
and Sam
. kars.an.a presumably began as the focus of local cults and only
gradually became upgraded and provided with Vedic pedigrees (or, if you
prefer, assimilated to Vedic gods). However there is a marked difference
in tone; these are clearly warrior gods, and in time gods of universal rule,
rather than local protectors. The iconography contrasts too with the yaks.a
protectors who tend to be presented as corpulent and benevolent, even when
serving as ‘generals’ of yaks.a armies (Kubera, Pāñcika). The new deities tend
to be young and muscular, with affinities perhaps to the cult of youthful
virility that we have noted in the Brahmanical ascetic complex in Chapters 7
and 8.
We can perhaps distinguish between three general areas in the pre-Kus.ān.a
and Kus.ān.a period, though the limited range of sculptural material from
this period means that much may be a question of chance survivals. On
the whole, the new deity cults seem to have been focused in the north and
in parts of the Deccan.11 In the Central Gangetic and Eastern region, the
dominant deities at this period seem still to have been old-fashioned yaks.a
deities such as Man.ibhadra (patron of travellers and traders; Thapan 1997)
or Vajrapān.i (protective deity of Rājagr.ha, and now a major Buddhist protective deity; Lalou 1956, Lamotte 1966), along with Laks.mı̄ and her various
equivalents (such as Ambikā, Manasā and Padmāvatı̄). Laks.mı̄’s cult seems
to have been spreading into the northern area, with Chandraketugarhlike reliefs found extensively at Mathurā, a major crossroads and point of
confluence for different traditions.
Meanwhile, in the Gandhāra region, one finds confluences of Indic and
Central Asian gods, such as the cult of Hāritı̄ (with her consort Pāñcika)
as deities of fertility and prosperity, overlapping with the cult of Ardoxsho
11
For Śiva, see Srinivasan 1984 and fig. 1. If the Gud.imallam liṅga can really be dated as early as third
or second century BCE, as Srinivasan and Mitterwallner 1984 propose, this would provide another
site in the southern Deccan.
206
Origins of yoga and Tantra
Figure 9.4. Anointing of Kumāra as Divine General (India, Gupta Period)
and Pharro (e.g. Lerner 1984: 186). There also appear to have been overlaps
between Skanda or Kumāra (a young warrior god, who became identified
as Śiva’s son Karttikeya) and the Iranian deity Sraoša (Pal 1986–88: vol. 1
no. S41; see also Czuma 1985: No. 93), between Śiva/Oesho, the Iranian wind
god Oado and Greek Herakles, and between the Buddhist protective deity
Vajrapān.i, Indra and Herakles (e.g. Klimburg-Salter 1995: 122 & pl.xv).
These mutual influences may have had a significant effect on later Indian
conceptions of some of these deities, both in terms of the iconography
which was only just beginning to develop at this period and the actual
conception of the deity. Mathurā was closely connected to the Gandhāra
region and the Hāritı̄/Pāñcika cult was important there as well (Lerner
1984; Czuma 1985; Harle and Topsfield 1987: 11 no. 13; Srinivasan 1989).12
12
As noted in Chapter 5 the evidence for identifying the Gandhāran-Mathurān goddess with Hāritı̄
is limited.
The classical synthesis
207
Deities such as Skanda, Śiva and the early Bhāgavata deities were clearly
something other than benevolent city-protectors; what we are seeing here
goes well beyond the confines of the yaks.a religion. For all of the philosophical depth and complexity of Śaivism and Vais.n.avism in later times,
these deities can perhaps be best understood in this early period as divine
patrons of aggressive rulers out to expand their kingdoms.13 This probably
holds for much of the early Gupta period as well.
This helps to explain why, unlike the yaks.a deities, these new – or at any
rate newly-important – gods are much less tied to a specific local presence.
They may and do have local cult centres, but their role is not that of a
local deity. They are claimed to have supreme status among the gods. I
do not think it is too crude or simplistic to link this with a new model of
Brahmanical kingship in which the god is a divine reflection of the king.
Indeed, the divine status of the king is a cliché of the Indian political theory
of this time, and is clearly related to the supreme status of these new deities.
Thus we see a move from a polytheism, whether Kuru-Pañcāla or Central
Gangetic, in which a variety of gods corresponds to a variety of coexisting
powers, to a monotheistic or monistic universe in which the king is the god,
and everybody and everything else is simply a projection or part of him.
The Brahmanical law codes, in mandating the subordination of women,
children and younger males to the male head of the household, replicated
the same principle at the social level. This generates what the anthropologist McKim Marriott famously referred to as the model of people as
‘dividuals’ – subordinate parts of a greater whole, and sharing their substance with the wider whole – rather than individuals (Marriott 1979).
This emphasis on the social, and on the individual as subordinated to it,
continued to be a strong ideological emphasis within Brahmanical religion
into modern times. While one can exaggerate the extent to which it was
internalised to produce a radically different personality structure – I think
Marriott is taking the texts a little too seriously here – it was undoubtedly
also very influential on how people saw themselves in relation to family
and community.
the sansk rit epics: m a h ā b h ā r ata and r ā m ā ya n. a
The Mahābhārata and Rāmāyan.a, our most substantial literary documents
from the period, are an important source for these political and religious
13
This is not to deny that, as Srinivasan has pointed out, the iconography and other aspects of the
conception of these deities may have owed a considerable amount to the yaks.a deities (Srinivasan
1997).
208
Origins of yoga and Tantra
developments. In Chapter 4 I considered some of the mythological material
that presumably formed the basis for these two great epic poems, particularly the bodies of legends regarding the Lunar and Solar Dynasties. Here
I consider the two epics as finished literary creations. Recent scholarship
seems to confirm the picture of the Sanskrit text of the Mahābhārata in a
form fairly close to the Critical Edition as a relatively integrated work (e.g.
Hiltebeitel 2001; Fitzgerald 2004; Hiltebeitel 2004). As far as the Rāmāyan.a
is concerned, while there have been plausible attempts to identify earlier
and later parts of the text, most notably in recent years by John Brockington (Brockington 1985), my present concern is with the Sanskrit text as a
whole. There is no real agreement at this stage on either the absolute dates
of these two completed works or on their relative dates. Fitzgerald suggests a
shorter version of the Mahābhārata in pre-Gupta times, probably between
c. 150 CE and 0 CE, followed by a rewriting during the Gupta Empire
to produce a version close to the Critical Edition, whereas Hiltebeitel
argues for a pre-Gupta (and pre-Common Era) composition for the entire
work (Hiltebeitel 2004). There are arguments for placing the Rāmāyan.a
before the Mahābhārata and also for placing the Rāmāyan.a after the
Mahābhārata.
The Rāmāyan.a appears to be the more straightforward of the two works.
The central character, Rāma, in pre-Rāmāyan.a legend perhaps an epic hero
and ideal king of the old Central Gangetic variety, is transformed in the
finished Rāmāyan.a into a model for the new Brahmanical god-king. The
opening and closing books, generally considered to be among the latest
components for those who regard the work as incorporating earlier and
later elements, are extensively concerned with articulating this model of
kingship.
The Rāmāyan.a’s position, however, though clearly Brahmanical, is a
kind of compromise, of which the central marriage between Rāma and
Sı̄tā is a key image. Values of wisdom, renunciation and connection with
the female powers of the earth are given considerable recognition here,
despite the ambiguity created by Rāma’s later treatment of Sı̄tā. A similar
message is conveyed in Book I of the Rāmāyan.a (the ‘Bālakān.d.a’) through
the treatment of those old Vedic adversaries, Vāsis.t.ha and Viśvāmitra, who
frequently serve as a proxy for opposing positions, particularly concerning
relations between Brahmins and ks.atriya. Vāsis.t.ha is the court Brahmin
of Rāma’s father Daśaratha, but it is Viśvāmitra who arrives to request
the young Rama’s assistance in dealing with some troublesome rāks.asas, so
bringing about Rāma’s marriage with Sı̄tā and setting the story in motion
(I, 17, vv.26ff.). Viśvāmitra, as we are told at length in a later section of
The classical synthesis
209
Book I of the Rāmāyan.a (I, 60–4) is a ks.atriya who has attained the status
of a Brahmin and a .r.si through his austerities and sacrifices. It hardly seems
accidental that he acts, in effect, as matchmaker between Rāma and Sı̄tā,
daughter of the wisdom-king lineage of Mithila.14
The Mahābhārata is the more complex of the two epics. Deciphering
its exact relationship to a political context regarding which we have only
fairly vague ideas is not easy. The Mahābhārata incorporates elements of
the ‘wisdom king’ model which is so central in the Rāmāyan.a but the
implicit relationship to Brahmanical authority is different and gives
Brahmins a higher place. Yudhis.t.hira, the main ‘wisdom king’ figure in the
Mahābhārata, is a less central and more problematic figure than Rāma, and
the renunciate king theme, though present, is also treated in a less positive
way.15 Fitzgerald has suggested that the five Pān.d.avas can be seen as a divine
raiding-party, sent to punish the Ks.atriyas for their rejection of Brahmanical
authority.16 Where the Rāmāyan.a presents a compromise between the old
Kuru-Pañcāla-style Brahmins and Central Gangetic wisdom-king values,
the Mahābhārata has little room for conciliation. At a more general level
again, however, both epics endorse a move from the old Vedic model
towards a new model based on acceptance of a single supreme deity, and
both also present a new conception of kingship, though this is more strongly
articulated in the Rāmāyan.a.
The new model incorporates elements of the old Vedic-Brahmanical
model of kingship, such as the aśvamedha or horse-sacrifice, and there is
an obvious concern to legitimise what is going on in terms of the older
Brahmanical texts. In reality, though, the model of the tribal chieftain of
a pastoralist people that underlay the old Vedic conception, even in its
Kuru-Pañcāla recasting, has now been left drastically behind.17 The Vedic
chieftain hoped, through the proper performance of sacrificial rituals on his
behalf, for the blessings and favour of Indra and a variety of other members
14
15
16
17
See White 1991: 78–80 for more on Vāsis.t.ha and Viśvāmitra. Another interesting episode in Book
I is Rāma’s defeat of the archetypal Bhargava hero, Rāma Jāmadagnya (I, 73–5), which culminates
in Rāma Jāmadagnya recognising Rāma as the god Vis.n.u (I, 75, vv.19–20).
Yayāti, the renunciate king who is the subject of an extended episode in Book 1 of the Mahābhārata, is
an ambivalent figure and hardly presented as a role-model. His renunciation follows on the excessive
indulgence of his relationships with Devayānı̄ and Śarmis.t.hā and his thousand years of youthful
pleasure at the expense of his son, Pūru, and he is subsequently thrown out of Indra’s heaven for
his excessive pride (cf. van Buitenen 1973: 175–210).
See Hiltebeitel 2004: 212 and n. 31. Hiltebeitel appears to endorse this view of the Mahābhārata
himself, though disagreeing with Fitzgerald in relation to the chronology of the text’s composition
among other matters.
See also Gitomer 1992, which explores the character of Duryodhana as articulating old and new
conceptions of masculinity and leadership.
210
Origins of yoga and Tantra
of a divine pantheon. The new Brahmanical king was a god himself, in the
image of a single supreme god of whom other deities were at most aspects
or derivatives.
Elements of the Central Gangetic dharmarāja model are clearly also
reflected in the new conception. The new Brahmanical king incorporated
and enforced dharma and he, along with his divine patron and reflection,
became its source. The newly devised law codes, such as the Laws of Manu,
provided the basis for these developments.
Obviously, all this had to go along with a new version of religion for
the masses of the population, at least those who lived in the towns and
cities and the upper social levels of the rural population. The new forms
of theism focused on the Vais.n.avite deities and on Śiva provided the basis
of this new mass religion. They provided the successors of the innovative
Brahmins discussed by Lubin (see Chapter 7), with new, more powerful
and persuasive concepts and ritual techniques.
The classic statement is of course the Bhagavadgı̄tā, often regarded as
one of the later additions to the Mahābhārata. Here I would follow Norvin
Hein, who suggests that the Gı̄tā reflects the ‘concerns and worries of the
Indo-Aryan social leadership’ who had become ‘distressed . . . by secessions
into non-brāhmanical religions and by the flight of disaffected householders from their civic duties into the monastic life’ (Hein 1986b: 298). He
describes the Gı̄tā as
essentially a great sermon calling the alienated young back to their duties in the
brahmanical social order. In it Kr.s.n.a argues that the renunciation that brings
salvation is not the repudiation of the world’s work but the elimination of selfinterest while remaining steadfast in one’s earthly duties. (Hein 1986b: 298)
Certainly the Gı̄tā is explicit that withdrawal from society is not the way
to follow:
He who does the work which he ought to do without seeking its fruit he is the
sam
. nyāsin, he is the yogin, not he who does not light the sacred fire, and performs
no rites.
What they call renunciation [sam
. nyāsa], that know to be disciplined activity, O
Pān.d.ava (Arjuna), for no one becomes a yogin who has not renounced his (selfish)
purpose. (Bhagavadgı̄tā 6,1–2, trans. Radhakrishnan 1963: 187)
While the concept of bhakti in the sense of devotion to a personal deity
is very much part of this new picture, it did not yet carry the emotional
content that it would bear at a later stage. This seems, as Fred Hardy
has argued in his book Virāha-Bhakti, to be a primarily South Indian
development, and to belong to a somewhat later period (Hardy 1983; see
The classical synthesis
211
also Hart 1979). Its progressive adoption in the North can perhaps be linked,
as Norvin Hein has argued, to the tensions associated with the enforcement
of the new Brahmanical social order (Hein 1986b). Hein has a two-stage
model here, with the Kr.s.n.a of the Bhagavadgı̄tā and the later sections
of the Mahābhārata being part of the Brahmanical rethinking which was
to bear fruit in the Gupta state and other Hindu monarchies, while the
later emphasis on Kr.s.n.a as child or as young lover was a response to the
constraints of life after the new Brahmanical state model had been imposed.
The later development of these ideas can be seen in the Purān.as, the great
assemblages of mythology, history, ritual and other material that developed
from about the fifth century CE onwards in various parts of South Asia.
These texts also developed a new religious geography, prefigured in the
Mahābhārata and Rāmāyan.a, in which the sacred landscape of India was
refigured in terms of the new versions of Brahmanical religion, centred on
the new pantheon of great Brahmanical deities. The Purān.ic framework
allowed for a great deal of absorption and reworking of local religious
material, and the multiplicity of these texts also lead to a de facto acceptance
of religious coexistence. Whether by choice or necessity, this was a model
of religious pluralism, at least between different Brahmanical traditions,
and implicitly in relation to the non-Brahmanical traditions, primarily the
Buddhists and Jains. This certainly did not exclude local conflict, forced
conversions, or worse, but the basic pattern appears to have been one of
coexistence between a variety of religious traditions, with families and caste
groups having traditional linkages with one or another priestly lineage or
order, as indeed they still do in much of modern South Asia.
budd hist d evelopments
On the Buddhist side, the Mahāyāna Sūtras with their visions of cosmic
Buddhas seem to belong to very much the same world. While Buddhism’s
fundamental stance in relation to the samsāric world of everyday life meant
that many things which Brahmanical religion saw as its legitimate concern (such as marriage and the other sam
. skāra or life-cycle rituals, with the
exception of death) fell outside the realm of operation of Buddhist clergy,
one can see an increasing concern with matters of secular life, of politics and the state. Sūtras such as the Suvarn.aprabhāsa discuss the proper
behaviour of kings (cf. Emmerick 1970); holy men such as Nāgārjuna
wrote, or at least were held to have written, letters of advice to kings (the
Suhr.llekha, see Kawamura 1975; Fynes 1995: 46). We will see more of these
texts in Chapter 12, in relation to their adoption in East Asia. Alexander
212
Origins of yoga and Tantra
Studholme has suggested that the Kāran.d.avyūha Sūtra, dating perhaps
from the late fourth or early fifth century, is written in conscious dialogue
with contemporary Purān.ic material. This text proposes the Buddhist deity
Avalokiteśvara as an appropriate figure for worship as a supreme deity and in
particular propounds the teaching of his basic prayer-formula, the famous
mantra om
. man.ipadme hūm
. (Studholme 2002).
The Buddhist model of kingship seems to have remained for a long
time however that of the dharmarāja – the king who rules according to the
Buddhist dharma, so bringing prosperity to his country, and who patronises
the Buddhist saṅgha with massive offerings. Deborah Klimburg-Salter has
argued this in detail for the Kingdom of Bāmiyān in Afghanistan – the site
of the colossal Buddhas destroyed recently by the Taliban (Klimburg-Salter
1989).
Precisely what was meant by the term ‘Mahāyāna’ is a complex issue,
historically clouded by polemic on both the Mahāyāna and Theravāda sides
(cf. R. Cohen 1995; Silk 2002; Skilling 2004; Hallisey 1995). The Mahayāna
was never a distinct organisational entity within Indian Buddhism; monks
belonged to one or another of the Nikayas or ordination lineages associated
with the various early Buddhist schools. The Chinese pilgrim Yijing notes
in his Record of Buddhist Practices, dating from 691, that ‘Those who worship
the Bodhisattvas and read the Mahāyāna Sūtras are called the Mahāyānists,
while those who do not perform these are called the Hinayānists’ (Silk
2002: 360). In reality, though, there is no single clear defining feature that
enables us to distinguish Mahāyānists from others, and it seems likely that
many monasteries in India continued through most of the history of Indian
Buddhism to have both Mahāyānists, non-Mahāyānists and perhaps also
others who would not have seen themselves as clearly within one of the two
camps.
At the same time, the Theravāda, as we know it in later times, developed
gradually as a conservative movement in Sri Lanka and South India. It
appears to have added extra rules to the Vinaya and it undoubtedly edited
its early texts in the light of its own developing position, as did all Buddhist
schools (Prebish 1996; Collins 1990). As I noted in Chapter 2, there is no
reason to regard the version of the early texts redacted in the Pali Canon as
any more authoritative or representative of early Buddhism as the parallel
texts from other, non-Theravādin Nikāyas. In addition, as Skilling has
pointed out, Mahāyāna texts are not necessarily later than Theravāda texts:
the earliest bodhisattva sūtras ‘were probably composed in the second or
first century BCE’ (Skilling 2004: 147), and developed Mahāyāna sūtras
were already being translated into Chinese in the second half of the second
The classical synthesis
213
century CE. It is not unreasonable either to suppose that tendencies and
modes of thinking that we now see as ‘Mahāyāna’ may have gone back to
early days of the Buddhist tradition (cf. Lindtner 1991–93).
I focused in Chapter 6 on the emergence of monasteries in or near urban
centres and it is clear that many of the major Buddhist centres were of this
kind. Taxila (Taks.aśı̄lā), which was a major academic centre for Brahmanical
scholarship (Pān.ini and Kaut.ilya are both traditionally associated with it)
was also a major Buddhist centre with numerous monasteries and stūpas
(Dar 1993), and the important urban centre of Vidiśā had a large number
of monastery and stūpa sites within relatively close proximity (Shaw 2000;
Shaw and Sutcliffe 2001). As I noted at the start of the chapter, most of the
archaeological material refers to sites of this kind; there is little evidence
for the extensive presence of Buddhism at the village level, a development
which perhaps never really took place in South Asia.18
At the same time we should take into account, as Reginald Ray has
argued in his Buddhist Saints in India (1999), the importance of a tradition
of forest-monk practitioners, perhaps centred mainly around the major
pilgrimage sites. Certainly such practitioners have been important in more
recent times in most Theravādin countries (Tambiah 1984; Carrithers 1983;
Taylor 1993; Tiyaranich 1997, 2004; see also Keyes 1987; Lehman 1987;
Tambiah 1987). Ray suggests that much of the initial growth of ‘Mahāyāna’
tendencies may have taken place in these circles, and that the ‘Mahāyāna’
might be seen as part of a forest response to the process of monasticisation. Thus Ray suggests that we need a threefold model of Buddhism
in India, which incorporates urban monastics, forest renunciates and lay
people.
Ray’s model seems to me to be an important contribution to the debate
and to make general sense, although it is difficult to be sure at this distance how clear-cut the distinction was between urban monastics and forest renunciates, and exactly how these divisions might have worked out
‘on the ground’. The Chinese reports suggest that by the seventh century ‘Mahāyāna’ tendencies could be found in a wide variety of monasteries, many of which would presumably have been urban. However, given
the apparent importance of visionary techniques and experiences in many
‘Mahāyāna’ sūtras one might imagine that they would have originated in
contexts that were strongly committed to meditative practice, a situation
unlikely to occur in a busy urban monastic centre with close links to lay
18
As noted earlier, it seems not to have taken place even in Southeast Asia until the thirteenth century
onwards (Keyes 1987: 133). It may have developed earlier in Sri Lanka.
214
Origins of yoga and Tantra
affairs (cf. Beyer 1977; Prebish 1995; McBride 2004). We will look in a little
more detail at some of these techniques later in the chapter.
Ray suggests that what we see in the forest traditions is a continuation of
a forest-bodhisattva ideal which is represented by the Buddha Śākyamuni
himself and which existed among his immediate followers and contemporaries. In this sense, the ‘Mahāyānist’ forest-renunciates discussed by Ray,
and the Mahayāna itself as an attempt to recapture that early ideal, could
be seen as deriving from ‘the earliest and most authentic Buddhism’ (Ray
1999: 417).
Leaving aside questions of authenticity, Ray’s study points to a bifurcation in the Buddhist movement which undoubtedly goes back to early
times. This bifurcation was between those parts of the tradition which
engaged with urban centres, wealthy lay patrons and local rulers, and in
effect developed a form of Buddhist practice which was part of the ongoing process of everyday social life, and those which maintained a renunciate ideal and saw themselves as essentially outside of society. In the later
Buddhist texts, the figure of Devadatta, the Buddha’s cousin and rival, stands
for the extreme form of this position, in which only ascetic forest practice
is seen as valid. Initially recognised as a saintly figure, Devadatta appears
to have become progressively vilified until he becomes an embodiment of
evil (Ray 1991: 162–73). However, forest asceticism was never completely
prohibited, and the forest renunciate ideal continued to modern times.
Part of the point here can be seen in the association between forest practice, meditation and the acquisition of supernatural power. Some forest
monks in more recent times have been charismatic figures whose image of
purity, asceticism and magical ability can be channelled and accessed by the
state and by the urban elite but which can also be seen as a threat to the state.
I have suggested elsewhere that the constant concern with the ‘purification
of the saṅgha’ in Theravāda Buddhist history can be seen as being largely
about the need to keep charismatic figures under control, and that the
gradual marginalisation and prohibition of Tantric approaches in societies
dominated by Theravāda Buddhism reflects this need (Samuel 1993: 24–
36, 2005: 52–71). In retrospect, I think that I understated in these analyses
the extent to which supernatural power remained a major concern within
‘Theravāda’ societies even in modern times, but the overall argument here
still seems to me to be valid. In Indian Buddhist history, as in other Indian
renunciate traditions, we can see a continual dialectic between forms of
Buddhism which have become more or less assimilated to the needs of the
state and of ‘respectable’ society, and others which position themselves at
the edges or outside society as a whole. It seems sense to me to see some of
The classical synthesis
215
the early Mahāyāna material in these terms, although certainly later forms
of the Mahāyāna developed into forms of religion fully compatible with the
state and urban society. The same process occurs with the developments
of ‘Tantric’ (Vajrayāna) forms of Buddhism, which initially position themselves in large part outside the world of the large monastic centres and of
respectable society, but become progressively integrated within it.
The large monasteries in or near cities would doubtless have been more
closely integrated with the political order, in those states where the ruling
dynasty patronised Buddhism to a greater or lesser degree. It seems that at
Bāmiyān, as in other Buddhist kingdoms of the time, the role of the king
as patron of Buddhism was acted out annually through massive festivals in
which huge presentations were made to the Buddhist saṅgha or monastic
community. These were the so-called pañcāvars.ika rituals that have been
documented widely in the Buddhist states of this period. At Bāmiyān,
they were in fact depicted by paintings, which Klimburg-Salter reproduces.
Klimburg-Salter suggests that this Buddhist ceremony may have evolved
in part ‘as a response to the revitalization of ancient Vedic state ceremonies
by the Gupta kings’ (Klimburg-Salter 1989: 125). In them, the king as a lay
Buddhist ‘legitimises his own authority as protector of society, including
the saṅgha’.19 This is a pattern which has continued into modern times
with the Buddhist states of Southeast Asia, though the use of Brahmin
priests in Southeast Asian kingdoms, and the popularity of the story of
the Rāmāyan.a (and to a lesser extent the Mahābhārata) in Southeast Asia
indicates that the new Brahmanical model had considerable success even
with Buddhist states. The Buddhist Kings of Thailand, after all, chose to
name a later capital city, Ayutthaya, after Rāma’s capital city of Ayodhyā,
as did the rulers of Java (Jogja = Ayodhya).20
As at Bāmiyān, one way to reinforce the message on the Buddhist side
was through colossal Buddha statues that insisted visually on the majesty
and supremacy of the Buddha. We find these in China too, where gigantic
images of the cosmic Buddha were constructed at Yungang in the late fifth
century (c. 460–7, during the Northern Wei period) and at Lungmen in
the late sixth century, during the reign of the pro-Buddhist Empress Wu.
19
20
Walter has discussed the evidence for alternative sets of state ritual procedures on the Buddhist side,
such as the agricultural rituals, involving the rāja ploughing a furrow around the city, described
in many Mahayana Sutras as well as in the Dı̄pavam
. sa and Mahāvam
. sa (Walter 2000). These are
linked to the pañcāvars.ika rites, and to the idea of the king as a cakravartin (see also Aung-Thwin
1981 for Burma). But in practice even states which were in other respects committed to the support
of Buddhism tended to accept much Brahmanical ritual in relation to the kingship, as in Thailand,
Laos, Cambodia or Burma into modern times.
In fact, Ayodhyā still forms part of the ‘official’ full name of the current Thai capital of Bangkok.
216
Origins of yoga and Tantra
These statues are generally taken to represent the cosmic Buddha Vairocana,
whom we will return to a little later. Amitābha in the Saddharmapun.d.arı̄ka
(‘Lotus Sūtra’) has a similarly cosmic role, and became the basis of the
major form of established state Buddhism in the Far East (‘Pure Land’
Buddhism).
philosophical schools
On the whole I have kept philosophy or theology in the background in
this book. This is not because I think that it is unimportant, but because
it has been so much at the centre of many other accounts that it helps to
focus elsewhere for once. It is useful however to cast a very brief glance at
the evolution of philosophical and theological conceptions throughout this
period.21
Looking across the Brahmanical and non-Brahmanical traditions, there
is an understandable similarity between what is happening in both areas
at similar times. Brahmanical schools such as Sām
. khya and Vaiśes.ika, like
the various Buddhist Abhidharma traditions or Jain philosophy, work in
terms of enumeration of the constituents of experience; essentially, a refined
form of list-making. The Sām
. khya school, said to have originated with the
mysterious asura-sage Kapila,22 was evidently particularly close to yogic
circles.23 All these approaches are based in analyses of experience into a
discrete number of fundamental categories, which are regarded as providing
the basis of a truer and more spiritually beneficial understanding of the
world than is provided by our ordinary perception.
The similarities reinforce the impression we have gained from other
material that everybody was operating within the same general climate of
thought. There is also a strong impression in many of these systems that
experience in the context of meditative practice is a primary element in
the development of philosophical concepts, a point made many years ago
by Edward Conze (1962: 17–19) and reinforced recently in the case of early
Buddhist philosophy by Sue Hamilton (1996, 2000).
The significant resemblances between the two major later Buddhist
philosophical schools or groups of schools, Mādhyamika and Yogācāra24
21
22
23
24
On the whole, in India, it is hard to draw a rigid distinction between theology and philosophy, unless
one restricts the term theology to philosophical understandings of the more theistic traditions.
On Kapila, see e.g. Bronkhorst 1998a: 69–77. A chapter of the Jaina Uttarādhyayana Sūtra is
attributed to him and consists on advice for renunciates (Jacobi 1985: 31–5).
In the Mahābhārata, the goals of Sam
. khya and Yoga are often said to be identical: ‘that which the
yogins see, is the same as is understood by the Sām
. khyans’ (Jacobsen 2005b: 9).
The Yogācāra are also referred to as Cittamatrā (‘mind only’) and Vijñānavāda.
The classical synthesis
217
and the Brahmanical school of Advaita Vedanta have long been noted.
Here we see a move away from enumeration towards the assertion of an
underlying basis of unity. This move is undoubtedly prefigured in the early
Upanis.ads, where brahman has already become a universal principle, but
its rigorous working out in philosophical terms, whether negative or positive, is a new development and one which seems to have first taken place
with Nagarjuna’s theoretical reworking of ideas from the Prajñaparamita
Sūtras and with the development of alternative perspectives by subsequent
Buddhist theoreticians. Nāgārjuna is traditionally dated to the second century CE and the first major Yogācāra scholar, Asaṅga, to the fourth century
CE. Possibly the Yogācāra innovations can be linked to the new political situation, though they seem to be prefigured in a group of sutras
from the third century, such as the Avatam
. saka, Sam
. dhinirmocana and
Laṅkāvatāra (Warder 1970: 423). Asaṅga himself is said to have been born
in Gandhāra: Warder, who dates him from c. 290 to c. 360, suggests that he
lived
in a period of political instability before the new imperial system was completely
established, though he may have lived to see the weak remnant of the Kus.ān.a
kingdom invaded by Samudra Gupta and attached to his empire by a rather tenuous
form of vassalage. (Warder 1970: 436–7)
Thurman has suggested more recently that Asaṅga and his brother
Vasubandhu ‘presided over a sizeable expansion of the Buddhist movement in Gupta-dominated India of the fifth and sixth centuries CE’ and
that a key issue here was that the emphasis on consciousness in the new
school made the bodhisattva vow, the vow to follow the Buddha’s example
and achieve enlightenment in order to relieve all beings of their suffering,
seem a more meaningful and achievable ideal:
How difficult it seems to make such a vow in the mind and heart if one is thinking
of the ‘universe of beings’ as an infinite, substantial, external, dense, and heavy
bunch of objects! [. . .] But, if the nature of all beings and things is mental, mindconstituted, and mind-created, then a radical transformation of the inner mind, in
intersubjective interconnection with other minds, could very well be able to effect a
total transformation of everything that exists with a semblance of greater ease. One
could then take up the bodhisattva vow with a sense of possibility empowering the
compassionate emotion and the messianic determination. (Thurman 2004: xliii)
Advaita Vedānta seems to be a much later development, with Gaud.apāda,
the first major Vedantic teacher, dating perhaps from the eighth century CE
and referring back to a variety of Buddhist teachers. What is notable though
is the similarity between these schools, all of which tend in one way or
218
Origins of yoga and Tantra
another to collapse phenomenal reality into a single substrate or underlying
principle. T. R. V. Murti argued some years ago that the three groups
of schools were saying essentially the same thing is different languages
(Murti 1960). This was an oversimplification, but it points to a significant
resemblance.
The shift in ideas from Sām
. khya, Abhidharma and Jaina theory to these
new kinds of approach can surely be correlated with a longer term social,
political and religious process in which pluralism (the mahājanapadas, the
numerous yaks.a cults associated with particular cities and places) has been
replaced by an assertion of unity (the king claiming universal rule, the
single supreme deity). A polytheistic universe, with a multiplicity of deities
and powers with whom the individual must negotiate and deal, was being
replaced by a monistic one, in which the individual increasingly became
simply a separated part of the whole.
I am not suggesting here that we treat philosophy or theology as a simple
derivative from politics. If this were the case, then we would have got
to Rāmānuja’s modified nondualism and Madhva’s dualism much more
quickly. In any case, it is clear that the philosophers had other things on
their mind besides providing an ideology for the newly-emerging state. The
Buddhists in particular had reason to avoid too strong a commitment to
a supreme inclusive deity-figure. Arguably, they did not really reach this
point until the Ādi-Buddha concept came in with the Kālacakra Tantra in
the early eleventh century.
We can see another aspect of the same developments, however, in the
realm of ritual and meditation. Here I turn to look at developments in
Buddhist and Brahmanical meditation procedures, and also to consider
one of the themes that would later be a central component of the Tantric
synthesis: the man.d.ala.
m editation in the bud dhist, brahmanical,
and j aina trad itions
The classic manual of the Theravāda tradition, the Visuddhimagga (‘Path
of Purification’) of the early fifth-century scholar Buddhaghosa, presents
a masterly synthesis of meditation as understood in his day in relation to
the Pali Canon (Ñān.amoli 1991). As for the Mahāyāna, there is a considerable amount of material on early Buddhist meditational procedures in
the Mahāyana sūtras, but it is often difficult to decipher; these texts were
not necessarily meant as instructional manuals. They can be supplemented
The classical synthesis
219
however by other material, such as the fragmentary Sarvāstivāda meditation
manual found at Turfan (Schlingloff 1964; see also Karlsson 1999).
Theravāda meditation in the fifth century, if the Visuddhimagga can be
taken as representative, placed considerable emphasis on the development
of samādhi (concentration or trance, perhaps better described as a focusing
of the body-mind). Samādhi here is treated mainly as equivalent to the
four jhānas25 and is developed through concentration on simple meditation subjects: a disk of clay (the earth-kasin.a) or representations of the other
elements; the foulness of corpses; recollections of the Buddha, Dharma and
Sangha; the brahmavihāras, and so on (Ñān.amoli 1991: 85–368). The earthkasin.a gets most extensive treatment (1991: 118–65); meditation on breathing, the most widespread technique among modern Theravādins along
with its modern Burmese variation, awareness of abdominal movement, is
also discussed though more briefly (1991: 259–85). The kasin.a is used as a
device to stimulate an internal visual image or nimitta that becomes the
actual object of concentration (cf. Karlsson 1999: 80). A simple and ascetic
lifestyle and quiet and peaceful surroundings are an important basis for the
practice.
Outside the Theravāda, it seems that a wider range of visualisations
developed, including visualisations of the presence of one or another
of the Buddhas, with the intention of being in that Buddha’s presence
and listening to his teachings. These fall within the general category of
buddhānusmr.ti or ‘recollection of the Buddha’. This kind of procedure
can be seen in one of the first few Mahāyāna sūtras to be translated into
Chinese (in the first century CE, perhaps a century or so after it was written), the Pratyutpannabuddhasam
. mukhāvasthitasamādhi Sūtra or ‘Samādhi
of Direct Encounter with the Buddhas of the Present’ (Harrison 1990,
cf. also Karlsson 1999: 68–70), a text which is one of our earliest witnesses for the practice of meditation on the Buddha in front of images or
paintings (Karlsson 1999: 69).26 Texts such as the Amitāyurdhyāna Sūtra
and Sukhāvatı̄vyūha Sūtra, which build up elaborate visualisations of the
25
26
There is a second series of formless jhānas mentioned in the Sūtras, but it receives much less attention
from Buddhaghosa, and one has the impression that these were not very real practices in his day.
It is interesting to note that the samādhi which is the central theme of this text is defined not simply
in terms of the five standard factors whose presence or absence are diacritic marks for the four jhānas
(vitakka, vicāra, pı̄ti, sukha, ekaggatā) but in terms of a long list of factors – Harrison counts 154 –
which form an entire conspectus of the Buddhist path: desire to hear the dharma, unbroken faith
in the Buddha’s teaching, seeing one’s sons and daughters as enemies and one’s wife as a demonness
(!), not causing schism in the Sangha, not boasting of one’s high birth, and so on (Harrison 1990:
26–30). The implication may be that if one achieves this samādhi, these qualities and attributes will
arise spontaneously.
220
Origins of yoga and Tantra
presence of the Buddha in a heavenly realm, can also be seen as a witness to
this kind of practice (cf. also Beyer 1977). Paul Williams and Robert Mayer
have discussed how visionary procedures such as those found within the
Pratyutpannabuddhasam
. mukhāvasthitasamādhi Sūtra may have provided
the mechanism by which the Mahayana Sūtras in general were received
and held to be the authentic word of the Buddhas (Williams 1989: 30;
Mayer 1996: 74–5).
It is unclear how far there was an active meditation tradition in the
urban monastic setting,27 but it seems evident that there was a continuing
meditation tradition in the ‘forest’ context and that this focused at least in
part on the establishment of visionary contact with the Buddhas of the four
directions (particularly Amitābha and Aks.obhya), with the future Buddha
Maitreya, perhaps also increasingly with deceased masters of the past (e.g.
Nāgārjuna) and with ‘Vajrasattva’ seen as the essence of all Buddhas. Legends such as that of the revelation of texts to Asaṅga suggest something
of this kind of meditation, and indicate that it may also have been associated with long-term ascetic practice in solitude, doubtless conducive to the
establishment of visionary contacts of this kind. In other words, this would
have been part of Ray’s forest monk component, not of the urban context.
The recitation of dhāran.ı̄ and mantra (ritual formulae) would probably also
have played a significant role in such practices.
Whether we choose to call this kind of meditation ‘Tantric’ or ‘protoTantric’, or prefer to avoid the label altogether, is somewhat arbitrary. There
is no reason to suppose the employment of sexual practices, let alone the
‘transgressive’ aspects of kāpālika-style practice (see Chapter 10). Nor are
there indications of actual identification with the Buddha or other deity.
The practice of visualising the presence of one or more Buddhas and receiving instruction in some form from them can be seen as a natural development from earlier forms of buddhānusmr.ti meditation. To put this in
other words, such visualisation and invocation practices formed a significant component to what would later become developed Buddhist Tantra,
but they were only one component. (A second component, the man.d.ala,
is discussed in a subsequent section of this chapter.) The kriyā and caryā
classes of Buddhist Tantras can be seen as a later extension and formalisation
of these practices in monastic ritual contexts.
27
Edward Conze famously commented that the treatment of the path to enlightenment in the Abhidharmakośa was ‘not a guide to action, but to the reverent contemplation of the achievements of
others’ (1962: 177). For all of his personal preference for the Mādhyamika, however, he was well
aware of the extent to which the Yogācāra tradition derived from ongoing meditative practice (1962:
251–7).
The classical synthesis
221
Turning to Brahmanical forms of meditation in this period, the Yogasūtra
of Patañjali, the classic meditational text of the Brahmanical meditational
tradition, has generally been regarded as our chief witness. It has become
clear in recent years that it needs to be supplemented by material in other
texts such as the Mahābhārata and Sam
. nyāsa Upanis.ads (Olivelle 1992;
Bronkhorst 1993; Brockington 2003b; White 2006).
The references to yoga in the Mahābhārata are of particular interest and
have been re-examined recently by White (2006) in relation to Upanis.adic
and early medical material. White argues against the frequent modern
interpretation of yoga, based in large part on Vivekananda’s selective reading
of the Yogasūtra, as ‘a meditative practice through which the absolute was
to be found by turning the mind and senses inward, away from the world’
(White 2006: 6). Yogic practices are about linkages between the microcosm
and macrocosm, and they postulate an ‘open’ model of the human body,
not a closed one.
In particular, White notes that the commonest use of the term ‘yoga’
in the narrative sections of the Mahābhārata is to refer to a dying warrior
transferring himself at death to the sphere of the sun through yoga, a practice
that links up with Upanis.adic references to the channel to the crown of the
head as the pathway by which one can travel through the solar orb to the
World of Brahman (2006: 7). This channel is called sus.umn.ā in the Maitri
Upanis.ad, a term that recurs some centuries later in the Tantric context.
So do the practices of transferring consciousness at the time of death,
still a major concern within Tibetan Tantric practice, where a variety of
techniques have been developed and are widely taught for this purpose
(’pho ba in Tibetan; see e.g. Guenther 1963: 197–201; Brauen-Dolma 1985;
Kapstein 1998).
Another theme in the Mahābhārata which recurs in later contexts,
including both Tibet (Guenther 1963: 201–2) and a well-known Jain meditation manual, the Yogaśāstra of Hemacandra (Qvarnström 2002), is that of
transferring one’s consciousness temporarily or permanently into another
human body (White 2006: 8–10).
The Yogasūtra of Patañjali in fact refers in passing to both these uses
of yoga (White 2006: 10), though the central concerns of this text are
more with yoga as a process of isolation from the world. The Yogasūtra
has noticeable affinities to the dhyāna or jhāna meditations of the Buddhist
tradition and has generally been regarded as strongly influenced by Buddhist
meditational procedures, with samādhi seen, as in the Buddhist practices,
as a state of withdrawal from external concerns and focusing of the bodymind. Rather than seeing this in terms of ‘Buddhist influence’, we should
222
Origins of yoga and Tantra
perhaps again see this more in terms of participation within a shared ascetic
sub-culture.28 Thus the Yogasūtra 1.17 defines an initial meditational state
(sam
. prajñāta samādhi) in terms close to the Buddhist definition of the
first dhyāna state,29 while 1.33 recommends the practice of the four states
(friendliness, compassion, sympathetic joy, equanimity) known in Buddhist
texts as the four brahmavihāra states (Satyananda 1980: 33, 57).
The dating of the Yogasūtra is problematic; Ian Whicher suggests around
the second and third centuries CE, while Gavin Flood says ‘sometime
between 100 BCE and 500 CE’ (Flood 1996: 96). Its author, Patañjali, may
or may not be the same as the grammarian Patañjali (the general consensus
at present appears to be that they are different) and in any case, as we saw
in Chapter 2, the dating of the grammarian Patañjali is far from certain.
A second to third century CE dating, or a little earlier, would, however,
put it alongside the earlier Mahāyāna Sūtras, which makes some sense (see
below).
In a series of recent writings (e.g. 1997, 1998, 2002, 2002–3), Ian Whicher
has argued that the Yogasūtra’s overall position is not a rejection of the world
but a ‘responsible engagement’ with it:
Contrary to the arguments presented by many scholars, which associate Patañjali’s
Yoga exclusively with extreme asceticism, mortification, denial, and the renunciation and abandonment of ‘material existence’ ( prakr.ti) in favour of an elevated
and isolated ‘spiritual state’ ( purus.a) or disembodied state of spiritual liberation,
I suggest that Patañjali’s Yoga can be seen as a responsible engagement, in various
ways, of ‘spirit’ (purus.a = intrinsic identity as Self, pure consciousness) and ‘matter’
( prakr.ti = the source of psychophysical being, which includes mind, body, nature)
resulting in a highly developed, transformed, and participatory human nature
and identity, an integrated and embodied state of liberated selfhood ( jı̄vanmukti).
(Whicher 2002–3: 619)
As Whicher himself implies, this is somewhat at odds with how later
Brahmanical tradition, particularly its modern variants, tends to view the
Yogasūtra, and his reading has been criticised (e.g. Pflueger 2003). However, Whicher’s position has some attractive aspects. For one thing, the
third of the Yogasūtra’s four parts explains how to develop samyama, here a
focused and exclusive identification with a specific object of concentration,
28
29
Thus the list of six components (s.adaṅga) of yoga, pran.ayāma, pratyāhāra, dhyāna, dharana, tarka, and
samādhi, first found in the Maitrı̄ Upanis.ad, is adapted and used variously in Brahmanical, Buddhist
and Jain contexts, eventually becoming a key component of the Buddhist Guhyasamājatantra
(Zigmund-Cerbu 1963).
Sam
. prajñāta samādhi is characterised by the presence of vitarka (deliberation), vicāra (reflection),
ānanda (joy) and asmitā (sense of individuality), while the first dhyāna state is characterised by
vitarka, vicāra, prı̄ti (joy) and sukha (bliss) (Bronkhorst 1993: 22–4, 71–2, 88).
The classical synthesis
223
and how to apply this state to a series of ‘objects’ (mental factors, the
body, the brahmavihāra states, etc.) in order to develop various powers and
knowledges. Many of these have obvious pragmatic utility; there seems
little point in developing mind-reading (3.19), the strength of an elephant
(3.25), knowledge of the position and movement of the stars (3.28–29) or
knowledge of the internal structure of the body (3.30) if one is simply
going to reject the world for a ‘disembodied state of spiritual liberation’.
It is here too that we find the ability to enter another body (3.38) and
to travel upwards out of the body, implicitly at the time of death (3.39,
cf. White 2006: 10). The achievement of the state of kaivalya in the fourth
and final section of the Yogasūtra also seems to refer to an internal reversal
rather than necessarily involving a cessation of existence in the world.
More significantly, perhaps, such a reading brings the Yogasūtra and its
treatment of Sām
. khya categories closer to Sue Hamilton’s reinterpretation
of Buddhist categories as primarily significant in relation to practice (see
Chapter 6, and Hamilton 1996, 2000). The yogic practitioner is less concerned with the ontology of ‘Spirit’ (purus.a) and ‘matter’ (prakr.ti) than
with the development of yogic states and the consequent inner transformation or reversal. A consistent emphasis on world-rejection is certainly
found in major figures of later Indian philosophy such as Śaṅkara but does
not need to be read into this earlier period.
The various psychic powers achieved by samyama on different objects in
the third part of the Yogasūtra are reminiscent of the numerous different
samādhis referred to in some of the Mahāyāna Sūtras dating perhaps from
around the same period or slightly later, such as the Saddharmapun.d.arı̄ka
Sūtra (‘Lotus Sūtra’). The terminology is slightly different, since the
Yogasūtra regards samādhi as one of three stages of samyama.30 The process
and result may be quite similar however; compare the practice of samyama
to achieve a vision of the siddhas or masters of the yogic tradition (Yogasūtra
3,33) with the Pratyutpannabuddhasam
. mukhāvasthitasamādhi Sūtra’s development of the vision of the Buddhas (see above). More generally, the cultivation of these states suggest a pragmatic aspect to the yogic tradition,
Brahmanical or Buddhist, in which renunciates, like modern Indian sādhus,
are expected to have achieved special powers which enable them to help
their lay followers and clients.
Little has been written on Jaina meditational practices in this period. As
we saw in Chapter 6, the orientation of the Jaina traditions was towards the
30
The others are dhāranā (concentration on the object) and dhyāna (exclusive focus on the object).
Samādhi, defined as identity with the object of concentration, is the third and highest stage of the
three.
224
Origins of yoga and Tantra
cessation of all activity, above all that of the mind. A series of four meditational states (śukladhyāna) seems to have formed part of the path, at least
in ideal terms, from early times (Qvarnström 2003: 131–2), although the
Jaina appear to have emphasised external ascetic practice as their primary
technique; Qvarnstrom notes that the theoretical requirements for undertaking the śukladhyāna are such as to exclude their employment (2003: 140
n. 10). Dundas comments that Jainism ‘never fully developed a culture of
true meditative contemplation’ (Dundas 2002: 166) and that
It is difficult to avoid the conclusion that later Jain writers discussed the subject only
because participation in the pan-Indian socio-religious world made it necessary to
do so. (Dundas 2002: 167)
This is perhaps unfair to Hemacandra, who claims in his great twelfthcentury work on the Jain path, the Yogaśāstra, to be writing at least in
part from his own experience. The techniques he describes, however, show
signs of the Tantric approaches of Abhinavagupta and of the Nāth Siddha
tradition, rather than deriving from earlier Jaina practices (Qvarnstrom
2002: 12–13, 2003: 134–8). I will return to the Yogaśāstra in Chapter 12.
the m a n. d. a l a
I turn now to what was eventually to become one of the most conspicuous
elements of Tantric practice, the man.d.ala. The man.d.ala is mostly familiar to
us now in the Tantric Buddhist form, particularly in the elaborate versions
of the ‘Anuttarayoga’ practices,31 culminating in the Kālacakra with its 722
deities (e.g. Bryant 1992; Leidy and Thurman 1997: 98–9). The yantras
of Śaiva Tantra form a close parallel, and, as we will see, there were also
Vais.n.ava equivalents. However, as Stanley Tambiah, Ronald Davidson and
others have pointed out (Tambiah 1985; Davidson 2002a), the man.d.ala
is also a social model – it retains the meaning of administrative district
in modern Indic-derived states such as Thailand (cf. Hanks 1975) – and
its early versions were not ‘Tantric’ in any of the usual meanings of that
term.
In fact, man.d.ala-type structures appear to go back to quite an early date
in the history of Indic religions, and similar structures can be found in many
other cultures. I noted this in my book Mind, Body and Culture, where I
cited Australian Aboriginal examples among others (Samuel 1990: 90, 170
31
As I explain in Chapter 10, the term ‘Anuttarayoga’ appears to be based on an incorrect re-translation
from Tibetan back into Sanskrit. However, it has become so widely established in the Westernlanguage literature that it would create more confusion not to use it.
The classical synthesis
225
n. 4). Nevertheless, there are differences between these various man.d.alatype structures, and an examination of these will be of some significance.
The most critical difference, it seems to me, is the extent to which the
man.d.ala-type structure has a centre, and the degree to which the noncentral components are treated as emanations of that centre and reducible
to it. This sense of the outer elements as emanations or projections of an
underlying unity is an essential component of the fully-developed man.d.ala
scheme.
Thus, deities associated with the four directions can already be found
in early Vedic material. In the Br.hadāran.yaka Upanis.ad however, we find
something different: an underlying unity behind the deities of the directions. Interestingly, it seems to be signalled as new and unusual in the text:
Yājnāvalkya explains that he is able to defeat the Brahmins of Kuru and
Pañcāla because he knows ‘the quarters together with their gods and foundations’ (Olivelle 1998: 49 = Br.hadāran.yaka Upanis.ad 3.9.19). He explains
that the Sun (Āditya) in the East is founded on sight, which is founded
on visible appearances, and so on the heart (hr.daya); Yama in the South
is founded on the sacrifice (yajña), which is founded on the sacrificial gift
(daks.ina), which is founded on faith (śraddha) which is again founded on
the heart; Varun.a God of the West is founded on water, semen and again
on the heart; the Moon (Soma) in the North on the sacrificial consecration
(dı̄ks.ā), which is founded on truth (satya) and again on the heart. Finally
Agni (Fire) at the zenith or centre is founded on speech (vāc), which is
founded on the heart.
This proto-man.d.ala, if we can call it that, is from a Brahmanical context.32 The earliest Buddhist proto-man.d.ala known to me, which is also
the first reference to the Buddhist scheme of the four Buddhas of the four
directions, is considerably later. It occurs in a well-known and important
Mahāyāna sutra (not a specifically ‘Tantric’ text), the Suvarn.aprabhāsa Sūtra
or ‘Sutra of Golden Light’. This dates to some time before the beginning
of the fifth century, since it was translated into Chinese between 414 and
421 CE. In the second chapter of this sūtra, the Bodhisattva Ruciraketu (or
Śraddhāketu) is meditating on the Buddha – specifically, on why it was that
the Buddha Śākyamuni only had an ordinary human life-span of eighty
years.
As he does so, his house undergoes a visionary transformation into a vast
building made of beryl and with divine jewels and celestial perfumes. Four
32
For another early man.d.ala scheme from a Brahmanical context, of a rather different kind, see the
discussion of the Atharvaveda’s Vrātyakān.d.a in Chapter 10.
226
Origins of yoga and Tantra
lotus-seats appear in the four directions, with four Buddhas seated upon
them: Aks.obhya in the East, Ratnaketu in the South, Amitāyus in the West
and Dundubhiśvara in the North. They explain to him that the life-span
of the Buddha Śākyamuni was not eighty years; its true limit is far beyond
the understanding of any human or divine being. (J. Huntington 1987; also
Emmerick 1970: 3–4.)
If this is a reflection of actual meditation practices, as seems likely, then
these can be seen as a logical extension of the meditative procedure involving the summoning of the presence of the Buddha that we saw in the
Pratyutpannabuddhasam
. mukhāvasthitasamādhi Sūtra. This text from the
Suvarn.aprabhāsasūtra and another text translated into Chinese at around
the same time (Matsunaga 1977a: 174) provide the earliest references so far
known to a set of four Buddhas in the four directions, although Huntington suggests that we can see hints of such structures in iconography back to
the second or first centuries BCE. This schema occurs in many more sūtras
over the next couple of centuries. If we add a central Buddha, this yields the
five-fold structure that later became familiar as the five-tathāgata man.d.ala,
and forms a basic core for most later Buddhist man.d.alas (Matsunaga 1977a:
174–5).
On the Vais.n.ava side, the sculptural expressions of the vyūha scheme
of the Pāñcarātra system, where deities in the four directions are seen as
emanations from an underlying more complete divine nature, would seem
to have much the same structure. Our earliest evidence for this scheme may
be an image that Doris Meth Srinivasan dates to the first half of the second
century BCE on stylistic grounds. As with Huntington, her interpretation
may be optimistic, and Hans Bakker has cast some doubts on it (Srinivasan
1979 and 1997; Bakker 1999). The Pañcaratra tradition did, however,
develop a man.d.ala-type structure based around the vyūhas, with the four
directional deities corresponding to four of the Bhāgavata deities, including
Kr.s.n.a Vāsudeva himself, Sam
. kars.an.a, Pradyumna and Aniruddha.
The Bhāgavata deities, as I mentioned earlier, are generally supposed
to have begun as tribal deities of the Vr.s.n.is and were later patronised by
the Kus.ān.a kings. In this case one can again the see political point of the
imagery as well as its philosophical meaning, particularly if we assume that
the Bhāgavata deities started off as deities of different Vr.s.n.i groups and
were only gradually incorporated into the vyūha scheme.
In any case, the idea of a man.d.ala in which the central figure represents
a supreme deity, and the directional figures are subordinate aspects, seems
a natural development in view of the political situation of the time. It
closely reflects the idea of a supreme king at the centre in relation to whom
The classical synthesis
227
lesser kings are expected to be local projections rather than independent
rulers (cf. Davidson 2002a). We can see this at a somewhat later stage, I
would suggest, in the imagery of Buddha Vairocana (sometimes replaced by
Śākyamuni or Amitābha) at the centre of a man.d.ala of eight bodhisattvas,
which rapidly became one of the commonest iconographical schemes in
China and Tibet from the sixth century onwards (Heller 1994; Leidy 1997:
26–8). Vairocana’s imperial imagery clearly proved tempting to rulers of
the time.
This is perhaps enough to establish my line of argument. The fullyfledged Tantric man.d.ala is a model of a specific kind of state, what Stanley
Tambiah has called a galactic polity (Tambiah 1985). In such a state, in
principle, the centre is reduplicated at the various regional capitals, where
local rulers imitate on a lesser scale the splendour of the royal court at the
main centre. In reality, of course, the allegiance of the lesser rulers to the
main centre may change quite rapidly if the centre is unable to maintain
its role. Much of what one is seeing is a kind of theatrical performance, but
it is a theatrical performance that suits the interests of both the central and
lesser rulers.
One point is worth emphasising though. While these kinds of schemes
have conventionally been called ‘tantric’, we are in a different world here
from the kind of wild, aggressive, antinomian, and deliberately shocking
imagery that is often associated with Tantra. The origins of this kind of
Tantric material are the subject of Chapter 10. Tantra in the sense of the
Pāñcarātra or of the Buddhist Tantric texts we are discussing here is about
the ritual worship of divine beings who – at least once the new state has
been established – are very much aligned with the respectable moral order.
On the Brahmanical side, this kind of imagery allowed for the creation of a whole series of new ritual procedures for the new deities associated with the Brahmanical states of this period. In the Buddhist context,
things are not all that different in relation to the ‘early’ Tantras, those
later allocated to the kriyā and caryā classes. Many of the Tantric texts of
this class are not much different from the Mahāyāna Sūtras, except for a
greater focus on ritual procedures, and in fact it is not that easy to draw
the distinction between sūtra and tantra at this stage. Several texts that
are classified as tantra in, for example, the Tibetan canonical collections,
are in fact described as sūtra in their title, including the principal text
for the Vairocana cult, the Mahāvairocanābhisam
. bodhi Sūtra (Yamamoto
1990; Wayman and Tajima 1992; Hodge 2003). The Mahayana Sūtras from
quite an early date placed great emphasis on the cultivation of visionary
modes of meditation, and themselves give the appearance of being the
228
Origins of yoga and Tantra
result of such modes, as suggested above. These sūtras also provided the
basis for a whole series of new modes of ritual, in which Buddhist deities
who partook directly in the power of the Buddha could be called on for
assistance – rather than just yaks.as who acted under Buddhist auspices. In
other words, one could access the power of the Buddha himself.
I am not suggesting that this was all some vast conspiracy on the part of
self-interested Brahmin priests and Buddhist monks to ensure the patronage
of the new rulers. One cannot dismiss such possibilities altogether, but as
noted elsewhere I would go rather for a model based on what Max Weber
called many years ago an ‘elective affinity’ (Weber 1976). He was talking
about the relationship between the Protestant Ethic and the development of
early capitalism, and saw these as developing a relationship on the basis of a
kind of mutual sympathy. One can imagine something similar here: one can
see a wide variety of religious models being elaborated at this time in both
Brahmanical and non-Brahmanical contexts. Some of these would prove
attractive to the new rulers, and would be supported. However, as rulers and
wealthy patrons saw resources that they could use, or religious modes that
appealed to their self-image, these resources and modes would gain new
prominence, and become more and more part of people’s consciousness,
in a process of mutual reinforcement. Alternatively, one could regard it as
a kind of resonance or mutual harmony, in which increasingly everybody
ended up singing the same song, or finding themselves excluded.
These new religious modes allowed for a much closer fit between religion
and the new social and political forms that were developing. Over time, the
religious, political and social modes of the new Brahmanical state combined
into a powerful and convincing mixture that enabled Indic cultural patterns
to be widely exported and adopted, indeed to seem the natural way to run
a state throughout much of South and Southeast Asia for many centuries.
The ‘respectable’ and positive aspects of kingship and the state that we
have seen in this chapter were, however, only part of the picture. In Chapter
10 we turn to their dark complement, the forces of disorder and misfortune.
These too were an important part of the evolving model of the Indic state.
chap t e r 10
Tantra and the wild goddesses
If much of Chapter 9 dealt with the cult of respectable, high-status deities,
linked with the new image of kingship in the new Brahmanical and
Buddhist states of the post-Mauryan and the Gupta periods, on both
śraman.a (Buddhist) and Brahmanical sides, Chapter 10 introduces religious forms of a less respectable and more ambiguous kind. The practice of
these forms of religion was for the most part, initially at least, the preserve
of more marginal groups. It will become evident, however, that they were
not unimportant for all that, and they were to become more important
over the following centuries.
Tantric studies, and especially Tantric historiography, as Urban (1999),
Wedemeyer (2001) and others have noted, has, like other fields of study,
been historically dominated in various ways by Western fantasy (partly
shared by Indian scholars under Western influence). The term itself is
problematic, particularly in the Śaiva, Vais.n.ava and Jaina contexts, though
less so for the Buddhists where the Tibetans had an explicit category of
works in the canon labelled ‘Tantra’ (rgyud). Many of the Śaiva works that
are now customarily labelled ‘Tantric’ might have been called kaula or by
other labels, particularly before the tenth and eleventh centuries; Tantra as
a standard category of Śaiva texts emerged only at a fairly late period.
Defining Tantra remains a problematic business; the articles in Katherine
Harper and Robert Brown’s collection The Roots of Tantra (2002) spend
quite a few pages on this topic without providing much enlightenment.
Indigenous terminology is not necessarily very useful: tantrika in the Brahmanical context can mean not much more than ‘non-Vedic’, whereas the
Tantra sections of the Tibetan Buddhist canon contain a very wide variety of material, varying from the invocation of peaceful deities by quite
conventional ritual procedures to practices which, at least in a literal interpretation, are highly transgressive of any kind of conventional morality. It
is primarily these transgressive styles of practice with which I am concerned
in this chapter.
229
230
Origins of yoga and Tantra
The first Western scholars to discuss Tantra had little doubt about what
the category referred to, or about their attitude to it. The initial theme
and one of the most dominant until modern times was outright rejection
and condemnation. The oft-quoted comments of William Ward (writing
in 1817) sets the tone:
Many of the tantrus [sic] [. . .] contain directions respecting a most [. . .] shocking
mode of worship. [. . .] These shastrus [sic = śāstras] direct that the person [. . .]
must, in the night, choose a woman as the object of worship [. . .] Here things too
abominable to enter the ears of man, and impossible to be revealed to a Christian
public, are contained. [. . .] The priest [. . .] behaves towards this female in a manner
which decency forbids to be mentioned. (Ward, cited in Urban 1999: 128)
What upset Ward above all was presumably the use of sexual practices in
Tantric ritual. Ward was a Baptist minister, though from his other writings
I suspect that he may not have been quite as prudish as this quote suggests.
He may have felt some need though to be cautious in the face of the possible
reactions of his ‘Christian public’.
Horace Wilson, writing fifteen years later for a more scholarly readership
perhaps, takes the other main line, which is polite distance. Certainly he
deals with the subject much more calmly, though leaving most of the details
safely in Sanskrit. He notes that Ward’s account is based on hearsay, but
observes that while such practices are ‘certainly countenanced by the texts’
they are in his opinion ‘practised but seldom, and then in solitude and
secrecy’ (H. Wilson 1832: 225). He comments that
although the Chakra [or Tantric ritual gathering] is said to be not uncommon [. . .]
it is usually nothing more than a convivial party, consisting of the members of a
single family, or at which men only are assembled, and the company are glad to eat
flesh and drink spirits, under the pretence of a religious observance. (H. Wilson
1832: 225–7)
He also adds that
In justice to the doctrines of the sect, it is to be observed, that these practices, if
instituted merely for sensual gratification, are held to be as illicit and reprehensible,
as in any other branch of the Hindu faith. (H. Wilson 1832: 227)
Wilson was right on this point, though whether Tantric cakras in the early
nineteenth century were really no more than an excuse to indulge in meat
and drink is hard to tell at this distance.
In more recent years, of course, the level of scholarship has increased
significantly – though Horace Wilson’s account as a whole is impressively
knowledgeable for 1832, and reminds one that early colonial writers can
Tantra and the wild goddesses
231
provide valuable perspectives on the Indian religion of their times, which
preceded the great reform movements of the nineteenth century. The main
additional option in contemporary discourse however is that we now have
a minority of Western yogic and Tantric followers, with affiliations to a
variety of Asian traditions, often in forms that have been thoroughly
repackaged and redigested for today’s world – either in the Asian or Western
context. This leads to another kind of rejection of Tantra as presented in
the earlier Indian texts, by contemporary scholars who wish to present the
tradition in forms more in accord with what they have learned from their
lamas or gurus.
The issue of sexual practice remains a problem, both for those who find
it objectionable and for those who find it attractive. The problem here is in
part that sexuality undoubtedly has different meanings for a contemporary
Western or Asian scholar, or a contemporary practitioner of Buddhist or
Śaiva Tantra, from those that it had in the period when these practices
developed. The physiology of sex is undoubtedly much the same now as it
was then, but the meanings attached to the process are not. In particular,
while our contemporary concerns tend to focus around immorality, a more
important issue at the time was probably that of impurity and pollution.
From this point of view, also, there are wide differences between various
kinds of sexual practices, and we need to be careful about lumping them
all together.
Arguably, while the early kula practices considered later in this chapter
were regarded as transgressive in their own time, the sex itself was a relatively
secondary consideration. What was really at issue was the impurity and
pollution attached to sexual substances and the magical power that might
in consequence be generated by using them in a transgressive fashion.
Later Tantric sexual practices, in both Buddhist and Śaiva traditions, work
differently, and sexual practice is used as a way of generating particular kinds
of experience that are seen as conducive to the generation of liberating
insight. These practices will be discussed in more detail in Chapter 11.
Arguably, though, Tantra did not become primarily concerned with the
spiritualisation of sex until it came to the West (Samuel 2005a: 345–66).
Modern Western Tantra, for all its dubious lineage and frequent commercialisation, is not necessarily a bad thing, and I have argued elsewhere
that it deserves to be taken seriously as a cultural development within
Western society (Samuel 2005a: 357–61). However, it is operating in a very
different context to the Indic societies we are considering here, and it
is at least as much a product of the West and of modernity as of Indic
traditions.
232
Origins of yoga and Tantra
We have quite a lot of material, one way or another, on ‘transgressive’
styles of Tantra. Historically, these transgressive styles of ‘Tantra’ developed
in both Brahmanical (mainly Śaiva) and in Buddhist contexts. Modern
Hindu scholars have often tended to blame Tantra on the Buddhists, while
Buddhist scholars have either returned the compliment, or argued with
varying degrees of plausibility that Buddhist Tantric practice (unlike that
of the Hindus) was pure and compassionate, with only the most elevated
of aims. In fact, it is easy to find both the elevated pursuit of enlightened consciousness, and the pragmatic exercise of power, on both sides
of the Hindu-Buddhist Tantric divide. In any case, Śaivas and Buddhists
borrowed extensively from each other, with varying degrees of acknowledgement (Sanderson 1994, 1995, 2001; Bühnemann 1996, 1999, 2000a; see
also Davidson 2002a: 202–3, 206, 386 n. 105, 388 n. 120).1 Vais.n.ava and
Jaina traditions were also involved with ‘Tantric’ material, although in a
more selective way; neither appears to have adopted Tantric practices in
transgressive and antinomian form as a pathway to spiritual realisation,
although both incorporated elements associated with these practices in less
controversial forms.2
I doubt that we will ever be able to trace the precise origins of the wild,
transgressive, antinomian strain in Tantra in detail. However, we can follow
a considerable part of the development of this religious style. The question,
I think, is not just that of magical ritual for pragmatic ends. This has been
a commonplace in virtually all human societies, and the Vedas, particularly
the Atharvaveda, have plenty of examples, as do the Brāhman.as and even
the Upanis.ads. What I want to understand is rather more specific.
While, as noted above, I have generally avoided the tricky question of
defining Tantra, what I am interested in here can, however, be specified
relatively easily. We have a body of ritual practices and traditions, in both
Buddhist and Śaiva sources, that present themselves as sophisticated and
elevated means for the attainment of exalted spiritual goals, yet contain constant reference to practices that seem deliberately transgressive and bizarre:
1
2
For other examples of material apparently shared between Śaiva and Buddhist Tantra, see Benard
1994 (and Orofino 1998); Stablein 1976, which presents clearly Śaiva material in a Buddhist context;
Schaeffer 2002; Walter 1992, 1996, 2003; White 1996: 72–3. See also White 2005: 9.
For Pāñcarātra body practices, see Flood 2006: 99–119; for equivalent Jaina practices, see Qvarnström
2002, 2003. In later centuries, the Vais.n.ava Sahajiyā tradition of Bengal employed sexual practices
(e.g. Hayes 2003), as did the Bauls, whose theology was part Vais.n.ava, part Muslim (see e.g. Salomon
1991; Das 1992). The Jainas used a variety of fierce goddess rituals; for a survey see Cort 1987:
255. There were other traditions apart from those mentioned here that might be included within
an extended definition of ‘Tantra’, for example the Gan.apatyas, some of whom also engaged in
‘transgressive’ practices, but our knowledge of them is much more limited (see e.g. Courtright 1985:
218–20; Thapan 1997, esp. 176–97).
Tantra and the wild goddesses
233
night-time orgies in charnel-grounds, involving the eating of human flesh,
the use of ornaments, bowls and musical instruments made from human
bones, sexual relations while seated on corpses, and the like.
There is little doubt that at least some contemporary practitioners of
these traditions are what they claim to be: people whose lives are indeed
dedicated to reaching high levels of spiritual attainment, and whose lives
are guided by high moral principles. Having said this, there remains, in
the Indian context at least, a divide between Tantric practitioners who are
now respectable members of the wider society and those who are not. One
can find South Indian or Orissan Brahmins who carry out rituals derivative
of earlier ‘transgressive’ modes of practice in ways that are not in the least
transgressive, but one can also find naked Aghoris, smeared with ash, eating
and drinking from a human skull-cup and dwelling in cremation-grounds
and other such unsavoury places. In Indian parlance, this is figured in
terms of ‘right-hand’ and ‘left-hand’ styles of practice, terms which are
rather more loaded in India than they are in the West, since the right hand
is used for eating and pure functions, and the left for impure and polluted
tasks.
This suggests that, in India, if one looks from outside the tradition,
transgressive ‘Tantric’ practices were often viewed quite negatively. This is
evident from the mostly negative depictions of early ritual practitioners in
the kāpālika style as evil, horrifying or comic characters in Sanskrit drama
and literature from the sixth or seventh centuries CE onwards.3 The same
cliché is alive and well in Bollywood movies (such as Nagina, made as
recently as 1986).4 This negative stereotyping does not, however, mean that
people did not consult Tantric ritual practitioners in the past, or that they
do not consult them today. Even in societies, like contemporary Tibet or
Bali, where Tantric ritual practitioners are indeed highly respected by the
lay population among whom they live, much of what they seem to do in
practice, at least in relation to lay people, is the performance of pragmatic,
this-worldly ritual for the health and well-being of their following.
All this may seem a contradictory situation, but it is perhaps not quite
as anomalous as it might seem. At least one component of the mixture,
the ambivalent status towards those who are believed to possess magical
power, is familiar from recent studies of shamanic practitioners in many
different societies. I give a couple of instances. The shamanic practitioners
of the Peruvian Amazon, studied a few years ago by Michael Fobes Brown,
3
4
With occasional exceptions, as in the sympathetic portrayal of Bhairavacarya in Ban.a’s Hars.acarita
(Lorenzen 1991: 20).
For a summary of the plot, see Kakar 1990.
234
Origins of yoga and Tantra
are certainly thought of as possessing power, but those who consult them
for healing are never sure if they can be trusted to use it on their behalf.
Brown transcribes a consultation in which the tension between clients and
shaman is palpable, with the husband of one of the women seeking healing
threatening the shaman with violence if he does not heal his wife (Brown
1988).5
Again, there are tens of thousands of shamanic practitioners in modern
Korea, but people consult them, according to a recent study by my former
student Kim Chongho, with reluctance and distaste, and only when no
other solution is available for their problems (C. Kim 2003). Nobody would
admit in respectable company to visiting a shaman or to taking what she has
to say (most Korean shamans are women) seriously. Shamanic practice is
shameful, dirty and disreputable. To be compelled by the spirits to become
a shaman is a real misfortune, and condemns one to the lowest level of
Korean society.
Other anthropologists, working, perhaps significantly, for the most part
with shamans rather than with their clients, have produced a more positive
picture of shamanic ritual and its meanings, although presenting the same
basic outlines of the situation (e.g. Kendall 1985, 1988 for Korea). There has
been a recent tendency to celebrate shamanism as a Korean art form, and to
emphasise the high level of accomplishment of shamanic drama, song and
music, although it has had little impact on the situation of most Korean
shamans, who continue to be regarded as disreputable and of low status.
Nevertheless, the myths told in the songs of the shamans, such as that of
Princess Pari, who is a kind of founding ancestor of the shaman tradition,
emphasise their compassionate and altruistic nature, and their willingness
to undergo suffering for the good of others (T. Kim 1988).
Readers should not be too concerned by the fact that the practitioners in
both my examples were called ‘shamans’ by the anthropologists who studied
them. It would be possible to find similar cases with a variety of other labels:
marabouts in North Africa, perhaps, diviners or nganga in Southern Africa,
brujos or curanderos in Latin America. The significant issue is that these are
all, indeed, practitioners, with high levels of performative skills, who are
carrying out pragmatic services for the surrounding population, yet are not
really trusted by them. Also, there are suggestions in many of these cases
that there is an ‘internal’ perspective, held by the practitioners themselves,
in which what they do is seen as characterised by high levels of compassion.
5
Yet Michael Harner, who worked with a closely-related group of shamans, also in the Peruvian
Amazon, presents them simply as repositories of tribal wisdom and transmitters of a valuable body
of knowledge (Harner 1982).
Tantra and the wild goddesses
235
These people see what they are doing as spiritually rather than materially
motivated.6
I think that for the earlier, transgressive stages of Tantra we are indeed
dealing with magical practitioners who were both valued and feared. What
is more surprising, perhaps, is that much Tantric practice, both in India and
elsewhere, has become a respectable and valued mode of spiritual activity.
I will try to explain how this came about in the course of the following
chapters. As the above material suggests, and as I have commented earlier
in the book, it is worth taking both ‘internal’ and ‘external’ perspectives.
In other words, we need to understand why practitioners might see these
forms of practice as worthwhile, and we have to understand why those who
supported and paid for them thought they were worth supporting.
the state and the powers of d arkness
I start here with the ‘external’ perspective, more specifically with the question of why kings and rulers might see this involvement with the dark
powers as appropriate and useful. George Hart’s suggestions regarding the
origins of caste in South India provide a convenient starting point (Hart
1987). I referred to Hart’s work briefly in Chapter 4. Hart was dealing
with hereditary practitioners, not people who have chosen to take on a
renunciate lifestyle, but as we will see the boundaries here are not clearcut;
Śaiva Tantric texts speak of kula traditions (hereditary or clan traditions)
and kaula (traditions derived from hereditary or clan traditions), though
the precise nature of the kula or clans in question is unclear (White 1998,
2003).7 In fact, kula and kaula are more common labels for these traditions
than ‘Tantra’ itself in the earlier Śaiva context, though I shall continue to
use ‘Tantra’ below for convenience.8
Hart argued that the South Indian king was surrounded by a number of ritual specialists from low-status castes (drummers, bards, spiritmediums, funeral priests, perhaps also priests of the Tamil deity Murugan)
6
7
8
As do such devalued and underpaid secular practitioners as the dāı̄ or traditional birth assistants in
North India and Bangladesh (Rozario 2002, and Chapter 4 above).
The kula in later times is a spiritual clan or lineage rather than one of hereditary practitioners, and it
may be possible that this was also true of these early kula, though the shift from kula to kaula would
seem to imply a move, as White suggests, from hereditary to spiritual lineage. The idea of a hereditary
‘substance’ passed down a genealogical lineage may also have existed and was certainly present in
Tibet in later times (gdung brgyud, cf. Aziz 1978: 53; Samuel 1993: 279, 284). Cf. also Wayman 1961,
who sees elements of ‘totemic’ clans in the Buddhist Tantras.
White prefers to reserve the term for what he refers to as ‘high Hindu Tantra’, the result of the reforms
of Abhinavagupta and others in the eleventh century (White 1998: 173).
236
Origins of yoga and Tantra
who played a crucial role in relation to his power. The king was dependent
on these people to maintain his power, and there was a kind of economy
of power in which death and misfortune were continually processed into
auspiciousness:
The king can be seen as a sort of massive converter of power: he is supplied with
disordered and dangerous power, and, with the proper ceremonies and institutions,
he converts that dangerous power to its ordered and auspicious analogue, thus
keeping his land in a flourishing condition. This explains one of the most common
similes in the early war poems: the comparison of the process of war to agriculture.
The enemy army is watered with showers of arrows, then cut down like grain; their
bodies are heaped up like haystacks; then their bodies are scattered and trampled
with elephants just as the grain is scattered and threshed by the tread of buffaloes.
The most unimaginable scenes of death and horror are, like the process of sowing
and harvest, the source of future fertility and prosperity. (Hart 1987: 473)
The king’s royal battle-drum (muracu) was a critical element of his power, so
much so that its possession was held actually to confer title to the kingdom.
Hart notes that ‘the most important element connected with the drum was
the god – or spirit – who was supposed to inhabit it, and who had to be
kept in the drum and made happy with blood and liquor sacrifice’ (Hart
1987: 475).9
Thus Hart argues that
[t]he role of groups ‘of low birth’ in ancient Tamilnad was to control, or order,
dangerous power. [. . .] [they] had an inborn ability to deal with dangerous power,
whether in the form of dead spirits or just dangerous power. [. . .] This is why the
performers worked in symbiosis with the king: the king had need of the forces
that the performers could summon; they fed his nature and made him a proper
king, able to lead an army and kill in battle. But only the king could transmute
those forces into the beneficial aura that kept disaster away from his kingdom.
(Hart 1987: 476–7, 478)
Hart regards the Brahmins, when they arrived in the south in around the
third and fourth centuries CE, as having provided a ‘new and revolutionary’
element; like the low castes, they could be sources of auspicious power, but
their power was ‘ordered and benevolent’ and did not need transmuting
and controlling by the king. In addition, it did not depend on the constant
warfare that had been endemic in the South until that period; it offered
the possibility of a more peaceful social order, resembling that established
by the Guptas and other northern dynasties.
Hart’s arguments are persuasive in relation to South Indian preBrahminical concepts of power and kingship. Where I might differ is in
9
It is worth comparing the hymns to the battle-drum in the Atharvaveda (5.20, 5.21). These provide
no indication of ideas of this kind (Stutley 1980: 71–2; Griffith 1916: vol. 1, 220–3).
Tantra and the wild goddesses
237
suggesting that such ideas of the necessity of channeling disordered and
transgressive power for the good of society as a whole may already have
been present elsewhere in South Asia at this period. In modern times, ritual drummers with clear parallels to those described by Hart are significant
within Tibetan societies in Ladakh and Northern Nepal and again form a
distinct low-status grouping (see Chapter 4).
In addition, there is plenty of evidence that such ideas were present
in North India as well, at least from Gupta times. The Arthaśāstra, which
was probably compiled between the mid-second and fourth century, clearly
endorses the use of sorcery in support of political ends. This appears to have
been in particular the role of the royal purohita, an Atharvavedan priest who
was entrusted with protective rituals on behalf of the king and also with
the employment of ‘secret practices’ in support of his political ends.10 We
will shortly see evidence for Brahmanical ritual of the Gupta period that
employed transgressive and destructive powers for political and/or military
purposes.
exercising power: the v r ā t ya k ā n. d. a
of the atharvaved a
First, however, I return to the vrātyas, who seem to develop into a ‘Vedic’
equivalent of Hart’s South Indian low castes. As I explained in Chapter 5, the
vrātya tradition appears to have begun as the performance of a pragmatic
and somewhat transgressive ritual by a sub-group of late Vedic society,
perhaps an age-grade of young men. We might date this phase to the tenth
to eighth centuries BCE on Witzel’s chronology. By the time of the Laws
of Manu (around first century CE), the vrātyas seem to have become a
permanent outcaste group. If they still existed as a social reality rather than
an abstract category, they presumably continued to make a living as ritual
performers.
An Atharvavedic text relating to the vrātyas perhaps relates to a stage some
way along this hypothetical development. This section of the Atharvaveda,
the so-called Vrātyakān.d.a, forms Book Fifteen of the Atharvaveda in the
Śaunaka version11 and consists of a single short text in prose, in a style resembling that of the Brāhman.as. It is generally regarded as a late part of the
Atharvaveda, so might be dated between the sixth and third centuries BCE.
10
11
Michael Willis (personal communication, 2004). I would like to thank Michael Willis for giving
me access to some of his research work in this area, which is to be published shortly. On the date of
the Arthaśāstra, see Willis 2004: 25 n. 114.
The Atharvaveda has been transmitted in two rather different versions, the Śaunaka and Paippālada
recensions. Only the first section of the Vrātyakān.d.a is included in the Paippālada version.
238
Origins of yoga and Tantra
The first part of the text describes a vrātya who ‘rouses’ Prajāpati and
becomes Mahādeva (presumably meaning Śiva). After this, he is referred to
as the Chief vrātya and the holder of Indra’s bow. In the next section he is
described as ‘moving’ in the four directions, and various occult connections
or correspondences are described, much in the style of early Upanis.ads such
as the Br.hadāran.yaka or Chāndogya. The vrātya then sits on his ritual seat,
and further correspondences are stated in the subsequent sections to the
components of his seat and to the six seasons of the year. He is provided with
six gods as attendants in the four directions, zenith and nadir, and then
‘moves’ into a series of further directions, with further correspondences
including the earth, gods, sacrificial fires and such.12
The remaining sections describe how a vrātya knowing this knowledge
should be treated by kings and other people – in both cases, very respectfully – and provide some additional correspondences, including sets based
on his various internal breaths (prān.a) and on parts of his body.
Early Western translators of the Atharvaveda were rather dismissive of
this text: Whitney (1905: vol. 2, 770) calls it ‘puerile’ while Griffith describes
it as a ‘unique and obscure Book’ whose purpose is the ‘idealization and
hyperbolical glorification of the Vrātya’ (Griffith 1916: vol. 2, 185). A more
recent translation by Sampurn.ānand13 (1956) treats it essentially as a meditation on the nature of Śiva. I would suggest that it might be both ‘glorification of the Vrātya’ and ‘a meditation on the nature of Śiva’; in other words
that it is essentially, at least as far as the first five sections are concerned, a
liturgy for the self-identification of a vrātya practitioner with Śiva. If read
in this way, the resemblance with the much later deity yoga practices of
Śaiva and Vajrayāna Tantra is quite striking. The remainder of the text sets
up further correspondences that might be used in specific ritual contexts,
and also prescribes the appropriate behaviour for kings and other people
whom the travelling vrātya visits.
If this interpretation of vrātya ritual as involving the vrātya’s selfidentification as a deity is plausible,14 then this is the first example known
to me of this significant theme. We can perhaps also see in it an initial
occurrence of an important structural relationship that will come up again
in later chapters. This is the situation of the renunciate or outsider who is
12
13
14
Thus the text provides another version of an early man.d.ala scheme, though of a rather different
kind from those discussed in Chapter 9.
I.e. the Hon. Sri Sampurn.ānand, who was the Chief Minister of Uttar Pradesh at the time.
Such an interpretation is made more plausible if one thinks of the process of incorporating animal
spirits into the body for healing purposes in some South Asian folk healing (‘shamanic’) traditions
(e.g. the Magar, de Sales 1991).
Tantra and the wild goddesses
239
viewed ambivalently or negatively by the surrounding ‘respectable’ society,
while at the same time valued or even viewed as necessary for their ritual
services. The members of the outsider group themselves, however, have a
different and more positive self-image focused on their possession of spiritual power, and the self-identification with Śiva that seems to be at the core
of this text suggests how this might work.
This is a widespread cultural pattern both in South Asia and elsewhere,
and it can be seen clearly in cases such as the Pāśupatas and the kāpālikastyle ascetics (see below). In such contexts, the vocabulary of ‘magical power’
may be in part a question of conscious trickery, but it is unlikely that it
is only this. The performance carries conviction because the members of
the outsider group themselves take it seriously to at least some degree, and
because they have acquired the genuine skills needed to carry it off.
So far, we have an opening for ritual practitioners, and we have hereditary
groups (low caste groups in South India, vrātya in the North) filling the role.
What is also significant is that we find renunciate ascetic groups taking over
similar functions over the following centuries. This is perhaps not entirely
surprising. As we have already seen, the renunciates have placed themselves
outside society. I suggested in Chapter 6, in relation to the Buddhists
and other early śraman.a orders, that renunciate ascetics placed themselves
outside ordinary society in formal terms, however closely they might in fact
interact with it subsequently. In practice, there was an ongoing dialectic
between forms of renunciate practice that became more or less assimilated
to the needs of society, and others that positioned themselves at the edges
or outside society.
For renunciates who felt it appropriate to operate in the latter, more
marginal position, the status of a low-caste practitioner of sorcery and
dubious magical practices made some sense as a new identity. It capitalised
on the powers associated with ascetic practice, while positioning the renunciate firmly outside respectable society. The early willingness to take over
the cult of dead spirits in marginal places outside the cities (see Chapter 6
and deCaroli 2004) suggests that Buddhists initially occupied a somewhat
similar role.15
15
This is reflected only to a limited degree in the early Buddhist and Jaina texts. One might explain this
in a variety of ways, such as the fact that these aspects of the role were largely taken for granted, or
the desire to offer a relatively ‘respectable’ account at a later point in time when monastic Buddhism
was more fully integrated into society. One would not realise from a random sample of modern
books on Buddhism that perhaps the most important role of Buddhist monks in many parts of the
Buddhist world remains that of dealing with death. Arguments from silence are always risky, but in
this case the texts do include at least some indications of such a role (see deCaroli 2004). We might
also note that such a role would help to explain the possible derivation of the term shaman from
240
Origins of yoga and Tantra
In the next section I look at some of the Śaivite ascetic groups that came
to occupy this space.
ś a i va ascetics: p ā śupatas and k ā p ā l i k a -st yle ascetics
Early ascetic practice in South Asia appears to have been mainly performed
by devotees of Śiva and of various forms of the Goddess, all of whom
would perhaps have identified generically as Śaivite. This is perhaps not
surprising; Śiva, for all of his appropriation as a respectable state deity, has
always retained a more transgressive side to his image, later on increasingly
expressed through subsidiary gods such as the Bhairava deities, whom we
will consider later in this chapter.
Ascetic practice in modern South Asia is still closely linked with Śaivite
traditions, although Vais.n.ava forms of asceticism have become more important in recent centuries. Modern Śaiva ascetics are generally linked to one
or another branch of the so-called Daśanāmı̄ tradition, which is regarded as
going back to the time of Śaṅkara. In reality, the Daśanāmı̄ order appears
to have taken shape at a considerably more recent time, probably in the
late sixteenth or seventeenth century (Clark 2004). Śaiva ascetics however
go back to a much earlier period, well before the time of Śaṅkara, who
probably lived in the eighth or ninth century CE.
Leaving aside the vrātya themselves, who might be regarded as Śaivite
or proto-Śaivite ascetics, the earliest group of Śaivite ascetics known to us
are the Pāśupatas. The Pāśupatas do not seem, initially at any rate, to have
been involved with ‘Tantric’ religion as such, but they were practitioners of
a style of spirituality that involves deliberately shocking behaviour and the
conscious courting of disrepute, elements that we also find in the Tantric
context.16
The Pāśupatas are first mentioned in the Mahābhārata17 and are commonly referred to in the Purān.as. By the seventh and eighth centuries,
we find them being patronised by the rulers of Orissa, where a series of
temples in Bhubaneswar have Pāśupata iconography, and also in Southeast
16
17
śraman.a (cf. Harold Bailey, cited in Blacker 1986: 321–2; Menges 1989: 240–1), and the success of
Buddhist and Jaina monks as ‘middlemen’ in relation to marginal territories (Bailey and Mabbett
2003).
For general introduction to the Pāśupatas see Lorenzen 1991: 173–92; Ingalls 1962; Donaldson 1987:
284–9; Dyczkowski 1989: 19–26. For more recent material, see Bisschop and Griffiths 2003; Bisschop
2005; Clark 2004: 137–8.
In the Śānti Parvan, 349.64, though this passage is generally regarded as a late addition, perhaps
c. 300–500 CE (Clark 2004: 137).
Tantra and the wild goddesses
241
Asia, where they served as court ritualists for the Khmer kings and, perhaps
somewhat later, in Java.18 Their somewhat legendary founder, Lakulı̄śa or
Nakulı̄śa, who is regarded as an incarnation of Śiva, is generally placed
in around 100 CE, mainly on the grounds of the number of generations required for the lineages of Pāśupata teachers cited in later dateable
sources. If Lakulı̄śa is a legendary rather than historical figure, this becomes
a weaker argument, and in any case orally-transmitted lineages can easily lose some early members. Thus the Pāśupatas might well have existed
somewhat earlier.
In fact, Daniel Ingalls pointed some years ago to the intriguing
resemblances between the Paśupatas and the classical Greek cult known
as the Cynics, who appeared in Athens at around the beginning of the
fourth century BCE, though they continued well into the Roman Empire
(Ingalls 1962: 281). Both Cynics and Pāśupatas adopted practices that were
deliberately intended to attract rejection and dishonour and lead them to
be regarded as crazy. These involved wandering from place to place and
indulging in deliberately shocking and improper acts. Hercules, who was
on earth ‘the most trouble-ridden and wretched of men’ (Ingalls 1962:
283) but eventually became a god, was regarded as the first teacher of the
Cynic tradition: Ingalls notes the similarity in both name and iconography
between Hercules (Heracles in Greek) and Lakulı̄śa. Both are typically
depicted holding a club, an attribute that makes little sense for a renunciate
ascetic, though it links to Lakulı̄śa’s name (lakula = club) (Lorenzen 1991:
173).19
Ingalls avoids suggesting any direct connection between the Pāśupatas
and the Cynics, but the idea does not seem inherently absurd. David White
considers it as a possibility in his book Myths of the Dog-Man, noting that
the Black Sea region from which the Greek Cynics traced their tradition
had trading links to the port of Broach on the Arabian Sea in the first
centuries of the common era (White 1991: 104; see also Lorenzen 1991:
178–9). One might also consider the Indo-Greek presence in India from
the late fourth century BCE onwards, which could have provided a channel
18
19
For the extensive Pāśupata/Lakulı̄śa iconography in Orissan temples from the seventh century
onwards, see Mitra 1984; Donaldson 1987: 288–9; for references to Pāśupatas in Cambodia and Java,
see Chapter 12 and Becker 1993.
There are also intriguing overlaps between Hercules imagery and that of the Buddhist deity Vajrapān.i
(e.g. Klimburg-Salter 1995: 122, abb.XV) and between later Lakulı̄sa imagery and depictions of the
Buddha Śakyamuni (Mitra 1984, plates 82, 8396, 106, 108, 110, 125 etc.). I do not have any explanation
for these similarities, but they suggest that there were linkages between some of these apparently
distant or separate groups of which we now know little or nothing.
242
Origins of yoga and Tantra
at a somewhat earlier date. Clearly, though, if there was a common element
to the Cynics and Pāśupatas, there was also a specifically Indic element, and
whatever the vrātya had become by this time might well have contributed
to this.
While we do not have much material directly from Paśupata circles,
there are enough references to give a reasonable picture of their practices,
which are classified in terms of a series of stages (avasthās) in the practitioner’s spiritual development (Lorenzen 1991: 185). The first or ‘marked’
(vyakta) stage involves bathing in ash, living in a temple and performing
specified acts of worship, which include laughing, dancing and singing. It
is at the second or ‘unmarked’ stage (avyakta) that he leaves the temple and
deliberately courts criticism and censure from the public. Ingalls gives an
example from Kaun.d.inya’s commentary on the Pāśupata Sūtra (ch. 3):
He should take up his stand by a group of women neither too close nor too far
away but so that he falls within their sight. Turning his attention to one of them
that is young and pretty he should stare at her and act as though he were setting
his desire upon her and honouring her. When . . . she looks at him, he should
act out the symptoms of love such as straightening his hair, etc. Then every one,
women, men and eunuchs will say, ‘This is no man of chastity; this is a lecher.’
By this false accusation their merit comes to him and his bad karma goes to
them . . . (Cited in Ingalls 1962: 290)
We can see the continuity here with practices such as the ox-vrata and
dog-vrata discussed in Chapter 7. The point of this kind of behaviour is
to court disrepute, and the logic is given in the last sentence; since the
Pāśupata practitioner is not really a lecher, and is being falsely accused, he
acquires the merit or good karma of the accusers and his own bad karma is
transferred to them.
This ‘unmarked’ stage is transitional, and another way of viewing what
is happening here is that the would-be Pāśupata practitioner is placing
himself20 firmly outside conventional society. At later stages, the Pāśupata
practitioner is held to achieve the magical powers of a siddha or ‘accomplished one’ (cf. Bisschop 2005). Siddha is a term that we will meet again in
its specifically Tantric context, but it is significant that the Pāśupata practitioner is expected to gain magical ritual powers that could support a career
as an itinerant ritual practitioner or as part of a court ritual establishment.
The powers achieved are essentially those of Śiva himself (Bisschop 2005:
542). Thus we seem to have, as with the vrātya passage in the Atharvaveda, a
20
It is unclear whether there were significant numbers of female practitioners among the Pāśupatas,
although, as we will see later, kāpālika-style practitioners include women as well as men.
Tantra and the wild goddesses
243
ritual identification with Śiva that enables the practitioner to exert various
kinds of magical power.
A variety of further Śaiva ascetic groups are described as existing over
the next few centuries. In addition to the Pāśupatas and Lakulı̄śas,21 we
hear of Kapalikas, Kālamukhas, Bhairavas and other terms; it is not always
clear how far these are distinct groups and how far they are labels applied
to these ascetics by others (Lorenzen 1991). The term kāpālika in particular
seems not to have referred to a distinct order of renunciates, and we should
be wary of over-reifying it.22 However, it was evidently used to refer to a
distinct mode of practice and one that seems to have been closely associated
with the growth of the more ‘extreme’ forms of Tantra. These ‘kāpālikas’
were regarded as associated with the god Bhairava, a violent and criminal
transform of Śiva who appears to continue the inauspicious and dangerous
aspects of the late Vedic Rudra-Śiva. It is here too that we appear to find a
growing involvement with fierce and dangerous female deities.
These kāpālika-style practitioners, as mentioned above, were worshippers
of Bhairava. In the later Purān.as at any rate he is described as an emanation
or projection of Śiva, just as Bhairava’s consort, the fierce goddess Kālı̄, is
described in the probably eighth-century Devı̄māhātmya as an emanation of
Durgā (Coburn 1982, 1988, 1991; Yokochi 1999). Bhairava, like his followers,
the kāpālika-style ascetics, is generally depicted as an explicitly transgressive
figure, smeared in ashes, living in cremation-grounds, accompanied by
dogs, and carrying a skull (kapāla), which is where the term kāpālika or
‘skull-man’ comes from. Bhairava does this because, like the kāpālika-style
ascetics, he is performing the mahāvrata, the great vrata, vow or mode of
behaviour, which is imposed as a penance according to the Brahmanical
law codes on the most evil of criminals: the person who has murdered
a Brahmin. In Bhairava’s case, he is carrying the guilt for Śiva’s having
destroyed the fifth head of the god Brahmā.
In fact, there are three related sets of ideas here, which occur to some
extent separately: We have the mahāvrata for Brahmin-killers in the lawcodes, which involves wandering around wearing ashes, carrying a skull,
living in cremation-grounds and the like. We have the kāpālika-style ascetic,
who is performing this mahāvrata but presumably, in general, has not killed
a Brahmin. Lastly, there is Bhairava, who is believed to have performed the
21
22
For the later history of the Pāsupata/Lakulı̄sa grouping in India, see Lorenzen 1991: 107–9. They
appear to have moved to the South after the Muslim conquests in the North of India.
White has noted that there ‘are no Hindu Tantric scriptures that contain the term kapalika in their
title, or whose authors call themselves Kāpālikas, and there is not a single inscription in all of South
Asia that names the Kāpālikas in a way that would indicate an actual sectarian order’ (2005: 9).
244
Origins of yoga and Tantra
Figure 10.1. Bhairava. Vidisha Museum
Tantra and the wild goddesses
245
mahāvrata because he cut off Brahmā’s fifth head at Śiva’s request and so
‘killed a Brahmin’.
The law-code is the first of these three to be attested in the literature,
but it would seem more natural to suppose that the practice of performing
the mahāvrata for religious purposes already existed and that the law-code
was, as in other cases such as the ‘explanations’ of various caste-groups as
originating from improper marriages between various groups of persons,
simply legitimating something that already existed. If Gombrich is right
about Aṅgulimāla in the Pali Canon being intended as some kind of early
version of the kāpālika-style ascetic (Gombrich 1996: 135–64), this is further
evidence pointing in the same direction.23 The Bhairava myth comes from
relatively late Purān.ic sources but in a sense closes the circle: the kāpālikastyle ascetics are approaching the deity Śiva by imitating his aspect Bhairava
who is performing the penance.
What the whole complex of ideas does is to ‘justify’ a stance which is
outside society and which involves a radical rejection of social norms. The
kāpālika type of ascetic behaves like the most despised and contemptible
of people in Brahmanical terms, the murderer of a Brahmin. In a sense,
there is nowhere further that he can go, and nothing worse that he can
do, so he is free to behave more or less outside all social constraints. In
reality, kāpālika-style ascetics, like everyone else, developed their own modes
of social behaviour, which we know mainly from the caricatures in the
Brahmanical plays and stories of the times (see e.g. Lorenzen 1991), but
also from sources such as the Buddhist caryāgı̄ti songs where the Buddhist
Tantric yogin takes on the kāpālika style of behaviour and consorts with a
real or metaphorical low-caste woman (d.ombı̄ or can.d.ālı̄):
Outside the town, o D
. ombı̄, is your hut; the shaven-headed brahmin boy goes
constantly touching you.
Ho D
. ombı̄! I shall associate with you, I Kān.ha, a kāpāli-yogin, shameless and
naked.
One is the lotus, sixty-four its petals; having mounted on it, the poor D
. ombı̄
dances.
Ho D
. ombı̄! I ask you earnestly: in whose boat, o D
. ombı̄, do you come and go?
Strings you sell, o D
ombı̄,
and
also
baskets;
for
your
sake I have abandoned the
.
actor’s box.
Ho! You are a D
. ombı̄, I a kāpāli. For your sake I have put on the garland of
bones.
23
I am not fully convinced, since, as Sanderson observes (cited in Gombrich 1996: 152 n. 7), we do not
know of any later kāpālika-style ascetics, as distinct from Śakta devotees, who indulged in murder
as part of their religious practice. Aṅgulimāla might conceivably have been intended as some sort
of hostile parody of an extreme Śaivite ascetic.
246
Origins of yoga and Tantra
Troubling the pond, the D
. ombı̄ eats lotus-roots – I kill you, o D
. ombı̄, I take
your life! (Kvaerne 1986: 113)
∗∗∗
I dispelled the three worlds by means of amorous play, having fallen asleep in the
sport of sexual union.
How, o D
. ombı̄, is your coquetry? The noble-born is outside, the kāpālika inside.
By you, o D
. ombı̄, everything has been disturbed; for no purpose the Moon has
been agitated.
Some there are who speak ill of you; (but) those who discern do not remove you
from their neck.
Kān.ha sings of the amorous Can.d.ālı̄; there is no greater harlot than the D
. ombı̄.
(Kvaerne 1986: 150)
These songs, as the commentators point out, are coded descriptions of the
practices of sexual Tantra (the lotus-roots are bodhicitta; killing the D
. ombı̄
is purifying avadhūti, the central psychic channel), but, like the ‘Bāul’ songs
of modern Bengal that are their lineal descendants, the imagery is realistic,
and it presupposes a world in which there are d.ombı̄ women living outside
the town and selling petty goods, and kāpālika-style yogins consorting with
them.
If our information can be taken at face value, these kāpālika-style yogins
felt free to indulge in sex and alcohol as part of their spiritual path. What
also comes through these accounts is that kāpālika-style practitioners were
regarded as claiming or actually exercising magical powers, including flying
through the air, and that their powers were linked to their relationships with
the female spirits and/or human female practitioners known as yoginı̄s,
d.ākinı̄s and by other names. These were frightening and dangerous spirits,
and they form part of an ‘alternative pantheon’ of dark and scary female
deities associated with Bhairava and his various transforms. We turn now
to them. They include such deities as the skeletal witch-figure Cāmun.d.ā,
and the goddess who was eventually to become the best known of them all,
Kālı̄.
wild goddesses and d isease goddesses
In later contexts, these deities have been somewhat tamed and civilised. Kālı̄
in modern Bengal is seen as a loving mother, though she retains an ambivalent side to her image (S. Gupta 2003; McDermott 1996; McDaniel 2004).
In Tibet, the reversal has been more radical; the Tibetan word for d.ākinı̄
(khandroma) is a highly positive term when applied to a woman, implying a high level of spiritual attainment. In most North Indian vernaculars,
Tantra and the wild goddesses
247
however, the term d.ākinı̄ or its modern derivatives (such as d.āin in Hindi)
means ‘witch’, in the sense of ‘woman who harms people through malevolent spiritual power’. Certainly, in the earlier material, as we will see, these
goddesses were seen as both dangerous and malevolent.
I have argued elsewhere that a key element in the growth of ‘Tantra’ was
the gradual transformation of local and regional deity cults through which
fierce male and, particularly, female deities came to take a leading role in
the place of the yaks.a deities (Samuel 2002d; see also Chapter 12). Much of
this seems to have taken place between the fifth to eighth centuries CE.
While the new universal deities which I referred to in Chapter 9 became
the object of state cults closely related to the role of their new royal patrons,
texts such as the Mahāmāyūrı̄ (Lévi 1915; Sircar 1971–72) or the various
Purān.ic accounts suggest that the yaks.a-type deities remained important as
patron deities of cities and regions well into the Gupta period. Increasingly,
from around the sixth century onwards, we see the emergence of a new
deity type, of whom the best-known example in the West is undoubtedly
Kālı̄: fierce, violent deities, wearing flayed human skins, garlands of human
heads and the like, surrounded by flames, often portrayed in cremationgrounds or similarly terrifying backgrounds. Not all these deities are shown
in such extreme forms, and Kālı̄ herself became a dominant figure at a
somewhat later period, but the violent, transgressive nature of the imagery
is characteristic.24
We can start by asking where these wild, dangerous goddesses came
from. The goddess iconography that I referred to in Chapter 5 shows no
real sign of a cult of wild or fierce goddesses. Early goddess iconography
is relatively gentle. Even the early Kus.ān.a-period images of what Doris
Meth Srinivasan calls the ‘Warrior Goddess’ (mostly showing the theme
later known as Mahis.amardinı̄, Durgā killing the buffalo demon), are quite
mild (Srinivasan 1997, plates 20.1–7, plates 20.16–19; Divakaran 1984, plate
251) and this continues throughout much of the first millennium CE (e.g.
Divakaran 1984, plate 252–6; Harle and Topsfield 1987: 23–4, no.31). Later
Durgā imagery is a little more forceful, but it is still a long way away from
the violence and horror of some of the imagery of the new deities (e.g.
Cāmun.d.ā; Dehejia 1979: 13, 24).
The textual record shows more indication of the fierceness of the new
deities, and there have of course been plenty of references to female demons,
24
Gombrich has pointed to possible early evidence for the goddess Kālı̄ in the Pali Canon (1996:
158–62).
248
Origins of yoga and Tantra
raks.ası̄s, and to yaks.inı̄s who also seem quite demonic in their behaviour,
in the Buddhist sūtras and jātakas and similar sources (Sutherland 1991).
Some of these stories are of considerable interest, such as the well-known
story of the goddess Hāritı̄, who has already been mentioned in Chapters 5
and 9. Haritı̄ in the Pali Canon and other Buddhist sources is a demoness
who kills and eats small children. One day she loses one of her own five
hundred children, and the Buddha brings her to a realisation of the suffering
she has caused the mothers of the children she has killed. She becomes a
protector of children, and is given regular offerings in the monasteries (e.g.
Strong 1992: 36–7; Samuel 2002b: 1–2).
Hāritı̄’s role as child-murderer may be in part Buddhist propaganda
against the locally-important cult of a respected deity – Elinor Gadon and
Janet Chawla have suggested something similar for S.as.t.hı̄, who plays a
parallel role in the Brahmanical context (Gadon 1997; Chawla 1994). If
the large number of images of a mother goddess of fertility and prosperity
from Gandhara, Mathura and elsewhere have been correctly identified as
representations of Hāritı̄, she would seem to have been a respected deity;
her iconography overlaps with both Laks.mı̄ and the Iranian goddess of
prosperity, Ardoxsho, and it is not at all terrifying (see Chapter 5).
The death of young children, however, was a common enough occurrence in pre-modern societies, and there certainly were female demons
linked to the illness and death of children at a slightly later date, since they
are discussed in early medical texts. In fact, the dangerous new female
deities seem closely related to these goddesses of illness who have retained an
important role in Indian folk religion into modern times. Often these occur
as a set of similar goddesses, described as mothers or sisters, and five, nine
or, most often, seven in number (Kosambi 1960; Freed and Freed 1998: 122–
36; Mayaram 2000). Some individual goddesses of this class have achieved
independent status, the best known being Śı̄talā, the goddess of smallpox
and cholera (e.g. Dimock 1982). Such diseases are seen as a sign of possession by the goddess, who may or may not be willing to let her victim go. In
South India, Śı̄talā’s equivalent, Mariamman, is one of a range of amman
goddesses who are very important in village religion in Tamil areas of India
and Sri Lanka. An amman goddess is often married to a male god who is the
village protector and who also seems to be related to the buffalo killed in
village sacrificial rituals to the goddess and in the well-known iconography
of the goddess killing the buffalo demon (Biardeau 1989; Brubaker 1983).
Kosambi attempted to trace Brahmanical references in North India for
a cult of these mother goddesses and came up with little that is substantial
for an early Brahmanical cult (Kosambi 1960). They may well have been
Tantra and the wild goddesses
249
present in non-Brahmanical, village religion. The situation in South India,
where goddesses with a strong local identity have clearly been assimilated
to Durgā/Pārvatı̄, suggests this, though whether it can be generalised to
the rest of South Asia is unclear. It is unclear too how far back the diseasedemonesses in the medical literature can be traced back. In Book Three of
the Mahābhārata, they occur in the context of a narrative about the god
Skanda (here the son of Agni, not Śiva), and appear to be conflated with
the Pleiades or Kr.ttikā sisters, wives of the seven .r.sis (3.214–17, 3.219; see
van Buitenen 1975: 650–6; 656–9; White 2003: 35–49). The same narrative
also discusses other graha (‘graspers’) who are both male and female and
are emanated by Skanda himself (3.219 = 1975: 658–9). Graha is also the
word for planet, and the planetary deities, who are all male in Brahmanical
thought, are again seen as possible sources of illness. All in all, I do not
think that it is possible at present to say when the idea of female diseasedemons arose, though if it were significant in the Vedic period one would
expect more reference to it in sources such as the Atharvaveda, which is very
concerned with countering diseases of all kinds. If, as Richard Gombrich
has suggested, the move to living in cities was associated with a substantial
increase in the incidence of epidemic disease, this may have provided a
context for such ideas to originate or become more significant.25
In any case, the disease-goddesses and demonesses seem to give us both
a plausible source for the female deities of the proto-Tantric complex, and
a sense of how they could be seen as powerful and dangerous, feared but
also invoked as a protectress. The question of illness might also underlie the
idea of the d.ākinı̄; if the human d.ākinı̄s were folk healers from low-status
castes such as the modern birth attendants or dāı̄, constantly suspected also
of being sorceresses, then their ambivalent role as partners for the male
Tantric practitioners, and their elevation to high status in Tibet where this
whole complex of ideas was more or less irrelevant, might make sense.26
brahmanical equivalents
In fact, there are a couple of sets of Brahmanical goddesses that are worth
some attention in relation to this question. These are the Pleiades or
Kr.ttika sisters, mentioned above in relation to Skanda (hence his name
25
26
The Nine Witch Sisters of the Magar shamans and neighbouring groups provide a ‘tribal’ equivalent
but the presence of influence from literate and urban Tantric traditions here is evident and it would
be inadvisable to take current Magar practices as reflecting early conceptions (see de Sales 1991;
Oppitz 1978–80, 1981; Maskarinec 1995).
On the concept of the d.ākinı̄ see also Herrmann-Pfandt 1992.
250
Origins of yoga and Tantra
Karttikeya), and the Saptamātr.kā (‘Seven Little Mother’) deities. The
Kr.ttikā sisters seem to be mainly significant in relatively early texts such as
Mahābhārata. The Saptamātr.kā goddesses are another matter, and seem
to represent a Brahmanical adaptation of a set of seven or nine fierce
disease-causing goddesses who are found in different forms throughout
South Asia from Tamil Nadu to Nepal (Coburn 1988; Meister 1986; K.
Harper 2002). They occur extensively in both Śaivite and Buddhist Tantric
contexts.
In these contexts, the Mātr.kā deities, often expanded to eight so as to
form the eight members of a circle in the man.d.ala, are subordinated to
the power of the male deity (or male-female couple) at the centre of the
man.d.ala, a pattern that would appear to reflect the idea of a male Tantric
magician, with or without a female partner, controlling the powers of these
deities for magical ritual purposes. A Śaivite example is the Kulacūd.āman.i
Tantra (Finn 1986). Finn dates this text to between the ninth to eleventh
centuries CE, probably towards the earlier end, in part on the grounds that
the cult of the Saptamātr.kās was at its height between the fifth to eighth
centuries, though in the light of Harper’s arguments (see below) it may well
include earlier elements. The Kulacūd.āman.i Tantra deals at some length
with both Mahis.amardinı̄ (Durgā as slayer of the buffalo-demon) and the
Mātr.kā deities, and is strongly oriented towards this-worldly and pragmatic
results. It requires that the female participants in the ritual be identified
with specific members of the Mātr.kā deities, which suggests an ‘acting out’
of the man.d.ala similar to that found in the Tantric ritual dance traditions
of Nepal and Tibet, which I shall discuss in Chapter 12.
For a Buddhist example, we could cite the Vidyottama Mahātantra,
discussed by Davidson (2002a: 300–2). Here the Seven Mothers (sapta
mātarah.) offer mantras to the Buddhist Tantric deity Vajrapān.i, the former
yaks.a protector of the Magadhan capital of Rājagr.ha and of the Buddha who
had become by this period a key representation of Tantric power (Snellgrove
1987: 134–41). They explain a series of rituals (mātr.vidhi) for protection,
including ‘overcoming Vināyaka demons and Graha spirits, ruining harvests, spoiling liquor, making an enemy ugly and diseased (but then turning
him back again), discovering who one’s enemies are, and expelling them’
(Davidson 2002a: 302). The ritual includes the making of a painting showing Vajrapān.i surrounded by the Seven Mothers, whose appearance should
be ferocious (2002a: 301). The general structure of a male deity or male–
female couple at the centre surrounded by a circle of eight female deities
(some of whose names are in this case those of low caste women) is also the
basis of the man.d.ala of the well-known Hevajra Tantra (Snellgrove 1959;
Tantra and the wild goddesses
251
Farrow and Menon 1992; Willemen 1982; Huntingdon and Bangdel 2003:
454–67).27
The Saptamātr.kā deities vary somewhat from list to list. In the earliest
versions known to us, from the early fifth century, they generally consisted
of six deva-śaktis, female consorts of major Brahmanical deities, along with
the fierce goddess Cāmun.d.a, and are often, as in the three Saptamātr.kā
reliefs at the important early Gupta site of Udayagiri, accompanied by the
war-god Skanda (Harper 2002: 117). Katherine Harper has argued that the
Saptamātr.kā deities first rose to importance in the context of state sorcery,
more specifically the performance of military rituals to aid the Gupta kings
in their campaign to expel foreign rulers from India (Harper 2002).
Her argument is based to a considerable degree on an interpretation
of the material at Udayagiri. This site is associated in particular with
Candragupta II, who visited there in 401–2 and was responsible for the
building of a shrine to Śiva that incorporated a relief of the Saptamātr.kās
holding a series of emblematic banners. Elsewhere at Udayagiri the banners, which Harper identifies as war-banners, occur twice as a separate set.
Harper argues that the Gupta kings kept their usage of such ritual veiled
in allusive and indirect language, since they did not want it to fall into
hostile hands. If she is right about the meaning of the Saptamātr.kā reliefs,
and the frequent occurrence of the Saptamātr.kā in later Tantric contexts
suggests that she may be, then this is very early evidence for the aggressive
and violent use of ‘Tantric’-style ritual.28 Whether the rituals were at this
point conducted by explicitly ‘Tantric’ priests is perhaps another question.
Michael Willis, who has studied this site in detail, has suggested, as noted
earlier, that this kind of ritual work would have fallen within the domain
of the Royal purohita, in origin and by training an Atharvavedic Brahmin
priest.29 We will see evidence in Chapter 12 for the role of the royal purohita
being taken over at a later stage by explicitly Tantric priests (Sanderson
2004).
If a site such as Udayagiri is the initial context of the Saptamātr.kās as an
explicit set of deities used for destructive ritual, how do these deities relate
to the disease goddesses? The Saptamātr.kās themselves have respectable
27
28
29
This man.d.ala consists of Hevajra and his consort Nairatmyā at the centre, surrounded by eight female
deities: Gaurı̄ (E), Śavarı̄ (SE), Caurı̄ (S), Can.d.alı̄ (SW), Vetālı̄ (W), D
. ombı̄ (NW), Ghasmarı̄ (N)
and Pukkası̄ (NE).
There is a problem in relation to her argument about the dating of the Devı̄māhātmya, which is
probably much too early (cf. Harper 2002, Yokochi 1999), but this is not critical in relation to the
point I am making here.
Willis (personal communication, 2004). For Udayagiri see also Willis 2001, 2004; Dass and Willis
2002.
252
Origins of yoga and Tantra
Brahmanical identities, as wives of the major Brahmanical gods, and only
one of the set, Cāmun.d.ā, is particularly terrifying in her appearance, but
if we assume that the disease-goddesses were around at or before this time,
then it would make sense to see the Saptamātr.kās as a Brahmanical transform of these goddesses, whose purpose was to control disease and misfortune but also to release them on the enemies of the state. This would also
fit with the defensive role of the set of eight Matr.ka deities, who are an
expansion of the set of seven Mātr.kās, in Nepalese state ritual, which I will
discuss in Chapter 12.
śaiva tantra and the yoginı̄ cults
These practices occur in both Śaiva and Buddhist contexts. I shall return
to the specific situation of this borrowing later, but to begin with we might
try to make sense of the Śaiva case.
I have suggested already that the Śaiva ascetics were consciously positioning themselves outside of respectable society and thus found themselves
among the low-caste and marginal groups such as the performers of magical
ritual. The Paśupata rationalisation of this, as we have seen, was in terms of
unloading bad karma.30 The kāpālika-style ascetics saw themselves as identifying with Śiva through his negative aspect of Bhairava. We can assume
however that a key element in both these cases and with other renunciates
is the issue that we have already discussed in Chapter 6, that of the liberating insight. For the ‘Tantric’ traditions, the transmission of the liberating
insight was the function of the guru or spiritual teacher, and a central step
was the process of initiation, in which the guru passed on his insight to
the initiate. As David Gordon White has pointed out, the initiation here
took on a sexual form, in which the initiate consumed the sexual substances
(semen and female sexual secretions) produced in ritual intercourse by the
guru and his consort, and thus also ingested the essence of the clan or spiritual family (kula) (White 1998, 2003). This might, as in modern Tibetan
practice, take place symbolically, but the question of ‘joining the family’
and thus of being in a specific lineage of practice is nevertheless a critical
one.
These sexual substances, along with other polluting bodily substances,
also provide the offerings needed by the yoginı̄s and other fierce female
spirits (Sanderson 1988: 671; White 2003). In modern times, both the ‘clan
30
One might expect them to reject karma altogether, and it is a little surprising that the passage cited
by Ingalls above speaks of taking on the good karma of those who reject them.
Tantra and the wild goddesses
253
essence’ and the offerings may be offered in symbolical form, although this
is not necessarily always the case.31
What, though, was the nature of the ‘liberating insight’ as understood
by these renunciate (and in some later cases also lay) practitioners? In a
significant conspectus of the Śaiva systems of Kashmir in the early mediaeval
period, Alexis Sanderson notes that Tantric Śaivites in Kashmir at that time
were divided between those who adhered to the dualist theology of the Śaiva
Siddhānta32 and those who followed the non-dualist theology of the Trika
system. Here one can see among other things the contrast between Tantric
practitioners who accepted the validity of antinomian and transgressive
practices (the non-dualists) and those who rejected such practices (the
dualists, see Sanderson 1995: 17). According to Sanderson, the Trika
adhered to a nondualistic theology according to which all phenomena are nothing
but the spontaneous self-projection of an all-encompassing divine consciousness,
so that the substance of the universe and its efficient cause are one and the same.
This absolute idealism was formulated and defended by Somānanda ( fl. c. AD 900–
950), Utpaladeva ( fl. c. AD 925–75), and the latter’s commentator Abhinavagupta
( fl. c. AD 975–1025). The principal works of this philosophical tradition came
to be known as the Śāstra of Recognition (Pratyabhijñāśāstra) after their central
concept, namely that liberation comes about as the recognition ( pratyabhijñā) that
the true identity of oneself and all phenomena is the Lord (ı̄śvarah.) defined as this
all-containing, autonomous consciousness. (Sanderson 1995: 16)
This is significant, in that it indicates that the identification with Śiva which
was originally perhaps intended as a mode of access to magical power is
here seen as in itself a liberating insight. What we have here, however, is
a discursive and verbal presentation of the nature of the liberating insight,
which could be paralleled to other such expressions in Śaiva and nonŚaiva traditions. What was significant for practitioners was not the words
(‘my true identity is ı̄śvarah.’, etc.) but the inner realisation which scholars
in the tradition attempt to describe through those words. It is this inner
realisation that the guru attempts to transmit, or rather attempts to provoke
the student into achieving for himself or herself.
This is an important point both for the Śaiva tradition and for other
traditions, including the Buddhists, and it links to the fact that these are
lineages of practice, and that the texts throughout are held to be secondary
to the practice lineage as conveyed through the teaching of the guru. A key
element here is the secrecy and power of the teaching; not all gurus and
31
32
See Hanssen 2002 for a recent account of the use of menstrual blood among Bauls in Bengal.
As Sanderson notes, Śaiva Siddhānta is nowadays generally used to refer to the South Indian
traditions of which this Kashmiri tradition was a predecessor.
254
Origins of yoga and Tantra
lineages are equal, and what we find is a succession of schools of practice,
each of which defines itself as superior to the one before.
A related element is that of the mantra, which can perhaps be defined in
this context as a ritual formula transmitted from guru to student through
which the presence of the deity is invoked or compelled. This is linked
in Śaiva tradition to a complex theology of the role of sound, seen as a
constitutive and creative force through which the manifest universe takes
place (e.g. Padoux 1990; Beck 1995).
For the nondualist Śaiva traditions the hierarchical succession of schools
of practice has been described in some detail by several authors, including
Sanderson himself (Sanderson 1988; see also Dyczkowski 1989; Dupuche
2003; Brooks 1992a). Thus the Śaiva tradition was divided by its textual
exponents into the Atimārga, the Outer path, intended for ascetics alone,
and the Mantramārga. The Atimārga consisted of the Pāśupata and Lākula
division (Sanderson 1988: 664–7); the Lākula here includes the kāpālikastyle and Kalamukha ascetics. The Mantramarga is regarded as a higher
teaching than that of the Atimārga. It can be practised by householders as
well as renunciate ascetics and includes both the Śaiva Siddhanta texts and
a series of yet ‘higher’ teachings, commencing with the Tantras of Bhairava,
which are themselves hierarchically subdivided several times further, culminating with the Tantras of the Trika tradition and the highest group of
all, the Tantras of the goddess Kālı̄ (Sanderson 1988: 667–9).
The progression from Pāśupata to the Trika and Kālı̄ Tantras involves
a progressive shift from male deities to female deities. Thus the Bhairava
Tantras are divided into the Mantrapı̄t.ha, which is centred around the cult
of the male deity Svacchandabhairava, and the Vidyāpı̄t.ha, which involves
the cult of the yoginı̄s or fierce female deities and culminates in the Trika
and Kālı̄ cults (Dupuche 2003: 13–14; Sanderson 1988: 672–8).33 Thus the
‘higher’ teachings move increasingly into the area of the Śākta cults or
Goddess cults and it is the Śākta pı̄t.has, the sacred sites of the Goddess cults,
which become the ideal sites for their performance (Sircar 1973; Banerji
1992: 308ff; Heilijgers-Seelen 1994: 152; Hartzell 1997: 1020–8, 1050–4; Pal
1988). There are strong suggestions of groups of travelling renunciates and
lay practitioners meeting at these pı̄t.has for ritual gatherings, but there are
also practices in which the pı̄t.has are relocated within the practitioner’s
body. As we will see, the same is true of the Buddhist Tantras.34
33
34
For other classifications see Dyczkowski 1989.
The role of the so-called yoginı̄ temples, a group of open-air, mostly circular temples in central
and east India (present-day Madhya Pradesh and Orissa) with images of sixty-four or eighty-one
yoginı̄ figures arranged in a circle facing towards the centre, within this development is unclear.
Tantra and the wild goddesses
255
It should be noted that this is a logical progression, as viewed from
within the tradition, and not necessarily a historical sequence, although to
a considerable degree it probably does correspond to a historical succession
in which each new school positioned itself as superior to preceding schools.
In reality, however, at any given time there were a number of competing
traditions, and schools that were of considerable importance at one stage
(for example the traditions associated with the deity Tumburu, to which
we will return in Chapter 12) might at a later stage be reduced to a marginal
position (Sanderson 1988: 669).
In the course of this movement from ‘lower’ to ‘higher’ teachings, the
role of the female deities also changes, from an entourage surrounding the
male practitioner to being the primary objects of worship and identification
in their own right. We will consider what this development might mean
after looking at what was happening in the non-Śaiva traditions.
It is worth noting here though that one further important development
takes place in these ‘higher’ Tantras. This is the introduction of a system
of ‘internal’ yogic practices based on a subtle anatomy of internal channels
(nād.ı̄) and meeting places (cakra). In the Śaiva context these practices are
explained in terms of the ascension of the deity Kun.d.alinı̄ through the
cakras and the consequent untying of knots in the nād.ı̄ (e.g. Silburn 1988).
These ideas are closely related to the practice of sexual yoga, and they
enable a reconceptualisation of the sexual practices so that these are no
longer primarily concerned with linking to the tradition or producing food
for female spirits but are seen as a practice able to lead to liberating insight
in its own right.
This important development also takes place in other Tantric traditions,
and will be considered in some detail in Chapter 11.
For the present, though, we might consider what has been something
of a point of contention in Tantric studies for many years: what was the
original context of the fierce goddess practices? The two broad positions
that have been taken are to consider them as part of a ‘pan-Indian religious
substrate’ of some kind, as postulated by Ruegg (1964), an assumption that
goes back to Paul Mus (e.g. 1975) and earlier scholars, or to regard them as
an essentially ‘orthogenetic’ development out of the Vedic material. In this
case, the first approach would seem to require some kind of pre-existing
goddess cult, a Śākta or at least ‘proto-Śākta’ tradition, which was gradually
incorporated into Śaiva and Buddhist practice.
They appear to represent a regional development promoted by ruling dynasties within this general
area, presumably with the intention of accessing the power of the yoginı̄s for State political ends
(cf. Dehejia 1986).
256
Origins of yoga and Tantra
The disadvantage of this first approach, as Davidson, following Sanderson, notes, is that evidence for such a ‘nonaffiliated religious system’
is lacking; ‘all the personalities, both human and divine, are associated
with specific institutionalized lineages and not simply free-floating forms’
(Davidson 2002a: 171–2). The problem with the second approach is that
the Vedic material simply does not provide the basis for such developments.
The Vedas undoubtedly include transgressive practices, sexual rituals and
some material about female deities, but the continuity between these and
later developments is limited, and easily explainable by the need for Brahmanical and Buddhist scholars to justify their practices in terms of their
own previous traditions.
In fact, we seem to see the progressive introduction of Goddess cult
material in texts such as the Harivam
. śa and the Skanda Purān.a, with
deities such as Ekānam
śa,
Kot
avı̄
and
Vindhyavāsinı̄
(Couture and Schmid
.
.
2001; Couture 2003; Yokochi 2004) showing every sign of having a substantial local history before they enter the Brahmanical mainstream. Thus
Yokochi, on the basis of the Harivam
. śa and the Skandapurān.a, reconstructs
a ‘proto-myth’ of Vindhyavāsinı̄, the goddess of the Vindhya mountains,
which she regards as initially independent from either Bhāgavata-Vais.n.ava
or Śaiva myths (Yokochi 2004: 83–96). This proto-myth already includes
Brahmanical elements in the form Yokochi reconstructs (the demons
Sumbha and Nisumbha, an abhis.eka by the Brahmanical gods, Indra’s
adoption of Vindhyavāsinı̄ as his sister), but it would seem entirely plausible to suppose that it might derive from a pre-Brahmanical cult of a
Vindhya mountain goddess.35
There are also clear indications, as we have seen above, of potentially
malevolent female deities as agents of illness in the Mahābharata and in the
medical treatises, as well as in the continuing popular religious tradition
from all over South Asia. The existence of this popular religious tradition
is worth emphasising as a general issue, as well as in the specific context
of the cult of disease-goddesses. Its uniformity can be exaggerated, though
I think that Paul Mus was by no means entirely mistaken in pointing to
common themes throughout and beyond South Asia (Mus 1975, cf. also
Samuel 2001a). Its existence though can scarcely be doubted. Even in the
modern period, after more than two thousand years of Brahmanical and
Sanskritic presence, much of the religion of Indian village society has a
strongly local character and is clearly not a simple transform of Vedic and
Brahmanical cults (see e.g. Marriott 1969b). This is particularly true in
relation to Goddess cults, although a systematic process of assimilation to
35
On the present-day cult of Vindhyavāsinı̄, see Humes 1996.
Tantra and the wild goddesses
257
Brahmanical ideas, such as the widespread Durgā-Mahis.amardinı̄ mythology, is apparent (e.g. Biardeau 1989).
It is important here, as elsewhere, to disaggregate the category of ‘Tantra’
(and for that matter ‘Śākta’) since there are several different elements
involved in each case. As I have said, local Goddess cults, however limited our direct sources of information, seem to me to be a near-certainty.
What is perhaps a different question is the organised presence of pilgrimage
sites associated with a widespread Goddess mythology, the so-called Śākta
pı̄t.has.
As is well-known, these are now associated on the Śaiva side with the myth
of Daks.a’s sacrifice and the self-immolation of Śiva’s consort Satı̄ (=Pārvatı̄)
(e.g. Sircar 1973). The Buddhist mythology associated with these locations
is considered briefly later in this chapter. Each of the Śākta pı̄t.has today
consists of a goddess temple, associated with a named form of the goddess,
regarded as deriving from a specific part of Satı̄’s body that is held to have
fallen to earth at that point, along with a temple of a named aspect of Śiva.
Often the Goddess’s body-part is associated with a specific natural feature,
such as the ‘tongue’ of flame at Jwalamukhı̄ in the Punjab.36
Today, these sites include many of South Asia’s most celebrated pilgrimage sites, such as Jwālamukhı̄ (Jālandhara), Hiṅglāj in Pakistan
(Baluchistan), Kāmarūpa in Assam or Kalighat in Calcutta (White 1996:
121; Desani 1973; S. Gupta 2003; Morinis 1984). In the Śaiva and Buddhist
Tantras from the eighth or ninth centuries onwards we find lists of twentyfour sites, of which four are regarded as of particular importance. The
names mentioned vary from list to list, and are not always identifiable,
although certain sites are common to most lists (e.g. Davidson 2002a:
206–11; Heilijgers-Seelen 1994: 152). Several cremation-ground locations
are also included.
We do not have a clear picture of how this network of pilgrimage sites
arose. As Davidson notes, they are not necessarily sites associated with
kāpālika-style practice, nor are they generally sites with a previous history
of Buddhist religious activity. At this stage, I think that we have to leave the
question open of the milieu within which this network of sites came into
being, and whether it can meaningfully be labelled ‘Śākta’ in some sense
distinct from the general Śaiva milieu with which the sites later became
associated. Assuming, however, that these sites had a material reality in the
eighth and ninth centuries, it was probably within this general context that
many of the specific features of kaula and anuttarayoga practice developed.
36
In relation to Jwālamukhı̄, I am indebted to the unpublished research of Rosemarie Volkmann
(personal communication, 2003).
258
Origins of yoga and Tantra
Figure 10.2. Cakreśvarı̄. Mount Abu
It is possible that we are mistaken in any case in looking for a primarily
Śaiva or Buddhist identity. These places may have been from an early stage
location where practitioners from different traditions, Śaiva, Buddhist and
perhaps also early Sufis, along with ‘hereditary’ low-caste kula specialists,
might have gathered for ritual purposes.
buddhist, jain and vais. n. ava tantra
In any case, the three major components we have seen, the identification
with a fierce male deity, the cult of the fierce goddesses, and the internal
‘subtle body’ yoga practices, also developed in the Buddhist traditions, and
within much the same chronological frame as within the Śaiva traditions.
Here, again, there is a hierarchical arrangement of successively more and
more profound or powerful styles of practice starting with the kriyātantra
and caryātantra classes and moving on to yogatantra, mahāyogatantra,
etc.37 It is mainly in the later stages that the new elements enter, in the mahāyogatantras and yoginı̄tantras.
37
For the various schemes in which these revelations came to be arranged, cf. Snellgrove 1987, 1988;
Orofino 2001; Dalton 2005.
Tantra and the wild goddesses
259
As we saw in Chapters 6 and 9, the basic orientation of the practice
of Buddha recollection (buddhānusmr.ti) as it appears to have developed in
‘Mahāyāna’ circles was to summon up the Buddhas and communicate with
them, but increasingly also to identify with them as a mode of obtaining
insight.38 This perhaps happens more naturally in Buddhism, which is nondualistic, and where the Buddha is clearly an ideal which people sought to
realise from early on.
The early phases of the Buddhist Tantras develop fairly smoothly out
of the later Mahāyāna sūtras and as mentioned earlier the same text may
in some cases be given either label. This is still true, for example, for the
Mahāvairocanābhisam
. bodhi Sūtra or Tantra (Yamamoto 1990; Wayman and
Tajima 1992; Hodge 2004), which, Stephen Hodge suggests, appeared in
around 640 CE. This text was of major importance both in the Tibetan
tradition, where it is the main text of the caryātantra class and in East
Asia where it forms one of the two principal scriptures of the Shingon or
Japanese Tantric tradition.39
From the perspective of the contemporary Tibetan tradition of the
‘new’ or ‘reformed’ traditions that grew up from the eleventh century
onwards, the important subsequent developments are the yogatantras, for
which the most significant text is the Sarvatathāgatatattvasam
. graha (probably end of seventh century, Matsunaga 1977a: 177–8), and the so-called
anuttarayogatantras, which are much the most important texts for contemporary Tibetan practitioners.
In fact, the reconstructed Sanskrit term anuttarayogatantra is not
entirely appropriate, since it is apparently based on a Tibetan misunderstanding.40 Initially these Tantras fell into two main groups. The first was known
as mahāyoga, and included texts such as the Guhyasamāja Tantra (perhaps
38
39
40
As Hartzell notes, the transition from the situation in the Suvarn.aprabhāsa Sūtra (translated into
Chinese in the early fifth century CE so probably dating from the fourth century) where the meditator
is surrounded by Buddhas in the four directions, and that in the Mahāvairocanābhisam
. bodhi Sūtra,
where the deity Vairocana takes the central place, still needs to be traced and explained (Hartzell
1997: 278–9). The Mahāvairocanābhisam
. bodhi still feels a need to justify and explain the use of a
visualised man.d.ala of deities (Hodge 2003: 89) but clearly includes the self-identification of the
practitioner with the deity (e.g. Hodge 2003: 100, where the Vajra-master transforms himself into
Vajrasattva).
The Mahāvairocanābhisam
. bodhi Sūtra is known as the Dainichi-kyô in Japanese and is one of the
two fundamental sūtras of Shingon, the other being the Kongôchô-gyô (Vajraśekhara), a generic term
for a number of texts of which the most important is the Sarvatathāgatatattvasam
. graha, translated
into Chinese by Amoghavajra in 753. This is regarded as the principal text of the yogatantra class.
(Yamasaki 1988: 83–6; Linrothe 1999: 151–8).
See e.g. Isaacson 1998: 28 n. 11. Isaacson suggests yoganiruttara as the original form; Orofino gives
yogānuttara (Orofino 2001: 545). See also English 2002: 5, again citing Isaacson but suggesting the
form yogottara.
260
Origins of yoga and Tantra
seventh century, though both later and much earlier dates have been suggested)41 and Guhyagarbha Tantra (perhaps early eighth century).42 The
second, probably slightly later, group was known variously as yoginı̄tantra,
bhaginı̄tantra (‘sister’) or prajñā Tantras.43 A variety of classificatory terms
were used in India and Tibet over the next two centuries (cf. Isaacson 1998;
Orofino 2001; Dalton 2005). These Yoginı̄ Tantras include the Hevajra
Tantra (perhaps late eighth or early ninth century), Cakrasam
. vara Tantra
(sometimes referred to simply as Sam
vara
Tantra,
also
probably
from the
.
late eighth or early ninth century) and Kālacakra Tantra (mid-eleventh century, see Chapter 12).44 In most of these cases the Tantric cycle is named
after the central deity.
We should be wary in all these cases about identifying a particular Tantra
with its core text or ‘root Tantra’; what is significant for the tradition is
more the ongoing practice lineage. The core texts were not necessarily stable, with additional sections being added and earlier components modified
for some time, and in some cases, as with the Guhyasamāja Tantra, divergent traditions of practice developed (Matsunaga 1977b; Wayman 1977;
Wedemeyer 1999, 2002, in press).
The Guhyasamāja Tantra was classified as part of the mahāyogatantra
group, as was another important work that has only survived in Tibetan
translation, the Guhyagarbha Tantra (Dorje 1987). As mentioned, the dating
of the Guhyasamāja Tantra is controversial. It gives the impression of being
an early text. The central deities of the Guhyasamāja Tantra (Fig. 10.3) are
relatively normal Buddha-forms with extra arms and attributes, unlike the
innovative and clearly Śaiva-influenced deities of the later yoginı̄ Tantras.
The teachings in the Guhyasamāja Tantra seem to be presented in such a
way as to maximise their transgressive and antinomian nature:
[T]hose who take life, who take pleasure in lying, who always covet the wealth of
others, who enjoy making love, who purposely consume faeces and urine, these
are worthy ones for the practice. The yogin who makes love to his mother, sister
or daughter achieves enormous success in the supreme truth (dharmatā) of the
Mahāyāna. (Snellgrove 1987: 170–1)
41
42
43
44
See discussion in Linrothe 1999: 241–3; Wedemeyer 2001.
On the Guhyagarbha, which contains an early version of the man.d.alas of the Peaceful and Fierce
Deities familiar from the later Tibetan bar-do visualisations, see Dorje 1987; Blezer 1997: 39–56.
Bhaginı̄ means ‘sister’. Prajñā means ‘wisdom’ or ‘insight’, but is probably also in this context a code
term for the female partner.
Complete or partial English translations are available of all of these texts. Fremantle 1971 for the
Guhyasamāja Tantra (there is also a German translation, Gäng 1988); Snellgrove 1959, Farrow and
Menon 1992 and Willemen 1982 for the Hevajra Tantra, Gray 2001 and 2007 for the Cakrasam
. vara
Tantra; Newman 1987, Wallace 2004, Andresen 1997 and Hartzell 1997 for various sections of the
Kālacakra Tantra.
Tantra and the wild goddesses
261
Figure 10.3. Guhyasamāja. Vajrabhairava Temple, Tsaparang, Western Tibet
Perhaps not surprisingly, the audience of Bodhisattvas is terrified, falls into
a swoon and has to be magically revived by the Buddha.
These references to apparently antinomian behaviour, however, do not
appear to be meant literally; the Guhyasamāja Tantra certainly teaches practices of sexual yoga, which were perhaps already current in pre-Tantric
262
Origins of yoga and Tantra
lay Buddhist circles (see Chapter 11) but the references to taking life,
lying, consuming faeces and urine, making love to one’s mother, sister or
daughter, etc., are undoubtedly meant metaphorically or symbolically. The
Guhyasamāja Tantra root text was not interpreted in a radically antinomian
sense by the later Indian or Tibetan tradition (Wedemeyer 2002).
In fact, we can see a series of developments that took place over the
seventh to tenth centuries that led to the introduction of these new deities.
The first of these was the idea of identification with a fierce male deity.
Initially this was Vajrapān.i (also known as Vajrasattva or ‘Vajra being’
although in later Tibetan tradition these are distinct deities), the one-time
fierce yaks.a protector of the Buddha.
Linrothe’s book Ruthless Compassion is a study of these fierce deities, on
the basis of both iconographic and textual material. Linrothe refers to these
deities as krodha-vighnāntaka or ‘fierce destroyers of obstacles’. He sees the
idea of identification with a wrathful deity at the centre of the man.d.ala as
happening in three phases (Linrothe 1999). In the first phase, fierce protective deities (Hayagrı̄va, Yamāntaka) appear in secondary roles in iconography. In the second phase (around the seventh century CE?), they appear as
independent figures: the basic five Tathāgata man.d.ala is generally expanded
in the texts of this period to include the consorts of the deities and four
krodha-vighnāntaka guardian deities, but the consorts or the fierce krodhavighnāntaka can also substitute in secondary man.d.alas for the principal
deities (the Buddhas or Tathāgatas), and a krodha-vighnāntaka may occasionally be found at the centre of the man.d.ala. Phase Two texts include the
Mahāvairocanābhisam
. bodhi Sūtra and the Sarvatathāgatatattvasam
. graha.
In the third phase, appearing in iconography from the late tenth or early
eleventh century, the krodha-vighnāntaka deity normally appears, with or
without a female consort, at the centre of the man.d.ala, often surrounded by
an entourage of eight or a multiple of eight goddesses (the pattern I referred
to earlier in the chapter). The Hevajra Tantra (Fig. 10.4) and Cakrasam
. vara
Tantra would be representative of this phase.
Thus the Buddhist man.d.ala model was progressively adapted by the
incorporation of the wild goddesses and Bhairava-type deities, initially as
guardians and protectors at the edges of the man.d.ala but increasingly as
major figures, as we can see in the Hevajra man.d.ala. The circle of fierce
goddesses comes in this last stage, as does the question of identifying directly
with the goddess (rather than a god) at the centre of the man.d.ala (English
2002).
As implied by my earlier comments, the sexual practices also take on a
new and different significance in the course of this development. I am not
Tantra and the wild goddesses
263
Figure 10.4. Hevajra. Abu Cave, Piyang, Western Tibet
clear how far this is parallel to the Śaiva development. There, as we have
seen, it seems that the initial point is that these practices are there to connect
one to the family (in a suitably transgressive manner) and also to feed the
female spirits. In the Buddhist context, as we will see in Chapter 11, there
are suggestions of sexual yoga at an earlier stage, but they are not necessarily
transgressive.
The Mahāyoga Tantras however include both the transgressive initiation
rituals and the idea of a sexual yoga that is intended to produce a state of
264
Origins of yoga and Tantra
experience of enlightenment.45 Dalton discusses this in some detail on
the basis of Dunhuang meditation handbooks apparently linked with the
Guhyasamāja Tantra and Guhyagarbha Tantra (see Chapter 11). The more
complex subtle-body practices however develop at a somewhat later phase,
and are first fully attested in the texts of the mid to late eighth century
onwards (Hevajra Tantra, etc.; see Dalton 2004).
The sexual practices raise the question of how far practitioners were
celibate. This is a complex issue, but it seems that in the Indian context these
late Tantras with the associated sexual yogas were not regularly practised in
the monastic context. Certainly the iconography is mostly small-scale and
apparently considerably later than the surviving texts. This may suggest
that the travelling siddha model that we find in the tradition was not
entirely inaccurate, and that Tibetan descriptions of the great monasteries
of northeast India as centres of Tantric teaching, if accurate, refer to a very
late stage.46 We might also imagine however that these were private interests
of particular monks rather than being the basis of large-scale monastic
ritual.47
If Sanderson is right in arguing that the Yoginı̄ Tantras represent a
Buddhist appropriation of Śaiva Tantra, how did this take place? In fact, as
I have noted above, one can ask how far this really is an adaptation of ‘Śaiva
Tantra’. For both Buddhist and Śaiva Tantra the rise to primacy of female
45
46
47
Nāgārjuna, who is traditionally associated with the Guhyasamāja Tantra lineage, is regarded in
later Indian thought as an alchemist and a specialist in sexual techniques. This however raises the
problematic topic of whether there was more than one author using the name Nāgārjuna (Ray 1997;
Mabbett 1998). As Wedemeyer notes, the historiographical issues are certainly not fully resolved
(Wedemeyer 2001), but to assume that the Madhyamika philosopher Nagārjuna was associated
with the Guhyasamāja Tantra raises more problems than it solves. Wedemeyer has recently argued
that the attribution to Nagarjuna should be understood more in terms of visionary inspiration by
Nāgārjuna and/or the assertion that the tradition represents a valid continuation of Nāgārjuna’s
teaching; this seems to me to make considerable sense (Wedemeyer 2006).
In Tibet, the Bengali monastic teacher Atiśa seems to have hedged his bets somewhat on the issue,
saying that the sexual practices should not be undertaken by monastics, but inserting a proviso that
they could be performed if the practitioner had knowledge of śūnyatā (Samuel 1993: 471). Davidson
however translates this passage differently to imply that a celibate yogin can attain knowledge
of śūnyatā without recourse to the sexual practices (Davidson 1995a: 301). The sexual practices
already had great prestige, and even the school of the great fourteenth century lama Tsong-khapa, known for its emphasis on celibate practice and opposition to the sexual practices, accepted
that they had to be performed in some form for the attainment of Buddhahood (Samuel 1993:
276, 511).
The rDzogs-chen tradition also avoided the sexual practices. This tradition developed mostly in
Tibet, but seems to go back to India and perhaps to earlier origins outside the subcontinent. It
presented an alternative path to the attainment of the central liberating insight of the tradition, with
its own specific methodology, and initially at least rejected most of the Tantric techniques (Germano
1994, 2005).
Tantra and the wild goddesses
265
divinities and the sexual yogas seem to have been a new development.
These developments seem to have taken place more or less at the same time
among Buddhists and among Śaivas, and borrowings have been traced in
both directions.48
The female divinities may well best be understood in terms of a distinct
Śākta milieu from which both Śaivas and Buddhists were borrowing; certainly the historical relationship between Śakta and Śaiva practices is far
from clear. I shall discuss the sexual yogas in more detail in Chapter 11;
in this case there are both suggestions of an earlier history in a Buddhist
context and of borrowings from or connections with circles outside South
Asia.
Other elements, such as the kāpālika-style practice, and perhaps also the
use of sexual and other polluting substances in initiation and ritual magic,
seems more clearly to have had a Śaiva prehistory before their Buddhist
adoption. It is apparent, too, that these elements were seen as problematic
in the Buddhist context, and as needing some kind of justification; we will
see below how this was undertaken.
The pı̄t.has, the Śākta cult centres associated with the new practices for the
Śaivas were centres for the Buddhists as well. As mentioned above, we find
more or less the same lists of sites in sources from both traditions, including
four main sites and various subsidiary sites, yielding for both the Śaiva
Kubjikā tradition and the Buddhist Cakrasam
. vara three circles of eight
sites. Many of the sites are also common to both lists. It is not clear however
how far we should accept the picture of a cultic underground presented by
the texts, in which practitioners are meeting at the pı̄t.has, recognising each
other through special signs and coded language, and coming together for
ritual gatherings. It is not intrinsically unlikely that such things happened.
In fact, we have seen something like this with the shared early BuddhistJain-Ājı̄vika-Brahmanical renunciate scene, and at a later stage there were
sharings between Sufis and Hindu yogis, still visible today in the dual
identities of Nāth gurus (van Skyhawk 1993, Khan 2004) and in the Bengali
Bāul traditions (Openshaw 1997, 2002; McDaniel 1989, 1992; Salomon
1991). However, there was also an alternative mode of practice in which
practice was largely internalised. I turn further to this in Chapter 11.
The incorporation of all this material within the Buddhist tradition
involved a certain amount of reconceptualisation of Tantric thought within
the Buddhist frame. Some things could be taken over fairly directly.
48
See references given at beginning of chapter.
266
Origins of yoga and Tantra
The Buddhist equivalent of Bhairava, Mahākāla, has virtually the same
iconography and became the protective deity of Buddhist Tantrics just
as Bhairava was the protector of Śaiva Tantrics (see Stablein 1976; Snellgrove 1987: 150, 317; Linrothe 1999: 123–4).49 However, it had to be made
clear that the Buddhist path was pre-eminent, and that Tantra was not
simply borrowed from the Śaivas. Sandy Macdonald (1990), Rob Mayer
(1996), Toni Huber (1999), Ronald Davidson (1991, 1995b, 2002a) and
others have helped to show how this was achieved, in particular through
the Cakrasam
. vara myth, which legitimates the Buddhist takeover of Śaiva
sacred sites. Mayer summarises this conversion myth as follows:
A variety of evil non-human spirits occupy and rule over the 24 power places; they
invite Śiva in extreme kāpālika form as Bhairava and his consort Kālarātri/Kālı̄ to
be their lord, and they rule the 24 power-places in Bhairava’s name. Bhairava is too
busy making love to Kālarātri/Kālı̄ to visit the sites in person, so he sets up liṅgams
to represent himself and receive worship in each of the 24 power places. In this
way, Bhairava and his following establish control over all who move in the entire
triple world (khecara, bhūcara and nāgaloka), and encourage transgressive Tantric
practices that lead beings to hell. In response, Buddhist herukas emanate in forms
identical to Bhairava, kill Bhairava, revive him, convert him to Buddhism, absorb
all his wealth and power, and seize the power places and Bhairava’s accoutrements,
consorts and retinues and even mantras for their own Buddhist use. Bhairava and
Kālarātri joyfully offer themselves in devotion as the Heruka’s seats, as do all their
following. (Mayer 1996: 119, referring to Kalff 1979: 67–76; Huber 1999: 40–41
and 236 n.5)
Mayer connects this sequence by which Śiva and other deities of his
entourage are killed, liberated and converted into devotees of the Buddha
to the well-known Indian pattern of ‘demon devotees’, a term introduced
by Alf Hiltebeitel (Hiltebeitel 1989). Śiva becomes, in effect, the Buddha’s
‘demon devotee’, the symbolic representation of the attachment to self
which the Buddhist practitioner has to overcome (Mayer 1996: 109ff ). He
and related deities (e.g. Mahākāla) then take on the role of ‘divine policemen’, protectors and assistants of the peaceful and elevated Buddhist Tantric
deities.
The conversion process also made it appropriate for Buddhist Tantric
deities such as Vajrakı̄laya or Vajrabhairava themselves to take on the fierce
and aggressive nature of the Śaivite Tantric deities. Here the foremost and
prototype figure was Vajrapān.i, who was now upgraded from his status as
49
Though Rhie says of Mahākāla only, ‘Possibly there has been some relation to or assimilation of
certain aspects of the Hindu god Shiva’ (Rhie 2004: 44).
Tantra and the wild goddesses
267
the Buddha’s yaks.a-protector to be the key representative of the new kind
of terrifying and aggressive deity who was nevertheless identical with the
ultimate nature of the Buddha’s enlightenment.50
This Buddhist adoption of Tantra also created the opportunity for the
development of a new and powerful set of meditational techniques, in
which negative emotions and behavioural modes such as anger could be
transformed directly into the divine anger of the fierce Tantric deity and
directed towards the destruction of negative forces within the practitioner.
The visual and dramatic performance of these techniques, in the masked
ritual dances which are still an important practice in most major Tibetan
monasteries, maintained the dual orientation or dual interpretation towards
the destruction of obstacles to the successful pursuit of the goals of everyday life, and the destruction of obstacles for the practitioner’s progress to
enlightenment.
The above discussion has focused on Buddhist (Vajrayāna) and nondualist Śaiva Tantric traditions. There were other Tantric traditions, both
among Brahmanical and non-Brahmanical schools, but to the extent that
these are known to us in any detail they have steered clear from the antinomian and transgressive elements conspicuous in the nondualist Śaiva and
the later Vajrayāna traditions. They include Śaiva traditions (principally
Śaiva Siddhānta) and the Vais.n.ava Tantric school known as the Pāñcarātra
Āgama, which was mentioned briefly in Chapter 8. These two traditions
are major sources of later Indic ritual procedures, and indeed ‘Tantric’ in
this context primarily implies that these are sources of non-Vedic ritual
procedures (Colas 2003). They also share, however, much of the ritual
structure, techniques and theoretical presuppositions of the nondualist
Śaiva traditions (Sanderson 1995; S. Gupta 1972, 1989, 1992; Flood 2006:
99–145).
The Jaina tradition too shared some of these techniques and procedures, and developed a specialist class of ritualists performing Tantric-style
ritual primarily to the various yaks.a and yaksı̄ deities associated with the
tı̄rthaṅkaras (Jaini 1991: 196). These deities appear to have progressively
developed as male and female ‘protectors’ of the tı̄rthaṅkaras. The earliest
iconographic and textual evidence is from around the sixth century CE,
50
The Śaivas, naturally, developed their own response to these myths of Buddhist supremacy. Sanderson
narrates a ‘well known Śaiva myth’ according to which Br.haspati invented Vajrayāna Buddhism,
complete with its images of Buddhist deities standing on the heads of Śaiva deities, as a ruse to
distract the demons from their devotion to Śiva so that they could be destroyed (Sanderson 1994:
93).
268
Origins of yoga and Tantra
and there appears to be a progressive development from an original malefemale pair to separate male and female deities for each of the twenty-four
tı̄rthaṅkara. These yaks.a and yaksı̄ deities offered a context in which various
existing deity-cults could be acknowledged within Jaina practice (Cort 1987;
J. Sharma 1989). Some of these deities, for example the goddesses Padmavatı̄
and Jvālāmālinı̄, have elements in common with the fierce goddesses known
from the Śaiva and Buddhist traditions, and would appear to derive from
a similar context. The same would appear to be true of Vāgı̄śvarı̄, a fierce
form of Sarasvatı̄ whose worship was popular in the mediaeval period, and
the so-called vidyādevı̄s, such as Cakreśvarı̄ (Fig. 10.2), associated with magical powers attained through ascetic practice (Cort 1987: 237–40; cf. also Pal
1994, Cat.No.67). Rituals of Padmāvatı̄ can be used to perform a set of six
ritual actions (s.āt.karman) closely similar to sets of four and six ritual actions
in Buddhist and Śaiva sources (pacification, subjugation, etc.) (Cort 1987
245–6; see also Jhavery 1944). The Digam
. bara and Śvetām
. bara Jains have
somewhat different sets of yaks.as, yaksı̄s and vidyādevı̄s (Cort 1987: 253–4
nn. 30, 33).
None of these cults have a soteriological aim, however, and it appears
that the Jains employed fierce goddess rituals only for pragmatic purposes.
There does not seem to be any cult parallel to that of the major Buddhist
or Śaiva Tantric deities, with whom the practitioner identifies as a way of
accessing the central liberating insight of the tradition. However, as Olle
Qvarnström has noted, much research still remains to be done before we
have a clear picture of Jaina Tantra (Qvarnström 2000: 595).
Thus it seems to have been principally among the two groups I discussed above, the nondualist Śaivas and the Mahāyāna Buddhists, that the
development of ‘Tantra’ as a transgressive spiritual path took place.
conclusion
I began this chapter, after some preliminary discussion of Western attitudes
to Tantra, with the low-caste ritual practitioners who supported the role
of South Indian kings and perhaps had equivalents elsewhere outside the
Vedic regions of early India. These people were specialists, in a sense, in
death and misfortune, and their low and polluted status, however necessary
to society, derived from this involvement.
I then suggested that early Indian renunciants, particularly the Pāśupata
and those who had adopted a kāpālika style of practice, had in a sense
made a choice to live in this world of death and misfortune rather than the
everyday world of the householder. We then turned to the fierce goddesses
Tantra and the wild goddesses
269
of illness and disease and I suggested that the Saptamātr.kā deities used in
state ritual by the Gupta kings represented a Brahmanical appropriation of
these forces of misfortune. It is these dangerous goddesses who come to take
a leading role in the more transgressive aspects of ‘Tantra’, and the kula or
‘family’ traditions by which they are initially approached may derive from
earlier hereditary low-caste practitioners.
All this suggests a convergence of interests in the early stages of the transgressive styles of Tantra. This convergence is between low-caste, polluted
ritualists from hereditary backgrounds, cult-groups of the Pāśupata and
kāpālika kind who were already close to them, and a new style of ritual for
which the script was perhaps written in large part by the court Brahmins of
the Gupta kings and others of the new Brahmanical regimes of the fourth
and fifth centuries onwards.
In this case, the significance of pollution and impurity, as David White
has suggested (White 1998, 2003) was primarily to do with the power of
the ritual. Thus the initial rationale for the ritual use of sexual substances
was that they were extremely polluting, powerful and therefore effective
when used transgressively, rather than because sex as such was seen as a
source of spiritual fulfilment. Given the logic of Brahmanical thought, the
more polluted, extreme and transgressive the ritual and its performers, the
more powerful it might be. Thus kāpālika-type practitioners, to the extent
that they made a living by selling ritual services to local rulers and other
wealthy people, had some interest in both exaggerating their transgressive
behaviour, and in developing a mythic justification for it.
From within the practice tradition, the perspective was quite different.
For their employers, the important things about these early Tantric practitioners is that they were unsavoury, dangerous and therefore effective at
bringing about the this-worldly ends which the employers wanted. For
the practitioners themselves, however, this rejection of everyday life had a
positive spiritual value; it was a path to a liberating insight.
This may already have been the case to some degree for the low-caste
groups who took on this role in the South Indian state. Anthropologists
have found that most low Indian castes today have, at the least, caste myths
which justify a higher status than that assigned to them by the surrounding
society (e.g. Lynch 1969). Many, too, have adopted religious perspectives,
such as that of the low-caste saint Ravidas, which see the caste hierarchy
and the value systems of respectable society as irrelevant to human worth
(e.g. Khare 1984).
In any case, it makes sense both that the initiation by which renunciates
took on this new identity was specifically centred around pollution (the
270
Origins of yoga and Tantra
ingestion of the sexual fluids of guru and consort) and was also seen as an
occasion for the direct perception of the liberating insight.
The self-identification as a deity which is a key element of Tantric practice
may have originated primarily as a way of accessing divine power for ritual
purposes, for example by the ritualist seeking to control a circle of fierce
goddesses, but this too could easily be read in spiritual terms. Certainly, as
Paul Muller-Ortega has shown, ‘becoming Bhairava’ by the tenth century
CE could be reinterpreted by Abhinavagupta in purely spiritual terms, as a
mode of access to nondual consciousness. ‘Here, Bhairava comes to mean
the unencompassable and exquisitely blissful light of consciousness that is
to be discovered as the practitioner’s true inner identity’ (Muller-Ortega
2002: 213).
The possibility for such interpretations, which may well go back to quite
early stages, helps to explain why practices of this kind could appeal to Buddhists as well. It is generally assumed that these transgressive practices in
Buddhism were initially carried out in non-monastic contexts and only at
a fairly late stage – perhaps tenth or eleventh century – started to become a
significant part of the monastic scene. By this stage there is evidence, particularly outside India proper, for the Pāśupatas and kāpālika-style practitioners having long taken on a role as court ritualists (see Chapter 12). This
would have brought them into direct competition with the Buddhists, and
one can see a double attraction of the new approaches for the Buddhists:
they offered new, more powerful ritual techniques that would be competitive with those of their Śaivite competitors, and they also offered new, more
effective and direct, modes of accessing the central goal of Buddhism itself,
that of Buddhahood.
What we have not yet examined in any detail is how the sexual practices
might have operated as a form of meditation to achieve liberating insight.
This question is tied up with the emergence of the subtle body practices
that form the main subject of Chapter 11.
chap t e r 11
Subtle bodies, longevity and internal alchemy
As we have seen, a striking innovation in both Śaiva and Buddhist traditions in the course of the eighth to tenth centuries was a new series of yogic
practices and techniques based on a subtle anatomy of the practitioner’s
body. Specifically, these practices assume an internal ‘subtle physiology’
of the body (or rather of the body-mind complex) made up of channels
(nād.ı̄ ) through which substances of some kind flow, and points of intersection (cakra) at which these channels come together. The practices typically
involve moving these substances through the body, clearing ‘knots’ in the
flow, and directing the substances into a central channel flowing along the
spine (Fig. 11.1). I refer to these practices, as do a number of other authors,
as ‘subtle body’ practices (Samuel 2005b; White 1996).1 What I shall do in
this chapter is to track this new pattern and explain how it developed.
A key element of the ‘subtle body’ practices is their close connection with
sexual practices. Sexual intercourse, real or imagined, is used as a way to
stimulate the flow of substances along and within the body, a process which
is associated in various ways with meditative techniques for the attainment
of health, long life and/or liberating insight. The subtle body practices thus
provide a new and different purpose and significance for sexual practices
within the Tantric context and increasingly take over from earlier views of
sexual practice as being about initiation or feeding fierce female spirits.
sexual practices in the gupta period
An obvious place to look for the development of sexual practices in early
mediaeval India would seem to be the kāmaśāstra literature, the technical
1
The term ‘subtle body’ is also used in English to translate sūks.ma-śarı̄ra, which is a key concept in the
Vedantic philosophy of Śan.kara. This is the second of a series of three increasingly subtle ‘bodies’, the
material or physical body (sthūla-śarı̄ra), subtle body (sūks.ma-śarı̄ra) and causal body (kāran.a-śarı̄ra),
beyond which lies identity with the ultimate self, itself identical with Brahman (Samuel 2005b). It
should be noted that I am using the term here in a wider sense.
271
272
Origins of yoga and Tantra
Figure 11.1. Meditator with Cakras and Kun.d.alinı̄
literature on erotics, of which the best-known text is of course the Kāmasūtra
of Vātsyāyana (e.g. Upadhyaya 1961). Vātsyāyana is variously dated between
the first and fourth centuries CE; Daud Ali suggests the early third century
(Ali 1998: 164). Vātsyāyana represents himself as the latest in a sequence
of authorities on kāmaśāstra, often referring back to the opinion of his
Subtle bodies, longevity and internal alchemy
273
predecessors. None of the works of these earlier authorities survive, but
there are many later works on kāmaśāstra, mostly considerably shorter than
Vātsyāyana and reflecting the changed social circumstances of later times
(Archer 1964; Comfort 1964).
Anyone looking at the Kāmasūtra, or for that matter at later kāmaśāstra
texts, for enlightenment on sexual yoga and the subtle body will, however,
be disappointed. It is clear from Indian and Tibetan texts on sexual yoga
that their authors had some familiarity with the kāmaśāstra literature (e.g.
George 1974: 71–3; G. Dorje 1987: 900–7; Guenther 1963: 77–8), which was
evidently seen as relevant knowledge by practitioners of these techniques.
The influence, however, is all one way; there is nothing in the work of
Vātsyāyana or his successors to suggest any knowledge of or concern with
sexual yoga. The kāmaśāstra literature is about the pursuit of sexual pleasure
(kāma), not the quest for liberation (moks.a).2
At the same time, there are aspects of the Kāmasūtra that are worth our
attention. As Daud Ali noted, the Kāmasūtra is part of what Foucault has
called a ‘technology of the self’ (Foucault 1988a, 1988b).3 In other words,
this is a text that presents a ‘code’ or set of rules on how to live one’s
life. Only one of the seven sections of the Kāmasūtra deals with the actual
physiology and technique of sex; most of the book is concerned with laying
down the appropriate codes of behaviour for men and women within and
outside marriage. Ali suggests that as a ‘technology of the self ’, this courtly
code of behaviour bears a notable resemblance to the monastic code as
laid down in the Buddhist monastic Vinaya. He suggests that despite the
different states of life at which these codes are directed, ‘the technologies
deployed for the achievement of these states share a common and novel
conception of reality as a complex system of “signs”, “marks”, and “surfaces”
to be engaged and disengaged with, manipulated and deconstructed’ (Ali
1998: 159). This novel conception of reality was part of the classical pattern
discussed in Chapter 9. Ali goes on to argue that ‘the technology of Buddhist
monastic discipline was [. . .] constituted in “complementary opposition”
to the practices of the urban courts’ (Ali 1998: 163).
Buddhist treatises, like those of other renunciate traditions, needed to
bring about this reversal, to seek to transform the code of everyday life into
2
3
While the Kāmasūtra begins with a homage to dharma, artha and kāma, the three classical aims of
life (purus.ārtha) in Indian thought, excluding moks.a, a later verse recognises moks.a as an appropriate
pursuit for the closing years of life (Rocher 1985).
Technologies of the self according to Foucault ‘permit individuals to effect by their own means or with
the help of others a certain number of operations on their own bodies and souls, thoughts, conduct
and way of being, so as to transform themselves in order to attain a certain state of happiness, purity,
wisdom, perfection, or immortality’ (1988a: 18). See also Samuel 2005a: 335–8, and Chapter 14.
274
Origins of yoga and Tantra
that of the renunciate who has stepped outside the concerns of sam
. sāra,
whether or not he or she is still physically present within it. The passage in
the Mahāyānasūtrālam
. kāra (IX, 41–7) of Asaṅga, a classic Buddhist treatise
dating probably from the fifth century CE, which Snellgrove has pointed
out as containing indications of sexual yoga can be understood in these
terms (Snellgrove 1987: 126–8).4 It forms part of a series of reversals of
ordinary experience:
Supreme self-control is achieved in the reversal of the five sense-organs with regard
to the universal operation of all of them, associated wth the manifestation of twelve
hundred good qualities.
Supreme self-control is achieved in the reversal of mental activity with the
consequent self-control with regard to knowledge which is free of discriminating
thought and thus totally immaculate.
Supreme self-control is achieved in the reversal of appearances and their (imagined) significance in a (Buddha-)realm that is thus purified for the blissful vision
just as desired.
Supreme self-control is achieved in the reversal of discriminating thought resulting in the nonobstruction at all times of all knowledge and acts.
Supreme self-control is achieved in the reversal of substrata resulting in that
imperturbable state of the Buddhas, nirvān.a without any substratum.
Supreme self-control is achieved in the reversal of sexual intercourse in the
blissful Buddha-poise and the untrammelled vision of one’s spouse.
Supreme self-control is achieved in the reversal of spatial perceptions resulting
in the supernatural production of thought-forms and in material manifestation in
phenomenal spheres (gati). (Snellgrove 1987: 127–8)
Snellgrove suggests, plausibly, that the reference to the ‘reversal of sexual
intercourse’ here indicates the deliberate avoidance of male ejaculation in
a form of sexual yoga:
It is by no means improbable that already by the fifth century when Asaṅga was
writing, these techniques of sexual yoga were being used in reputable Buddhist
circles, and that Asaṅga himself accepted such a practice as valid. The natural
power of the breath, inhaling and exhaling, was certainly accepted as an essential
force to be controlled in Buddhist as well as Hindu yoga. Why therefore not the
natural power of the sexual force? [. . .] Once it is established that sexual yoga
was already regarded by Asaṅga as an acceptable yogic practice, it becomes far
easier to understand how Tantric treatises, despite their apparent contradiction of
previous Buddhist teachings, were so readily canonized in the following centuries.
(Snellgrove 1987: 127)
4
On the interpretation of this controversial passage see also Thurman 2004: 89.
Subtle bodies, longevity and internal alchemy
275
The point is well taken, but one should also note that there is nothing particularly transgressive in Asaṅga’s text. The reference is explicitly to practices
carried out in the context of a marital relationship, not by supposedly celibate Buddhist practitioners.5 There is no suggestion here of a context such
as that of the kāpālika-style practitioner, with its transgressive use of sexual
substances in frightening and dangerous places to attract dangerous female
spirits. We need to be careful here of the temptation to lump all sexual
practices together.
At most we might note that sexuality is closely tied up with sam
. sāric
existence, so that its occurrence in this context seems somewhat paradoxical.
However the Mahāyānasūtrālam
. kāra elsewhere notes (XIII.11, 13):
There is no element (dharma) apart from the elemental sphere (dharmadhātu),
So passion, etc. (viz., wrath and delusion) serve as their own extrication in the
opinion of the Buddhas. [. . .] In that one has recourse to them, passion and the
rest, at the source (yon.iśah.) one is released by their means; thus they are their own
extrication. (Snellgrove 1987: 126)6
This is essentially the same justification that was later to be offered by the
Tantric texts themselves: through passion one becomes free from passion.
Here we might return to Ali’s discussion of the Kāmasūtra and the wider
courtly culture of the period, which regarded sexual pleasure both as an
entirely legitimate object of pursuit for men and for women but also as
a source of danger if one did not maintain an appropriate degree of selfdiscipline and self-mastery in one’s approach to it (Ali 1998, 2004: 240–1).
Over-attachment to sexual pleasure could lead to the ‘royal disease’,
rājayaks.man (Ali 2004: 242–5), a weakening and wasting away of the
body caused less, Ali suggests, by physical over-indulgence than by ‘an
excessive predisposition or inclination of the mind towards sexual desire’
(2004: 243).7 One could imagine a ready opening in a culture preoccupied
with such matters for teachings that involved male seminal control and
5
6
7
Snellgrove’s ‘spouse’ translates dāra, ‘wife’, so it is clear that the partners are married. The lack
of reference in the Chinese pilgrims’ accounts suggests that such practices, and indeed any distinctively ‘Tantric’ observances, were absent from monastic contexts up to the early seventh
century.
Thurman translates: ‘Since no thing is found outside of the ultimate realm, the buddhas consider that
passions and so on constitute their own transcendence [. . .] Thus, one properly engages with the passions and so forth, and thereby becomes liberated from them; such is their transcendence’ (Thurman
2004: 169–70). Snellgrove suggests that yon.iśah. (‘at the source’) here may have an intentional double
meaning ( yon.i = vagina) (Snellgrove 1987: 126).
The disease of rājayaks.man is described in Caraka’s medical treatise (Ali 2004: 242–3). See also David
White, who points out the connections between ideas of rājayaks.man and the underlying logic
of rasāyana (elixir or rejuvenation therapy) and vājikaran.a (sexual rehabilitation therapy), major
concerns for both the Āyurvedic tradition and the related discipline of medical alchemy (1996: 24–5).
276
Origins of yoga and Tantra
self-mastery in the sexual context. Our evidence for such practices is, however, only indirect.
There is at least a possibility, then, that some kind of sexual yoga existed
in the fourth or fifth centuries. Substantial evidence for such practices,
however, dates from considerably later, from the seventh and eighth centuries, and derives from Śaiva and Buddhist Tantric circles. Here we see
sexual yoga as part of a specific complex of practices. On the Śaiva side
this is associated with a series of named teachers in South and North India,
the Cittar (Siddha) teachers in the south, including Tirumūlar and Bogar,
and the so-called Nāth teachers in the north, where the principal names
are Matsyendra (Matsyendranāth) and Gorakh (Gorakhnāth). On the
Buddhist side, it is associated with so-called mahāyoga Tantras. These developments appear to be happening at more or less the same time in all three
areas, again suggesting a shared ascetic culture. We begin with the Śaiva
developments.
the tamil cit tars and the n āth reforms
The Tamil Cittars in the South and Matsyendra(-nāth) and his companions
in the North are all legendary figures in the form that we know them.
Their names are however attached to a coherent body of approaches and
techniques. The elements of the pattern include practices aimed at health,
long life and immortality, sexual practices, internal ‘subtle body’ practices,
and alchemical practices (rasāyana) associated with mercury-based elixirs.
Sanderson and White associate Matsyendra with the translation from
the kula traditions, with their cremation ground associations and symbolism, to the reformed, domesticated, kaula traditions (Sanderson 1985: 214
n.110; White 1998: 173). This enabled the techniques developed among
kāpālika-style ascetics to be adopted by the ‘wider community of married
householders’ (Sanderson 1988: 679). In these cults sexual ritual has been
‘aestheticised’:
The magical properties of the mingled sexual fluids are not forgotten: those seeking
powers (siddhis) consumed it and even those who worshipped for salvation alone
offered the products of orgasm to the deities. However the emphasis has now moved
to orgasm itself. It is no longer principally a means of production. It is a privileged
means of access to a blissful expansion of consciousness in which the deities of
the Kula permeate and obliterate the ego of the worshipper. The consumption
of meat and alcohol is interpreted along the same lines. Their purpose, like that of
everything in the liturgy, is to intensify experience, to gratify the goddesses of the
senses. (Sanderson 1988: 680)
Subtle bodies, longevity and internal alchemy
277
Something else however can also be seen in the Cittar and Nāth teachings
than the blissful expansion of consciousness, and this is the emphasis on the
practices as giving rejuvenation and immortality. This is a major theme, for
example, in the Kaulajñānanirn.aya, a text often ascribed to Matsyendranāth
and apparently dating from the ninth or tenth century (White 1996: 73):
Devi said – O Grace Giver, Mahādeva, speak freely of the conquering of death. [. . .]
Speak from Your inmost being, Mahādeva, (of that method by which) a person
may roam on earth immortal! [. . .]
Dear One, listen one-pointedly to the details (of this knowledge) I speak of.
(One should meditate) inwardly on a pure, white, cool, celestial, fragrant and vast
effulgence, the heavenly cause of all lunar refreshment, like the centre of space,
flowing within oneself, through all the many channels. Doing this one becomes
long-lived and conquers death. (Kaulajñānanirn.aya 5, 1–6, Bagchi and Magee 1986:
text, 13–14)
The remainder of this pat.ala or chapter of the Kaulajñānanirn.aya gives
further details, again focusing on the imagery of the inner moon and its cool
moist white rays;8 subsequent chapters give further methods for defeating
old age and disease. Later chapters present another key element of the
Nāth tradition, the sequence of cakras along the spinal column and the
importance of the ‘raising’ of the goddess (śakti, Kun.d.alinı̄) from the lowest
cakra up through the cakras, so ‘piercing’ through the upper cakras and
attaining the liberating insight of this tradition. This practice is closely
associated with sexual yoga, and the earliest descriptions of it in Śaiva
sources come from this text and the Kubjikātantra, a text belonging to a
slightly later branch of the kaula reform movement (cf. Sanderson 1988:
686–8; Heilijgers-Seelen 1994; White 1996: 134–5).
The emphasis on defeating old age and disease extends beyond these
inner yogic practices, since these Siddha traditions were also involved in
physical alchemy with the aim of producing elixirs of immortality. These
were mostly based on mercury, which shared the cool white lunar associations visible in the above quotation from the Kaulajñānanirn.aya. As White
has pointed out, the imagery of elemental mercury also links up with soma
(another name for the moon, as well as the sacrificial substance of Vedic
ritual), and seminal fluid, whose preservation and use is a central issue for
the sexual yogic practices. Mercury is identified with the seed of the god
8
Variants of this cool white light practice are common in the Śaiva and Buddhist literature; a dateable
example from a literary source comes from Ratnakara’s Haravijaya (between 826 and 838; D. Smith
1985: 265). The Vajrasattva purification practices of present-day Tibetan Vajrayana have a similar
structure, involving visualisations of purifying white nectar pervading the practitioner’s body (e.g.
Beyer 1973: 434–5).
278
Origins of yoga and Tantra
Śiva, caught in the mouth of Agni the fire god and scattered in deep wells in
the earth, while the other principal substance of Indian alchemy, sulphur,
is associated with the menstrual blood of the Goddess (White 1996: 191–3,
213). Outer or physical alchemy (rasāyana) was most fully developed by the
rasāyana siddhas of Western India, but it was important in the Tamil Cittar
tradition and appears in association with inner alchemy throughout these
traditions.9
t h e orig ins of the sex ual and alchemi cal practices
Where though did this new body of complexes come from? How did it
originate? Here we can examine individual elements of the complex, and
we can also look at the complex as a whole. If we take the second option,
the similarities with Chinese alchemical practices and ‘technologies of the
self’ are striking and have long been observed (Filliozat 1969; White 1996:
53–4, 61–5). These Chinese practices, closely associated with Daoism, have
been described in detail by a number of scholars (for recent accounts see
Schipper 1994; L. Kohn 2006a), and there is little doubt that they predate
the Indic versions by several centuries. Early versions of these practices are
found in the Mawangdui manuscripts, recovered from a burial dated to
168 BCE (D. Harper 1987, 1997). They include sexual practices based on
the internal movement of ‘subtle’ bodily substances (qi), are aimed at long
life, health and immortality, and are again linked to alchemical practices
associated with mercury-based elixirs (e.g. Schipper 1994: 174–81).10
The similarities here are quite close; thus one can parallel the Daoist practise of ‘living without grains’ (bigu), involving extended fasts during which
the practitioner eats only specially-prepared alchemical compounds (Shawn
2006, Jackowitz 2006, Schipper 1996: 167–70), with similar rasāyana practices, still carried out in the Tibetan tradition (bcud len).11 One might also
9
10
11
For Buddhist versions of these alchemical practices, see Fenner 1979; Walter 1980; Samuel 2006b.
Mercury-based medicines are mentioned in the Āyurvedic treatise of Vāgbhat.t.a, so it appears that
they were already entering the Āyurvedic pharmacopoeia at around the same time. Apart from
possible Chinese relationships, in this case there are also possible connections with Islamic alchemy,
whose rather more historical founding figure Jabir ibn Hayyan also dates from the eighth century.
There was certainly a developed alchemical tradition in Western India at a slightly later point.
One also might note that the original mise-en-scène of the Chinese ‘bedroom arts’ is said to be the
need of Chinese emperors to maintain the sexual abilities needed to deal with their large harems and
produce enough descendants, as well as to attain longevity (Winn 2006: 153). Early Chinese texts
on sexology, such as the Sunü Jing, are often presented in the form of a dialogue between a female
sexual adept and the Yellow Emperor, the mythical founding-figure of the first Chinese imperial
dynasty (see also Mussat 1978; Wile 1992). Here we might note the significance of rājayaks.ma or
‘royal consumption’ as a disease in Indian medical and literary convention (White 1996: 24–6).
Subtle bodies, longevity and internal alchemy
279
point to the similarity between the Chinese meditational practice of the
‘microcosmic orbit’, in which qi is directed along a circuit which moves
up the spine from the perineum and returns down the front of the body
and the Indian kriyā practices which utilise the same circuit. In both cases,
a crucial link is made by contact between the tongue and the back of the
upper palate (known as khecarı̄ mudrā in Sanskrit).12 In both cases, Chinese
and Indian, physical exercises are also used to bring the body as a whole
into a suitable condition to perform the internal practices. These are the
exercises that later become known in various versions as daoyin or qigong in
the Chinese context, as hat.ha yoga in the Śaiva context, and as ’phrul ’khor
in the context of the Tibetan Vajrayāna.13
At the same time, there is no doubt that the Indian version of these
practices is as thoroughly Indian in its vocabulary and conceptual structure
as the Chinese version is Chinese. Compare the internal landscape of the
Chinese practices according to the Huangting Jing (‘Book of the Yellow
Court’):
In the Yellow Court sits someone dressed in scarlet.
The door is locked, its two leaves tightly closed.
The Dark Towers rise to vertiginous heights.
In the Cinnabar Field, semen14 and breath subtly mingle.
Above, the clear water of the Jade Fountain flows abundantly,
Making the Divine Root sturdy and hard; it will not ever weaken.
In the Center lake a noble person, dressed in red.
Below lies the Field, three inches away; that is where the god lives.
Lock the passage between the Inner and Outer with a double lock.
Always keep the Hut spotlessly clean.
When suddenly you receive through the tunnel of the Mysterious
Meridian the semen’s signal,
You should quickly retain semen and hold yourself together.
(Schipper 1994: 140–1)
with that of the S.atcakranirūpana (‘Description of the Six Cakras’):
12
13
14
However, rājayaks.ma was already a matter of concern for Āyurvedic texts before the development
of Chinese-style sexual yogic practices.
For modern versions of these practices, see Chia 1983 for China; Saraswati 1985 for India. On the
importance of the khecarı̄ mudrā in India, see White 1996: 252–4.
Of these, the ’phrul ’khor exercises are the least known to Western scholars; the only substantial
academic publication so far is Loseries-Leick 1997. On the Tantric origins of hat.ha yoga, see Alter
2005. Kohn 2006b describes daoyin and compares them with yoga, but discounts the possibility of
mutual influence because of yoga’s ‘strong otherworldly and theistic orientation’ (2006b: 129). Here
she is perhaps misled by popular Vedantic-influenced Western accounts of yoga. It is noticeable
however that she sees hat.ha yoga as developed by the Nāth yogis as being closest in orientation to
daoyin (L. Kohn: 2006b: 141–2).
According to Schipper, the term used can also refer to female sexual secretions (Schipper 1994: 140).
280
Origins of yoga and Tantra
Near the mouth of the Nād.ı̄ called Vajrā, and in the pericarp (of the Ādhāra Lotus),
there constantly shines the beautifully luminous and soft, lightning-like triangle
which is Kāmarupa, and known as Traipura. There is always and everywhere the
Vāyu called Kandarpa, who is of a deeper red than the Bandhujı̄va flower, and is
the Lord of Beings and resplendent like ten million suns.
Inside it (the triangle) is Svayam
. bhu in His Liṅga-form, beautiful like molten
gold, with His head downwards. He is revealed by Knowledge [jñāna] and
Meditation [dhyāna], and is of the shape and colour of a new leaf. As the cool
rays of lightning and of the full moon charm, so does His beauty. The Deva who
resides happily here as in Kāśı̄ is in forms like a whirlpool.
Over it shines the sleeping Kun.d.alinı̄, fine as the fibre of the lotus-stalk. She is
the world-bewilderer, gently covering the mouth of Brahma-dvāra by Her own.
Like the spiral of the conch-shell, Her shining snake-like form goes three and a
half times round Śiva, and Her lustre is as that of a strong flash of young strong
lightning. Her sweet murmur is like the indistinct hum of swarms of love-mad
bees. She produces melodious poetry and Bandha and all other compositions in
prose or verse in sequence or otherwise in Sam
. skr.ta, Prākr.ta and other languages.
It is She who maintains all the beings of the world by means of inspiration and
expiration, and shines in the cavity of the root (Mūla) Lotus like a chain of brilliant
lights. (S.atcakranirūpana verses 8–11, trans. Woodroffe 1974: 340, 343, 346–7)
The Indian account represents the lowest of the series of cakras through
thoroughly Indian imagery, including a homology with the Tantric pı̄t.ha of
Kāmarupa and a Śiva-liṅga that is compared to that at Kāśı̄. One might also
refer to the way in which mercury, the prime alchemical substance, is treated
in Indian alchemy in terms of a complex series of symbolic connections
that identify it with the Moon, semen and with Vedic soma, also seen as a
source of immortality.
It is not surprising that the question of orthogenetic explanation comes
up again in this context. David Gordon White argues both sides in his
Alchemical Body. He notes the likelihood of contacts via the sea route from
South India to China (which was in fact the chief supply route for mercury,
an import from China):
Since India’s original fascination with alchemy most probably arose out of early
contacts with a China [. . .] whose Taoist speculative alchemical tradition had
been developing since the second century A.D.,15 one might conclude that such
traditions reached south India via a maritime route. (White 1996: 53)
Elsewhere, he points to the similarities between the South Indian Cittar
tradition and the physiological alchemy (neidan) of Daoist traditions
(1996: 55). However, he wants to deny that we are dealing with the import
of the whole complex of ideas:
15
The Mawangdui texts would now enable us to push this back to the second century BCE.
Subtle bodies, longevity and internal alchemy
281
While Chinese (Taoist alchemy) and Persian (the Shi’a Jabirian school) traditions
no doubt interacted with Tantric alchemy, the Indian material is so specifically
Indian–as much in the subcontinental provenance of its materiae primae as in its
nearly exclusive Hindu religious and metaphysical presuppositions – as to preclude
any possibility of this being a matter of wholesale borrowing. (White 1996: 54–5)
What counts as ‘wholesale borrowing’ is perhaps arbitrary; it is partly a
matter of how one defines ‘the Indian material’. It is certainly possible to
make a good case for individual components of Tantric ideas, including
aspects of the subtle-body practices and even sexual rituals, going back to
the ‘Vedic’ culture of Kuru-Pañcāla, or at least to the Upanis.ads. However,
in the eighth-century material these elements have come to be combined
in a pattern of behaviours, a cultural complex, that is recognisably not
Vedic or even Upanis.adic or Buddhist. The elaborate internal structure of
the subtle body, with the series of cakras along the central channel, has no
real precedent in Indian material, for all that the language in which it is
described is entirely Indic.
It seems also that links between South India, probably the earliest Indian
location for the complex of sexual practices, inner and outer alchemy, and
China, where the same complex had taken form at a considerably earlier
date, are plausible. The Cittar tradition has stories of teachers with Chinese
connections, but our sources for these appear to be relatively late (White
1996: 61–3). They mostly centre around the figure of Bogar, described
either as a Indian alchemist who studied in China or as a Chinese adept
who came to India. He is connected with the famous Murugan shrine at
Palani in South India, where he is credited with making the main shrineimage of Murugan from a compound of nine poisonous metals. The image
is bathed with substances that are then regarded as medically potent and
are sold off to visitors to the pilgrimage site. Bogar himself is supposed to
have gained the siddhi of long life and still dwells somewhere inside the
hill of Palani (White 1996: 61; Samuel 2001a). A realistic picture, as White
suggests, might be of an ongoing exchange along the India-China route,
though it seems that most of the innovations came from the Chinese end
(White 1996: 63).16
In this connection, it is worth pointing out that the material we are
looking at has a degree of conceptual unity in its early Chinese context that
is less obvious in its later Indian versions. In the Mawangdui texts both
medical and sexual prescriptions are essentially to do with maintaining the
16
Other material suggestive of Chinese connections includes the story of Vasis.t.ha and the Cı̄nacāra
practices (e.g. Bharati 1965: 65–79) and the Chinese associations of the rDzogs-chen teacher Śrı̄
Siṅgha (Guenther 1974).
282
Origins of yoga and Tantra
optimal health of the human organism. It has been argued persuasively
that early concepts of qi also grew out of medical practice, and that the
whole idea of learning sensitivity to qi derives from the doctor learning to
be sensitive to the movement of sensation in a patient’s body (Hsu 2005).
In later forms of Chinese medicine this is best known in the context of the
sophisticated techniques of pulse-reading. These are now a standard part of
Āyurvedic practice, but they almost certainly spread from China to India,
since they do not occur in the early Āyurvedic texts. They came to form
part of the Siddha medical tradition in South India (Daniel 1984), and are
also a principal means of diagnosis in Tibetan medicine (Samuel 2001c).
In the next section, however, I turn to look briefly at precedents for these
practices in Indic traditions.
ve di c an d in dic components in the tantric synthesis
Firstly, we should note that ‘sexual practices’ in a generic sense were present
in the Vedic and Upanis.adic material. I have already discussed some of these
in relation to the ‘transgressive’ aspects of the horse-sacrifice (aśvamedha)
and the Vrātyas, and considered the close relationship between sexual
energy and ascetic practice in the Vedic material (see Chapter 7). As is well
known, late Vedic texts treat sexual intercourse as symbolically equivalent
to the Vedic sacrifice, and ejaculation of semen as the offering. This theme
occurs in the Jaiminı̄ya Brāhman.a (Hartzell 1997: 86–7) and the Chandogya
Upanis.ad (5.8.1), as well as in the Br.hadāran.yaka Upanis.ad, which has a
particularly elaborate version:
Her vulva is the sacrificial ground; her pubic hair is the sacred grass; her labia
majora are the Soma-press; and her labia minora are the fire blazing at the centre.
A man who engages in sexual intercourse with this knowledge obtains as great a
world as a man who performs a Soma sacrifice, and he appropriates to himself
the merits of the women with whom he has sex. The women, on the other hand,
appropriate to themselves the merits of a man who engages in sexual intercourse
with them without this knowledge. (Br.hadāran.yaka Upanis.ad 6.4.3, trans. Olivelle
1998: 88)
The Br.hadāran.yaka Upanis.ad goes on to specify a series of sexual rituals
and practices, mostly directed to obtaining a child of a particular kind,
but also including rituals to be carried out before and after the birth of a
child (6.4.9–28 = Olivelle 1998: 88–93). The concern in this text with male
loss of virility and power is striking, and as is well known this has been a
Subtle bodies, longevity and internal alchemy
283
continuing theme in Indian thought up to modern times (e.g. Carstairs
1957; Alter 1997).
What we are dealing with in the new sexual rituals of the seventh and
eighth centuries is different, although there is a certain relationship, in that
avoidance of ejaculation fits into the logic of the need for seminal continence
to preserve male vitality and virility (see Chapter 8; the same theme is found
in Chinese material). Rather than sexual intercourse being homologised
with the sacrifice, the process of sexual excitation is now homologised with
the movement of internal substances or energies within the body.17 As is
well known, the most familiar Śaiva imagery here is of the arousal of the
serpentine goddess Kun.d.alinı̄, conceived of as dwelling in the lowest of the
cakras within the human body, and her ascent through the higher cakras
(e.g. Silburn 1988). Silburn’s account is particularly useful because it is
based on the earlier Trika texts, particularly the work of Abhinavagupta
and his commentator Jayaratha, and so predates the development of the
elaborate imagery of deities, multi-petalled lotuses and so on familiar from
later material:
During the rising of Kun.d.alinı̄, since the yogin experiences a vigorous whirling at
the level of the centres located along the central axis, the latter are called ‘whirling
wheels’. From there the divine energies spread out and become active in the body.
Each wheel has a definite number of spokes; fifty in all have been listed for the
whole body. [. . .] In ordinary persons these wheels neither revolve nor vibrate,
they form inextricable tangles of coils, called accordingly ‘knots’ (granthi), because
they ‘knot’ spirit and matter, thus strengthening the sense of ego. [. . .] Together
they constitute the unconscious complexes (sam
. skāra) woven by illusion, and the
weight and rigidity of the past offers a strong opposition to the passage of the
spiritual force. Each knot, being an obstruction, must be loosened so that the
energy released by the centres can be absorbed by the Kun.d.alinı̄ and thus regain
its universality. These wheels are by no means physiological and static centres of
the gross body, but centres of power belonging to the subtle body, centres that the
yogin alone, during the unfolding of Kun.d.alinı̄, can locate with as much accuracy
as if they belonged to the body. (Silburn 1988: 25–6)
Abhinavagupta’s scheme employs five cakras rather than the seven more
familiar from more modern accounts (Silburn 1988: 27–30):18
17
18
It is worth noting that there are also statements in the Brahman.as where orgasm is identified with
divine bliss (ānanda), see Olivelle 1997. I would see these more as indicative of a theme that was
taken up in the later sexual practices rather than as a direct precursor of those practices.
Another early source, the Kaulajñānanirn.aya, discusses a number of different schemes, including one with eleven cakras (ch.5, vv.25–7) and one with eight (ch.10; cf. Bagchi and Magee
1986).
Origins of yoga and Tantra
284
19
bhrūmadhya
kan..tha
hr.dayacakra
nābhicakra
mūlādhāra or mūlabhūmi
Between eyebrows
Base of neck or back of throat
Heart
Navel
Base of spine
While there is no direct precedent for such a scheme in earlier Brahmanical
sources, there are certainly suggestions of an internal ‘subtle physiology’.
These include the five-body structure of the Taittiriya Upanis.ad II.1–5.20
These bodies were later developed by Vedantic writers into five ‘sheaths’
or kośa obscuring the inner self (e.g. in Śan.kara’s Vivekacūd.āman.i). The
Taittiriya Upanis.ad is generally regarded as one of the earliest Upanis.ads;
Olivelle dates it to the fourth or fifth century BCE, though the dating is
far from certain (Olivelle 1996: xxxvii).21
Apart from the discussion of the five bodies, the Taittiriya Upanis.ad also
includes a passage (I.6) suggestive of ideas regarding an internal anatomy of
the subtle body, with a central channel through the body and the possibility
of movement in different directions from that central channel. Similar ideas
can be found in other probably early texts in the Upanis.adic corpus, as well
as in Greek thought at what was perhaps around the same time (McEvilley
2002).22 The idea of five breaths or forms of prān.a within the body is also
found in the later Vedic material, and homologised in various ways with
the sacrificial ritual (e.g. Hartzell 1987: 100–7). Connolly comments on the
concept of prān.a in the Śatapatha Brāhman.a that
It exists in two principal modes, a unitary one, when it is the foundation of all
existence and the inner controller of the individual, and diversified one, when it
is the various cosmic forces and the breaths and faculties which exist in the body.
[. . .] The unified prān.a enters the body by way of the head and then spreads
19
20
21
22
The bhrūmadhya centre is also often known as bindu (point) ‘because, when this centre is pierced,
the pent-up energy that has accumulated there is realeased, and a dot of dazzling light appears, “a
subtle fire flashing forth as a flame”. This is the “bindu”, a dimensionless point – free therefore
from duality – in which a maximum of power is concentrated’ (Silburn 1988: 29.) Bindu (thig-le in
Tibetan) is a key term in the Buddhist versions of these practices as well, though it is there given a
somewhat different meaning.
These start from the physical body formed by food (anna-maya) and proceed through a series of
other, subtler bodies or selves, each one shaped like a human being and pervading the one before:
the body or self made of vital breath (prān.a-maya); the body or self made of mind (mano-maya); the
body or self made of consciousness or intellect (vijñāna-maya), and finally the body or self made of
bliss (ananda-maya).
I have moved Olivelle’s dates in the main text forward a century in accordance with his note at 1996:
xxxvi n.21, since I have assumed the ‘later’ chronology for the Buddha in this book.
McEvilley cites the Chāndogya and Maitri Upanis.ads, Plato’s Timaeus, and the early Greek medical
tradition (2002: 93–6). I should add that I do not endorse some of the later, more speculative sections
of McEvilley’s article.
Subtle bodies, longevity and internal alchemy
285
throughout, infusing every limb. In doing so it nourishes and vitalizes the body.
Those parts not reached by the prān.a dry up and wither away. The distribution of
prān.a appears to be effected by means of definite pathways, though the text is not
clear on this. (Connolly 1997: 24)
However, the idea of breath/ prān.a as a substance to be moved consciously
around the body does not appear to be present in any of the early sources
discussed by Connolly. All of this material is suggestive in various ways
of later developments, and it can hardly be denied that the language of
the internal yogas as it was later formulated was influenced by these ideas,
but there is nothing here which implies yogic practices in which prān.a or
other similar substances are consciously moved along internal channels. It
is certainly not predictive of them.
We are unlikely ever to know precisely how the ‘internal yogas’ of Tantric
Śaivism and Buddhism developed. All we can really say at this stage is (1)
we have evidence for similar practices in China many centuries before our
earliest Indian evidence, and we know that there had been several centuries
of active interchange along the China-India trade routes by the time that
these practices appeared in India, and (2) from their earliest appearance in
India, these practices were conceptualised (or reconceptualised) within a
specifically Indic vocabulary.23
buddhist and j ain d evelopments
The earliest explicit versions of these practices in North Indian Śaiva material (in the Kaulajñānanirn.aya and Kubjikātantra) probably date from the
ninth and tenth centuries. If the Tirumantiram, ascribed to the Tamil saint
Tirumūlar, can really be dated to the seventh century CE, it would seem to
be the earliest Indic account of these practices (see Zvelebil 1973: 73–80, esp.
78–9; 1996: 121), but the dating of this work is by no means certain.24 On
the Buddhist side, the Hevajra Tantra, perhaps dating from the late ninth
century in its present form, represents a stage similar to that discussed above
by Silburn on the Śaiva side. Book One, Chapter 1 of the Hevajra describes
thirty-two named nād.ı̄s (channels) of which the three principal ones are
Lalanā, Rasanā and Avadhūti, and four cakras (the lowest cakra of the Śaiva
scheme is omitted):
23
24
It is worth mentioning that we have clear evidence for such radical reconceptualisation of Tantric
yoga in at least one more recent context: in the songs of the nineteenth century Bengali saint Lalon
Fakir and his fellow ‘Bauls’ we find the internal practices rewritten in Islamic terms (Salomon 1991).
Venkatraman argues for a tenth to eleventh century date on the grounds that the text refers to the
Nine Nāths, the teachings of Matsyendranātha and Goraknāth, and the Kālacakra (1990: 42–8; see
also N. Subrahmanian’s comments in his foreword, 1990: xi–xiii).
286
Origins of yoga and Tantra
mahāsukhacakra (‘Centre of Great Bliss’)
sam
. bhogacakra (‘Centre of Enjoyment’)
dharmacakra (‘Centre of Essential Nature’)
nirmān.acakra (‘Centre of Creation’)
Head25
Throat
Heart
[Navel]
As the names suggest, these cakras are named after the four bodies of the
Buddha (an expansion of the more familiar set of three bodies, nirmān.akāya,
sam
. bhogakāya, dharmakāya). They are then correlated with various further
components of the Buddhist teachings (Farrow and Menon 1992: 16–21).
The last verse of the chapter presents the equivalent to the arousal of
Kun.d.alinı̄ in a highly condensed form:
Can.d.alı̄ blazes up in the navel. She burns the Five Buddhas. She burns Locanā and
the others. Aham
. [‘I’] is burnt and the Moon flows down. (Farrow and Menon
1992: 21)
Can.d.alı̄, another fierce goddess, here occupies the place of Kun.d.alinı̄ in the
Śaiva system.26 The ninth century Yogaratnamālā commentary interprets
the name in terms of Can.d.a = ‘fierce’, but Can.d.alı̄ also means a woman of
the Can.d.ala low-caste group, suggesting a connection with ‘polluted’ social
groups. Can.d.alı̄ is also one of the eight goddesses of the Hevajra man.d.ala,
discussed in Chapter 10; others bear names relating to other polluting
occupational groups (D
. ombı̄) or tribal populations (Śavarı̄).
As I mentioned in Chapter 10, we can to some degree at least trace the
adoption and development of these practices within the Buddhist tradition
from the eighth century onwards. There appears to be a gradual evolution
from the tantras classed by the later tradition as kriyā and caryā, which are
primarily about ‘external’ ritual practices, to the yogatantra (c. 700–c. 750
CE) where the emphasis becomes much more on the achievement of
inner purity, and we first find the identification of the practitioner with a
Tantric deity, a form of the Buddha. From then on, as Dalton has noted,
there is a progressive interiorisation of the ritual, leading eventually to the
major Tantric cycles of the so-called anuttarayoga Tantra, the Hevajra,
Cakrasam
. vara, etc., in the late ninth century.
The Tantric interiorization of Buddhist ritual was not a rejection of ritual. Nor
was it a psychologization [. . .]. This shift took place in the physical realm. Its
beginnings can be traced to the first half of the eighth century, and the ritual
technologies it spawned continued to develop through the ninth century. By the
25
26
The locations of the four cateras are not stated directly in the root text of the Hevajra Tantra. The
upper three are specified in the ninth century Yogaratnamālā commentary (Farrow and Menon 1992:
14–16). Verse 32 makes it clear that the lowest centre is at the navel.
The Tibetan for Can.d.alı̄ is gtum-mo, the term for the well-known ‘psychic heat’ or ‘mystic heat’
practices which derive from the yogic processes discussed here (e.g. Guenther 1963: 53–60; Huber
1999: 86–90).
Subtle bodies, longevity and internal alchemy
287
end of these two crucial centuries, a new ritual discourse of the bodily interior was
in place. The Tantric subject had become the site for the entire ritual performance;
the body’s interior provided the devotee, the altar, the oblations, and the buddha
to be worshipped. (Dalton 2004: 2)
Dalton’s recent study focuses on the period between the yoga Tantra and
the anuttarayoga Tantra. The key category here, mahāyoga Tantra, has been
elided from the dominant ‘New Tantra’ classification in Tibet, but was
important in India and remains part of the ‘Old Tantra’ (rNying-ma-pa)
classification. Dalton notes that it was the mahāyoga Tantras, such as the
Guhyasamāja and Guhyagarbha, which introduced ‘the ritualized sexual
practices for which tantra has since become so notorious. They focused on
the body’s interior, on the anatomical details of the male and female sexual
organs and the pleasure generated through sexual union’ (2004: 3).
The development of the initiation ceremony, and specifically the introduction of the second initiation, the guhyābhis.eka where the initiate consumed a drop of the combined sexual fluids of the Tantric master and his
consort, was a particularly significant part of this development. As noted
in Chapter 10, this was a key component of Śaiva tradition as well, where
it was seen as ingesting the essence of the clan or lineage of practice.
While the sexual practices hinted at by Asaṅga in the Mahāyānasūtrālam
. kāra were not necessarily transgressive, there is little doubt that
this initiation would have been a transgressive act for those taking part.
What is also striking is that it was seen as an occasion for the direct experience of the central liberating insight of the tradition, and also as a model
for later sādhana or ritual practice, in which the disciple would himself
re-enact the rite with a female consort.
This is most explicit in ritual manuals from Dunhuang in Central Asia,
which appear to present a picture of the mahāyoga Tantras as practised in the
ninth and tenth centuries, than in the later Indian and Tibetan commentarial material, particularly that of the ‘Ārya tradition’, where the sexual aspects
are often significantly downplayed (Dalton 2004: 9; cf. Wedemeyer 2002).
In these Dunhuang meditation texts we find the identification of bodhicitta,
the classic altruistic motivation to achieve Buddhahood in order to relieve
all sentient beings of their suffering, with the seminal drop held at the
tip of the penis in the Perfection Stage, the second of the two stages of
the practice.27 Dalton suggests that the symbolism here is of procreating
27
In later Tibetan versions of the Perfection Stage, male and female sexual substances, both seen as
internal to the body of either male or female practitioner, are directed into the central channel of
the body (Huntingdon and Bangdel 2003: 240–51). As Beyer comments, however, the ‘experiential
dimensions of Great Bliss and Clear Light’ were always balanced with the ‘intellectual categories of
Emptiness’; ‘bliss or light without Emptiness was simple sensual indulgence’ (Beyer 1973: 133).
288
Origins of yoga and Tantra
the Buddha; in the first (generation) stage, the practitioner ‘has imagined
himself as a son of the Buddha, generated out of the buddha’s seed syllable,
[w]hereas in the second stage the Buddha is generated out of the practitioner’s own seed, a syllable that arises at the tip of his penis inside the
vagina of the female partner’ (Dalton 2004: 9):
The practitioner is then instructed to worship the man.d.ala, using the blissful
sensations flowing through his body. Such instructions seem to reflect an early
prototype of the subtle body systems that were articulated in more complex forms in
later works. The later systems involved intricate arrangements of cakras and energy
channels mapped across the body’s interior. In the early Mahāyoga texts, however,
the technologies are simpler, the descriptions limited to the energies associated
with sexual pleasure which rushes through the practitioner’s torso. (Dalton 2004:
10–11)
Dalton suggests that these more complex systems only appear in the Perfection Stage yogas in the late ninth and tenth centuries. At this earlier
stage, the sexual energy is more ‘a raw physical force that is used to worship
and energise the buddhas’ (2004: 12). The ritual ends with the consumption of a sacrament (samaya; the term also means ‘vow’ or ‘commtment’),
presumably the combined sexual fluids of the two participants, by both
partners, taken from the female partner’s vagina, thus re-enacting the initiation rite in the form of a self-consecration (2004: 15–16). Dalton argues
that this practice was the ‘defining characteristic’ of mahāyoga practice at
this period, although it later dropped out of the tradition.28
In fact, while mahāyoga practice (at least as interpreted by Dalton)
includes the moderately transgressive and polluting element of the consumption of sexual fluids from the female partner’s vagina,29 it seems in
other respects rather close to the kind of early sexual yoga implied by Asaṅga,
with the seminal restraint being a key element during the ‘Perfection Stage’
of the practice.30 The imagery becomes much more transgressive with the
28
29
30
Dalton also suggests that the initial context of the term ‘Great Perfection’ (Tibetan rdzogs chen), also
described in the Guhyagarbha Tantra as a samaya or sacrament, was the state of liberating insight
associated with this self-consecration (2004: 8, 17–20 and n.56). Note here Isaacson’s discussion of
the various interpretations of the four consecrations associated with the initiation proper (1998: 32).
Doubtless somewhat more transgressive if the male practitioner was high caste and the woman was
from a low caste background. The Guhyasamāja Tantra, as noted in Chapter 10, presents itself as a
shocking and radical new teaching, but the radically antinomian language for which it is known is
not for the most part intended to be taken literally.
If the sacrament is made up of the combined sexual fluids of the two partners, the male obviously
does ejaculate before this stage; Dalton cites another Dunhuang text in which the fall of the bodhicitta
(i.e. semen) is seen as an offering to the goddess (PT841, see Dalton 2004: 16). See also Guhyasamāja
Tantra 7.33–4: ‘What is that meditation on recollection of the pledge [samaya]? In accordance with the
pledge, he who desires the fruit should pour out his seed and drink it according to the rite; he should
Subtle bodies, longevity and internal alchemy
289
late ninth and early tenth century anuttarayoga or yoginı̄ tantras such as the
Hevajra and Cakrasam
. vara, with their cremation-ground symbolism, the
central role of fierce goddesses, and similar kāpālika or Śākta material.
As far as the ‘internal’ yogic practices are concerned, there is a significant
step from the kind of practice Dalton is suggesting, where sexual practice
is used to energise the ritual and perhaps create the state of bliss and loss of
personal identity which is homologised with liberating insight, to the cakra
and nād.ı̄ practices. One could read the development of the new system in
either of two ways, as an indigenous inspiration suggested by continued
yogic practice in mahāyoga style, or as a set of concepts and processes
incorporated from another source into the more basic sexual yoga discussed
here. To me, the second alternative seems more likely, especially given
the presence of such ideas in the cognate Chinese practices. If Buddhist
practitioners were already employing sexual yoga of the kind suggested
here, the cakra system would provide both a structure to aid practitioners
in directing their sensations, and a grid through which associations could
be built up with other elements of the symbolic world within which the
practice was taking place. There are no obvious grounds here for deciding
whether the cakra-based sexual practices first arose in a Śaiva or Buddhist
context, but it is evident that they were being actively developed in both
by the tenth century.31
Some Jainas also appear to have adopted the internal yogic practices
involving the circulation of prān.a, but as far as I know there is no evidence
for Jaina use of sexual yoga. Since the principal evidence for Jaina meditation
practices derives from the writings of Hemacandra in the twelfth century I
shall leave discussion of them to Chapter 13, where I shall be dealing with
that period.
conclusion
As we have seen, these new Tantric traditions are a confluence of a variety
of different factors and components: man.d.ala and mantra practices going
back to the late Mahāyana sūtras and early Tantras, internal sexual yogic
practices of the kind discussed in this chapter, with the incorporation of
fierce male and female deities and cremation-ground symbolism apparently
31
kill the host of Tathagatas and attain supreme perfection. What is that meditation on recollection
of the pledge of transcendent wisdom? All are naturally luminous, unarisen, uninfluenced; there is
no enlightenment, no realisation, no end and no origin’ (Fremantle 1990: 109, 113).
For a summary of the developed Tibetan approach to Guhyasamāja practice, see Thurman 1997:
135–43.
290
Origins of yoga and Tantra
from kāpālika and other Śaiva contexts, all within a philosophical and conceptual frame which goes back to the early Mahāyāna and in some respects
to the common Indian background of the early renouncer traditions. It is
natural to wonder why Buddhist practitioners might choose to put together
this particular complex of ideas and practices. But before we look further
at this question, we should turn to look at the people who employed these
practitioners, above all the rulers and states of the time.
chap t e r 12
Tantra and the state
As we saw in Chapters 10 and 11, there were striking innovations in
both Śaiva and Buddhist practices from the seventh and eighth centuries
onwards.1 While these innovations have been frequently lumped together
under the label of ‘Tantra’, they include a variety of different components
on both Śaiva and Buddhist sides.
To summarise, on the Śaiva side, a ‘transgressive’ tradition of rituals
involving cremation grounds, polluting substances associated with sex and
death, fierce gods and particularly goddesses appears to have been originally
carried out as ritual sorcery by hereditary caste groups (kula). The initiation
rituals for these traditions involved the consumption of the ‘clan essence’,
the mixed sexual secretions of male guru and female consort.
These practices were adopted by renunciate practitioners in the so-called
kāpālika style and gained increasing importance in the seventh to ninth centuries (the kaula lineages associated with the early Nāth siddhas). The more
extreme elements were mostly dropped and the ‘external’ practices were
progressively substituted by ‘internal’ yogic practices which have a marked
sexual component and involve a subtle body physiology of cakras and internal flows and channels. Within these practices, sexual ritual became a mode
of access to a state of the body-mind in which the liberating insight central
to the tradition could be directly perceived.
On the Buddhist side, early ‘Tantric’ practices (sixth and seventh centuries) were mostly a further development of established deity-visualisation
practices in the later Mahāyāna sūtras, with fierce, mostly male deities initially introduced as secondary figures in the man.d.ala. In the eighth to
tenth centuries, fierce Śaiva-style male and female deities, and couples in
sexual union, took over as primary figures at the centre of the man.d.ala,
kula-style initiation rituals were adopted, and there was a progressive shift
1
The Vais.n.ava Tantric tradition of the Pāñcarātra seems by and large to have stayed clear of these
developments, and Jaina involvement also appears to have been relatively limited.
291
292
Origins of yoga and Tantra
to ‘internal’ yogic practices closely parallel to those which were developing
at around the same time or a little later in the Śaiva tradition. Again, sexual ritual became a mode of access to a state of the body-mind in which
the liberating insight central to the tradition could be directly perceived.
Circles of fierce Śaiva-style deities were incorporated into the structure of
Buddhist man.d.alas. These new practices formed the basis of the Mahāyoga
and Yoginı̄ (‘Anuttarayoga’) Tantras.
It is not clear to what extent practices involving sexual intercourse were
undertaken literally by supposedly celibate practitioners,2 though there are
certainly stories of formerly celibate Buddhist practitioners abandoning
their vows to take on these new ‘Siddha’ practices. In the later Tibetan
Buddhist context, where these Tantric lineages were performed by celibate
practitioners, they were mostly undertaken in symbolic form. This is also
true of much later Śaiva practice. In the eighth to tenth century Indian
context, however, it would seem that they were mostly performed literally, by both Śaiva and Buddhist non-celibate practitioners, who might be
renunciates or (particularly in the later stages) ordinary lay people.
I will consider some of the later history of these practices in Chapter 13.
In the present chapter, I look at the initial development of these practices
and attempt to relate it to the wider social and political context in which
they took place. I begin with a brief introduction of political developments.
i ndic states in the eighth to tenth centuries
Gupta and Vākāt.aka rule came to an end in the early sixth century; the
precipitating factor in the Gupta case may have been the invasions of the
Huns (more properly, the Hephthalite Hūn.a peoples). Ronald Davidson
has attempted to catalogue the numerous, often short-lived, dynasties and
the political interplay of the succeeding centuries; the results understandably verge on the unreadable, but give some indication of the complex and
chaotic politics of the period (Davidson 2002a: 30–62). States between the
early sixth and mid eighth centuries were for the most part relatively shortlived and unable to maintain control of substantial areas for sustained
2
Sanderson notes that ‘[t]he tradition of Abhayākaragupta and Darpan.ācārya remained true to the
early tradition, insisting that any Buddhist, layperson or monk, may take the Tantric vows and receive
all the consecrations, including the problematic consecrations involving sexual intercourse, provided
he has achieved insight into the doctrine of emptiness’ (1994: 97). This also appears to have been
the position of the Indian guru Atı̄sa, who came to Tibet in the mid-eleventh century (Samuel 1993:
470–1). In the longer run, however, as Sanderson again notes, the rituals were generally modified to
avoid transgressing vows of celibacy, as in Tibet, or practised only by non-celibates, as among the
Newars (Sanderson 1994: 97).
Tantra and the state
293
periods of time. In the early seventh century, much of North India was
temporarily unified under King Hars.a (c. 606–47 CE); we are particularly
well informed about this period through the accounts both of Hars.a’s court
poet Bān.a and the Chinese Buddhist visitor Xuanzang, who visited India
between approximately 629 and 645 CE. However, Hars.a’s empire fell apart
at his death, and another century of rapidly changing short-lived regimes
followed.
The mid-eighth century marked the emergence of a more stable situation, with three major empires, the Gurjara-Pratı̄hāra dynasty, based at
Ujjain and extending over much of North-West India, the Pāla state, based
in Bengal and Bihar, and the Rās.t.rakūt.a kings, based in the southern Deccan. North India was contested territory between the Gurjara-Pratı̄hāra
and Pāla states. In South India, the Cola dynasty also established a stable
and long-lasting state, centred at Tanjore in modern Tamilnadu (Fig. 12.1).
What is more significant than the details of dynastic conflicts and shift
in territorial control from the sixth to eighth centuries are the new social
and cultural patterns that emerged during this period. Davidson speaks
of a culture of ‘military adventurism’, leading to ‘decentralisation and the
coalescence of fiefdoms’; military defence became a prime imperative, with
a rapid increase of castles and other other fortifications (Davidson 2002: 67).
The key political concept of the new order was the feudal lord (sāmanta),
with each major state ideally surrounded by a circle of lesser states whose
rulers paid homage to the king at the centre and replicated his style at
the local level. This political structure was referred to as a man.d.ala, and
formed a conscious model for the Tantric Buddhist man.d.ala, with its central
divinity surrounded by lesser deities in the four or eight directions. As
Davidson notes, the ‘samantisation’ of Indian society was mirrored by the
‘sāmantisation’ of the deities, who increasingly came to be viewed on the
model of feudal lords and ladies (Davidson 2002a: 71–2).
The ‘wisdom king’ ethic of the Rāmāyan.a had been in part a literary
fantasy, but it had also been an ideal to which rulers in the first few centuries
CE paid at least lip-service. Davidson points to a new trend emerging from
the late fifth century onwards: the ‘divine erotization of kingship and,
ultimately, of warfare’ (2002a: 68). War was the domain of Śiva and of the
fierce goddesses, and it is not surprising that rulers turned increasingly to
these cults in both non-Tantric and Tantric forms.
The Pāla dynasty was pro-Buddhist, and was responsible for the creation
of a major series of new monastic centres in the late eighth and early
ninth centuries. The Gurjara-Pratı̄hāra, Rās.t.rakūt.a and Cola were proBrahmanical, as were most of the smaller powers of the period, and while
Origins of yoga and Tantra
294
Limit of areas allegedly under
nominal Pāla suzerainty c.790–800
P
Maximum extent of
areas under Pratihāra
rule (including feudatories)
Ā
L
Kānyakubja
A
R A - P R A T Ī H Ā R A S
RJA
GU
S
Mudgagiri
Ujjayinı̄
Acalapura
Maximum extent of
areas at some time
under Pāla rule
(including feudatories)
RĀS.T. RAKŪT. AS
Mānyakhet.a
Maximum extent
of areas at some
time under firm
control of Rās.t.rakūt.as
(including feudatories)
C O
L
¯ A S
Limit of areas at some time
tributary to Rās.t.rakūt.as
Outer limit of Col-a
domains, c.949
Tān̄jāvūr
Core areas
Pālas
Gurjara Pratihāras
Rās.t.rakūt.as
Colas
¯
Figure 12.1. Map of South Asia, eighth to tenth centuries
Tantra and the state
295
a general pattern of religious pluralism continued, state support for the nonBrahmanical traditions was decreasing. Jainas were occasionally successful at
gaining royal protection and support, as in Gujarat during the time of King
Kumaragupta and the Jaina teacher Hemacandra (mid-twelfth century), but
they too were increasingly marginalised during this period. The weakening
of long-distance trade consequent on the Muslim expansion into Iran,
Afghanistan, and the northwestern parts of South Asia, and the general
disturbance of urban life due to the constant warfare of these centuries, may
also have affected the economic base on which the Buddhists and Jainas,
with their close links to merchant and artisan communities, depended.
Buddhism was by now becoming established in parts of Southeast Asia,
but the sāmanta model and the accompanying religious cults were also
increasingly exported from South Asia. By the eighth and ninth centuries,
the major states in the region were Pagan in present-day Burma, the Khmer
state (Kambuja), Champa in present-day Vietnam and Śrı̄vijaya in peninsular Malaysia, Sumatra and Western Java (Fig. 12.2). These regions were
increasingly part of the wider Indic cultural orbit, and we find, for example,
Buddhist scholars such as Atiśa travelling to Southeast Asia to study with
local Vajrayāna Buddhist teachers (Chattopadhyaya 1967).
Northwest India and Central Asia were still important parts of the Indic
cultural sphere up to the Muslim conquests, and there remained a scattering of significant Śākta and Tantric centres in the Northwest, including
the important pilgrimage site of Hiṅglāj (White 1996: 66, 121–2) and the
major Tantric centre of Od.d.iyāna, which is now generally accepted as being
located, as the Tibetan pilgrims believed, in the Swat Valley (Davidson
2002b: 160). By the ninth century, however, these regions were increasingly
cut off from direct contact with the surviving Buddhist communities in
India, and they were significant more as legendary locations for the origins
of Tantric teachings than as real places (Davidson 2002b: 160–3). The culmination of these imaginings was the mythic realm of Śambhala, origin of
the Kālacakra Tantra which arrived in northern India in the early eleventh
century (Davidson 2002b: 166–8; Bernbaum 1980; Newman 1998). Those
parts of Central Asia that remained Buddhist, such as the oasis state of
Dunhuang, the Tangut (Xixia) kingdom at Karakhoto, or the Mongols,
turned to the Tibetans and Chinese for continuing contact and renewal.3
3
As we saw in Chapter 11, Dunhuang includes a large body of Buddhist texts in Tibetan representing the
latest developments in Tantric Buddhism in Tibet at the time. Shen Weirong has recently shown how
the Buddhist scholars of Karakhoto in the eleventh and twelfth centuries were producing translations
of works by sGam-po-pa and other Tibetan scholars within a few years of their composition (Shen
2006).
Origins of yoga and Tantra
296
Probable outer limit
of areas at some time
under control of Pagan
Pagan
PA G A N
Probable outer limit of areas at
some time under control of Champa
CH
AM
KAMBUJADEŚA
PA
Yaśodharapura and
Angkor Thom
Indrapura
(Dong-duong)
Vijaya
Probable outer limit of areas at
some time under control
of Kambujadeśa
Probable outer limit
of area at some time
under control of
Śrı̄vijaya
S´
R
I¯
V
I
J
A
A
Core areas
Y
Śrı̄vijaya
Kambujadeśa
KADIRI
Kadiri
Śrı̄vijaya
Pagan
Daha
Possible outer limit of area at some
time under the control of Kadiri
Figure 12.2. Map of Southeast Asia, eighth to tenth centuries
politics and tantra in south asia
To return to South Asia proper, the process of ‘sāmantisation’ and military
adventurism described by Davidson doubtless goes a long way to explain
the increasing interest in fierce deity cults and Tantric ritual in general by
the rulers of this period. We saw what was probably an early example of this
Tantra and the state
297
in the relief carvings of the Saptamātr.kā deities at Udayagiri, patronised
by Candragupta II in 401–2, which I discussed in Chapter 10. It is unclear
whether Saptamātr.kā rituals at Udayagiri were carried out by an old-style
royal purohita from an Atharvavedic Brahmin background, as suggested
by Michael Willis, or by non-Vedic priests who specialised in these new
rituals. If the director of the ritual was a Vedic Brahmin, then he was
certainly employing techniques that go well beyond anything found in the
Atharvaveda, and which we later see in specifically kaula and Tantric contexts. Within the next few centuries, at any rate, Tantric-type deity-cults
employing fierce goddesses such as the Saptamātr.kā deities would seem
to have progressively taken over in state ritual throughout South and
Southeast Asia.
In this connection, Sanjukta Gupta and Richard Gombrich argued some
years ago that
for at least the last thousand years, perhaps longer, the concept of power in its political and social application has been intimately connected with Tantric theology –
so intimately, one might suggest, that the one cannot be adequately understood
apart from the other. (Gupta and Gombrich 1986: 123)
They noted that images of Durgā ‘start to show many royal insignia from
about 500 AD on’ (1986: 132). Durgā is the most common exoteric representation of śakti or power in Tantric terms, and she is particularly associated
with battle. Gupta and Gombrich imply that this shift in symbolism reflects
the importance Durgā increasingly had for Hindu kingdoms from this time
onwards. The king was symbolically married to Durgā when he ascended
the sim
. hāsana or ‘lion throne’, and the annual Durgā festival of navarātrı̄
was a ‘major political occasion’ and an ‘annual ceremony of reintegration’ in
Hindu kingdoms into modern times (1986: 133). Typically, kings were also
symbolically married to other more local or esoteric Tantric goddesses; the
king was regarded as an earthly representative of whichever male god was
currently regarded as most important (typically a form of Vis.n.u or Śiva)
and so as married to whichever local goddesses were treated as that deity’s
consort (Gupta and Gombrich 124–5; Vergati-Stahl 1979).
The Saptamātr.kā goddesses lost much of their importance after the
eighth century, to be replaced by a variety of other female deities, including
Durgā, Gaurı̄, Laks.mı̄, Pārvatı̄ and others. They were, however, absorbed
into sets of eight or more goddesses who formed one of the circles of
the man.d.ala. Ritual forms based on these goddesses have survived into
modern times, and have been studied in relation to the Newar cities
of Kathmandu Valley, primarily Bhaktapur, Patan or Kathmandu. Each
298
Origins of yoga and Tantra
of these Kathmandu Valley cities, formerly the capitals of small Newar
kingdoms, is surrounded by a protective circle of eight Mātr.kā temples
(Gutschow 1982, 1993; Levy 1987, 1992). These goddesses, accompanied by
Bhairava and Gan.eśa, and generally referred to now as Navadurgā (the ‘Nine
Durgas’), assume human form through masked low-caste trance-dancers in
annual ritual dances throughout the valley (Gutschow and Bāsukala 1987;
Iltis 1987, van der Hoek 2005; Korvald 2005; Sardar n.d.). We find maskeddance traditions of this kind in many parts of South Asia and there seems
every reason to assume that they go back to the early mediaeval period that
we are discussing here. I consider these practices in a little more detail later
in this chapter.
First, however, we look at further evidence from Indian textual material.
One of the most striking pieces of evidence is provided by the Netra Tantra,
a text probably composed in Kashmir in the eighth or (more likely) early
ninth century. In a recent article, Alexis Sanderson has argued that this
text constitutes, in effect, a manual for a Tantric Śaivite priest to take over
virtually the complete role of the royal purohita and rājaguru. This includes
the entire cult of non-Tantric deities such as Vis.n.u, Sūrya, Brahma, the
Buddha and Gan.eśa, along with deities from a wide variety of other Tantric
cycles, all of whom are regarded by the Neta Tantra priest as transforms of his
own central Tantric deities, Amr.teśvara (also known as Amr.teśabhairava)
and Śrı̄ (Amr.talaks.mı̄) (Sanderson 2004: 245). It is no great surprise to
find a set of eight mātr.kā goddesses in a circle around Amr.teśvara and
Amr.talaks.mı̄, though they are worshipped in this case in order to ward off
ills rather than invoke them on one’s enemies (2004: 245–6). Sanderson
argues that
[The Śaiva Mantramārga] succeeded in forging close links with the institution
of kingship and thereby with the principal source of patronage. I see four main
elements here: (1) the occupying by Śaiva officiants of the office of Royal Preceptor
(rājaguruh.) and in this position their giving Śaiva initiation (dı̄ks.ā) to the monarch
followed by a specially modified version of the Śaiva consecration ritual (abhis.ekah.)
as an empowerment to rule beyond that conferred by the conventional brahmanical royal consecration (rājyābhis.ekah.); (2) the promoting by Śaiva officiants of
the practice of displaying and legitimating a dynasty’s power by their officiating
in the founding of Śaiva temples in which the new Śivas that they enshrined
bore as the individuating first half of their names that of the royal founder or,
where complexes of royal Śiva temples were established, those of the founder and
any kin that he might designate for this purpose; (3) the provision of a repertoire of protective, therapeutic and aggressive rites for the benefit of the monarch
and his kingdom; and (4) the development of Śaiva rituals and their applications
Tantra and the state
299
to enable a specialized class of Śaiva officiants to encroach on the territory of
the Rājapurohita, the brahmanical expert in the rites of the Atharvaveda who
served as the personal priest of the king, warding off all manner of ills from him
through apotropaic rites, using sorcery to attack his enemies, fulfilling the manifold
duties of regular and occasional worship on his behalf, and performing the funerary and other postmortuary rites when he or other members of the royal family
died. (Sanderson 2004: 232–3)
Apart from taking over the office of official royal chaplain and purohita,
Tantric officiants might be independent practitioners called upon at times
of need. Sanderson cites a twelfth-century example in a Col.a inscription, where the ruler called on a Tantric specialist to destroy an invading Sri Lankan army (2004: 233–4), and also discusses the well-known
Sdok kak Thom inscription from Cambodia, which I will consider later in
this chapter (see also Sanderson 2003–4). We will see further examples in
Buddhist contexts from East Asia.
From Sanderson’s account, it seems that the Kashmiri kings adopted the
new Tantric procedures. The popularity of kaula thought in Kashmir is
also demonstrated by the extensive reference to kaula ideas in the Haravijaya, a Sanskrit court epic composed in Kashmir between 826 and 838 CE
(D. Smith 1985: 262–5).
Elsewhere in South Asia, kaula ideas were also widely adopted (e.g. Gupta
and Gombrich 1986; A. Cohen 1992, 1997). David Gordon White has noted
that
In their seventh-to-eleventh century heyday, these forms of Kaula theory and
practice were so compelling, as direct paths to gnosis, power, and godhead, that they
won the adherence of some of the great royal houses of the period: the Somavam
. śis,
Chandellas and Kalacuris, whose kingdoms stretched across the Vindhya range and
beyond, from Rajasthan to Assam. It was these royal patrons who constructed many
of the Yogini temples whose unusual architectural ruins dot this swathe of central
India, who built the ‘erotic’ temple complexes at such sites as Bhubanesvara and
Khajuraho, and who undoubtedly sought out Kula and Kaula specialists for their
expertise in both the sacred and secular spheres. (White 1998: 198)
White’s reference to the erotic temple complexes is significant, and this
theme requires further discussion. We have seen something of the internal
alchemical aspects of Tantra in Chapter 11, and of their close association
with sexual rituals. It seems likely enough that the promise of health, long
life and magical power through sexual practices was an appealing mixture
for members of the ruling dynasties of these times.
300
Origins of yoga and Tantra
I referred earlier to Davidson’s comment on the eroticisation of warfare
during this period; kingship was being made over in the image of a heroic
male figure whose sexual powers were homologised with his military power
and his very ability to fill the royal role. Thomas Donaldson notes the poetic
use of double meanings at this time, in which a king may be eulogised at
one and the same time ‘as a lover and as a great warrior’ (Donaldson 1986:
315). He cites as an example the Somavam
. śı̄ king Yayati II Chandı̄hara
Mahaśivagupta III (c. 1015–40 CE) of Orissa, of whom one of his charters
states that he,
is a winking sensualist in (curing) the love-fever (the war-fever) of Karn.āta, Lāt.a
and the lord of Gurjjara; [. . .] an expert voluptuary in removing the jingling girdles
of women (in removing the warring Kāñcı̄, i.e. the Colas); [. . .] a fierce wind in
opening and taking away the clothes of Gaud.a and Rad.hā ([. . .] a fierce wind in
the sky of Gaud.a and Rad.hā for capturing and exposing) and [. . .] a full moon
in the clean clothes of the love-afflicted body of Vaṅga ([. . .] a full moon in the
bright sky of the peaceful Vaṅga). (Donaldson 1987: 315)
In this series of double-entendres, the king’s wars against neighbouring countries are portrayed as a series of erotic encounters.
The hypereroticism of Indian temple architecture at this period is well
known. The relatively restrained mithuna figures which we find as auspicious devices on earlier Buddhist shrines and temples (Agrawala 1983,
e.g. plates 73–83; Stone 1994: plates 128–32, 168–71, 214–15) now became
occasions for increasingly exaggerated erotic display, culminating in the
well-known temples of Khajuraho (mid-tenth to mid-eleventh centuries)
and Konarak (mid-thirteenth century); Rabe’s analysis of the Khajuraho
temples as a kind of celestial transform of the royal harem, focused around
the king as their implicit male partner, seems entirely appropriate (Rabe
1996).4 The earlier auspicious function of the mithuna images seems to
have been buried under exuberant sexual display.
4
A charter of an earlier Somavam
. śı̄ ruler, Yayāti I Mahaśivagupta (c. 922–55 CE) provides an extraordinary description of the royal city, Yayātinagara, as a kind of sexual paradise, as ‘Kāmadeva’s pleasure
garden’ in Donaldson’s phrase, ‘where the enjoyment of love is being continually intensified and still
more intensified by the close embraces (of lovers), by which fatigue is removed, in which hissing sound
often appears and in which hairs stand on their ends, although such enjoyment suffers interruptions
as the ardent young couple show their skill in the various processes of conjugal enjoyment with their
eyes dilated (with excitement) and with their minds subdued and fascinated by amorous thoughts;
where, even in the midst of quarrels arising from jealousy, lovers, beaten by lotuses from the ears of
women who have cast the beauty of the celestial damsels into shade by the greatness of their endless
and peculiar charms, have all their mental anguishes roused to action by the entrance of the sharp
arrows of Cupid’, and so on for several more lines (Donaldson 1987: 316; see also Dehejia 1979: 203
n. 4).
Tantra and the state
301
The poetry and the temple architecture I have referred to are for the most
part erotic (and by implication martial) rather than Tantric, though explicit
Tantric themes are occasionally found in temple imagery (e.g. White 2003:
97–9).5 One can begin to see here though how Tantra provided both a
ritual celebration of the new ideal and a set of techniques that promised
to preserve and strengthen the sexual ability and personal power of the
initiate, in part by giving him control over a pantheon of terrifying (but
also enticing) female deities (see also White 2003: 123–59).
One can also imagine the vulnerability of rulers in such a climate to those
who offered these techniques. A king who could not fulfil the conjoined
military and sexual demands of his role, after all, would feel under considerable threat of being replaced by a younger and more virile man. Consider in
this connection the description of the Kashmiri king Hars.a (reigned 1089–
1101) in Kalhan.a’s Rājatarangin.ı̄, in which the king is being ridiculed for
his addiction to what were clearly Tantric sexual and alchemical practices
aimed at health and long life:
Others brought slave girls before him and said they were goddesses. He worshipped
them, and abandoning his exalted position and wealth was laughed at by the people.
These [slave girls], instructed by the parasites, who taught them [to give] counsels
etc. [pretended to have obtained] from conversations with the gods, confused
his mind. Some among these [slave girls] showed themselves eager for amorous
intercourse at those occasions, and the king forsook his good fortune by touching
them with his own body.
As he was anxious to live for a very long time, they [the goddesses] granted
him, when in his foolishness he asked for a long life, hundreds of years to
live.
When he desired to give magic perfection to his body (pin.d.asiddhi) some d.omba
[low-caste laundress] made him swallow a drink which he pretended was an elixir
having that power. [. . .]
What respectable man could relate the other even more shameful practices of
his which he followed to obtain strength and beauty? (cited in Rabe 1996)
It is easy enough perhaps to make fun of king Hars.a. No doubt most
sizeable Indian courts of the time had their quota of snake-oil merchants.
However, Kalhan.a’s comments, or those of the twelfth-century Kashmiri
satirist Ks.emendra (Gupta and Gombrich 1986: 132; see also Wojtilla 1984,
1990) would make little sense if the idea of kings seeking long life and bodily
5
The small group of yoginı̄ temples which have survived in Central and East India, which appear to
have been explicitly designed as locations for kaula-style Tantric ritual (Dehejia 1986; White 2003
and see Chapter 10) are by contrast more restrained; the images here are more serious and central to
the business at hand. They are figures of power, not of erotic display.
302
Origins of yoga and Tantra
perfection through sexual practices was not in fact common currency at
the time.6
Much of the imagery of Khajuraho and Konarak has rightly been valued
as among the world’s finest artistic achievements in the visual celebration
of sexual love, but there is an uncomfortable undertow to the royal cult of
Tantra that gave rise to it. For all of the exotic imagery of King Yayāti II’s
royal charter, its combination of martial violence and eroticism strikes a
note that is both profoundly disturbing and all too familiar from our own
time, in which the fantasy world of contemporary cinema constantly retails
much the same story. Doubtless the association between sex and violence
is grounded in some basic elements of human ethology, but a culture that
allows itself to indulge too far in such fantasies runs the risk of distorting
the basic relationships between men and women, and between men and
men, in damaging ways.
Certainly, if perhaps ironically, as women were increasingly displayed
as sexualised and spiritualised objects on the one hand, or invoked as
powerful protective goddesses on the other, their role and autonomy in
everyday life was progressively diminishing. Davidson has pointed to the
striking decrease in female donors, especially autonomous female donors,
at Buddhist sites (Davidson 2002a: 91–8). At Sāñcı̄, an early site, there are
almost as many female donations as male, many of them from nuns. At
Nālandā, founded in the sixth century, personal sealings and inscriptions
attributable to women are a tiny minority. Inscriptions on Pāla and Sena
period sculptures (late seventh to twelfth centuries) include only a small
number of references to women. The few women who are mentioned are
either members of the royal family, or otherwise defined by their relation
to men (Davidson 2002a: 93–5).
Miranda Shaw has suggested that women took a central role in Tantric
Buddhist practice but in reality there is little concrete evidence for this
(M. Shaw 1994). There are indications of women acting as Tantric gurus
in their own right, and as originating particular Tantric lineages, but they
are relatively few in number. It is true that Vajrayāna Tantric vows required
respectful behaviour towards women, in a society where they were clearly
second-class citizens, and that the role of Tantric consort might legitimate
a degree of independence which was increasingly difficult to find in the
wider social context, but the evidence we have suggests that the Buddhist
6
On other occasions, rather than imported slave girls, the members of the court might themselves take
on the roles of Tantric deities. This is the situation envisaged, for example, in the doubtless legendary
account of the physical enactment of the Vajradhātu man.d.ala at the court of King Indrabhūti of
Zahor (Davidson 2002a: 243).
Tantra and the state
303
Tantric cults, and doubtless also the Śaivite kaula and Tantric cults, were
male-directed and male-controlled.7
W. G. Archer suggested some years ago that the growth in mystical
and symbolic interpretations of sexuality, and in sculptural representations of sexuality, was a kind of compensatory development in a society
where morality was becoming ever more restrictive and women’s position
was increasingly subordinate (Archer 1964: 17–25): ‘What is condemned
in actual life is countenanced in sculpture and religion’ (1964: 25). No
doubt the increasingly patriarchal and Brahmanical influence associated
with ‘sāmantization’ progressively brought to an end the relatively liberal
gender climate I noted in the early period; purity and chastity become key
themes. It is among the new martial elite groups, the Rajputs, that the
cult of satı̄ developed (Harlan 1992). Women’s new situation, increasingly
enclosed and controlled by their menfolk, mirrored that which was found
in the Islamic states who were beginning to make inroads into South Asia
at the start of this period (seventh to eighth centuries) and would control
most of the subcontinent by its end (late twelfth and thirteenth centuries).8
It is unclear at this distance how far the sources of this increasingly
restrictive situation for women were internal or external. India already had
its own sources for notions of female purity, in texts such as the Laws of
Manu or myths such as that of Kannaki (Pandian 1982), and the process of
‘sāmantization’ was driven as much by internal factors as by external attack.
What one can say perhaps is that the conflict with Islam led to a polarisation
and a militarisation on both sides. The early Rajputs, increasingly militantly
Hindu and with Tantric overtones in many cases, were the classic Hindu
form of leadership to evolve.
7
8
Unfortunately, most of these indications, including the few female members in the Mahāsiddha lists,
are difficult to pin down. The Shangs-pa bka’-brgyud tradition of Tibet claims to go back to a series of
women teachers in India, including Sukhasiddhi, but their historicity is problematic (Kapstein 1980,
1997, 2005). Sukhasiddhi is however also named in a short text attributed to Tilopa, which survives in
a Tibetan version as the source of bar do (intermediate state) and consciousness-transference practices
associated with the ‘Six Yogas of Nāropa’ (Mullin 2006: 28–9). There are a number of other similar
examples (Shaw 1994). In the modern South Asian context, Tantric sexual practices seem to vary
from situations of relative equality (e.g. among the Bauls, McDaniel 1992) to those where women act
as Tantric consorts on a commercial basis.
Another sign of this change can be seen in the way in which, towards the end of this period, dress
codes for women become much more confining. Neither men nor women in the first millennium
routinely covered the upper part of the body; this remained the case in non-Muslim parts of rural
Southeast Asia (Thailand, Laos, Bali, etc.) into modern times, as in parts of South Asia itself (Kerala,
see Devika 2005). Court versions of women’s clothing were often sexually revealing, as a vast body
of evidence from Indian sculpture and literature makes clear. One can see a progressive change from
the late first millennium onwards, although the near-total body cover of the modern sari is in fact a
very late development (Fabri 1994).
304
Origins of yoga and Tantra
tum buru and his sisters in cam bodia
While the pattern of Śaivite Tantric ritual for state purposes in South Asia
is evident in many places, the textual documentation for such ritual is
fairly limited. What is probably the best-known historical example of such
documentation comes from Southeast Asia (Cambodia), in the form of
an inscription at Sdok kak Thom, written in Sanskrit and Khmer and
dated from 1052. This refers to the introduction of a new official cult by
Jayavarman II, a Khmer king of the early ninth century.
We have evidence of a Pāśupata presence in the early years of the Khmer
state, in seventh-century inscriptions (Wolters 1979: 431; Snellgrove 2000:
443) and Śaivite forms of religion were evidently established from quite
early on. By the time of kings Indravarman (877–89) and Yaśovarman (889–
900), extensive Śaivite temples and monasteries were being erected. There
are signs of support for other Brahmanical communities and for Buddhists
as well, but substantial signs of Buddhist support are mostly rather later and
on a smaller scale; the only Khmer king to construct large-scale Buddhist
monuments was Jayavarman VII (1181-?1219) (Snellgrove 2000: 458–87).
The dominant elements of the Buddhist pantheon were the triad of the
Buddha, Avalokiteśvara and the female deity Prajñāpāramitā (Felten and
Lerner 1989: 237–8 and no. 38), suggesting kriyā or caryā Tantra practice,
but there are also substantial indications of the presence of the Hevajra
Tantra and other anuttarayoga material.9
The Sdok kak Thom inscription however is one our most substantial
pieces of evidence for Śaiva state ritual. There have been a number of studies
of this inscription, among the most recent being those of Ronald Davidson
(2002a: 204–6) and Alexis Sanderson (2003–4, cf. also 2004: 234–5). It
describes how a Brahmin named Hiran.yadāma, an ‘expert in the science of
siddhis’ (here meaning Tantric powers) revealed a unique siddhi to the king.
At the king’s invitation, he performed a ceremony intended to accomplish
the total independence of Kambuja from Javā10 and to establish Jayavarman’s position as a cakravartin. The ceremony was performed according
to the Vı̄n.āśikha, a Tantric text that corresponds to and is spoken by one
of the four faces of the deity Tumburu. Again with the king’s permission,
Hiran.yadāma taught the siddhis of the Vı̄n.āśikha Tantra and the other three
Tantras corresponding to the other three faces, along with the methods of
realising them, to Śivakaivalya, the king’s own religious teacher.
9
10
Cf. recent work by Peter Sharrock (personal communication, 2004).
It is not clear what ‘Javā’ meant at this period, but the reference may be to the kingdom of Śrı̄vijaya
in peninsular Malaysia and Sumatra (Sanderson 2004: 234 n. 8).
Tantra and the state
305
The inscription was actually put up in 1052 CE by a descendant of
Śivakaivalya. Śivakaivalya’s family claimed to have continued as court ritualists to the Khmer kings, and managers of the Tumburu cult, as well as of
the devarāja, which was perhaps a moveable cult-image associated with the
cult, in the intermediate two and a half centuries. In reality, as Davidson
(2002a: 204–6) has pointed out, our evidence refers to the mid-eleventh
century, not to the early ninth, but it still provides definite evidence for the
use of the Vı̄n.āśikha Tantra and the cult of Tumburu at this later date.
In 1973, Teun Goudriaan summarised the then-available sources on
Tumburu (Goudriaan 1973). In brief, Tumburu is known from three
main contexts. He occurs in mainstream Brahmanical texts as a gandharva or minor celestial deity. He is described in several Śaivite and Tantric
texts (Yogavāsis..tha, Vis.n.udharmottara Purān.a, Agni Purān.a, Śāradātilaka,
S.at.karmadı̄pika, etc.) as a (usually) four-headed form of Śiva, always in association with four or seven Mātr.kā-type goddesses who are his consorts, and
he occurs in one Buddhist Tantric text (the Mañjuśrı̄mūlakalpa), where
he is accompanied by four goddesses described as his sisters. Goudriaan
summarised this material as follows:
From all these references a picture emerges of Tumburu as a healing god of
sovereignty seated in the midst of four fierce goddesses with allegorical names
expressive of various aspects of victory. The god commands these and occasionally
counteracts their evil influence. (Goudriaan 1985: 23)
Some years later, a copy of the Vı̄n.āśikha Tantra itself turned up in the
National Archives of Nepal, and Goudriaan subsequently edited and translated this text. It takes the usual form of a dialogue between Śiva and his
consort, who are seated on Mount Kailas, ‘surrounded by Ganas (headed
by Mahākāla), Siddhas, sages and other supernatural beings’ (Goudriaan
1985: 30).
The picture of the god that emerges in this text is much the same as that
in the previously known sources: a royal manifestation of Śiva, particularly
associated with healing, but clearly able to exercise more destructive powers.
The Goddess, interestingly, says that the three other Tantras Śiva has spoken
so far (which correspond to the texts associated with the three other faces
of Śiva by the Sdok kak Thom inscription) are about the realisation of
higher wisdom, whereas humankind needs techniques to deal with everyday
problems. The text goes on to describe preparatory rituals, the construction
of the man.d.ala and the worship of the deities, and the structure of nād.ı̄s
and cakras, etc. The goddesses and four further attendant goddesses are
described as fear-inspiring and the symbolism reinforces this in various
306
Origins of yoga and Tantra
ways (for example, one is seated on a corpse). These are still definitely wild
and dangerous goddesses who Tumburu controls – and allows the sādhaka
or practitioner to control and employ.
The Tantra goes on to explain how to use the Tumburu practice to achieve
various magical results: to attract a woman or man, to destroy a person
through sorcery, to subjugate someone, to eradicate or expel an enemy army,
to cause dissension among friends, to restore a person to health and protect them from danger, and so on. A subsequent section provides a further
series of rituals, many of them love magic of various kinds: to make someone impotent, to increase sexual potency, and the like. There are various
additional instructions on meditation and ritual, including directions for
performing a fire offering (homa).
These are all very typical contents and ritual purposes for both Śaivite
and Buddhist Tantra. Many texts also give techniques for the attainment of
higher knowledge, liberation or enlightenment. This particular Tantra, as
the Goddess’s opening request implies, is not concerned with such matters.
The significant thing about the Vı̄n.āśikha Tantra however is that we have a
dated historical reference stating that rituals according to this Tantra were
performed over a period of two and a half centuries on behalf of Khmer
royalty. While one can never be sure that a particular Tantric text which has
survived to modern times is actually the same as that which is mentioned in
an eleventh-century description, Goudriaan clearly felt that the similarities
here were close enough that we were probably dealing with the same or a
closely related text.
Tumburu’s healing powers occupy a rather small part in this text, but in
general healing and in particular the restoration of youth and sexual potency
appear to have been an important component of what Tantric priests had
to offer, often augmented by material medicines from the techniques of
physical alchemy. We have already seen that much of the appeal of the
possibly fraudulent Tantrics to King Hars.a was that they would provide
him with long life and a perfect body.11
Certainly in East Asia, which we turn to next, the ability of Buddhist
Tantra to provide healing seems to have been a major part of its appeal to
Chinese and Japanese rulers, as the frequency of the iconography of the
Healing Buddha, Bhais.ajyaguru, shows. There are also reports of incidents
where the successful performance of Bhais.ajyaguru rituals for members
of the ruling family led to state patronage of Bhais.ajyaguru and of the
11
Another Kashmiri text, Ks.emendra’s Samayamātr.kā, provides a range of further examples of the use
of Tantric ritual, again told in satirical style (Wojtilla 1984, 1990).
Tantra and the state
307
Tantric ritualists associated with his cult. However, it is worth noting that
Bhaisjyaguru himself is not only a healing deity. He has his own man.d.ala
of yaks.a generals who will protect those who reverence him (e.g. Kobayashi
1975: 119; Sugiyama 1982).
The conclusion of Goudriaan’s introduction to the Vı̄n.āśikha Tantra is
worth quoting:
Just like his deity, the Sādhaka is ambivalent: he can protect and destroy. It can be
imagined that such claims, if supported by a powerful and resourceful personality
(his mental power enhanced by concentration on the internal deity), could be
of interest to the South East Asian rulers who would be unable to find a similar
competence among native practitioners. (Goudriaaan 1985: 61–2)
There is a nice balance in this quote. Goudriaan sees meditation on the deity
as enhancing the mental power of the practitioner, but also gives full value
to the pragmatic demands and credulity of the Khmer kings. Goudriaan
goes on to say that ‘It is in any case a remarkable fact that a tradition
considered as inferior in India was able to strengthen its position in South
East Asia by royal support’ (1985: 62). There may well be some truth in this
picture of Tantric ritual being most prevalent at the frontier areas, but in
fact we would seem by now to have enough inscriptional, iconographical
and textual evidence to suggest that kings in the Indian subcontinent were
also calling on the services of Śaiva Tantric gurus.
As for Buddhist Tantra, our main evidence for mainland South Asia
comes from the Pāla state and from the Tibetan chronicles, since the
Tibetans acquired much of their own version of the Buddhist Tantric teachings from the great monastic universities of the Pālas in the eleventh and
twelfth centuries. Our evidence here is supplemented in part by the remains
of Buddhist monasteries and by surviving Buddhist images in South and
Southeast Asia; the Khmer state is particularly rich in Hevajra iconography.
The prevalence of man.d.ala schemes in temple and monastery architecture is one significant indicator of the importance of Tantric practice
for their builders. A particularly common temple form of the period, first
found in the Pāla and Samatat.a states from the late eighth century onwards,
consists of four ritual halls in the four directions, as at the great monastery
of Paharpur in northern Bangladesh (Samuel 2002c). Here we are probably
concerned with yogatantra and perhaps with the idea of four kinds of ritual
action. Other monasteries excavated in Bangladesh and East India also
seem to follow this pattern of a large man.d.ala-style temple in a courtyard.
The central temple of the monastery of Samye (bSam yas) in Tibet is also
built according to a man.d.ala scheme, and is said to have been modelled on
308
Origins of yoga and Tantra
another major Indian monastery, Odantapuri, the location of which has
not yet been identified.
Paharpur is in the core area of the Pāla state, which along with Nepal
and Kashmir was the main part of the subcontinent where Mahāyāna and
Vajrayana Buddhism continued to be patronised from the seventh century
onwards. While this region was later to be the source of many of the
Tantric teachers who brought Buddhism to Tibet and elsewhere, it seems
that Buddhism here was relatively conservative in the early Pāla period,
with a stress on the role of the historical Buddha Śakyamuni, and the sites
associated with him.12 It is only in the late ninth and tenth centuries that
anuttarayoga-type material, with the cult of the fierce goddesses and Śaivitederived male deities such as Hevajra and Cakrasamvara, begins to appear
in the iconographic record at sites such as the monastery of Ratnagiri in
Orissa (Hock 1987).
By this stage, the Pāla state was the only major Indian state where
Buddhism was still flourishing; Susan Huntington has described it as
‘an anomaly in the otherwise totally Hindu world of India at the time’
(S. Huntington 1990: 70), though this perhaps understates the survival
of Buddhist institutions elsewhere, particularly perhaps in South India,
Nepal and Kashmir. Huntington is probably right however to argue that
the expanding presence of Buddhism overseas was a major source of patronage which enabled the Buddhist institutions of the Pāla state to keep afloat:
[R]ecent research has demonstrated that only a few of the the Pāla kings were
exclusively Buddhist and that royal patronage constituted only one of the many
sources of support for the religious institutions of the day. Instead, it may be posited
that the survival of this island of Buddhism in India’s Hindu world was in great
measure due to the international activity of the period, which not only fostered
the religious institutions but must have contributed greatly to the economy of the
region. (S. Huntington 1990: 70–1)
Here, as earlier in its history, Buddhism was an internationalist religion,
doubtless preserving its close association with the mercantile class but also
as a consequence vulnerable to the disruption of long-distance trade and
to the need for traders to keep on good terms with the local political elites.
How far the eventual collapse of Buddhism in the region was a result of
direct military activity, how far to the loss of state patronage, and how far
to the fact that its traditional lay support base was no longer in a position
to support it is difficult to decide at this distance. Probably all these factors
12
Kinnard 1997 contrasts the conservatism of early Pāla Buddhism with Rās.t.rakūt.a-patronised
Buddhism in the Deccan, where the new man.d.ala schemes appear earlier.
Tantra and the state
309
played a part. International support was undoubtedly a factor, however,
and it is noticeable that by the twelfth and thirteenth century there appears
to be a shift away from Tantric man.d.ala-form temples, perhaps resulting
from an increasing reliance for patronage on the territories of Arakan and
Burma to the east, where Mahayana and Vajrayāna Buddhism was being
increasingly replaced by Theravādin orthodoxy (Samuel 2002c). In time,
what was left of Buddhism in East Bengal would be re-oriented towards
Arakan and the Theravāda (Tinti 1998).
bu dd hist tantric ritual in china, japan and korea
In the eighth to tenth centuries, despite the progressive incursions of Islam
in the West, the Buddhist support base continued to grow in East Asia.
Further evidence for the state use of Buddhist ritual procedures comes
from this region, where these procedures were at various times adopted by
the Chinese, Korean and Japanese states. Here, with the aid of the dated
Chinese translations and the generally better East Asian historical record,
one can see the progressive development of Buddhist rituals for the state
over several centuries.
Thus the Suvarn.abhāsottama or Suvarn.aprabhāsa Sūtra, the non-Tantric
text that I referred to in Chapter 9 as providing the earliest evidence
for Buddhist Tantric man.d.alas, gives considerable importance to political issues. The Suvarn.aprabhāsa Sūtra was translated into Chinese between
414 and 421. It was in fact classified in Japan as one of three ‘State Protecting
Sūtras’. The other two were the Saddharmapun.d.arika or ‘Lotus Sūtra’ and
the so-called Renwang Jing, a text in the style of a Prajñāpāramitā Sūtra that
is generally regarded as a Chinese production.13
Thus, in Chapter 6 of the Suvarn.aprabhāsa Sūtra, the Four Great Kings,
the four yaks.a-style deities of the four directions whom we encountered in
Chapter 5 of this book, approach the Buddha. They proclaim that should
a king of men who has heard this sūtra protect and support monks who
hold this and the other chief sūtras, they, the four Great Kings, along with
their twenty-eight yaks.a generals and numerous hundreds of thousands of
yaks.as, will protect and assist that king and ensure him peace and welfare.
13
In a recent book review, Whalen Lai has suggested that the Suvarn.aprabhāsa might have been used
as the basis of ‘a pact among early Buddhist kings dedicated to the propagation and protection of the
Dharma, in early Mahayāna Northwest India’. It was apparently used in this way in the Northern
Dynasties in China. The Renwang Jing was used in a similar way. (H-Net book review of Wang
Zhenping’s Ambassadors from the Islands of Immortals, published by H-Buddhism@h-net.msu.edu,
September 2006.)
310
Origins of yoga and Tantra
Similarly, if he makes gifts to the monks, nuns, laymen and laywomen who
hold the chief sūtras, the Four Kings will make his population prosperous.
(Emmerick 1970: 24–5) They also promise to cause dissension and trouble
for any neighbouring king who wants to invade his territory. Venerating
and reciting the sūtra will itself strengthen the Four Great Kings, while if
this is not done the region in question will suffer various kinds of trouble.
In subsequent chapters of the Suvarn.aprabhāsa the message is reinforced.
The earth-goddess promises to make the earth stronger and more fertile
where the sutra is preached and to produce more fragrant, moist, tasty and
beautiful leaves, flowers, fruits and crops, so that people who eat them will
be stronger, more beautiful and live longer. Chapter 12 is an ‘Instruction
concerning Divine Kings’ stressing that kings should act justly and that
disaster will follow should they fail to do so. A long series of yaks.as and
other spirits promise protection to those who hear the sūtra.
I shall leave aside the Saddharmapun.d.arika, though it was, of course,
to assume immense importance in East Asian Buddhism in later times,
and look briefly at the Renwang Jing. The Renwang Jing’s full title can
be translated as The Prajñāpāramitā Sūtra Which Explains How Benevolent
Kings May protect Their Countries. It is extant in two Chinese translations,
the first by Kumārajı̄va, dating from 401, and the second by Amoghavajra,
dating from 765. There were, however, doubts from early times about
the existence of a Sanskrit original for this text: Stanley Weinstein in his
Buddhism under the T’ang suggests that it originated in China, and that
‘[I]ts anonymous author was apparently attempting to dissuade Northern
Wei rulers from interfering with the autonomy of the [Buddhist sangha]’
(Weinstein 1987: 176 n. 6; see also De Visser 1928–35; Conze 1973, 1978;
Orzech 1998).
Whatever the case, this text, which is described as being preached to the
sixteen kings of the mahājanapadas, led by King Prasenajit of Śrāvastı̄, was
to become an important text both in China, and also in Japan. Chapter 5 of
this sūtra recommends that kings sponsor elaborate rituals, based around
the reading of the sutra, for times when ‘riots are imminent, calamities
are descending, or robbers are coming to destroy’. Chapter 7 prescribes a
further ceremony that can be performed to protect from other difficulties,
and promises that the Five Bodhisattvas of Great Power and various other
deities will protect the countries of kings who receive and keep the Buddha,
dharma and saṅgha.
A variety of other sūtras including rituals for various pragmatic ends
were translated into Chinese from the early fifth century onwards, but
the two texts which were to form the basis of East Asian Tantra, the
Tantra and the state
311
Sarvatathāgatatattvasam
. graha and the Mahāvairocanābhisam
. bodhi Sūtra,
were not translated until 723 and 724 respectively. Both seem to have been
written considerably before that time, with the Mahāvairocana Sūtra probably being the earlier (Matsunaga 1997a: 177–8; Hodge 2003). In fact,
the Chinese Buddhist traveller Wuxing had already obtained a copy of
the Mahāvairocanābhisam
. bodhi Sūtra in India in 685 CE. The delay in the
Chinese translations was due to internal political factors in China.
Buddhism had reached China quite early and began to gain popularity
among the Chinese from the first century CE. By the fourth century it
had a widespread following among all levels of society, and the various
rulers of the Northern and Southern Chinese dynasties at this period were
all basically supportive of Buddhism. The Tang Dynasty was a different
matter, and the establishment of the Tang state was immediately followed
by a repressive edict against Buddhism, in 621 CE.
While Buddhism in South and Southeast Asia was primarily competing against Brahmanical religion, and in some areas also the Jain religion,
in China the competition was from the Daoists. Gaozu, the first T’ang
emperor, was pro-Daoist, declared Buddhism a foreign religion and moved
to reduce the power of the Buddhist clergy. The second T’ang emperor,
Taizong (620–49), was also by and large anti-Buddhist, and decreed,
for example, that Daoist monks and nuns should take precedence over
Buddhist monks and nuns, although he patronised the famous translator
Xuanzang towards the end of his life (Weinstein 1987).
Xuanzang continued to be influential under the early years of the following emperor, Gaozong (649–83). Xuanzang was unsympathetic to Tantric
traditions and apparently used his influence at court to block the introduction of Tantric teachings by an Indian pan.d.ita who arrived in China in 655.
It seems that some Tantric teachers continued to be active in China during
this period, all the same, since the Korean monk Myongnang is said to have
studied Tantric ritual in China at this time. After Myongnang’s return to
the Korean kingdom of Silla, he was responsible for the Silla state’s ritual
defence against attempted Tang invasions in 668 and 669. His rituals were
successful, and on each occasion the Tang fleet heading for Silla was sunk
at sea (Grayson 1989: 59).
After Xuanzang’s death, the emperor Gaozong blocked further translation activities. However, the emperor had a stroke and his wife, the famous
or notorious Empress Wu (Wu Zetian), who took over from him, was
strongly pro-Buddhist. A reaction followed after Empress Wu’s reign, and
the next significant Emperor, Xuanzong (712–56) again moved to restrict the
power of the Buddhist clergy. Xuanzong was, however, interested in Tantric
312
Origins of yoga and Tantra
Buddhism, and it was during his reign and that of his two immediate successors, Suzong (756–62) and Daizong (762–79), that the main translations
of Tantric texts were undertaken. These were made by the Indian Tantric
masters Śubhakarasim
. ha (arrived in China 716, died 735) and Vajrabodhi
(arrived in China 719, d.741) and by Vajrabodhi’s students Yixing (d.727)
and Amoghavajra (also known as Bukong, d.774) (Weinstein 1987). This
was a period of Tibetan and Uighur invasions and Chinese civil wars, and
there was plenty of demand for Tantric defensive ritual at court. As I mentioned above, the Mahāvairocanābhisam
. bodhi Sūtra and the Sarvatathāgatatattvasam
. graha Sūtra were translated in the 720s. These correspond to the
earlier stages of Indian Buddhist Tantra, those which Tibetans nowadays
classify as kriyā, caryā, and yoga Tantra, and the emphasis here is on peaceful
deities, although some fierce and violent deities are included. These two
texts were transmitted on to Japan within a few years and formed the basis
of Japanese Tantric or esoteric Buddhism (Yamasaki 1988).
I have elsewhere narrated the famous story of the defeat of a Tibetan, Arab
and Sogdian attack in 742 by Amoghavajra, using rituals derived from the
Renwang Jing (Samuel 2002d). Weinstein (1987: 170 n.41) regards the story
as suspect, since it is not in the early biographical sources on Xuanzong’s
reign, but there is little doubt that, as Lokesh Chandra has argued, the
translations of Tantric texts at this time were carried out in large part ‘for
the security and stability of the State against the attacks of Tibetans and
Arabs . . . Reverses on the battlefield were followed by renewed vigour in
translating the Tantras’ (Chandra 1992b: 266; see also Chandra 1992c).
Some more translation activity followed in China in the late eighth
century, but the increasingly chaotic conditions of the later Tang were
unfavourable to Buddhism. A massive purge of the Buddhist saṅgha took
place in 835, followed by a period of brutal suppression under the emperor
Wuzong (840–6), beginning in 842 and climaxing in 845 and 846 with the
closing and destruction of almost all monasteries, the burning of scriptures
and the forcible disrobing of the saṅgha. The Emperor’s death brought
this to an end but the extent of the destruction, followed by a devastating
rebellion some thirty years later (875–84), effectively brought all schools
of Buddhism in China except Chan and Pure Land to an end (Weinstein
1987). The later ‘Anuttarayoga’ Tantric material, in which the fierce Śaivite
gods and goddesses move to the centre, thus did not reach China or Korea
until it arrived with the Mongol court some four centuries later, and it was
never really assimilated into Chinese Buddhist practice.
The Mahāyāna State Protecting Sūtras and the earlier phases of Tantric
Buddhism continued to be influential in Korea and particularly Japan,
Tantra and the state
313
and again I have narrated some of this story elsewhere (Samuel 2002d). In
Japan, the Renwang Jing, or the Ninno-kyō as it is called in Japanese, was
extensively used in state-protecting rituals. Marinus De Visser’s Ancient
Buddhism in Japan refers to many occasions on which such rituals were
in fact performed in Japan, for example in 940 CE by the Tendai priest
Jōzō, when the Japanese state was threatened by the rebellion of Taira no
Masakado (de Visser 1928–35; Chandra 1992a).
The well-known Naritasan Shinshōji temple, in the small town of Narita
close to Tokyo’s main airport, owes its legendary origins to another set of
rituals held to suppress this same rebellion. In this case, they involved a
specifically Tantric deity, Fūdō (Acala), who is an important figure in the
Mahāvairocanābhisam
. bodhi Sūtra (Dainichi-kyō in Japanese) (see Linrothe
1999: 151–3; Yamasaki 1988: 12, 83–4, 172–5). The Shingon priest Kanjo was
directed by the Emperor to bring an image of Fūdō to Narita in order
to defeat the rebellion. This image had been carved and consecrated by
Kōbō Daishi (c. 774–835 CE), the founder of the Shingon tradition. After
three weeks of continual fire offerings, the leader of the rebellion, Taira no
Masakado, was killed by the Emperor’s forces and peace was restored. At
least some of the credit for this was given to the Fūdō rituals. The image
itself refused to move back to Kyōtō (Acala means ‘immoveable’ in Sanskrit
so it was just living up to its name), and Fūdō directed Kanjo in a vision to
build a temple for the image at Narita. This was the origin of the Naritasan
Shinshōji temple.
Perhaps this is enough to make the point: both sūtras and tantras were
employed to defend the Chinese, Korean and Japanese states against attack,
and it seems clear that the translation of these texts into Chinese and the
establishment of monasteries to maintain the ritual traditions associated
with them was motivated in large part by their believed usefulness to the
state. We have little direct evidence to demonstrate that the same was true
of Buddhist Tantras in South or Southeast Asia, but it seems likely that
this was the case. We might consider, for example, the Buddhist and Śaiva
establishments supported at a slightly later date by the Javanese court of
Mataram (Becker 1993). The Javanese pattern of dual support for Śaiva and
Buddhist Tantric ritualists continued into modern times in Bali, which still
has both Śaiva and Buddhist hereditary Brahmin priests.
th e new pat tern: kathmand u valley as exempl ar
In Chapters 6 and 7 I gave the cities of Chiang Mai in Thailand and
Madurai in Southern India as examples of the Theravada Buddhist and
314
Origins of yoga and Tantra
Brahmanical Hindu patterns as they developed in later times. If the Tantric
pattern represents a third option, the place that comes closest to representing
it in modern times is probably the Kathmandu Valley. The valley today is
nearly a thousand years distant from the period of which we are speaking,
as are Chiang Mai and Madurai, and it has undergone two centuries of
strongly pro-Hindu rule since the Gurkha conquest, but it is still suggestive
in many ways of how a mediaeval Tantric polity might have operated.14
The Valley is a pluralistic society in both ethnic and religious terms, with
major pilgrimage sites associated with both Buddhism (the two great stūpas
at Svayambhū and Bauddha, along with many lesser sites) and the ŚaivaŚākta traditions (including Paśupatināth, Guhyeśvarı̄ and a host of other
temples). The main urban population, the Newars, have a strongly mercantile and artisan component15 and have both Buddhist and Brahmanical
groupings, with religious identity being mainly a question of whether one’s
family traditionally calls on a Buddhist or Brahmanical Tantric priest for
family rituals (G. S. Nepali 1965; Gellner 1992; Toffin 1984; Gellner and
Quigley 1995).
Buddhist Tantric priests have traditional associations with one or another
‘monastery’ (bāhāh. or bahı̄, Locke 1985; Gellner 1992), but monasticism
as such ceased to be practised some centuries ago in traditional Newar
society. Boys of the appropriate high Buddhist castes spend a week as monks
before formally renouncing the monastic life.16 Tantric priests are initiated
in the anuttarayoga Tantra traditions of Vajrayāna Buddhism. Vajrayāna
Buddhism has explicitly exoteric and esoteric components, with separate
shrines and also multiple layers of meaning and interpretation associated
with individual shrines and images (Gellner 1992). The same would seem
to be true on the Śaiva side, although there has been less work on this.
Major valley cults such as that of the deity Bum
. gadyah. (Lokeśvara) again
have multiple meanings and multiple identities, in this and some other
cases including both Buddhist and ‘Hindu’ identities (Locke 1980; Owens
1995; Tuladhar-Douglas 2005).
Another example here is the cult of the Buddhist deity Hāritı̄, whom
we met in Chapter 5 as converted yaks.a, and possibly also as an important
deity of Gandhāra and Mathurā. She remains a significant figure in the
14
15
16
The Hindu-Buddhist society of Bali in Indonesia has many parallels to Newar society (Samuel 1982).
While Vajrayana Buddhism is also the principal religion of Tibetan societies, they have there become
part of a quite different social structure, for reasons sketched in Samuel 1982 and discussed at length
in Samuel 1993.
See e.g. Lewis 2000: 50–4.
In the early twentieth century, the Theravāda monastic tradition was imported into the Valley, and
has now become quite significant.
Tantra and the state
315
Kathmandu Valley, where she is linked to a Tantric healing cult which is
still of some importance in the valley (Iltis 2002; Merz 1996). The principal
temple of Hāritı̄, on the hill at Svayambhū, nowadays has a second identity
for Hindu worshippers as a temple of Śı̄talā, one of the principal diseasegoddesses of the Brahmanical pantheon.
At the time of the Gurkha conquest, the Valley had been divided for some
time into three small Newar kingdoms, each with a chief town (Kathmandu,
Patan, Bhaktapur) and a palace complex centred on the temple of the royal
deity, Taleju, a fierce goddess regarded as a form of Durgā (Vergati-Stahl
1979). Each of these cities is surrounded by circles of temples, including sets
of the eight Mātr.kā deities (Slusser 1982; Gutschow 1982, 1993; Gutschow
and Bāsukala 1987). As mentioned earlier, these deities are brought to life
in annual cycles of ritual dances that form part of an extremely complex
ritual ‘choreography’ that encompasses the whole Valley throughout the
year (Levy 1987, 1992; Toffin 1996, 1999; Gutschow 1996; van den Hoek
2005; Korvald 2005). The dancers are collectively known as Navadurga (the
Nine Durgās). Alongside the various Goddesses (Fig. 12.3), they also include
Gan.eśa and Bhairava. These rituals are mostly sponsored by associations
(guthi) linked to one or another religious institution, a pattern that goes
back in some respects to the guild structure of mediaeval India and has
largely disappeared elsewhere in the subcontinent (Vergati 1986).17
This is a highly simplified sketch of a complex and fortunately wellstudied society, but it may serve to suggest some of the lines along which
the mediaeval Tantric polity functioned, both in terms of the way it was
instantiated in civic and royal ritual, and in regard to how sites maintained
multiple meanings and interpretations for the various social, religious and
ethnic groups within the Valley.
ma s ked ritual dance traditions in the nepal valley
and elsewhere
I mentioned the masked ritual dances through which the As.t.amātr.kā deities
are embodied in the Kathmandu Valley. Even today, these are not entirely
safe folkloristic performances. The dancers are in a trance, they carry swords
and burning torches that may be thrown at the audience, and they drink the
blood of sacrificed animals. There are legends of human sacrifice associated
17
For the rather different tradition of the Newar Buddhist town of Sankhu, in which the three dancers
(Devı̄, Can.d.ı̄, Bhairava) are seen as representing the Buddhist goddess Vajrayoginı̄, see Shrestha
1999. The role of Vajrayoginı̄ in Sankhu is discussed further in Zanen 1986.
Figure 12.3. (left to right) Indrayān.ı̄, Brahmayān.ı̄, Cāmun.d.ā
Tantra and the state
317
with the dances; Toffin notes that in Theco village, where the dancers are
brought out for two village festivals each year, their emergence
is much feared. Every year, it is believed that a person will die in the village or in
the neighbourhood during these two festivals. The fierce goddesses, it is said, “eat”
this ill-fated villager. Kumārı̄ and Vis.n.udevı̄ [. . .] are the most dangerous deities.
They have to swing their faces continuously during the dance and the processions
in the streets since if they fixed a devotee for a few seconds without moving their
head he would die at once. (Toffin 1996: 233)
A major festival is held every twelve years when the masks are renewed, and
there is a strong belief that human sacrifice formed part of this ritual in the
past and occasionally still does so today.
Even today, a feeling of insecurity and fear spreads around the village and the
neighbouring localities during the twelfth year festival. The families do not walk
out at night, and they watch over their children with special attention. (Toffin
1996: 245)
As this suggests, children are seen as particularly at risk, and the disappearance of small children in the recent past has been associated with genuine
fears that they may have been killed as offerings to the dancers. This is acted
out in a sequence in another set of these dances today in which a dancer,
representing a form of Bhairava, chases and tries to catch the local children
(Sardar n.d.).
Obviously there is an element of play here, but there is also an underlying
presence of something else. I would suggest that the dangerousness or perhaps
one should say rather the ‘scaryness’ of these deities – the stories of human
sacrifice and of villagers who die during the annual dances, the frightening
aspects of the dance performances – is an important part of how they
operate now and an important indication of how they operated in the past.
A significant issue raised by practices such as this is what could be called
the emotional and psychic force of divine forms in popular consciousness. Tantric ritual is about powerful and dangerous forces, which must be
encountered and dealt with for the good of the community. These forces
can only be manipulated by specialist priests and ritualists, and even then
there is still a risk that things can ‘go wrong’. The seven wild and dangerous goddesses who dance through the villages of the Kathmandu Valley in
the company of Bhairava and Gan.eśa are participants in one of a series of
Tantric rituals that undergirded and maintained the Newar state and which
were continued by the Nepali state that followed it. One can imagine that
in a pre-modern society, without the corrosive presence of modernity or
318
Origins of yoga and Tantra
Figure 12.4. ’Cham dancer at Zangs mdog dpal ri dgon pa, Kalimpong
the competing visual claims of the cinematic or television fantasies, such
rituals would have some real purchase on the collective psyche.
We can see versions of these masked ritual dance traditions in many places
throughout South and Southeast Asia, and it is worth referring to a few
other examples. Of these traditions, the Tibetan masked ritual drama of the
Tantra and the state
319
’cham, which has also received the most intensive study, is the most sophisticated and complex (Fig. 12.4). This is normally performed by monks, and
in fact almost all large monasteries perform ’cham annually. Most of the
deities are male, including transforms of Mahākāla and forms of Padmasambhava as well as Tibetan local gods, though in some versions, such as the
Bon-po cham of the goddess Srid pa’i rgyal mo (Schrempf 1999, 2001), there
are important deities of the ‘wild goddess’ class. As I pointed out in Civilized
Shamans, relying in large part on the work of Stephan Beyer (1973), the ’cham
is structurally a modification of a standard Tibetan ritual procedure of Indic
origin, the tshogs or gan.apūja ritual, in which the power of Tantra is directed
to the destruction in symbolic form of ‘obstacles’ to the spiritual progress of
the monks and the well-being of the associated village community (Samuel
1993: 265–7). The dancers represent in external physical form the deities of
the man.d.ala who would otherwise simply be visualised internally by the
performers of the ritual (see also R. Kohn 1988, 2001).
There is little sense here of possession by the deity or of the element
of risk and danger which goes along with it. Perhaps at the other extreme
are the teyyam rituals of Kerala and the bhuta rituals of southern Kannada.
In both of these cases, ferocious deities, mostly of the ‘wild goddess’ type,
are held to possess masked low-caste dancers. Normally in these cases one
dancer appears at a time, wearing the elaborate costume of a specific named
deity. He is expected to be possessed by the god or goddess. The deities here
mostly have Sanskritic identities and include many familiar figures, such as
Bhairava, Kālı̄ and Cāmun.d.ā. Richard Freeman and Gavin Flood, who have
studied the teyyam rituals in detail, note the overlap here between spiritpossession (or more accurately spirit-mediumship) and local Tantric traditions. The dance of the teyyam cannot be simply classed as a form of popular
spirit-possession ritual, nor as a learned Tantric procedure (Freeman 1993,
1994, 1999; Flood 1997; Pallath 1995). As Flood makes clear, it needs to
be analysed at several levels. However, it is undoubtedly an occasion of
heightened emotion, deriving from the controlled violence at its heart:
‘[t]he emotion evoked in the crowd is [. . .] integral to the structure and
process of the ritual’ (Flood 1997: 177).
The bhuta dances of South Kanara (Tulunad, Fig. 12.5) seem close to
teyyam (e.g. Claus 1973, 1979, 1984, 1993; Nichter 1977). One can see obvious
resemblances too in the yak tovil, the Sri Lankan tradition of masked exorcistic dances, but there are notable differences here too – yak tovil is normally
performed today as a healing and exorcistic ritual, in which specific malevolent deities are persuaded to stop afflicting a patient. There are, I think,
important commonalities, but these rituals require more extensive treatment than I can give them in this context (Kapferer 1979, 1983; Vogt 1999).
320
Origins of yoga and Tantra
Figure 12.5. Bhuta dancer, South Kanara District
Other dance traditions that appear to share some of the same fundamental
assumptions can be found in Orissa (Emigh 1984) and elsewhere in South
Asia.
We can also see significant resemblances with some of the Indonesian
masked dance traditions, such as or the well-known Rangda-Barong dances
of Bali (e.g. Belo 1949; Geertz 1975).18 While there is a dramatic structure
of sorts to the Rangda-Barong dances, they are essentially centred around
18
Another Indonesian example is the warok tradition of Ponorogo in Eastern Java, where the association
of male Tantric power remains quite strong (I. Wilson 1999). On the use of Tantric power in
Tantra and the state
321
a struggle between two contending powers, a male being who has the form
of a lion and is, like Bhairava, explicitly identified as a transform of Śiva,
and a female figure, Rangda, who takes the form of a terrifying witch, and
is closely linked to the Balinese form of Durgā (see Stephen 2001 for these
identifications). Leo Howe says of these dance-dramas:
The drama is usually put on in times of trouble for the village or as part of the
celebrations of the annual festival of the ‘death’ temple. It is often performed
near the graveyard. It begins quietly with much joking and slapstick, but around
midnight, when the anticipation and tension in the audience is high, Rangda erupts
into the arena causing mayhem and fear. Rangda’s acolytes possess the Barong’s
supporters, who try to stab themselves with daggers while the Barong tries to
safeguard them. Sometimes Rangda shoots off into the crowd scattering it in all
directions, and people are genuinely afraid. The performance ends in a stand-off
between Rangda and Barong [. . .], but there is always the fear that the dangerous
magical forces released may get out of control. (Howe 2000: 71–2)
As with the Newar and Keralan dances, when performed as a serious part of
a temple festival, for all that everybody knows what is going to happen, this
can be a genuinely scary occasion even today. There is also a transgressive
element, both in Rangda’s introduction of the unsavoury powers of night,
death and sorcery, and in the risk of spectators losing control and being
caught up in the possession-state that Rangda invokes on Barong’s followers.
One can sense on such occasions the ways in which Tantric styles of ritual
generated the conviction necessary to perform their state- and societysupporting function.19
The reported use of human sacrifice in Tantric rituals of state in the
past can also be read in this way. I have already mentioned the association
of the Newari Navadurgā dancers with rumours of child-sacrifice. The
use of human sacrifice in foundation rituals for cities and other major
state enterprises in the past seems well-attested, though again one cannot
be sure how much of this is rumour and how much reality. Examples
include foundation sacrifices in Burma in the early nineteenth century, for
the construction of the city of Mandalay (Houtman 1996; see also Jordaan
19
contemporary Bali by policemen, tax-collectors and others who wish to bolster their personal abilities
with Tantric magic, see Connor 1995.
I have not attempted to speculate on when these masked dance practices may have arisen. There
are indications of participants in ritual gatherings taking on the role of Tantric deities, for example
with the Saptamatr.kas in the Kulacūd.āman.i Tantra (Finn 1986), and there are also indications in
early South Indian material and elsewhere of spirit-possession cults and dance rituals (e.g. Hart
1975). The Tantric masked dances seem to integrate elements from both, and might be seen as a
progressive development under the influence of court ritualists. For all we know, they could go back
to the Gupta Saptamātr.kā cults. We do not however appear to have any iconographic record of early
versions of such practices.
322
Origins of yoga and Tantra
and Wessing 1999), routine foundation sacrifices of low-caste men in premodern Kerala (Uchiyamada 2000), and human sacrifice in nineteenthcentury Bali to empower the local rāja’s magical weapons (Wiener 1995).
conclusion
One can portray a picture in which Tantric ritual dominates mediaeval
South Asian society, but this would perhaps be exaggerated. This is in
part a question of balance. The mass of the population were committed
to the slowly ‘Brahminising’ village cults, while the official royal cults were
for the most part Vais.n.nava or ‘orthodox’ Śaiva. Tantric practice, in the
sense we are discussing here, was part of the ritual technology of the state,
but probably a relatively minor part in most cases; most state ritual did
not employ ferocious goddesses or transgressive imagery or inhabit this
dangerous fringe territory between the respectable and ordered world of the
social community and the realm of misfortune, terror and death. The use of
Tantric practitioners for pragmatic purposes was doubtless widespread, but
hardly everyday; most people would approach such ritualists, rather like
the mudang of contemporary Korea (C. Kim 2003) unwillingly and only
when other more straightforward solutions to their problems had failed.
There was also a continuing issue about whether these practices were
legitimate and appropriate, which is perhaps hardly surprising because their
position on or beyond the edge of the legitimate was intrinsic to their power.
Śaṅkara, the great Indian Vedāntic philosopher, who probably lived in the
early ninth century, is portrayed in his biography, the Śaṅkaravijaya, as
condemning the approaches of various kinds of Tantric practitioners and
defeating them through argument or spiritual power (e.g. Lorenzen 1991:
31–48; Courtright 1985: 218–19). He is also associated with the replacement
of fierce goddess cults with those of benign female deities through the
encouragement of the Śrı̄vidyā tradition, and thus with the ‘establishment
of a chaste Vedic (vaidika) goddess worship untainted by Tantric elements
(tāntrika) and the elimination of “objectionable” practices such as oblations of blood, alcohol, human flesh and sexual rites’ (Wilke 1996: 123–4).
Whether the historical Śaṅkara actually campaigned against transgressive
Tantric practices is far from certain, but he came to stand in later Indian
tradition for the ‘purification’ of ritual and the removal of these practices.
Further north, we have already met Kalhan.a and Ks.emendra’s attacks
on the Kashmiri kings who indulged in Tantric practices. In Buddhist
Southeast Asia, King Aniruddha (Anawratha) of Burma is supposed to have
disbanded the ‘Tantric’ ari monks (Ferguson and Mendelson 1981), and
Tantra and the state
323
Theravādin monarchs elsewhere similarly opposed Tantric-style practices
(Tambiah 1984). These practices never completely disappeared, but they
were increasingly marginalised. Much the same was true in large parts of
South Asia.
In my earlier book, Civilized Shamans, I suggested that the anomalous
character of the Tibetan religious and political system in comparison with
the Buddhist societies of Southeast Asia was because, unlike those societies,
it had never had an effective centralised state that was able to control and
marginalise Tantric practices (Samuel 1993: 24–36). My argument did not
imply that the state was incompatible with Tantra as such, and I had previously discussed the Newars and Balinese as examples of centralised states
where Tantric practice had retained a relatively central position (Samuel
1982). It is clear though that Tantra was a double-edged sword when wielded
by a king. It offered the possibility of strengthening his power, but might
also constitute a threat to his legitimacy. As in Kashmir or Burma, a new
ruler might justify his kingship in terms of the inappropriate use of Tantric
power by his predecessors. Those states that maintained a transgressive
Tantric element to their ritual life often tended to keep it relatively undercover, as with the ‘secret’ Tantric meanings of the devadāsı̄ rituals of Puri
(Marglin 1985b: 217–42).
In Chapter 13 I will discuss developments in Śaiva and Buddhist Tantra,
at the end of our period, the tenth to twelfth centuries, and will also sketch
the transformations undergone by these practices in subsequent centuries
and in recent times.
chap t e r 13
The later history of yoga and Tantra
Chapter 12 emphasised the ways in which the transgressive Tantra of the
fierce goddesses, criminal gods and destructive magic, in both Brahmanical
and Buddhist variants, was a part of state policy, adopted and supported
by Indic and Indic-influenced states both in South Asia and overseas. In
the Brahmanical states, as Gupta and Gombrich have noted, royal power
and Tantric theology became closely intertwined, so much so that, as they
note ‘one cannot be adequately understood without the other’ (Gupta and
Gombrich 1986). The Buddhists provided similar services to the East Asian
states, while in Southeast Asia we find elements from both Buddhist and
Brahmanical sides.
It seems scarcely deniable that these pragmatic uses of Tantra provided
a major context for the growth and development of Tantric religion. We
can see early signs of this accommodation of state needs in Buddhist texts
by the start of the fifth century, with the proto-Tantric man.d.ala and the
state-supporting rituals of the Suvarn.aprabhāsa Sūtra and the Renwang
Jing. We can see the development fully under way by the eighth and ninth
century, with Buddhist Tantric rituals for state purposes being translated
into Chinese and transmitted on to Japan, court patronage of Śaiva Tantra
in Kashmir, and Śaivite rituals being used for state consecrations and stateprotective rituals by the Khmer kings. Doubtless there was plenty of smallerscale employment of Tantric religion, at any rate by those wealthy enough
to employ Tantric ritualists.
One can imagine a situation similar to that in Bali in recent times, where
Śaivite and Buddhist Brahmins officiate at important temple ceremonies
and for high-status families, and other priests, often using rituals derivative
from Śaiva or Buddhist Tantra, perform rituals in less important contexts
or for lower status families (e.g. Hooykaas 1973a, 1973b). For that matter,
one can see a relatively similar picture in Burma or Thailand, where the
label ‘Tantra’ has disappeared, but pragmatic ritual procedures similar to
those of Tantra are still part of religious life at all levels (e.g. Spiro 1967, 1971;
324
The later history of yoga and Tantra
325
Mendelson 1961a, 1961b. 1963; Ferguson and Mendelson 1981 for Burma;
Tambiah 1984; Terwiel 1975; Mulder 1992: 15–41 for Thailand).1 In India
itself, while much of the more dramatic and transgressive varieties of Tantric
ritual, particularly at the state level, has disappeared along with the spread
of bhakti religion and dramatic political transformations since the time I
have been speaking of, it is not hard to find traces of pragmatic, resultsoriented Tantric ritual in all kinds of context (e.g. Diehl 1956 for South
India; McDaniel 2004 for Bengal).
All this is a very different picture of Tantra from that with which we are
familiar today, particularly as taught by contemporary Tibetan lamas, but
equally within the Śaivite ascetic traditions of India. Tantra, as we tend to
experience it with contemporary Tantric teachers, is primarily about the
individual quest for self-realisation: the performance of public ritual for
rulers or lesser clients is a distraction from the true nature of the Tantric
quest.
We can see the move to this view of Tantra as primarily a quest for
individual liberation or self-realisation prefigured in the Śaivite kaula and
the Buddhist mahāyoga traditions, with their progressive internalisation of
ritual. There is undoubtedly still a concern here with state patronage, and
in fact many of the examples of such patronage cited in Chapter 11 referred
to Tantric traditions with ‘internalised’ ritual components (e.g. the Netra
Tantra and the Vı̄n.āśikha Tantra). The progressive phasing out of the more
‘transgressive’ elements, such as the cremation-ground practices, prepared
the ground, however, for a move to a Tantric cult that could be presented
as primarily soteriological and which could be carried out on an individual
basis by both lay and monastic practitioners. The ‘privatisation of Tantra’
in the title of this chapter refers to these developments.
Thus, on both the Śaiva side and the Buddhist side, we can see by the
tenth and eleventh centuries a recasting of Tantric ideas in forms that are
much less open to negative stereotyping and where the more transgressive
elements are removed or practised only in secret. On the Śaiva side, this
has been associated above all with the activity of the great Kashmiri Tantric
master Abhinavagupta and his students and followers, and the Trika system
of which Abhinavagupta was the most eminent exponent (Silburn 1988;
Sanderson 1988, 1995; Dupuche 2003; Dyczkowski 1989; White 1998). On
the Buddhist side, it can be linked with the gradual appropriation of Tantra
by practitioners who were based or at least trained in a monastic milieu,
and who were concerned to incorporate or justify Tantric procedures much
1
See also Cousins 1997 on esoteric traditions within Southern (‘Theravāda’) Buddhism.
326
Origins of yoga and Tantra
more explicitly within the general context of Mahāyāna Buddhism. I begin
with the Śaiva developments.
the transf ormation of śaiva tantra
David White, following in part from Sanderson’s work, periodises Śaiva
Tantra into three phases. The first of these is kula (‘clan’) practice:
‘cremation-ground practice centred on the “terrible” worship of ŚivaBhairava together with his consort, the Goddess . . . and clans of Yoginı̄s
and/or the worship of the goddess Kālı̄, independent of a male consort but
surrounded by circles of female deities’ (White 1998: 173). This is where the
kāpālika-style practice and the magical uses of sexual substances are first
found.
Phase Two, the kaula (‘clan-derived’) phase, resulted from a reform
in about the ninth century, associated with the semi-legendary figure of
Matsyendranāth. Here we find more emphasis on the erotic element and
less on the cremation-ground location, the skulls and bones and so forth,
and a variety of cults associated with sexual initiation of the kind discussed
in earlier chapters. These kaula practices formed the basis of the Nāth
Siddha tradition (cf. Briggs 1989).
Phase Three results from a further reform which took place in Kashmir
in the eleventh century, leading to the Trika and similar systems. The
theoreticians of this school (above all Abhinavagupta) developed a sophisticated philosophical basis for kaula practice, based on reversing the process
of divine manifestation through light and sound so that the practitioner
could identify with its origins.
Here, in the socioreligious context of eleventh-century Kashmir, these reformers
of the Trika sought to win the hearts and minds of the populace by presenting a
whitewashed version of kaula theory and practice for public consumption, while
continuing to observe the authentic kaula practices in secret, among the initiated.
This eleventh-century development was Tantra, a religion of dissimulation and
of the progressive refinement of antinomian practice into a gnoseological system
grounded in the aesthetics of vision and audition. (White 1998: 173)
This was the basis of the Śrı̄vidyā system, which was exported soon after to
South India where it has remained an important tradition to modern times
(Brooks 1990, 1992a, 1992b). White prefers to reserve the label ‘Tantra’ to
this last phase, which he also refers to as ‘high Hindu Tantra’.
While this periodisation is a good starting point in general terms, there
are some problems with it. In particular, it should be noted that the kula
The later history of yoga and Tantra
327
period (Phase One) is largely conjectural; we have no direct textual material
regarding this period, although there are certainly plausible reasons to associate it with the widespread kāpālika trope in Sanskrit literature of the time.
Since the kaula phase (Phase Two) is already in part ‘reformed’, the question of whether the kula phase existed as a historical reality, and of what it
might have consisted, remains somewhat open. On the whole, I am sympathetic to White’s position here, given the early Tamil evidence for low-caste
ritualists and the indications at Udayagiri and in textual sources of early
destructive magical rituals on behalf of the state. Large gaps nevertheless
remain in our account.
On White’s analysis, the primary functions of sexual practice in Phase
One were to do with initiation and with generating magically potent substances. In the context of initiation, transgressive sexual practice served to
connect the practitioner with the origin of the lineage through the transmission of sexual fluids from male guru to student in conjunction with the
sexual secretions of a female partner. Offerings of sexual and other polluted
substances also served to gratify the fierce female deities and secure their
assistance. The use of sexual and transgressive substances was an ongoing
part of the ritual practice – the Kulacūd.āman.i tantra for example prescribes
that the basic yantra for the ritual should be drawn with a mixture of menstrual blood, sexual substances and various red and yellow pigments – but
Phase One was not concerned with the spiritualisation of erotic experience.2
The key element was transgression and impurity as generating power.
The significant innovation in the kaula phase (Phase Two) was the
internalisation of the practice through sexual yoga, perhaps, as I noted
in Chapter 11, through contact with similar Chinese practices. This would
have radically transformed the meanings associated with the whole system,
as it did in the case of the Buddhist Tantras as well. However, the idea
of transgression and impurity as generating power remained important.
It continued to be so even in the work of Abhinavagupta, the key theorist of White’s Phase Three. As Alexis Sanderson has put it, referring to
Abhinavagupta, ‘the Tantric’s sense of power was inseparable from his sense
of transgression’ (Sanderson 1985: 211 n. 61).
However, the incorporation of the internal sexual yogas in the kaula
practices meant that the transmutation of sexual experience was an increasingly central part of the process. This was particularly so as the third phase
involved the progressive internalisation of the Tantric ritual scenario. The
2
Nor, I think, was Tantra ever centrally about popular protest against religious orthodoxy, although
this can be found as a secondary theme in some Tantric writings, and one or two modern scholars
have been tempted to appropriate the Tantric siddhas for this purpose (e.g. Meenakshi 1996).
328
Origins of yoga and Tantra
deities and the ritual locations were increasingly seen as entities and locations within the practitioner’s body. With Kālı̄ and Bhairava now seen less
as cosmic powers that could be evoked and used for ritual purposes, and
more as symbols of the practitioner’s internal transcendence of the limits of
ordinary consciousness, the role of sexuality and transgression within the
system could also transform. Thus, Sanderson writes that ‘The act of sexual
union which produced the quintessences necessary to invocation, became
in this inner idealism the vehicle of illumination itself; for it was in orgasm
that the deity revealed itself as the transcendental core of the energies of
cognition and action, the unity of light and emptiness’ (Sanderson 1985:
202). This, Sanderson says, enables the tradition to ‘colonise the mental life
of the Kashmirian householder’ and indeed to deeply penetrate the court
and the intelligentsia (1985: 203).
buddhist and j aina developments
The Buddhist context was rather different, since a key issue here was the
relationship to the monastic context, although some level of adoption of
Tantric practices among court circles and wealthy householders in Pālaperiod Bengal is likely; late siddhas like Nāropā clearly built up a substantial following, and the local rāja felt it worthwhile calling to pay his
respects (Davidson 2002a: 317). We have little direct information regarding
developments among the laity, however, to put alongside such sources as the
satires of Ks.emendra (Sanderson 1985: 203; Wojtilla 1984, 1990; Meyer 1903)
and the history of Kalhan.a (Stein 1900) for lay Śaivite practice in Kashmir.
We have plenty of Buddhist ritual texts, and quite a lot of stories about the
various Tantric masters, but these are for the most part legendary rather
than historical.
However, it is clear enough that a similar transformation can be traced
on the Buddhist side, and at around the same time. If Matsyendranāth
and his close associates, such as Goraknāth and Cauraṅgināth, are the
semi-historical originators of kaula practice on the Śaiva side, other figures
among the so-called mahāsiddhas, or ‘Great Perfected Ones’, such as Sarāha,
Kr.s.n.ācārya (Kān.ha) and Nāgārjuna, played a similar role on the Buddhist
side (see Linrothe 2006). Our accounts of these figures, dating from the
late eleventh century or later, are, as noted above, more story than history,
but they give a picture of the context within which the sexual practices and
the fierce deities were progressively incorporated into Buddhist practice
(J. Robinson 1979; Dowman 1985; Yuthok 1990). These Tantric masters
are described as primarily concerned with the Guhyasamāja, Cakrasam
. vara
The later history of yoga and Tantra
329
and Hevajra Tantra cycles, in other words with the mahāyoga and yoginı̄
(anuttarayoga) phase of Buddhist Tantra.
The material in Chapter 12 suggests that the critical transformation,
corresponding to White’s Phase Two,3 took place with the Guhyasamāja and
other Mahayoga practitioners, and that it was already taken for granted in
the Cakrasam
. vara and Hevajra systems which dominate Abhayadatta’s lateeleventh or early-twelfth century hagiographic sketches of the eighty-four
Mahāsiddhas (J. Robinson 1979: 285–8).4
A Buddhist equivalent to White’s third Śaivite phase can be seen as
commencing with some of the latest figures among the mahāsiddhas, figures
such as Nāropā who lived in the late tenth and early eleventh century. It
is fully developed in the Kālacakra Tantra, the last major Indian Buddhist
Tantra, also from the early eleventh century, to which I shall turn shortly.
By the late eleventh century and twelfth century we find a variety of figures
such as Abhayākaragupta (Lee 2003) who are producing scholarly treatises
and commentaries on the basis of the new Tantras, as well as indications
of their adoption in the monastic context. As I have noted elsewhere,
the surviving icons are mostly small, suggesting that we are dealing with
individuals within the monastery who are specialists in the new practices,
rather than full-scale adoption as the major ritual practices of the monastery
as a whole (Samuel 1993: 412). We have textual material from this period,
however, which provides rituals based on mahāyoga-type procedures for
all kinds of major rituals, including the construction and consecration of
monasteries and stūpas (e.g. the Kriyāsam
. graha; Skorupski 1998).
It is primarily in Tibet, however, that we see the development of synthetic
schemes in which the new Tantric procedures are included as part of a
progressive scheme of teachings, commencing with the sūtra teachings.
The Bengali teacher Atı̄śa, who came to Tibet in the early eleventh century,
wrote a short text, the Bodhipathapradı̄pa, of this kind (Sherburne 1983;
Snellgrove 1987: 481–4) and this became the basis of the lam rim or ‘Gradual
Path’ teachings of the dGe-lugs-pa school in Tibet (Wayman 1978). In the
developed Tibetan version of these teachings, Vajrayāna, particularly as
presented in the anuttarayoga texts, was accepted as the central means to
the attainment of Buddhist Enlightenment, but its full practice was seen
3
4
Linrothe’s three-phase periodisation of Buddhist Tantra (or rather Esoteric Buddhism), which I
discussed in Chapter 10, is based on different principles (the evolution of the krodha-vighnāntaka
theme) so does not correspond to White’s three phases. The Guhyasamāja is an early text within
Linrothe’s third phase (1999: 226).
For the forty-one cases where Abhayadatta identifies the Tantric cycle involved, eighteen refer to
Cakrasam
. vara, sixteen refer to Hevajra, and seven to Guhyasamāja ( J. Robinson 1979: 285–8).
330
Origins of yoga and Tantra
as suitable only for an elite group of highly-qualified initiates.5 Part of
the problem here was how to reconcile a set of teachings in which sexual
practices played a central role with a context in which practitioners were
increasingly expected to be celibate monks. In the long term, the Tibetan
version of Vajrayana resolved this tension by reducing the significance of the
sexual practices. While these practices were never entirely eliminated, they
have a marginal position within contemporary Tibetan Buddhism (Samuel
1993: 206–7, 240–1, 275–8).
the k āl acakra tantra
The most complete expression of the ‘third phase’ on the Buddhist side
was the last major Buddhist Tantra, the Kālacakra Tantra. This is dateable
to the early eleventh century (Wallace 2001: 3; Newman 1998). The basic
structure of the Kālacakra practices does not differ dramatically from the
other anuttarayoga Tantras, though there are variations in matters of Tantric
physiology; the Kālacakra’s system of nād.ı̄s and cakras, internal channels
and psychic centres, is somewhat different from those of the earlier Tantras.
There is also a significant elaboration of the early and introductory stages,
with an extensive series of preliminary empowerments (e.g. Sopa 1985:
95–6). The main text of the Kālacakra Tantra, which is held to have been
compiled by Mañjuśrı̄ Yaśas, the eighth king of Śambhala, presents itself
as an abridged version of a full Tantra. The full Tantra is said to have
been originally revealed in India, at the Dhānyakat.aka stūpa at Amarāvatı̄
in Andhra Pradesh but preserved in Śambhala, a probably mythical land
to the West where it was transmitted onwards through the succession of
kings of Śambhala (Bernbaum 1980; Sopa, Jackson and Newman 1985).
This abridged Kālacakra Tantra text, along with the early Vimalaprabhā
commentary attributed to a later Śambhala king, Pun.d.arı̄ka, places the
Tantric practices in a different and new kind of social and intellectual
context.6 While the root texts of Tantras such as the Hevajra are primarily
ritual handbooks, and make little attempt to situate their teachings in a
wider context, the Kālacakra Tantra is notable for its broad and synthetic
5
6
The lam ’bras or ‘Path and Result’ cycle of teachings of the Sa-skya school (cf. Ngorchen 1987;
Yuthok 1990, 1997) formed another such scheme. The Indian root text attributed to the mahāsiddha
Virūpa, however, on which the lam ’bras is based, is almost entirely Tantric in content. The extensive
non-Tantric teachings which are now presented as part of the lam bras are based on three lines of the
original text (cf. Davidson 2005: 477 and Ngorchen 1987).
The commentary and the text seem to have appeared together in India and may well have been
composed together, despite being attributed to different (presumably both legendary or mythical)
figures.
The later history of yoga and Tantra
331
approach. It is not just a system for the advanced adept involved in Tantric
practice, but provides a virtual encyclopaedia of knowledge on a wide range
of topics.
Thus the first of the Kālacakra’s five sections, the ‘Outer Kālacakra’
or description of the world-system, includes an extensive treatment of
cosmological and astronomical topics (Newman 1987), while the second,
the ‘Inner Kalacakra’ or description of the individual, includes a detailed
description of embryology, an analysis of the person in terms of abhidharma categories, and a very elaborate account of the cakras and channels
of the body. All of this is related back to the Kālacakra man.d.ala, which
with its 722 deities is perhaps the most complex of all Tibetan man.d.alas
(Wallace 1995, 2001, 2004). The ‘Inner Kālacakra’ chapter also includes
practices to lengthen life, and alchemical practices, along with an extensive critique of other systems of thought, Buddhist and non-Buddhist.
The third and fourth chapters cover the more conventional topics of initiation and generation-stage processes respectively, while the fifth includes the
completion-stage processes and the path to Buddhahood (Hartzell 1997).
Vesna Wallace, in a recent book on the second section of the Kālacakra
Tantra, suggests that the author of the Kālacakra was openly aiming at
the conversion of heterodox groups, Buddhist and non-Buddhist. The text
refers extensively to non-Buddhist teachings, although it also disparages
and criticises them (Wallace 2001: 38–9). At the same time, features of the
system seem to have been deliberately designed to make it easier for nonBuddhists to approach Kālacakra practice (2001: 40–1). Wallace speculates
that the author of the Kālacakra Tantra ‘may well have sought to form a
united front of heterodox groups and Buddhists’ against the common foe
of the Islamic invasions (2001: 42; see also Orofino 1997).
Another notable feature of the Kālacakra regards its attitude to the sexual practices, which have a much more central and explicit place here
than in previous systems. The Vimalaprabhā comments that in previous
Tantras, including the Guhyasamāja Tantra and the Cakrasam
. vara Tantra,
the Buddha ‘taught the blissful state that arises from sexual union, but concealed it out of his great compassion for the sake of the spiritual maturation
of simple-minded people’. In the Kālacakra, the use of sexual practices as
part of the spiritual path is quite explicit. Thus Wallace notes (2001: 11–12)
that
The generation of sexual bliss without emission of regenerative fluids is regarded in
this Tantra as the most direct method of generating the mental bliss that refines the
mind by diminishing conceptualisations and thus makes it fit for the realization of
332
Origins of yoga and Tantra
the empty nature of phenomena. One who practices the generation of sexual bliss
without emission, which is referred to as sublime, imperishable bliss, is considered
to be like a young virgin. Such bliss is believed to empower one’s mind, just as
the mind of a young virgin, who has not experienced sexual bliss with emission,
can be empowered by deities and mantras that enable her to see appearances in a
prognostic mirror.
At the same time, the author of the Kālacakra was concerned to minimise the
conflict that might be caused by antinomian and transgressive practices with
the wider society. Thus ‘[t]he Tantric beginner is advised to eat, drink, and
have sexual relations according to the customs of his own country’ (Wallace
2001: 121) until he attains a suitably high level of realisation. The use of
impure substances is not to be prescribed to beginners, but only to yogı̄s
who ‘by the power of their mantras and meditative concentration are able
to transmute these poisons into ambrosias (amr.ta)’ (Wallace 2001: 121–2).
Here one can see the move to a ‘religion of dissimulation’ similar to that
described by Sanderson and White for the third phase of Śaiva Tantra. If the
watchword of ‘high Hindu Tantra’ was ‘privately a Śākta, outwardly a Śaiva,
among people a Vais.n.ava’ (White 1998: 173, quoting the Yon.i Tantra), the
Buddhist equivalent was ‘outwardly Hinayāna, inwardly Mahāyāna, secretly
Vajrayāna’.7
ja i na practices in the tenth and eleventh centuries
We have less information on Jaina developments during this period. Briggs
mentions two Nath subgroups who are named for two ‘sons’ of Matsyendranāth, Nı̄mnāth and Pārasnāth, in his classic account of the Nāths (Briggs
[1938] 1989: 72). These are the names of two of the Jaina tı̄rthaṅkaras, Nemi
and Pārśva. It is not clear from Briggs’ account whether the members of
these subgroups were actually Jains.8 There were certainly Jains involved in
alchemy, however (White 1996: 114, 372 n.16), and several important Jain
pilgrimage sites, including Girnar and Mount Abu, also have strong Nāth
Siddha connections (White 1996: 114–20, 331–4). It seems likely that there
were close relationships between Jain ascetics and Nāth Siddhas at these
locations.
One of the most famous of all Jaina writers, however, the Śvetām
. bara
scholar and ascetic Hemacandra (c. 1089–1172) gives extensive material on
Tantric-type practices in his Yogaśāstra (Qvarnström 2000, 2002, 2003).
7
8
A traditional Tibetan saying which I first heard from the late Geshe Ngawang Dhargyey (personal
communication, 1972).
White assumes so at 1996: 119, but Briggs may only be saying that Nı̄mnāth and Pārasnāth were Jains.
The later history of yoga and Tantra
333
The Yogaśāstra is a kind of general handbook on Śvetāmbara Jainism and
it was mostly known in later centuries for its first four chapters, which
present a detailed account of the Jaina path. The remaining seven chapters
deal with meditational and yogic practices. Chapters 5 and 6 cover such
topics as breathing, including the internal circulation and control of prān.a
and the prediction of death from breathing and other signs, as well as a
technique for taking over the body of another being (Chapter 6, 264–72).
Chapters 8 to 10 include various forms of meditation (dhyāna) including
elaborate internal cakra meditations closely paralleling those in the Śaiva
Tantra traditions (e.g. Chapter 8, 1–28, 47–56) and including techniques
to bring about the four kinds of ritual action (Chapter 8, 29–32), recollections of the qualities of the jina (Chapter 9, 1–15; Chapter 10, 1–6), and
so on. Hemacandra is clearly concerned to justify the validity of the practices he describes, none of which strictly speaking would be forbidden by
monastic vows. They are described as useful for converting and otherwise
aiding other beings, as reducing attachment (e.g. Chapter 9, 13; Chapter 10,
18–24) and also, in some cases, as leading to liberation (e.g. Chapter 8, 40,
43, 59, 78).
Chapter 11 covers the canonical topic of ‘pure meditation’ (śukladhyāna)
which as I noted earlier (in Chapter 9 of this book) was regarded as
beyond the capacities of ordinary human beings, but Chapter 12 provides a
somewhat abstract account of Hemacandra’s own meditational experiences
which according to Qvarnström shows clear signs of Nāth Siddha influence
(e.g. 2002: 11, 191 n. 4; 2003).
The cautious tone in which Hemacandra discusses these practices doubtless reflects his own status as a celibate monastic and perhaps also the
Yogaśāstra’s role as an official presentation of the Jaina teachings for the
Caulukya king of Gujarat, Kumārapāla (Qvarnström 2002: 2–5). It is striking that Hemacandra is nevertheless familiar with, and feels it appropriate
to expound, such a wide range of kaula and Nāth practices, suggesting that
these practices were widespread among the Jaina community in his time.
It is noticeable that he does not discuss any of the ‘Tantric’ vidyādevı̄ practices that appear to have been current in both Śvetām
. bara and Digam
. bara
circles at around the same time (Cort 1987). These were perhaps more specialist practices and less appropriate for inclusion in a general text of this
kind. Many of the practices given by Hemacandra are relatively simple and
straightforward and might be seen as suitable for a lay practitioner, though
some, such as the practices for taking over the body of another person,
or the various practices to predict the time of death, are hardly of this
type.
334
Origins of yoga and Tantra
l ay and householder practice
The whole question of lay practice at this stage is worth considering in
some more detail. The sexual practices clearly create problems for practice
by celibate monastics, whether Buddhist or Jaina, though these might be
justified in some way as not in fact breaking the rules,9 or by-passed through
symbolic performance. This aside, much of the clientele envisaged in the
Tantric texts or referred to in Sanskrit literature of the time appears to
consist of professional practitioners, whether from hereditary or renunciate
ascetic backgrounds. The state sorcery rituals at Udayagiri would surely
have been performed by professionals who were hired for the job, and the
various ritual performers at the Kashmiri, Khmer and East Asian courts
were clearly also full-time practitioners. Indeed, the sheer complexity of
much Tantric ritual would make it difficult to undertake without at least
one or more highly-trained experts. Tibetans such as Mar-pa who came
south across the Himalayas to acquire Tantric teachings also appear to have
taken it for granted that they were going back home afterwards to start a
family business as ritual practitioners (or, perhaps, to revive one with the
aid of some new techniques) (Samuel 1993: 459–60).
However, Tantric priests seem ready enough to initiate kings and members of royal courts into Tantric practice; apart from the examples of King
Hars.a of Kashmir and the legendary King Indrabhūti of Zahor, which have
been cited already, one could cite as distinguished a figure as Khubilai Khan,
who received the Kālacakra empowerment from O-rgyan-pa and Hevajra
empowerments from the Sa-skya hierarch ’Phags-pa (Shakabpa 1967: 64–71;
Roerich 1976: 701). Twenty-five of his ministers are also said to have been
initiated into the Hevajra cult.
In any case, the ‘second’ and particularly the ‘third phase’ Śaivite Tantric
teachings, and the new ‘internalised’ Buddhist practices both seem to be
oriented at least in part to the needs of lay (non-professional) followers. The
laity here are seen not just as clients for ritual services but as participants
within the ritual context. We could see this in instrumental terms, as part
of a move into a new market area, and there is doubtless some truth in this.
It would be a mistake, however, to reduce this to a purely instrumental
business. This would imply that Śaivite and Buddhist ritualists were living
in the same demythologised universe as most contemporary Western academics. This is unlikely; it makes much more sense to see Abhinavagupta’s
9
As in Atı̄śa’s argument that the practices could be undertaken by celibates provided that they had a
true understanding of śūnyatā (Samuel 1993: 471).
The later history of yoga and Tantra
335
circle or the propagators of the Kālacakra as people who genuinely felt that
they possessed something of great value that they wished to make available
to others. We might also wonder how far the Kālacakra mythos fed into a
general consciousness at the time of a society under threat from external
forces.
tan tric practice in l ater times
It is unclear how widespread Tantric practice became at the householder
level throughout Indian society in the twelfth and thirteenth centuries.
Today it exists in a number of places, including the Śrı̄vidyā practitioners
of South India (Brooks 1992a, 1992b), the householder Nāths of Rajasthan
(D. Gold and A. Gold 1984; A. Gold 1993), and the so-called Bāuls and
other lay Tantric practitioners of Bengal (McDaniel 1989, 2004; Openshaw
1997, 2002; R. Das 1992; Hanssen 2002).
Tantric or closely related practices also continued within the similar Sufi
context, where traditions of dissimulation and transgressive behaviour had
long had a role. In fact, contacts between Sufi or earlier Islamic mystical
traditions and the Śaiva and Buddhist yogic contexts long pre-dated the
Islamic takeover of most of North India in the twelfth and thirteenth
centuries. The Bengali Bāuls present their teachings and practices in both
Vais.n.ava and Islamic forms, and are able to translate freely between the two
(Salomon 1991). Similarly fluid identities can be found in Western India
and the Deccan, especially among the Ismailis, though increasingly under
threat from modern ‘purist’ religious forms on both the Hindu and Muslim
sides (Khan 2005; Van Skyhawk 1993).
New devotional styles of Brahmanical Hinduism were, however, to take
over as the dominant religious style of much of India over the succeeding centuries, including the urban upper and middle classes which would
have constituted the natural target group for the Phase Three Tantric practices. These new devotional styles of religion, with their emphasis on emotional submission to a supreme saviour-deity, whether Śaivite or Vais.n.avite,
were better adapted, perhaps, to the subaltern role of non-Muslim groups
under Muslim rule (Hein 1982). It is difficult to know how widespread
Tantric practice remained among the general population before the modernist movements of the nineteenth and twentieth centuries. White cites
reports from Bihar in the early 1800s suggesting that over 40 per cent of
the population were either Tantric or Śakta but his assumptions are optimistic and East India (Bihar, Bengal, Orissa, Assam) is an area where one
336
Origins of yoga and Tantra
might perhaps expect higher levels of support for Tantra than elsewhere in
South Asia.10
Ascetic practice of Tantra among renunciates was more widespread;
in fact most Śaivite ascetics appear to have adopted some version of Tantric
practice, both as their personal sādhana or spiritual practice and as the basis
for the ritual services they performed for their lay clientele. This is still the
case today, although Vais.n.ava ascetics now form a significant part of the
Indian renunciate scene and are less likely to be primarily Tantric in their
practice.
At a more exalted social level, Tantric ritualists continued to play a
significant role at some Hindu courts. We can see them, for example, in
Frédérique Marglin’s excellent ethnography of the temple dancers at the
Jagannāth Temple at Puri, which was closely associated with the Kings of
Orissa until modern times (Marglin 1985b).
The development of modernist forms of Hinduism in the nineteenth
and early twentieth centuries (e.g. Flood 1996: 250–73; King 1999) led to
a revival of yoga but also a radical reframing of yogic practices away from
the Tantric context. As Alter has noted, hat.ha yoga, the basis of modern
forms of health, medical and recreational yoga, originated as an adjunct
to Nāth Tantric practice (Alter 2005). The principal hat.ha yoga text, the
Hat.hayogapradı̄pikā, is an explicitly Nāth text directed towards the classic
Nāth internal yogic and sexual practices. Given the extremely negative
views of Tantra and its sexual and magical practices which prevailed in
middle-class India in the late nineteenth and twentieth centuries, and still
largely prevail today, this was an embarrassing heritage. Much effort was
given by people such as Swami Vivekananda into reconstructing yoga,
generally in terms of a selective Vedantic reading of Patañjali’s Yogasūtra
(de Michelis 2004). The effort was largely successful, and many modern
Western practitioners of yoga for health and relaxation have little or no
knowledge of its original function as a preparation for the internal sexual
practices of the Nāth tradition.
10
See White 2003: 5. Specifically, White assumes that when Francis Buchanan’s pandit informants
reported that ‘one-fourth of the population’s religion was “unworthy of the note of any sage”’ this
implied that ‘they consisted of cults of (predominantly female) village deities whose worship was
often conducted by the socially and culturally marginalized, in other words, Tantric cults’ (2003:5).
This sounds like a very broad definition of Tantra. To this he adds the one-quarter of the remaining
three-quarters (i.e. 18.75 per cent) who are described as Śakta, but, even leaving aside the rather
approximate numbers, Śakta covers a variety of folk religious forms (cf. McDaniel 2004), and
assimilating them into a general Śakta-Tantra category may be misleading. What White’s figures do
however demonstrate is that ‘orthodox’ Śaiva and Vais.n.ava cults were far from dominating village
religious life. My limited impressions of village religion from the nearby Maithili region of Nepal
in the 1990s suggests that this is still the case (Samuel 2001a).
The later history of yoga and Tantra
337
Within Buddhism, Tantric practice survived in a number of somewhat marginal contexts: as an elite tradition of hereditary Tantric priests
in the Kathmandu Valley (Gellner 1992; Stablein 1991) and in Bali, where
the remains of the Tantric-oriented court culture of Java regrouped after
the Muslim takeover of Java (Hooykaas 1973a, 1973b); as one of a variety
of ritual traditions in Japan (Yamasaki 1988); and in a transformed version
among the Tibetans and Mongolians, a story the Tibetan part of which I
have told elsewhere at considerable length (Samuel 1993).
In the eleventh to fourteenth centuries, there was still a significant Tantric
presence throughout much of Southeast Asia. In some ways, arguably, there
still is, in that sophisticated traditions of quasi-Tantric practice of magical
rituals have remained important in Thailand, Burma and Cambodia into
modern times. Stanley Tambiah has argued that the Thai practitioners of
this kind represent a ‘Tantric’ modality within Thailand (Tambiah 1984),
and if one can use the term of the forest monks and other magical ritualists
of Thailand, the occult masters of Burma would certainly deserve the title
as well (Mendelson 1961a, 1961b, 1963; Ferguson and Mendelson 1981; Spiro
1967, 1971). These practitioners no longer call themselves Tantric, however,
and while some, at least, of the specific Tantric initiation and teaching
lineages associated with the major Buddhist Tantric cycles were still present
in Southeast Asia as late as the fourteenth century, they appear to have
vanished over the succeeding centuries.
In part, this was a product of a consistent theme within the puristic
Theravāda-style Buddhism which became the norm in Southeast Asia and
Sri Lanka: the need for a regular ‘purification’ of a corrupt saṅgha by
Buddhist kings and rulers. One can see this as an ideological statement
of the morality and strength of new rulers, often associated with a blackening of the reputation of their predecessors as dabblers in dubious occult
arts and immoral practices. One can also see it, as Tambiah has noted, as an
assertion of state control over the aspects of Buddhism that are most likely
to lead to autonomous centres of power and resistance to royal authority.
Either way, the periodic purifications of the Buddhist community left little
space for the elaborate ritual structures and practices of the major Tantric
lineages (Tambiah 1984; cf. Samuel 1993). Magical ritual for pragmatic ends
never disappeared from Southeast Asia, and it remained linked to dhyānastyle meditation practices (Pali samatha = Skt. śamatha), associated in particular with the ascetic forest-monk tradition (Tambiah 1984; Tiyavanich
1997, 2004; Carrithers 1983), but the formal Tantric teaching lineages,
with their explicit connection to Buddhist soteriological ends, gradually
vanished.
338
Origins of yoga and Tantra
The modernist forms of Theravāda Buddhism that developed from the
late nineteenth century onwards (e.g. Gombrich and Obeyesekere 1990;
King 1999) tend to present Buddhism as a rational, scientific tradition and to
play down religious and magical aspects. Samatha meditation, which retains
an association with the attainment of magical power, has been generally
reduced to a brief introduction to vipassanā or ‘insight meditation’, with
a focus on awareness of breathing and other bodily sensation rather than
the systematic reshaping and transformation of experience involved in the
older techniques (Cousins 1984, 1996b). However, the increasing global
presence of Tibetan lamas in the aftermath of the takeover of Tibet by
the People’s Republic of China has meant that these techniques are now
competing in the global marketplace against the much more elaborate and
largely unreconstructed Tibetan practices, which have gained for at least
one form of Tantra a global following its original creators could hardly have
envisaged.
chap t e r 14
Postlude
In this final chapter, I reflect back on some of the material in the book as a
whole. I begin with the whole question of meditation, yoga and mind-body
processes, which has been central to this book. What do we make today
of these techniques, which are, as I noted in my introduction, rapidly
becoming a significant part of contemporary global society, if often in
much modified and adapted versions? Such techniques are of course not
unknown in other societies, and can generally be classed, in Foucault’s
phrase, as ‘technologies of the self’. Foucault introduced this term to refer
to methods by which human beings act upon their minds and/or bodies
(perhaps we should speak more generally of the mind-body complex) with
the intention of bringing about transformations of some kind (Foucault
1988a, 1988b; cf. Samuel 2005a: 335–7). Indic societies appear to have been
particularly rich in these techniques, at least as compared to other large-scale
literate cultures, and to have seen them as bound up with the acquisition
of some kind of liberating insight. I begin then by asking what we might
make of the historical evolution of meditation, yoga and Tantra within
Indic societies and religions.
If we look at the evolution of these techniques historically, we can see an
overall development from simple to more complex approaches. The early
(pre-Buddhist) śraman.a traditions seem to have had a concept of liberating
insight, but no systematic means for achieving it via training of the mindbody complex. Instead they relied on simple ascetic practices aimed at the
forceful stopping of physical and mental activities. A critical step forward
was that taken by, or at least attributed to, the Buddha, the systematic
cultivation (samādhi) of a specific state or series of states (dhyāna) of the
mind-body complex within which the liberating insight might arise. The
Jainas may also have had and then lost an early version of these dhyāna
practices.
Early Vedic procedures were centred more on the visionary revelation
of sacred knowledge in the form of hymns and statements of sacred truth,
339
340
Origins of yoga and Tantra
which could then be used in ritual contexts. Initially, these were for the
most part perhaps chemically induced (assuming that soma is interpreted
as a psychedelic substance or entheogen of some kind) and appear to
have had affinities to ‘shamanic’ procedures in many other societies, using
‘shamanic’ in a wide sense.1 This tradition was lost, perhaps because of the
non-availability of the original soma, and the Brahmanical ascetic tradition gradually developed a version of the Buddhist practices, as seen in the
Yogasūtra.
Another central emphasis of the Vedic-Brahmanical tradition was on
a right death and a proper transition to a heavenly afterlife; this was an
important issue within the Upanis.ads, and, as we have seen in Chapter 9,
was one of the primary contexts of the term yoga in the Mahābhārata,
along with the idea of yoga as a technique for entering the body of another
human being (White 2006). These were both seen as skills to be developed
by warriors as well as by sages, a connection that perhaps makes more
plausible the linkage postulated by Bollée and Dundas between the vrātyas
and the early śraman.a traditions.
In the later Indian Buddhist tradition, as represented by the Mahayāna
sūtras, from around the second or first century BCE onwards, the dhyāna
techniques were combined with increasingly elaborate visionary procedures, based for the most part around the imaginative summoning of one
or another Buddha-form. These techniques, described in terms of entry
into various specific samādhi states, provided the basis for a cult of visionary communication with the Buddhas and perhaps also for the production
of the Mahāyāna Sūtras themselves (Chapter 9). Early Buddhist man.d.ala
structures such as those in the Suvarn.aprabhāsa Sūtra appear closely related
to these procedures, as do the early stages of what are known in the Buddhist tradition as kriyā, caryā and yoga Tantras, which were developing
by the sixth and seventh centuries CE. Later versions of these practices
began to incorporate the self-identification of the practitioner with the
main Buddha-form as a central feature, and a number of ferocious deities,
mostly male to start with, as outer figures in the man.d.ala (cf. Linrothe
1999).
1
Even Eliade, who was insistent on the distinction between shamanic ‘ecstasy’ and yogic ‘entasy’,
seems prepared to concede a shamanic origin for yoga: ‘As a developed spiritual technique (we are
not discussing its possible “origins”), Yoga cannot possibly be confused with shamanism or classed
among the techniques of ecstasy’ (Eliade 1958: 339). I should note in this connection that, given the
close links between yoga and Tantra, and the presence of both ecstatic procedures and concepts of
magical flight in the latter (e.g. White 2003: 188–218), as indeed the vision-quest theme that runs
throughout Buddhist practices, Eliade’s distinction is difficult to maintain. Rather than insisting on
the identity of shamanism(s) and yoga, however, I would like to see a more subtle and fine-grained
understanding of the whole range of practices involved and of their underlying mechanisms.
Postlude
341
Meanwhile, the transgressive techniques of ritual magic, White’s kula
practices, may have been related to similar techniques of visionary communication with deities. Here, however, the deities with whom communication was sought were the fierce goddesses of illness and misfortune, and
the main purpose was, initially at least, sorcery and destructive magic. The
early history of these practices is impossible to trace at present, though it
may have been associated with a specifically Śakta milieu and with the Śakta
pı̄t.has, locations where the power of the fierce goddesses could be contacted
and utilised. By the fourth or fifth centuries CE these practices were being
adopted by transgressive kāpālika-style Śaiva ascetics, in combination with
man.d.ala structures incorporating the fierce goddesses. In these contexts,
they came to be seen as constituting a pathway to liberating insight in their
own right.
While there are suggestions of (non-transgressive) sexual practices in
Mahāyāna Buddhist contexts as early as Asaṅga (fifth century), a new set
of techniques, closely related to and perhaps influenced by the Chinese
qi cultivation and ‘inner alchemy’ practices, started to spread throughout
South Asia in the seventh or early eighth century. The Indic versions of these
internal practices involve the movement of prān.a through the channels of
the body and are closely linked to the conscious control of bodily processes
during intercourse, and so to the practices of sexual yoga. While building
on some old themes, in particular the idea of the movement of prān.a
through the channels of the body, they added both new understandings of
the meaning of prān.a and a much more elaborate internal structure for the
body (or mind-body complex).
These new techniques allowed for an internalisation of the deity practices. They were adopted both by kāpālika-style Śaiva ascetics (the kaula
practitioners) and by Buddhist practitioners of mahāyoga Tantra, who were
beginning to incorporate increasingly transgressive elements within their
own practices. With the yoginı̄ Tantras, in the late ninth century or thereabouts, the fierce goddesses and the Śaivite male deity-forms through
which they were controlled were adopted by Buddhist lay practitioners as well, leading to the ritual cycles of the Hevajra and Cakrasam
. vara
Tantras.
In the eleventh and twelfth century, modified versions of these practices
were developed in both Śaiva and Buddhist traditions (Abhinavagupta,
Kālacakra Tantra), in which they became ‘esoteric’ practices for advanced
practitioners, carried on in conjunction with non-transgressive practices for
beginners and ordinary lay-people. The transgressive components were not
entirely eliminated, but were largely internalised, and the sexual practices
were increasingly now seen in terms of access to enlightened consciousness
342
Origins of yoga and Tantra
and liberating insight. This is White’s ‘high Hindu Tantra’ and the anuttarayoga phase of Vajrayāna Buddhism. Elements of these practices were
also adopted by some Jaina practitioners and there were also close contacts
with Sufis and other Islamic traditions.
This body of techniques was also applied to the old questions of how
to transfer consciousness at the time of death and how to enter the body
of another human being. The procedures for consciousness-transfer at the
time of death are still important in the modern Tibetan context (generically
known in Tibetan as ’pho ba).
I have attempted to trace as much as possible of the social and institutional basis for these developments in the course of the book. The transition
to Muslim rule over most of India dramatically changed this institutional
basis and led to the marginalisation of these practices throughout South
Asia; other factors had a similar effect in much of Southeast and East
Asia. Versions of these practices survived, however, in a variety of contexts
throughout South, Southeast and East Asia.
I think that it is clear from the wide range of material that I have discussed
in the book why it can often seem hard to get a clear line on just what yoga,
meditation and Tantra are about. The connections back to the vrātyas, to
low-caste ritualists and early Pāśupatas are speculative, if plausible, but there
is certainly a clear sequence of historical development that includes most of
the later Śaiva and Buddhist practices. However, the practices involved are
very varied – from aggressive military ritual carried out by Gupta kings and
Chinese emperors to householders belonging to secret cults in eleventhcentury Kashmir and contemporary Tibetan monastics – and the extent
to which people performing these practices necessarily saw themselves as
tantrikas, or used the term at all, also varied. One can understand why it is
not at all easy to define or specify precisely what Tantra is. Nor have I done
so here: I think it is more useful to trace the genealogical connections of
practices and ideas than to try to shoe-horn the variety of historical instances
into a comprehensive definition. Yet we can ask what general conclusions might be drawn from the account sketched in this and the preceding
chapters.
societ y, politics and value systems
In looking at the social and political grounding of religious practice. I
have tried as far as possible to make sense of why specific religious forms
were practised and why they were patronised. As I have pointed out at
intervals, this is not intended as a reductionist exercise, but as emphasising a
Postlude
343
central component of the understanding of Indic religious traditions which
often gets marginalised or left to odd asides and comments, particularly
in our more general accounts. I wanted to see how far one could follow
through some of these themes on a larger scale to give an overall view of
the development of Indic religions over a long period of time.
A related theme in the book has been the question of value systems
within Indic society. I started off with the opposition between Hopkins’
‘two worlds’, the region of Kuru-Pañcāla on the one hand, which was the
centre for the expansion of Vedic-Brahmanical culture, and the surrounding
belt of evolving cities and states to its east (Central Gangetic region) and
south (northern and western Deccan) on the other. Connections with other
regions (Northwest India and Central Asia, the Bengal Delta, South India,
Southeast Asia) were brought in as the story proceeded. It seems to me that
an initial tension between the values of the Vedic society of Kuru-Pañcāla
and those of the Central Gangetic region can be sensed through much of
the early development of Indic religions, and in various ways continues
into much later times.
Here one could characterise the Central Gangetic model as that of the
agrarian society with rulers who are expected to have elements of the
philosopher and the renunciate, as contrasted with the Kuru-Pañcāla model
of the warrior-chieftain of a culture many of whose values derive from a
pastoralist context. I emphasise here again, to avoid misunderstanding, that
this is not a simple question of indigenous people and ‘Āryan invaders’.
We are talking about a cultural contrast that developed between populations who were already mostly speakers of Indo-Aryan languages and who
were genetically doubtless quite mixed throughout both regions. There is,
nevertheless, a sense in which the early Vedic material represents one pole,
whatever its origins, and the Central Gangetic region another.
In terms of value systems, I have suggested in Chapter 8 that the ideal
of the Vedic material, like that of many pastoralist societies, is the young
male warrior, transformed in later Brahmanical reworking to the image of
the brahmacārin or celibate young man. One can think, for example, of
the later imagery of such religious heroes of the Hindu tradition as Śam
. kara,
Rāmānuja, or, in a somewhat different vein, Caitanya and his associates,
all of them characteristically depicted as young men in celibate garb. These
men still have much of the young warrior in them, even if their warriorship
takes place for the most part on the inner stage of the self, and it is not
altogether surprising that the idea of the Hindu warrior-ascetic remained
throughout Indian history (Bouillier 1993) and also surfaced in Buddhist
contexts, particularly in East Asia (Mohan 2001; Victoria 2005).
344
Origins of yoga and Tantra
The central ideal of the agrarian societies, by contrast, was that of the
mithuna or marital couple, endlessly replicated as a decorative theme on
Buddhist stūpa railings and Brahmanical temples. The virtuoso eroticism
of Khajuraho or Konarak can, I think, best be read in terms of the royal
appropriations of Tantra that we discussed in Chapter 12. The earlier versions, by contrast, I would read as cultural representations of good fortune,
fertility and prosperity.
The Upanis.adic, Buddhist and Jain material suggests that the tension
between the two patterns was resolved initially by a cultural pattern in
which the celibate young warrior ideal became adopted as the basis of
the renunciate traditions, in the form of the Buddhist and Jain monastic
orders and the Brahmanical renunciates. These ascetic orders also provided
a context for the continuance of the wisdom traditions associated with the
Eastern Gangetic states. They co-existed with an agrarian society whose
values were essentially this-worldly, but they provided a counterbalancing
emphasis which gradually came to pervade village society. In effect, this is
the pattern that continued into modern times in the Theravādin societies
of Southeast Asia and Sri Lanka. The Jaina tradition still shares many of the
same assumptions, though since Jains in modern times have mostly lived
in a Brahmanical Hindu context, as part of a society with a large majority
of Hindus, Jains have adapted in various ways to this context.
The reassertion of the Brahmanical tradition provided an alternative
pattern, doubtless strongest, as Lubin notes, in the rural context but also
of increasing appeal to the royal courts and urban centres of South and
Southeast Asia. The elite and urban forms of this tradition were associated
both with the various sophisticated priestly ritual technologies of which
Śaiva Tantra is one, and also with the devotional Hinduism of the developing bhakti movement. Within this tradition, the Brahmin as teacher and
ritual performer could guide the spiritual life of the householder. This cultural pattern, which is the basis of modern Hinduism, dominates South
Asia today, though it is still contested among lower castes, in particular by
various heterodox movements, such as the nirgun.i devotional traditions.
Over the second half of the first millennium, one can see the two
cultural patterns, the one built around the śraman.a traditions, the other
around the evolving Brahmanical religion, developing throughout South
and Southeast Asia, in competition but also in constant interaction with
each other. The ultimate weakness of the Buddhist pattern2 in its Indian
version was probably its reliance on state and urban support for Buddhist
2
The Jain tradition survived somewhat better, through the ongoing patronage of the Jain merchant
community, and has retained a presence within South Asia, if on a limited scale. There seems no
obvious explanation beyond historical contingency for the different trajectories of the two śraman.a
Postlude
345
monasteries. The progressive loss of such support in Hindu and Muslim
kingdoms, along with the large-scale destruction of monasteries that apparently accompanied the Muslim invasions in the early thirteenth century,
meant that the major teaching centres where the Buddhist clergy at the local
level were trained were no longer there. The Brahmanical variant was better
adapted to survive in this situation, since the extensive land-grants made to
Brahmins over the centuries had provided them with a secure base at the
village level. Elsewhere, in Sri Lanka, Southeast Asia and Tibet, Buddhist
monasticism would in time develop a solid base at the village level, but it
appears that this never happened in mainland South Asia.
Within the developed Brahmanical pattern, the Brahmin-centred values
of purity and pollution became a major axis around which the village structure was built, while the this-worldly and pragmatic values of everyday life
continued as a parallel set of values and orientations, classed by anthropologists in recent years under the generic label of ‘auspiciousness’. This is a
theme I have discussed briefly in Chapters 7 and 8, and have also written
on elsewhere (e.g. Samuel 1997; Rozario and Samuel 2002a).
The contrast between these patterns is, I think, an important key to
understanding the differing structures of religion and of social life in South
Asia and Southeast Asia today. It is important, though, to recognise that
both of these variants, the Buddhist-Jaina pattern and the later Brahmanical
pattern, are equally Indic, and contain many of the same components
in differing arrangements. The Brahmanical solution, like the Buddhist,
incorporated the ideal of the wise king, the dharmarāja, and of the ascetic
teachings leading to transcendence, and it too allowed for the recognition
and sacralisation of the processes of everyday life.
Beyond these more general questions of value systems, there is the more
specific question of the role of ‘technologies of the self’ within society. In one
sense, such technologies are a common feature of very many human societies
(cf. Foucault 1988a: 18). Ideas of this kind are perhaps intrinsic in the rites of
passage that small-scale societies employ to manage the transitions through
life of their members, particularly where there is an element of initiation
into secret or private knowledge in such processes, and of choice in how far
one may progress through them. It is easy enough to point to the similarities
between the systems of initiatory knowledge that can be found in widely
traditions, though the higher profile of Buddhism at the time of the Muslim conquests may have made
it a more obvious target, and the existence of other Buddhist societies outside South Asia may have
led to Buddhist monks and teachers choosing to migrate rather than to remain and adapt. Buddhism
in South Asia today, leaving aside Sri Lanka and the newly-converted Ambedkar Buddhists, survived
only in Nepal and in parts of East Bengal, where it developed links with Arakan and Burma (Tinti
1998).
346
Origins of yoga and Tantra
different societies, but these similarities are also accompanied by radical
differences deriving from the contrasting social context and meaning of
initiation in, say, Melanesia and Bhutan (e.g. Barth 1990).
What we are dealing with in the Indic material has itself undergone
radical transformations over the long historical period considered in this
book. In the first half of the first millennium BCE we can perhaps see
traces of two rather different kinds of system of initiatory knowledge, that
of the Vedic priests and the vrātyas on the one hand, and that of the protośraman.a cults of the Central Gangetic region on the other. These traditions
undoubtedly interacted in all kinds of ways in many regions and places in
what was an, at least partly, shared milieu of ascetic practice, but it seems
worth exploring, as I have done in this book, the possibility that their origins
were distinct, and tied up with different cultural contexts. The Vedic system
was for the most part a system of hereditary ritual knowledge passed down
in Brahmin families, but there are also suggestions, in the vrātya material,
of a period of collective initiation for young men as a whole, based around
ritual and military activity in the forest away from the settled community. If
I am right in suggesting that the origins of the śraman.a cults might be found
in early initiatory cults in and beyond the Central Gangetic region,3 with
resemblances to the West African initiatory cults that played a significant
role in the growth of wider social and political networks in that region in
recent centuries, or perhaps to the early phases of the initiatory cults of the
Hellenistic world, we can glimpse here the historical origins of these two
different patterns and of their different emphases and approaches.
For the śraman.a cults, initiation is an elective process undertaken by a
limited sector of the population whose way of life or whose personal career
has made such a choice meaningful and desirable. The often-suggested links
between Buddhism and the trading communities of the newly-developing
North Indian states of that time provide one way in which this may
have worked, in that cult-initiation would have offered both a network
of potential trading partners in distant places, an ethical code that provided a basis for such transactions, and a world-view in which members
of distant communities shared with oneself a common human nature and
common potentialities for spiritual advancement. The cultural pattern built
around the śraman.a orders presupposes a background of local deity cults,
but not necessarily the development of large-scale Brahmanical religious
cult-centres which, for much of North India, may not have taken place
until a much later period, perhaps that of the Gupta kings.
3
I say ‘in and beyond’ with reference to Williams’ suggestions about Gujarat (R. Williams 1966).
Postlude
347
All this, it must be said, is speculative, but it provides a context in
which we can make sense of the teaching careers of the various semilegendary predecessors of the Buddha and Mahāvı̄ra and perhaps too of
the legendary renunciate kings of Mithilā and elsewhere. It might also
provide an interpretive background to the activities of the Buddhist and
Jain founding figures themselves, and to the interest of the historical kings
of the fifth and fourth century states and of the Mauryan Empire in their
teachings.
In the court culture of the second and third centuries and particularly of
the Gupta and Vākāt.aka kings, the idea of a consciously undertaken pattern
of life, with formalised bodies of rules, had become institutionalised in a
variety of contexts, including the legal handbooks (Dharma Sūtras), the
artha and kama śāstras, and the various formalised rules for yogic and
religious activity. Daud Ali is surely right to see the Buddhist Vinaya code
as part of the same world and as sharing many of the same assumptions
(Ali 1998). Some of these regulated modes of life were inherited, others
were voluntarily undertaken, but in all cases there was an expectation of
a controlled and disciplined way for the individual to follow. The various
traditions of meditation and yoga at this period need to be seen as specific
‘technologies of the self’ within such a context. They offered a particular
set of alternatives, both for lay followers who might be involved in yogic
and meditational practices to various degrees of seriousness and for those
who chose to commit themselves further and become full-time professional
renunciates.
At the same time, we need to recall that most people in South Asia were
not living in the courtly or urban middle-class milieu. Both in villages
and urban centres, the cycle of collective seasonal rituals and of contact
with the deities through spirit-mediums and shamanic-style practitioners
continued, and the renunciate career involved an ongoing interaction with
this milieu as well as with the courtly milieu.
The initial placing of the śraman.a community and of the Brahmanical
sannyāsin outside the ordered life of the city and village, in the social and
physical spaces associated with death and misfortune, provided one basic
reason why they needed to build a relationship with the wider community. As the increasing size and wealth of the Buddhist and Jain monastic
communities demonstrated, they developed other modes of support, centred around their spiritual appeal to the new urban elites and doubtless also
to the various pragmatic ways in which they became integrated into the
social, political and economic life of the urban centres (cf. Heitzman 1984;
O’Connor 1989; Bailey and Mabbett 2003). The ideal of the forest ascetic,
348
Origins of yoga and Tantra
on the edges or outside of the settled community, or living with other practitioners in remote locations, nevertheless persisted and, as Reginald Ray
has suggested, formed the basis for a vital ongoing dialectic between the
civilised and regulated urban monastics and a creative and visionary milieu
(Ray 1999). A similar dialectic perhaps took place between the ordered and
secure world of the settled Brahmins and the various renunciate practitioners, including the Paśupatas and the kāpālika-style ascetics, though it
is clear, as with the Buddhist forest-monks, that these groups developed
their own institutional base and their own connections with the courts
and cities. This dialectic between institutionalised and less regulated versions of the traditions probably provided a major source for innovation and
change. It was doubtless also a key element in the growth of the various
kaula and Tantric traditions which provided new ‘technologies of the self’
for the changing conditions of ninth and tenth century CE South and
Southeast Asia, as well as a body of practices which could be used to take
over the role of Atharvavedic purohits and other ritualists (cf. Sanderson
2004).
sexualit y and gender rel ations in tantra
This brings us to one of the most controversial issues about Tantra for
many modern readers, its use of sexuality and of ‘transgressive’ elements
in general. Why did Indian spiritual traditions develop in this particular
direction in the ninth and tenth centuries?
To begin with, we might ask how Tantric sexuality fits into the picture
of two cultural patterns that I drew up above, the brahmacārin-centred
Vedic pattern and the mithuna-centred Central Gangetic pattern. As we
have seen, the development of the kaula and mahāyoga practices took place
at a time when South Asian societies were becoming increasingly dominated by feudal and military values (Davidson’s ‘sāmantisation’ process,
see Chapter 12). The Southeast Asian states that adopted Tantric models
probably shared many of the same values and orientations. We have already
noted the conflation of warfare and eroticism that appeared to go along
with these values in parts of India in particular.
Whether the values of Tantric practice can be seen as a development
of such problematic emphases or as a potential counter or response to
them is a difficult question. It might have been some of each, at different
times. A variety of different meanings can also be given to the sexual and
transgressive elements within Tantra.
Postlude
349
Viewed from the point of view of the modern Tibetan tradition in particular, it is tempting to stress the elements of asceticism and control within
Tantra, and to marginalise the sexual and transgressive elements. This is
true enough for contemporary Tibetan practice, where these elements are
mostly a matter of visualisation or symbolic performance, but all the indications are that the sexual practices, at any rate, were seen as essential by
both Buddhist and Śaiva Tantric practitioners in the ninth to twelfth century Indian context, and that kāpālika-style practice, with its transgressive
and antinomian elements, also had some historical reality.
An alternative approach would be to focus on the use of Tantric ritual for
pragmatic ends, and to see the transgression purely in terms of pragmatic
power. As I have noted in earlier chapters, it seems clear that the use of sexual
practices and fluids, initially at least, derived from the ritual effectiveness
of impure and transgressive substances, both for creating the initiate’s link
to the lineage and for bringing about magically effective ritual, rather than
from any kind of spiritualisation of sexual experience as such. This phase
of Tantric practice adds little to a history of masculinity or sexuality in
South Asia, except perhaps to illustrate the power associated with breaking
the rules, and the need for religious practitioners to provide what their
employers are looking for. By the ninth and tenth centuries, it is clear
that both Buddhist and Śaiva practitioners were in the business of selling
aggressive and destructive rituals (and also, one might add, healing rituals
and rituals for prosperity) to their secular employers, and that there was a
market for both Buddhist and Śaiva versions.
Yet, as a series of contemporary scholars have made clear in relation
to the later, ‘spiritualised’, versions of Tantric practice associated with the
great Kashmiri scholar Abhinavagupta and his followers,4 and as is equally
clear from the practice of Buddhist Tantra in the late Indian context and
afterwards in Tibet, this is not all there is to the story.
These more spiritualised understandings of kaula and Tantric practice
may go back some way before the reforms of the eight to tenth centuries.
As I have commented elsewhere in this book, one needs to bear in mind an
‘internal’ understanding of what is going on in Tantric ritual as well as an
external understanding. As we have seen in relation to the Pāśupatas and the
kāpālika-style practitioners, the reasons why people undertook these apparently bizarre and transgressive practices may not bear much resemblance to
the image which the society as a whole had of them. Whether, as in the case
4
E.g. Sanderson 1985, 1995; Silburn 1988; White 2003; Muller-Ortega 2002; Dupuche 2003.
350
Origins of yoga and Tantra
of the low-caste hereditary practitioners, people were born to the career of
a Tantric ritualist, or whether, as with the Pāśupatas and the kāpālika-style
ascetics, they chose it voluntarily, the ritualists were likely to develop a
perspective on their activities which allowed them a degree of self-respect.
It is here perhaps that we find the sources of kaula practice as a path to
spiritual transcendence, along with the striking imagery that accompanies
it, in which the transgressive behaviour, the cremation ground and the fierce
deities become aids to achieving the breakthrough from ordinary worldly
experience which had already long been the central goal of the Indian ascetic
paths. In time, this provided the basis for a ‘spiritualised’ understanding of
Tantra, and a morally elevated view of the relationship of male and female
individuals that was at the base of its core practice.
Much the same would also be true on the Buddhist side, where these
practices were adopted within a tradition that already had a very strong ethical orientation in the bodhisattva ideal, and where the sexual practices were
connected directly with this ethical orientation through the identification
of male and female sexual substances with bodhicitta, the empathetic desire
to relieve the sufferings of other beings that is the central motive force of
the Buddhist quest for the liberating insight (Samuel 1989).
As for the politics of gender in South Asian societies, Tantra by itself
could scarcely have reversed the long-term processes by which the status
of women in South Asia became increasingly confined and restricted, but
it did at least enjoin and legitimate a more positive and equal relationship
between the sexual partners who were involved in its practice at the time.
Even this was a difficult and unstable position to uphold. Perhaps that is
as much as we can reasonably ask of it.
In the longer term, the sexual practices were marginalised both in India
and in Tibet, with the celibate practitioner, whether Hindu sannyāsin or
Buddhist monk, being increasingly seen as the ideal. Whether contemporary Western versions of Tantric sexuality have the potential to aid in the
all too necessary restructuring of sexual relationships within today’s society
is another question (Samuel 2005a: 357–61).
ph i losophy, spiritual realisation and social practice
Philosophical understandings of Indic religions have played a fairly small
part in this book.5 My interest here has been more on the techniques used
5
I apologise to those who would have liked it to have had a larger part, but there are plenty of good
books on Hindu and Buddhist philosophy already.
Postlude
351
to attain the ‘liberating insight’ within specific traditions, and on the wider
social context within which seekers of that insight have operated.
Indian traditions of spiritual realisation tend to assume that spiritual
practice is an ascetic process of one or another kind, and as we have seen
the amount of specific guidance as to technique varies from little or nothing, in the early period, to extremely elaborate internal visualisations and
mind-body transformations, in the later Tantric traditions. Indian spiritual
traditions also assume that one understands or perceives something as a
result of the successful accomplishment of spiritual practice, and that the
understanding or perception cannot be separated from the inner transformation brought about as a result of the practice. The liberating insight is
both understanding and inner transformation.
Even if the understanding could be understood as a logical proposition
outside from the inner transformation, and Indian traditions have been
far from united about whether this is possible, it is ineffective without the
transformation: the point is not to assert the logical proposition that one
is Śiva, or that all is Buddha-nature, but to directly experience the truth
to which those words refer. The liberating insight is thus not a logical
proposition but something intrinsic to a patterning or attunement of the
mind-body system as a whole to the wider universe of which it forms an
indissoluble part. This is one reason why the techniques employed to bring
this inner transformation about are of at least as much interest as the logical
propositions through which the resulting insight may be expressed.
Yogic and meditational procedures are taken more seriously now than
they were in Western societies even half a century ago, but the question
of how far we should regard these processes of inner transformation as
having real effects remains open within many parts of Western and global
society. In fact, this is a good question, though only partly answerable.
There is enough solid research to show that Indian techniques of yoga
and meditation, including the various internal yogas of Tantric practice,
can indeed affect the functioning, and so the health and vitality, of the
organism, and that the special awarenesses and sensitivities that are trained
by these techniques do have real correlates accessible to modern scientific
investigation. I have written a little elsewhere about these matters (Samuel
2006a, 2006c), and this is an area in which there is also beginning to be
some serious research in the natural sciences.
If we can accept this much, then we can also perhaps accept that the complex set of techniques and approaches which was assembled by the Tantric
practitioners of the ninth to twelfth centuries may be less arbitrary and
less bizarre than it initially appears. The fundamental Tantric techniques
352
Origins of yoga and Tantra
deal with sexuality and our connection with other human beings, with
the basic processes of life and vitality and with the ultimate breakdown of
the human organism in death. These aspects of human existence are still
part of our lives today. The encounters, real or visualised, with the ferocious Tantric deities and the horrific environment of the charnel-ground
allow for a confrontation and resolution of deep-rooted conflicts within
the human mind-body organism. The internal yogic practices valorise and
operate with the fundamental affective bond between sexual partners, and
treat it as a gateway to a liberating insight in which universal compassion
is a key element. They also confront the practitioner with the process of
the dissolution of personality in sleep and in death, and treat this too as an
opening to spiritual realisation. From this point of view, Tantra seems less
of a historical aberration, and more a set of techniques that might allow for
a healing and productive encounter with the basic problems of the human
condition.
With regard to the ultimate state or condition to which these practices
aim, that associated with the liberating insight itself, it is less possible to
speak in scientific terms. Persons who have achieved such conditions, if
they exist, are generally not available for scientific investigation. Perhaps in
the present context it is enough to say that these matters were clearly taken
seriously over thousands of years by many highly intelligent people.
One should go on, however, to ask a further question. What are the
consequences for contemporary society as a whole if it is choosing to take
such states seriously and to commit substantial numbers of people to their
pursuit? That question is in the background of much of what I have written
in this book, but we are still some way from being able to answer it fully.
Yet issues of this kind may be critical in relation to the future of our
planet. We can no longer treat the question of human feeling, motivation
and consciousness as something to be determined simply by the logic of
commercial exploitation and by the patterns of desire, dissatisfaction and
resentment that it inevitably creates. Nor do the various modernist forms
of religion, the so-called fundamentalisms, help the situation. They are part
of the same logic as the world of commerce. They operate with the same
crude levels of consciousness and emotion and they lead into much the
same resentments and destructive political directions. Only the enemies
against which the resentment and anger is focused are different.
In such a world, a body of traditions and techniques that claims to
detach the practitioner from excessive emotion, and to purify his or her
consciousness, is worth taking with some seriousness. In addition, much
of later Tantric practice, as we have seen, moved beyond this initial context
Postlude
353
to constitute a complex and subtle series of processes for transforming the
mind-body totality of the practitioner.
If we regard the techniques of yoga, meditation and Tantra as tools that
may still be worth investigating as of use in today’s global society, then
our ability to make sense of them and use them constructively can only
be assisted by an understanding of the historical context within which
these practices developed and out of which their imagery and language was
born. I hope that this book will provide some of the elements for such an
understanding.
References
Aasen, C. 1998. Architecture of Siam: A Cultural History Interpretation. Kuala
Lumpur: Oxford University Press.
Adriaensen, R., Bakker, H. T. and Isaacson, H. 1994. ‘Towards a Critical Edition
of the Skandapurān.a’. Indo-Iranian Journal 37: 325–31.
Adriaensen, R., Bakker, H. T. and Isaacson, H. 1998-. The Skandapurān.a. Critically edited with Prolegomena and English synopsis by R. Adriaensen, H. T.
Bakker, H. Isaacson. 2 vols. so far. Groningen: Egbert Forsten.
Agrawala, Prithvi K. 1983. Mithuna: The Male-Female Symbol in Indian Art and
Thought. Delhi: Munshiram Manoharlal.
Agrawala, V. S. and Motichandra 1960. ‘Yaksha Worship in Varanasi: Matsyapurān.a (Ch. 180)’. Purān.a 1: 198–201.
Ali, D. 1998. ‘Technologies of the Self: Courtly Artifice and Monastic Discipline
in Early India’. Journal of Economic and Social History of the Orient 41: 159–
84.
Ali, D. 2004. Courtly Culture and Political Life in Early Medieval India. Cambridge
and New York: Cambridge University Press.
Allchin, B. 1995. ‘The Environmental Context’, in Allchin 1995, pp. 10–25.
Allchin, B. and R., F. 1982. The Rise of Civilization in India and Pakistan.
Cambridge, London and New York: Cambridge University Press.
Allchin, F. R. 1963. Neolithic Cattle-Keepers of South India: A Study of the Deccan
Ashmounds. Cambridge: Cambridge University Press.
Allchin, F. R. (ed.) 1995a. The Archaeology of Early Historic South Asia: The Emergence of Cities and States. Cambridge and New York: Cambridge University
Press.
Allchin, F. R. 1995b. ‘Mauryan Architecture and Art’, in F. R. Allchin 1995a, pp. 222–
73.
Allchin, F. R. 1995c. ‘Language, Culture and the Concept of Ethnicity’, in F. R.
Allchin 1995a, pp. 41–53.
Allen, M. R. 2005. Anthropology of Nepal: Peoples, Problems, and Process. Reprint.
Kathmandu: Mandala.
Allen, N. J. 1987. ‘The Ideology of the Indo-Europeans: Dumézil’s Theory and the
Idea of a Fourth Function.’ International Journal of Moral and Social Studies
2(1): 23–39.
354
References
355
Allen, N. J. 1996. ‘Romulus and the Fourth Function’, in Indo-European Religion
after Dumézil, Polomé (ed.), pp. 13–36. Washington: Institute for the Study
of Man.
Allen, N. J. 1998. ‘The Indo-European Prehistory of Yoga’. International Journal
of Hindu Studies 2: 1–20.
Alter, J. S. 1997. ‘Seminal Truth: A Modern Science of Male Celibacy in North
India’. Medical Anthropology Quarterly 11: 275–98.
Alter, J. S. 2005. ‘Modern Medical Yoga: Struggling with a History of Magic,
Alchemy and Sex’. Asian Medicine: Tradition and Modernity 1: 119–46.
Ames, M. 1964. ‘Magical Animism and Buddhism: a Structural Analysis of the
Sinhalese Religious System’, in Religion in South Asia, Harper (ed.), pp. 21–5.
Seattle: University of Washington Press.
Ames, M. 1966. ‘Ritual Prestations and the Structure of the Sinhalese Pantheon’, in
Anthropological Studies in Theravada Buddhism, Nash (ed.), pp. 27–50. New
Haven, CT: Southeast Asia Studies, Yale University.
Andaya, B. Watson 1994. ‘The Changing Religious Role of Women in Pre-Modern
South East Asia’. South East Asia Research 2: 99–116.
Andresen, J. 1997. Kālacakra: Textual and Ritual Perspectives. PhD Dissertation,
Harvard University. UMI AAT 9733170.
Archer, W. G. 1964. ‘Preface’, in Comfort 1964, pp. 7–38.
Archer, W. G. 1974. The Hill of Flutes: Life, Love and Poetry in Tribal India: a
Portrait of the Santals. London: Allen and Unwin.
Atre, S. 1998. ‘The High Priestess: Gender Signifiers and the Feminine in the
Harappan Context’. South Asian Studies 14: 161–72.
Aung-Thwin, M. 1981. ‘Jambudipa: Classical Burma’s Camelot’. Contributions to
Asian Studies 16: 38–61.
Aziz, B. N. 1978. Tibetan Frontier Families: Reflections of Three Generations from
D’ing-ri. New Delhi: Vikas Publishing House.
Bagchi, P. C. and Magee, M. 1986. Kaulajñānanirn.aya of the School of
Matsyendranātha. Text Edited with an Exhaustive Introduction by P. C. Bagchi.
Translated into English by Michael Magee. Varanasi: Prachya Prakashan.
(Tantra Granthamala, 12)
Bailey, G. and Mabbett, I. 2003. The Sociology of Early Buddhism. Cambridge and
New York: Cambridge University Press.
Bakker, H. 1986. Ayodhyā. Groningen: Egbert Forsten.
Bakker, H. 1999. [Review of Srinivasan 1997.] Artibus Asiae 58: 339–43.
Bakker, H. 2001. ‘Sources for Reconstructing Ancient Forms of Śiva Worship’, in
Grimal 2001, pp. 397–412.
Bakker, H. (ed.). 2004. The Vākāt.aka Heritage: Indian Culture at the Crossroads.
Groningen: Egbert Forsten.
Balcerowicz, P. 2005. ‘Monks, Monarchs and Materialists’. [Review essay of Bollée
2002.] Journal of Indian Philosophy 33: 571–82.
Banerji, S. C. 1992. New Light on Tantra: Accounts of Some Tantras, Both Hindu and
Buddhist, Alchemy in Tantra, Tantric Therapy, List of Unpublished Tantras, etc.
Calcutta: Punthi Pustak.
356
References
Barth, F. 1990. ‘The Guru and the Conjurer: Transactions in Knowledge and
the Shaping of Culture in Southeast Asia and Melanesia’. Man 25: 640–
53.
Barua, B. 1934–37. Barhut. Books I to III. Calcutta.
Basham, A. L. 1951. History and Doctrines of the Ājı̄vikas. London: Luzac.
Bautze, J. K. 1995. Early Indian Terracottas. Leiden: E. J. Brill. (Iconography of
Religions, XIII, 17.)
Bayly, S. 2001. Caste, Society and Politics in India from the Eighteenth Century to the
Modern Age. Paperback edn. Cambridge: Cambridge University Press. (New
Cambridge History of India, IV, 3.)
Bechert, H. (ed.) 1991–97. The Dating of the Historical Buddha = Die Datierung
des historischen Buddha. 3 vols. Göttingen: Vandenhoeck & Ruprecht.
Beck, G. L. 1995. Sonic Theology: Hinduism and Sacred Sound. Delhi: Motilal
Banarsidass.
Becker, J. 1993. Gamelan Stories: Tantrism, Islam, and Aesthetics in Central Java.
Tempe, AZ: Arizona State University, Program for Southeast Asian Studies.
Bell, S. and Sobo, E. J. 2001. ‘Celibacy in Cross-Cultural Perspective: An Overview’,
in Sobo and Bell 2001, pp. 11–12.
Bellwood, P. 1999. ‘Southeast Asia Before History’, in Tarling 1999, pp. 55–136.
Belo, J. 1949. Bali: Rangda and Barong. Seattle: University of Washington Press.
Benard, E. A. 1994. Chinnamastā: The Aweful Buddhist and Hindu Tantric Goddess.
Delhi: Motilal Banarsidass.
Bentley, G. Carter 1986. ‘Indigenous States in Southeast Asia’. Annual Review of
Anthropology 15: 275–305.
Bernbaum, E. 1980. The Way to Shambhala. Garden City, NY: Anchor
Press/Doubleday.
Berreman, G. D. 1971. ‘The Brahmanical View of Caste’. Contributions to Indian
Sociology (N.S.) 5: 16–23.
Beyer, S. 1977, ‘Notes on the Vision Quest in Early Mahāyāna’, in Prajñāpāramitā
and Related Systems: Studies in Honor of Edward Conze, Lancaster (ed.),
pp. 329–40. Berkeley: University of California Press. (Berkeley Buddhist
Studies Series.)
Beyer, S. 1973. The Cult of Tara. Magic and Ritual in Tibet. Berkeley: University of
California Press.
Bharadwaj, O. P. 1989. ‘Yaksha-Worship in Kurukshetra’, in Prācı̄-prabhā: Perspectives in Indology: Essays in Honour of Professor B. N. Mukherjee, Bhattacharyya
and Handa (eds.), pp. 203–23. New Delhi: Harman.
Bharati, A. 1965. The Tantric Tradition. London: Rider.
Bhattacharyya, N. N. 1996. Ancient Indian Rituals and Their Social Contexts. Delhi:
Manohar.
Bhawe, S. S. 1939. Die Yajus’ des Asvamedha. Versuch einer Rekonstruktion dieses
Abschnittes des Yajurveda auf Grund der Überlieferung seiner fünf Schulen.
Stuttgart: W. Kohlhammer.
Biardeau, M. 1989. Histoires de poteaux: variations védiques autour de la Déesse
hindoue. Paris: Ecole française d’Extrême Orient.
References
357
Bisschop, P. 2005. ‘Pañcārthabhās.ya on Pāśupatasūtra 1.37–39 Recovered from a
Newly Identified Manuscript’. Journal of Indian Philosophy 33: 529–51.
Bisschop, P. and Griffiths, A. 2003. ‘The Pāśupata Observance (Atharvavedapariśis.t.a 40)’. Indo-Iranian Journal 46: 315–48.
Blacker, C. 1986. The Catalpa Bow: A Study of Shamanistic Practices in Japan. Second
edn. London and Boston: Mandala/Unwin Paperbacks.
Blench, R. and Spriggs, M. (eds.) 1998. Archaeology and Language II: Correlating
Archaeological and Linguistic Hypotheses. London and New York: Routledge.
Blench, R. and Spriggs, M. (eds.) 1999. Archaeology and Language III: Artefacts,
Languages and Texts. London and New York: Routledge.
Blezer, H. 1997. Kar gliṅ Źi khro: A Tantric Buddhist Concept. Leiden: Research
School CNWS, School of Asian, African, and Amerindian Studies, Leiden
University. (CNWS Publications, 56.)
Bollée, W. B. 1981. ‘The Indo-European Sodalities in Ancient India’. Zeitschrift der
Deutschen Morgenlandischen Gesellschaft 131: 172–91.
Bollée, W. B. 2002. The Story of Paesi (Paesi-kahān.ayam
. ). Soul and Body in Ancient
India: A Dialogue on Materialism. Text, Translation, Notes and Glossary.
Wiesbaden: Harrassowitz. (Beitrage zur Kenntnis südasiatischer Sprachen
und Literaturen 8.)
Bolon, C. Radcliffe 1992. Forms of the Goddess Lajjā Gaurı̄ in Indian Art. University
Park, PA: Pennsylvania State University Press.
Bouillier, V. 1993. ‘La violence des non-violents, ou les ascètes au combat’.
Purus.ārtha 16: 213–42.
Boyce, M. 1997. ‘Origins of Zoroastrian Philosophy’, in Companion Encyclopedia of
Asian Philosophy, Carr and Mahalingam (eds.), pp. 5–23. London: Routledge.
Brauen-Dolma, M. 1985. ‘Millenarianism in Tibetan Religion’, in Soundings in
Tibetan Civilization, Aziz and Kapstein (eds.), pp. 245–56. New Delhi:
Manohar.
Brereton, J. P. 1981. The R.gvedic Ādityas. New Haven, CT: American Oriental
Society. (American Oriental Series, 63.)
Briggs, G. Weston 1989. Gorakhnāth and the Kānphat.a Yogı̄s. Reprint. Delhi:
Motilal Banarsidass.
Brockington, J. L. 1985. Righteous Rāma: The Evolution of an Epic. Delhi: Oxford
University Press.
Brockington, J. L. 1998. The Sanskrit Epics. Leiden: Brill. (Handbuch der Orientalistik. Zweite Abteilung, Indien; 12. Bd.)
Brockington, J. L. 2003a. ‘The Sanskrit Epics’, in Flood 2003, pp. 116–28.
Brockington, J. L. 2003b. ‘Yoga in the Mahābhārata’, in Whicher and Carpenter
2003, pp. 13–24.
Brohm, J. 1963. ‘Buddhism and Animism in a Burmese Village’. Journal of Asian
Studies 22: 155–67.
Bronkhorst, J. 1993. The Two Traditions of Meditation in Ancient India. Delhi:
Motilal Banarsidass.
Bronkhorst, J. 1998a. The Two Sources of Indian Asceticism. Second edn. Delhi:
Motilal Banarsidass.
358
References
Bronkhorst, J. 1998b. ‘Did the Buddha Believe in Karma and Rebirth?’, Journal of
International Association of Buddhist Studies 21: 1–19.
Bronkhorst, J. 2000. ‘The Riddle of the Jainas and Ājı̄vikas in Early Buddhist
Literature’. Journal of Indian Philosophy 28: 511–29.
Bronson, B. and White, J. C. 1992. ‘Radiocarbon and Chronology in Southeast
Asia’, in Ehrich 1992, I, pp. 491–503; II, pp. 475–515.
Brooks, D. Renfrew 1990. The Secret of the Three Cities: An Introduction to Hindu
Śākta Tantrism. Chicago and London: University of Chicago Press.
Brooks, D. Renfrew 1992a. Auspicious Wisdom: The Texts and Traditions of Śrı̄vidyā
Śākta Tantrism in South India. Albany, NY: State University of New York
Press.
Brooks, D. Renfrew 1992b. ‘Encountering the Hindu “Other”: Tantrism and the
Brahmans of South India’. Journal of American Academy of Religion 60(3):
405–36.
Brown, C. Henning 1996. ‘Contested Meanings: Tantra and the Poetics of Mithila
Art’. American Ethnologist 23: 717–37.
Brown, M. Fobes 1988. ‘Shamanism and its Discontents’. Medical Anthropology
Quarterly 2(2): 102–20.
Brown, R. L. 1990. ‘A Lajjā Gaurı̄ in a Buddhist Context at Aurangabad’. Journal
of International Association of Buddhist Studies 13(2): 1–16.
Brubaker, R. L. 1979. ‘Barbers, Washermen, and Other Priests: Servants of
the South Indian Village and the Goddess’. History of Religions 19: 128–
54.
Brubaker, R. L. 1983. ‘The Untamed Goddesses of Village India’, in The Book of
the Goddess: An Introduction to Her Religion, Olson (ed.), pp. 145–60. New
York: Crossroad.
Brückner, H., Lutze, L. and Malik, A. (eds.) 1993. Flags of Fame: Studies in South
Asian Folk Culture. New Delhi: Manohar.
Bryant, B. 1992. The Wheel of Time Sand Mandala: Visual Scripture of Tibetan
Buddhism. San Francisco: HarperSanFrancisco.
Bryant, E. F. and Patton, L. L. (eds.) 2005. The Indo-Aryan Controversy: Evidence
and Inference in Indian History. London and New York: Routledge.
Bühnemann, G. 1996. ‘The Goddess Mahācı̄nakrama-Tārā (Ugra-Tārā) in Buddhist and Hindu Tantrism’. Bulletin of the School of Oriental and African
Studies 59: 472–93.
Bühnemann, G. 1999. ‘Buddhist Deities and Mantras in the Hindu Tantras: I The
Tantrasārasam
. graha and the Īśānaśivagurudevapaddhati’. Indo-Iranian Journal
42: 303–34.
Bühnemann, G. 2000a. ‘Buddhist Deities and Mantras in the Hindu Tantras: II
The Śrı̄vidyārn.avatantra and the Tantrasāra.’ Indo-Iranian Journal 43: 27–43.
Burghart, R. 1978. ‘The Disappearance and Reappearance of Janakpur’. Kailash 6:
257–84.
Cardona, G. 1976. Pān.ini: A Survey of Research. The Hague: Mouton.
Carman, J. B. and Marglin, F. A. 1985. Purity and Auspiciousness in Indian Society.
Leiden: E. J. Brill.
References
359
Carrithers, M. 1983. The Forest Monks of Sri Lanka: An Anthropological and Historical
Study. Delhi: Oxford University Press.
Carrithers, M. and Humphrey, C. (eds.) 1991. The Assembly of Listeners: Jains in
Society. Cambridge and New York: Cambridge University Press.
Carstairs, G. Morris 1957. The Twice-born: a Study of a Community of High-Caste
Hindus. London: Hogarth Press, 1957.
Castells, M. 1997. The Power of Identity (The Information Age: Economy, Society and
Culture. Volume II). Malden, MA and Oxford: Blackwells.
Chakrabarti, D. K. 1995. The Archaeology of Ancient Indian Cities. Delhi: Oxford
University Press.
Chakrabarti, D. K. 2001. Ancient Bangladesh: A Study of the Archaeological Sources
with an Update on Bangladesh Archaeology, 1990–2000. Dhaka: The University
Press Limited.
Chakrabarti, D. K. n.d. ‘The Historical and Archaeological Context of Chandraketugarh’. Downloaded from www.historyofbengal.com/articles.html,
4 Oct 2006.
Chandra, L. 1992a. Cultural Horizons of India. Delhi: International Academy of
Indian Culture. (Śata-Pit.aka, 366.)
Chandra, L. 1992b. ‘Tantras and the Defence of T’ang China’, in Chandra 1992a,
vol. 2, pp. 257–66.
Chandra, L. 1992c. ‘Emperor Hsüan-Tsung, Vajrayāna and Quarter of Vajras’, in
Chandra 1992a, vol. 2, pp. 267–76.
Chatterji, B. Chandra 1992. Anandamath. Translated and adapted from original
Bengali by B. Koomar Roy. New Delhi and Bombay: Orient Paperbacks.
Chattopadhyaya, A. 1967. Atisa and Tibet: Life and Works of Dipamkara Srijnana
in Relation to the History and Religion of Tibet with Tibetan Sources. Delhi:
Motilal Banarsidass.
Chawla, J. 1994. Child-bearing and Culture: Women-Centered Revisioning of the
Traditional Midwife, the Dai as a Ritual Practitioner. New Delhi: Indian Social
Institute.
Chawla, J. 2006. ‘Celebrating the Divine Female Principle’. Women’s Feature
Service. Downloaded from www.boloji.com/wfs/wfs082.htm, 23 September
2006.
Chia, M. 1983. Awaken Healing Energy Through the Tao: The Taoist Secret of Circulating Internal Power. Santa Fe, NM: Aurora Press.
Choong, M. 2005. ‘The Importance of Pali-Chinese Comparison in the Study of
Pali Suttas’. Khthonios: A Journal for the Study of Religion (Queensland Society
for the Study of Religion) 2(2): 19–26.
Clark, M. J. 2004. The Daśanāmı̄-Sam
. nyāsı̄s: the Integration of Ascetic Lineages into
an Order. PhD dissertation, University of London. [Revised version published
under same title by Brill, 2006.]
Claus, P. J. 1973. ‘Possession, Protection, and Punishment as Attributes of the
Deities in a South Indian Village’. Man in India 53: 231–42.
Claus, P. J. 1979. ‘Spirit Possession and Mediumship from the Perspective of Tulu
Oral Traditions’. Culture, Medicine and Psychiatry 3: 29–52.
360
References
Claus, P. J. 1984. ‘Medical Anthropology and the Ethnography of Spirit Possession’,
in Daniel and Pugh 1984, pp. 60–72.
Claus, P. J. 1993. ‘Text Variability and Authenticity in the Siri Cult’, in Brückner,
Lutze and Malik 1993, pp. 335–74.
Clothey, F. W. 1978. The Many Faces of Murukan: The History and Meaning of a
South Indian God. The Hague: Mouton.
Coburn, T. B. 1982. ‘Consort of None, Śakti of all: The Vision of the Devı̄Māhātmya’, in Hawley and Wulff 1982, pp. 153–65.
Coburn, T. B. 1988. Devı̄-Māhātmya: The Crystallization of the Goddess Tradition.
Delhi: Motilal Banarsidass.
Coburn, T. B. 1991. Encountering the Goddess: A Translation of the Devı̄-Māhātmya
and a Study of its Interpretation. Albany, NY: State University of New York
Press.
Coedès, G. 1975. The Indianized States of Southeast Asia. Edited by W. F. Vella;
translated by S. Brown Cowing. Canberra: Australian National University
Press.
Cohen, A. L. 1992. ‘The King and the Goddess: The Nol.amba Period
Laks.maneśvara Temple at Avani’. Artibus Asiae 52: 7–24.
Cohen, A. L. 1997. ‘Why a History of Monuments from Nol.ambavād.i?’ Artibus
Asiae 57: 17–29.
Cohen, R. S. 1995. ‘Discontented Categories: Hinayana and Mahayana in Indian
Buddhist History’, Journal of the American Academy of Religion 63/1 (Spring
1995): 1–25.
Cohen, R. S. 1998. ‘Nāga, Yaks.inı̄, Buddha: Local Deities and Local Buddhism at
Ajanta’. History of Religions 37: 360–400.
Colas, G. 2003. ‘History of Vais.n.ava Traditions: An Esquisse’, in Flood 2003,
pp. 229–70.
Collins, S. 1990. ‘On the Very Idea of the Pali Canon’. Journal of Pali Text Society
15: 89–126.
Comfort, A. 1964. The Koka Shastra: Being the Ratirahasya of kokkoka, and Other
Mediaeval Indian Writings on Love. Translated by A. Comfort. London: Tandem Books.
Connolly, P. 1997. ‘The Vitalistic Antecedents of the Ātman-Brahman Concept’,
in Connolly and Hamilton 1997, pp. 21–38.
Connolly, P. and Hamilton, S. (eds.) 1997. Indian Insights: Buddhism, Brahmanism
and Bhakti. Papers from the Annual Spalding Symposium on Indian Religions.
London: Luzac Oriental.
Connor, L. H. 1995. ‘Acquiring Invisible Strength: A Balinese Discourse of Harm
and Well-Being’. Indonesia Circle 66: 124–53.
Conze, E. 1962. Buddhist Thought in India: Three Phases of Buddhist Philosophy.
London: George Allen and Unwin Ltd.
Conze, E. 1973. The Short Prajñāpāramitā Texts. London: Luzac & Co.
Conze, E. 1978. The Prajñāpāramitā Literature. Second rev. and enlarged edn.
Tokyo: Reiyukai. (Bibliographia Philologica Buddhica. Series Maior; 1.)
References
361
Coomaraswamy, A. Kentish 1928–31. Yakshas. Parts I and II. Washington, DC:
Smithsonian Institution.
Coomaraswamy, A. Kentish 1956. La sculpture de Bharhut. Translation J. Buhot.
Paris: Vaneost. (Annales du Musée Guimet, Bibliotheque d’Art, nouvelle
série, 6.)
Cort, J. E. 1987. ‘Medieval Jaina Goddess Traditions’. Numen 34: 235–
55.
Cort, J. E. 2001. Jains in the World: Religious Values and Ideology in India. New
York: Oxford University Press.
Courtright, P. 1985. Gan.eśa: Lord of Obstacles, Lord of Beginnings. New York and
Oxford: Oxford University Press.
Cousins, L. S. 1973. ‘Buddhist Jhāna: Its Nature and Attainment According to the
Pali Sources’. Religion 3: 115–31.
Cousins, L. S. 1983. ‘Pali Oral Literature’, in Buddhist Studies: Ancient and Modern,
Denwood and Piatigorsky (eds.), 1–11. London: Curzon.
Cousins, L. S. 1984. ‘Samatha-yāna and Vipassanā-yāna’, in Buddhist Studies in Honour of Hammalava Saddhātissa, Dhammapāla, Gombrich and
Norman (eds.), pp. 55–68. Nugegoda, Sri Lanka: Buddhist Research Library
Trust.
Cousins, L. S. 1991. The ‘Five Points’ and the Origin of the Buddhist Schools’,
The Buddhist Forum Vol. II, Skorupski (ed.), pp. 27–60. London: School of
Oriental and African Studies.
Cousins, L. S. 1992. ‘Vitakka/Vitarka and Vicara: Stages of Samādhi in Buddhism
and Yoga’. Indo-Iranian Journal 35: 137–57.
Cousins, L. S. 1996a. ‘The Dating of the Historical Buddha: A Review Article’.
Journal Royal Asiatic Society Series 3, 6: 57–63. Available as e-text
from www.ucl.ac.uk/˜ucgadkw/position/buddha/buddha.html (downloaded
3 Aug 2006).
Cousins, L. S. 1996b. ‘The Origins of Insight Meditation’, in The Buddhist Forum
Vol. IV: Seminar Papers 1994–1996, Skorupski (ed.), pp. 35–58. London: School
of Oriental and African Studies.
Cousins, L. S. 1997. ‘Aspects of Esoteric Southern Buddhism’, in Connolly and
Hamilton 1997, pp. 185–207. London: Luzac Oriental.
Cousins, L. S. 2005. ‘New Discoveries and Old Baskets: Perspectives on the Development of Pali and Sanskrit Buddhist Canons in the Light of the Recently
Discovered Sanskrit Dı̄rghāgama’. Numata Lectures, School of Oriental and
African Studies, University of London.
Couture, A. and Schmid, C. 2001. ‘The Harivam
. śa, the Goddess Ekānam
. śa, and
the Iconography of the Vr.s.n.i Triad’. Journal American Oriental Society 121:
173–92.
Couture, A. 2003. ‘Kr.s.n.a’s Victory Over Bān.a and Goddess Kot.avı̄’s Manifestation
in the Harivam
. śa’. Journal of Indian Philosophy 31: 593–620.
Crangle, E. Fitzpatrick 1994. The Origin and Development of Early Indian Contemplative Practices. Wiesbaden: Otto Harrassowitz.
362
References
Cribb, J. 1985. ‘Dating India’s Earliest Coins’, in South Asian Archaeology 1983,
Schotsmans and Taddei (eds.), vol. 1, pp. 535–54. Naples: Istituto Universitario
Orientale, Naples.
Cribb, J. 1997. ‘Śiva images on Kushan and Kushano-Sassanian Coins’, in Studies
in Silk Road Coins and Culture, Tanabe, Cribb and Wang (eds.), pp. 11–66.
(Special volume of Silk Road Art and Archaeology 1997)
Curtis, V. Sarkhosh and Stewart, S. (eds.) 2005. Birth of the Persian Empire, Vol. I.
London and NY: I. B. Tauris.
Czuma, Stanislaw J. 1985. Kushan Sculpture: Images from Early India. With the
assistance of Rekha Morris. Cleveland, Ohio: Cleveland Museum of Art in
cooperation with Indiana University Press.
Dalton, J. 2004. ‘The Development of Perfection: The Interiorization of Buddhist
Ritual in the Eight and Ninth Centuries’. Journal of Indian Philosophy 32:
1–30.
Dalton, J. 2005. ‘A Crisis of Doxography: How Tibetans Organized Tantra during
the 8th –12th Centuries’. Journal International Association of Buddhist Studies
28: 115–81.
Daniel, E. Valentine 1984. ‘Pulse as Icon in Siddha Medicine’, in Daniel and Pugh
1984, pp. 115–26.
Daniel, E. Valentine and Pugh, J. F. (eds.) 1984 South Asian Systems of Healing.
Leiden: E. J. Brill. (Contributions to Asian Studies, 18.)
Daniélou, A. 1964. Hindu Polytheism. London: Routledge and Kegan Paul.
Dar, Saifur Rahman 1993. ‘Dating the Monuments of Taxila’, Spodek and
Srinivasan 1993, pp. 103–22.
Dass, M. I. and Willis, M. 2002. ‘The Lion Capital from Udayagiri and the
Antiquity of Sun Worship in Central India’, South Asian Studies 18: 25–45.
Das, R. P. 1992. ‘Problematic Aspects of the Sexual Rituals of the Bauls of Bengal’.
Journal American Oriental Society 112: 388–432.
Das, V. 1987. Structure and Cognition: Aspects of Hindu Caste and Ritual. 2nd edn.
Delhi: Oxford University Press.
Davidson, R. M. 1991. ‘Reflections on the Maheśvara Subjugation Myth: Indic
Materials, Sa-skya-pa Apologetics, and the Birth of Heruka’. Journal of the
International Association of Buddhist Studies 14/2: 197–235.
Davidson, R. M. 1995a. ‘Atiśa’s A Lamp For the Path to Awakening’, in Buddhism
in Practice, Lopez, Jr. (ed.), pp. 290–301. Princeton, NJ: Princeton University
Press.
Davidson, R. M. 1995b. ‘The Bodhisattva Vajrapān.i’s Subjugation of Śiva’, in
Religions of India in Practice, in Lopez, Jr. 1995, pp. 547–55. Princeton, NJ:
Princeton University Press.
Davidson, R. M. 2002a. Indian Esoteric Buddhism: A Social History of the Tantric
Movement. New York: Columbia University Press. [References to Davidson
2002 are to this.]
Davidson, R. M. 2002b. ‘Hidden Realms and Pure Abodes: Central Asian
Buddhism as Frontier Religion in the Literature of India, Nepal, and Tibet’.
Pacific World 3rd series, 4 (Fall 2002): 153–81.
References
363
Davis, R. 1984. Muang Metaphysics. Bangkok: Pandora Press.
de Michelis, E. 2004. A History of Modern Yoga: Patañjali and Western Esotericism.
London and New York: Continuum.
de Sales, A. 1991. Je suis né de vos jeux de tambours: La religion chamanique des Magar
du nord. Nanterre: Société d’Ethnologie.
de Silva, L. 1981. Paritta: A Historical and Religious Study of the Buddhist Ceremony
for Peace and Prosperity in Sri Lanka. Columbo: National Museums of Sri
Lanka. (= Spolia Zeylanica. Bulletin of the National Museums of Sri Lanka, vol.
36 part 1.)
de Visser, M. W. 1928–35. Ancient Buddhism in Japan: Sutras and Ceremonies in
Use in the Seventh and Eighth Centuries A.D. and their History in Later Times.
Paris: P. Geuthner. (Buddhica. Documents et travaux pour l’Étude du bouddhisme . . . 1. sér.: Mémoires, t. 3–4.)
DeCaroli, R. 2004. Haunting the Buddha: Indian Popular Religions and the Formation of Buddhism. Oxford and New York: Oxford University Press.
Deeg, M. 1993. ‘Shamanism in the Veda: the Keśin-Hymn (10.136), the Journey to
Heaven of Vasis..tha (R.V.7.88) and the Mahāvrata-Ritual’. Nagoya Studies in
Indian Culture and Buddhism (Sam
. bhās.ā) 14: 95–144.
Deeg, M. 2006. ‘Aryan National Religion(s) and the Criticism of Asceticism and
Quietism in the Nineteenth and Twentieth Centuries’, in Asceticism and
Its Critics: Historical Accounts and Comparative Perspectives, Freiberger (ed.),
pp. 61–87. Oxford and New York: Oxford University Press.
Dehejia, V. 1979. Early Stone Temples of Orissa. New Delhi: Vikas Publishing House.
Dehejia, V. 1986. Yoginı̄ Cult and Temples: A Tantric Tradition. New Delhi: National
Museum.
Desai, D. 1990. ‘12.6 Mother Goddess’. Encyclopedia of Indian Archaeology, ed.
A. Ghosh, pp. 267–9. Leiden: E. J. Brill.
Desani, G. V. 1973. ‘Mostly Concerning Kama and Her Immortal Lord’. Indian
Horizons 32: 1–44.
Deshpande, M. M. and Hook, P. E. (eds.) 1979. Aryan and Non-Aryan in India.
Ann Arbor, MI: University of Michigan, Center for South and Southeast
Asian Studies.
Deshpande, M. M. 1995. ‘Vedic Aryans, Non-Vedic Aryans, and Non-Aryans:
Judging the Linguistic Evidence of the Veda’, in Erdosy 1995a, pp. 67–
84.
Devika, J. 2005. ‘The Aesthetic Woman: Re-Forming Female Bodies and Minds
in Early Twentieth-Century Keralam’. Modern Asian Studies 39: 461–87.
Dhyansky, Y. Y. 1987. ‘The Indus Valley Origin of a Yoga Practice’. Artibus Asiae
48: 89–108.
Diehl, C. G. 1956. Instrument and Purpose: Studies on Rites and Rituals in South
India. Lund: Gleerup.
Dimock, E. C., Jr. 1982. ‘A Theology of the Repulsive: The Myth of the Goddess
Śı̄talā’, in Hawley and Wulff 1982, pp. 184–203.
Divakaran, O. 1984. ‘Durgā the Great Goddess: Meanings and Forms in the Early
Period’, in Meister 1984, pp. 271–89.
364
References
Dollfus, P. 1989. Lieu de neige et de genévriers: Organisation sociale et religieuse des
communautés bouddhistes du Ladakh. Paris, Éditions du CNRS.
Donaldson, T. 1987. Kamadeva’s Pleasure Garden: Orissa. Delhi: B. R. Publishing
Corporation.
Dorje, G. 1987. The Guhyagarbhatantra and its XIVth Century Commentary Phyogsbcu Mun-sel. PhD Dissertation, School of Oriental and African Studies,
University of London.
Dowman, K. 1985. Masters of Mahamudra: Songs and Histories of the Eighty-Four
Buddhist Siddhas. Albany, NY: SUNY Press.
Dreyfus, G. 1998. ‘The Shuk-Den Affair: History and Nature of a Quarrel. Journal
International Association of Buddhist Studies 21: 227–70.
Dumont, L. 1960. ‘World Renunciation in Indian Religions’. Contributions to
Indian Sociology 9: 67–89. (Reprinted in Dumont 1970, pp. 33–60.)
Dumont, L. 1970. Religion, Politics and History in India. Paris and Leiden: Mouton.
Dumont, L. 1972. Homo Hierarchicus: the Caste System and its Implications. London:
Granada.
Dumont, P. É. 1927. L’asvamedha, description du sacrifice solennel du cheval dans le
culte védique d’après les textes du Yajurveda blanc. Paris: P. Geuthner.
Dundas, P. 1991. ‘The Digambara Jain Warrior’, in Carrithers and Humphrey 1991,
pp. 169–86.
Dundas, P. 2002. The Jains. Second edn. London and New York: Routledge.
Dupuche, J. R. 2003. Abhinavagupta: The Kula Ritual As Elaborated in Chapter 29
of the Tantrāloka. Delhi: Motilal Banarsidass.
Durrenberger, E. P. and Tannenbaum, N. 1989. ‘Continuities in Highland and
Lowland Regions of Thailand’. Journal of the Siam Society 77: 83–90.
Dyczkowski, M. S. G. 1989. The Canon of the Śaivāgama and the Kubjikā Tantras
of the Western Kaula Tradition. Delhi: Motilal Banarsidass.
Ehrich, R. W. (ed.) 1992. Chronologies in Old World Archaeology. Chicago: University of Chicago Press.
Eliade, M. 1958. Yoga, Immortality and Freedom. Translated from the French by
W. R. Trask. New York: Pantheon Books. (Bollingen Series, 56.)
Elwin, V. 1947. The Muria and their Ghotul. London: Oxford University Press
Indian Branch.
Emigh, J. 1984. ‘Dealing with the Demonic: Strategies for Containment in Hindu
Iconography and Performance’. Asian Theatre Journal 1: 21–39.
Emmerick, R. E. 1970. The Sūtra of Golden Light. London: Luzac and Co.
English, E. 2002. Vajrayoginı̄: Her Visualizations, Rituals, and Forms: A Study of the
Cult of Vajrayoginı̄ in India. Boston: Wisdom.
Erdosy, G. (ed.) 1995a. The Indo-Aryans of Ancient South Asia: Language, Material Culture and Ethnicity. Berlin and New York: Walter de Gruyter. (Indian
Philology and South Asian Studies, ed. Albrecht Wezler and Michael Witzel,
Vol. 1.)
Erdosy, G. 1995b. ‘The Prelude to Urbanization: Ethnicity and the Rise of Late
Vedic Chiefdoms’, in F. R. Allchin 1995a, pp. 75–98.
References
365
Erdosy, G. 1995c. ‘City States of North India and Pakistan at the Time of the
Buddha’, in F. R. Allchin 1995a, pp. 99–122.
Eschmann, A., Kulke, H. and Tripathi, G. Charan (eds.) 1986. The Cult of Jagannath
and the Regional Tradition of Orissa. Second printing. Delhi: Manohar. (South
Asia Institute. New Delhi Branch. Heidelberg University. South Asian Studies
No. 8.)
Evans-Pritchard, E. E. 1940. The Nuer: a Description of the Modes of Livelihood and
Political Institutions of a Nilotic People. Oxford: Clarendon Press.
Evans-Pritchard, E. E. 1956. Nuer Religion. Oxford: Clarendon Press.
Fabri, C. 1994. Indian Dress: A Brief History. New Delhi: Orient Longman (Disha
Books.)
Falk, H. 1986. Bruderschaft und Würfelspiel: Untersuchungen zur Entwicklungsgeschichte des Vedischen Opfers. Freiburg: Hedwig Falk.
Farmer, S., Sproat, R., and Witzel, M. 2004. ‘The Collapse of the IndusScript Thesis: The Myth of a Literate Harappan Civilization’. Electronic
Journal Vedic Studies 11(2) (13 Dec. 2004): 19–57. Downloaded from
www1 shore.net/˜india/ejvs/issues.html, 22 Sept. 2006.
Farrow, G. W. and Menon, I. 1992. The Concealed Essence of the Hevajra Tantra
With the Commentary, Yogaratnamālā. Delhi: Motilal Banarsidass.
Felten, W. and Lerner, M. 1989. Thai and Cambodian Sculpture from the 6th to the
14th Centuries. London: P. Wilson Publishers.
Fenner, T. 1979. Rasāyana Siddhi: Medicine and Alchemy in the Buddhist Tantras.
PhD Dissertation, University of Wisconsin, Madison. UMI 80–08814.
Ferguson, J. P. and E. M. Mendelson 1981. ‘Masters of the Occult: The Burmese
Weikzas’. Contributions to Asian Studies 16: 62–80.
Ferreira-Jardim, A. 2005. ‘“Separated from Desires . . .”: New Light on the Historical Development of the Four Meditations [dhyāna] Pericope Amongst
Śrāman.ical Groups in Ancient India’. Paper given at the 14th Conference of
the International Association of Buddhist Studies, School of Oriental and
African Studies, University of London, 29 Aug.–3 Sept. 2005.
Ferreira-Jardim, A. 2006. ‘Some Notes Towards a History of Early Buddhist and
Jaina Meditation: vitakka, viyāra/-i and Terms Referring to Mental OnePointedness (egatta, egaggaman.a, etc.)’. Paper for the Australasian Association
of Buddhist Studies conference, University of Sydney, 16–17 June 2006.
Filliozat, J. 1969. ‘Taoisme et Yoga’. Journal Asiatique 257: 41–87.
Finn, L. M. 1986. The Kulacūd.āman.i Tantra and The Vāmakeśvara Tantra with
the Jayaratha Commentary introduced, translated and annotated by Louise M.
Finn. Wiesbaden: Otto Harrassowitz.
Fitzgerald, J. L. 2001. ‘Making Yudhis.t.hira the King: the Dialectics and the Politics
of Violence in the Mahābhārata’. Rocznik Orientalistyczny 54: 63–92.
Fitzgerald, J. L. 2002. ‘The Rāma Jāmadagnya “Thread” of the Mahābhārata: A
New Survey of Rāma Jāmadagnya in the Pune Text’, in Stages and Transitions:
Temporal and Historical Frameworks in Epic and Purān.ic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics
366
References
and Purān.as, August 1999, Brockington (ed.), pp. 89–132. Zagreb: Croatian
Academy of Sciences and Arts.
Fitzgerald, J. L. 2004. ‘The Many Voices of the Mahābhārata’. Journal American
Oriental Society 123: 803–18.
Flood, G. 1996. An Introduction to Hinduism. Cambridge: Cambridge University
Press.
Flood, G. 1997. ‘Ritual Dance in Kerala: Performance, Possession, and the Formation of Culture’, in Connolly and Hamilton 1997, pp. 169–83.
Flood, G. (ed.) 2003. The Blackwell Companion to Hinduism. Oxford: Blackwell.
Flood, G. 2006. The Tantric Body: The Secret Tradition of Hindu Religion. London
and New York: I. B. Tauris.
Fluegel, P. 2005. ‘The Invention of Jainism: A Short History of Jaina Studies’.
International Journal of Jain Studies 1: 1–14.
Foucault, M. 1988a. ‘Technologies of the Self ’. In Technologies of the Self: A Seminar
with Michel Foucault, ed. by L. H. Martin, H. Gutman and P. H. Hutton,
pp. 16–49. Amherst, MA: University of Massachusetts Press.
Foucault, M. 1988b. The Care of the Self: History of Sexuality vol. 3. New York:
Random House.
Fox, R. G. 1996. ‘Communalism and Modernity’, in Making India Hindu: Religion,
Community, and the Politics of Democracy in India, Ludden (ed.), pp. 235–49.
Delhi: Oxford University Press.
Freed, S. A. and Freed, R. S. 1998. Hindu Festivals in a North Indian Village. American Museum of Natural History, distributed by University of
Washington Press, Seattle. (Anthropological Papers of the American Museum
of Natural History, 81.)
Freeman, J. R. 1993. ‘Performing Possession: Ritual and Consciousness in the
Teyyam Complex of Northern Kerala’, in Brückner, Lutze, and Malik (eds.),
pp. 109–38.
Freeman, J. R. 1994. ‘Possession Rites and the Tantric Temple: A Case-Study from
Northern Kerala’. Diskus [Electronic journal of religious studies.] vol. 2 no. 2
(Autumn 1994).
Freeman, J. R. 1999. ‘Gods, Groves and the Culture of Nature in Kerala’. Modern
Asian Studies 33: 257–302.
Fremantle, F. 1971. A Critical Study of the Guhyasamāja Tantra. PhD Dissertation,
School of Oriental and African Studies, University of London.
Fremantle, F. 1990. ‘Chapter Seven of the Guhyasamāja Tantra’, in Skorupski 1990,
pp. 101–114.
Fruzzetti, L. M. 1990. The Gift of a Virgin. Delhi: Oxford University Press.
Fuller, C. J. 1992. The Camphor Flame: Popular Hinduism and Society in India.
Princeton, NJ: Princeton University Press.
Fulton, R. M. 1972. ‘The Political Structures and Functions of Poro in Kpelle
Society’. American Anthropologist 74: 1218–33.
Fussman, G. 1993. ‘Taxila: The Central Asian Connection’, in Urban Form and
Meaning in South Asia: The Shaping of Cities from Prehistoric to Precolonial
References
367
Times, Spodek and Meth Srinivasan (eds.), 83–100. Washington: National
Gallery of Art.
Fynes, R. C. C. 1991. Cultural Transmission between Roman Egypt and Western India.
D.Phil. dissertation, Oxford University.
Fynes, R. C. C. 1993. ‘Isis and Pattinı̄: The Transmission of a Religious Idea from
Roman Egypt to India’. Journal Royal Asiatic Society, Series 3, 3: 377–91.
Fynes. R. C. C. 1995. ‘The Religious Patronage of the Sātavāhana Dynasty’. South
Asian Studies 11: 43–50.
Gadon, E. W. 1997. ‘The Hindu Goddess Shasthi, Protector of Children and
Women’, in From the Realm of the Ancestors: An Anthology in Honor of Marija
Gimbutas, Marler (ed.), pp. 293–308. Manchester, CT: Knowledge, Ideas &
Trends.
Galaty, J. G. 1979. ‘Pollution and Pastoral Antipraxis: The Issue of Maasai Inequality’. American Ethnologist 6 (1979): 803–16.
Galvão, A. 1971. A Treatise on the Moluccas (c. 1544), probably the Preliminary Version
of Antonio Galvão’s lost Historı́a das Molucas. Trans. H. Jacobs, S. J. Rome:
Jesuit Historical Institute.
Gäng, P. 1988. Das Tantra der Verborgenen Vereinigung = Guhyasamāja-Tantra.
München: Diederichs.
Geertz, C. 1975. ‘Religion as a Cultural System’, in C. Geertz, The Interpretation
of Cultures, pp. 87–125. London: Hutchinson.
Geiger, W. 1964. The Mahāvam
. sa or the Great Chronicle of Ceylon. Translated into
English by Wilhelm Geiger. London: Published for the Pali Text Society by
Luzac & Co.
Gell, A. 1982. ‘The Market Wheel: Symbolic Aspects of an Indian Tribal Market’.
Man 17: 470–91.
Gellner, D. N. 1992. Monk, Householder, and Tantric Priest: Newar Buddhism and
its Hierarchy of Ritual. Cambridge: Cambridge University Press.
Gellner, D. N. and Quigley, D. (eds.) 1995. Contested Hierarchies: a Collaborative Ethnography of Caste among the Newars of the Kathmandu Valley, Nepal.
Oxford: Clarendon Press; New York: Oxford University Press.
George, C. S. 1974. The Candamahārosana Tantra. A Critical Edition and English
Translation. Chapters I–VIII. American Oriental Society, New Haven. (American Oriental Series, 56.)
Germano, D. 1994. ‘Architecture and Absence in the Secret Tantric History of the
Great Perfection (rdzogs chen)’. Journal International Association of Buddhist
Studies 17: 203–335.
Germano, D. 2005. ‘The Funerary Transformation of the Great Perfection (Rdzogs
chen)’. Journal International Association of Tibetan Studies 1 (October 2005):
1–54. Downloaded from www.thdl.org, 19 Aug 2006.
Gibson, T. 1997. ‘Inner Asian Contributions to the Vajrayāna’. Indo-Iranian Journal
40: 37–57.
Gitomer, D. 1992. ‘King Duryodhana: The Mahābhārata Discourse of Sinning and
Virtue in Epic and Drama’. Journal American Oriental Society 112: 222–32.
368
References
Gnoli, G. and Lanciotti, L. (eds.) 1988. Orientalia Iosephi Tucci Memoriae Dicata,
Roma: IsMEO (Serie Orientale Roma, LVI, 3.)
Gold, A. Grodzins 1993. A Carnival of Parting: The Tales of King Bharthari and King
Gopi Chand as Sung and Told by Madhu Natisar Nath of Ghatiyali, Rajasthan.
Delhi: Munshiram Manoharlal.
Gold, D. and Gold, A. Grodzins Gold 1984. ‘The Fate of the Householder Nath’.
History of Religions 24: 113–32.
Goldman, R. P. 1977. Gods, Priests and Warriors: the Bhr.gus of the Mahābhārata.
New York: Columbia University Press.
Goldman, R. P. 2005. Rāmāyan.a Book One: Boyhood, by Vālmı̄ki. Translated by
R. P. Goldman. New York: New York University Press and JJC Foundation.
Gombrich, R. F. 1971. Precept and Practice: Traditional Buddhism in the Rural
Highlands of Sri Lanka. Oxford: Clarendon Press.
Gombrich, R. F. 1988. Theravāda Buddhism: A Social History from Ancient Benares
to Modern Colombo. London: Routledge.
Gombrich, R. F. 1990. ‘Recovering the Buddha’s Message’, in Earliest Buddhism
and Madhyamaka, Seyfort Ruegg and Schmithausen (eds.), 5–23. Leiden and
New York: E. J. Brill.
Gombrich, R. F. 1992. ‘The Buddha’s Book of Genesis?’ Indo-Iranian Journal 35:
159–78.
Gombrich, R. F. 1996. How Buddhism Began: The Conditioned Genesis of the Early
Teachings. London: Athlone Press.
Gombrich, R. 1997. ‘The Buddhist Attitude to Thaumaturgy’, in
Bauddhavidyāsudhākarah.: Studies in Honour of Heinz Bechert on the Occasion
of his 65th Birthday, Kieffer-Puelz and Hartmann (eds.). Swisttal-Odendorf:
Indica et Tibetica Verlag.
Gombrich, R. 2003. ‘“Obsession with Origins”: Attitudes to Buddhist Studies in
the Old World and the New’, in Approaching the Dhamma: Buddhist Texts and
Practices in South and Southeast Asia, Blackburn and Samuels (eds.). Onalaska,
WA: Pariyatti.
Gombrich, R. and G. Obeyesekere 1990. Buddhism Transformed: Religious Change
in Sri Lanka. Delhi: Motilal Banarsidass.
Gonda, J. 1961. ‘Ascetics and Courtesans’. Adyar Library Bulletin 25: 78–102.
Reprinted in Gonda 1975 vol. 4.
Gonda, J. 1963. The Vision of the Vedic Poets. The Hague: Mouton.
Gonda, J. 1975. Selected Studies. Presented to the Author by the Staff of the Oriental
Institute, Utrecht University, on the Occasion of his 70th Birthday. Leı̄den: Brill.
Gonda, J. 1980. ‘The Śatarudriya’, in Sanskrit and Indian Studies: Essays in Honour
of Daniel H. H. Ingalls, Nagatomi, Matilal, Masson and Dimock, Jr. (eds.),
pp. 75–91. Dordrecht, Boston and London: D. Reidel.
Good, A. 1991. The Female Bridegroom: A Comparative Study of Life-Crisis Rituals
in South India and Sri Lanka. Oxford: Clarendon Press.
Goodison, L. and Morris, C. (eds.) 1998. Ancient Goddesses: The Myths and the
Evidence. London: British Museum Press.
References
369
Gottschalk, P. 2000. Beyond Hindu and Muslim: Multiple Identities in Narratives
from Village India. Delhi and Oxford: Oxford University Press.
Goudriaan, T. 1973. ‘Tumburu and His Sisters’, Wiener Zeitschrift für die Kunde
Südasiens 17: 49–95.
Goudriaan, T. 1985. The Vı̄n.āśikha Tantra. A Śaiva Tantra of the Left Current Edited
with an Introduction and a Translation. Delhi: Motilal Banarsidass.
Goudriaan, T. (ed.) 1992. Ritual and Speculation in Early Tantrism: Studies in Honor
of André Padoux. Albany, NY: State University of New York Press.
Gray, D. Barton 2001. On Supreme Bliss: A Study of the History and Interpretation
of the Cakrasam
. vara Tantra. PhD Dissertation, Columbia University, New
York. UMI AAT 9998161.
Gray, D. Barton 2007. The Cakrasamvara Tantra: A Study and Annotated Translation. New York: Columbia University Press for American Institute of Buddhist
Studies.
Grayson, J. Huntley 1989. Korea: a Religious History. Oxford: Clarendon Press.
Grenet, F. 2005. ‘An Archaeologist’s Approach to Avestan Geography’, in Curtis
and Stewart 2005, pp. 29–51.
Griffith, R. T. H. 1916. The Hymns of the Atharva-Veda Translated with a Popular
Commentary. 2 vols. Second edn. Benares: E. J. Lazarus & Co.
Griffiths, P. 1981. ‘Concentration or Insight: The Problematic of Theravada
Buddhist Meditation-Theory’. Journal of the American Academy of Religion
49: 605–24.
Grimal, F. (ed.) 2001. Les sources et le temps/ Sources and Time: A Colloquium,
Pondicherry, 11–13 January 1997. Pondicherry: Institut Francais de Pondichery:
École Français D’Extrême-Orient.
Guenther, H. V. 1963. The Life and Teaching of Nāropa. Oxford: Clarendon Press.
Guenther, H. V. 1969. Yuganaddha: The Tantric View of Life. Varanasi: The
Chowkhamba Sanskrit Series Office. (Chowkhamba Sanskrit Series, 3.)
Guenther, H. V. 1974. ‘Early Forms of Tibetan Buddhism’. Crystal Mirror 3: 80–92.
Guillon, E. 2001. Cham Art: Treasures from the Da Nang Museum, Vietnam.
Bangkok: River Books.
Gupta, E. M. 1983. Brata und Alpana in Bengalen. Wiesbaden: Steiner. (Beiträge
zur Südasienforschung 80.)
Gupta, S. 1972. Laks.mı̄ Tantra: A Pāñcarātra Text. Translation and Notes. Leiden:
Brill.
Gupta, S. 1989. ‘The Pāñcarātra Attitude to Mantra’, in Mantras, Alper (ed.),
pp. 224–48. Albany, NY: State University of New York Press.
Gupta, S. 1992. ‘Yoga and Antaryāga in Pāñcarātra’, in Goudriaan 1992, pp. 175–
208.
Gupta, S. 1999. ‘Hindu Woman, the Ritualist’, in Tambs-Lyche 1999, pp. 88–99.
Gupta, S. 2003. ‘The Domestication of a Goddess: Caran.a-tı̄rtha Kālı̄ghāt., the
Mahāpı̄t.ha of Kālı̄’, in McDermott and Kripal 2003, pp. 60–79.
Gupta, S. and Gombrich, R. 1986. ‘Kings, Power and the Goddess’. South Asia
Research 6: 123–38.
370
References
Gutschow, N. 1982. Stadtraum und Ritual der Newarischen Städte im
Kāthmāndu-Tal: eine Architekturanthropologische Untersuchung. Stuttgart:
W. Kohlhammer.
Gutschow, N. 1993. ‘Bhaktapur: Sacred Patterns of a Living Urban Tradition’, in
Spodek and Srinivasan 1993, pp. 163–83.
Gutschow, N. and Bāsukala, G. Mān 1987. ‘The Navadurgā of Bhaktapur:
Spatial Implications of an Urban Ritual’. In Heritage of the Kathmandu
Valley, Gutschow and Michaels (eds.), pp. 135–66. Sankt Augustin: VGH
Wissenschaftsverlag.
Gutschow, N. 1996. ‘The As.t.amātr.kā and Navadurgā of Bhaktapur’, in Michaels,
Vogelsanger and Wilke 1996, pp. 191–216.
Haaland, G. and Haaland, R. 1995. ‘Who Speaks the Goddess’s Language? Imagination and Method in Archaeological Research’, Norwegian Archaeological
Review 28: 105–21.
Hall, K. R. 1999. ‘Economic History of Early Southeast Asia’, in Tarling 1999,
pp. 183–275.
Hallisey, C. 1995. ‘Roads Taken and Not Taken in the Study of Theravāda
Buddhism’, in Curators of the Buddha: The Study of Buddhism under Colonialism, Lopez (ed.), pp. 31–61.
Hamilton, S. 1996. Identity and Experience: the Constitution of the Human Being in
Early Buddhism. London: Luzac Oriental.
Hamilton, S. 2000. ‘The Centrality of Experience in the Teachings of Early
Buddhism’. Oxford: Religious Experience Research Centre, Westminster
Institute of Education. (2nd Series Occasional Paper 24.)
Hanks, L. M. 1975. ‘The Thai Social Order as Entourage and Circle’, in Change
and Persistence in Thai Society, Skinner and Kirsch (eds.), pp. 197–218. Ithaca,
NY: Cornell University Press.
Hanssen, K. 2002. ‘Ingesting Menstrual Blood: Notions of Health and Bodily
Fluids in Bengal’. Ethnology 41: 365–79.
Haque, E. 2001. Chandraketugarh: A Treasure-House of Bengal Terracottas. Dhaka:
The International Centre for the Study of Bengal Art. (Studies in Bengal
Art, 4.)
Hardy, F. 1983. Virāha-Bhakti: the Early History of Kr..sn.a Devotion in South India.
Delhi: Oxford University Press.
Harlan, L. 1992. Religion and Rajput Women: the Ethic of Protection in Contemporary
Narratives. Berkeley and Oxford: University of California Press.
Harle, J. C. 1974. Gupta Sculpture: Indian Sculpture of the Fourth to the Sixth
centuries A.D. Oxford: Clarendon Press.
Harle, J. C. and Topsfield, A. 1987. Indian Art in the Ashmolean Museum. Oxford:
Ashmolean Museum.
Harman, W. P. 1992. Sacred Marriage of a Hindu Goddess. Delhi: Motilal
Banarsidass.
Harner, M. 1982. The Way of the Shaman. New York: Bantam Books (Harper and
Row).
References
371
Harper, D. 1987. ‘The Sexual Arts of Ancient China as Described in a Manuscript
of The Second Century B.C’. Harvard Journal Asiatic Studies 47: 539–92.
Harper, D. 1997. Early Chinese Medical Literature: The Mawangdui Medical
Manuscripts. London: Kegan Paul International; New York: Columbia
University Press.
Harper, K. A. 2002. ‘The Warring Śaktis: A Paradigm for Gupta Conquests’, in
Harper and Brown 2002, pp. 115–31.
Harper, K. A. and Brown, R. L. (eds.) 2002. The Roots of Tantra. Albany, NY: State
University of New York Press.
Harrison, P. 1987a. ‘Who Gets to Ride in the Great Vehicle? Self-Image and Identity
among the Followers of the Early Mahāyana’. Journal International Association
of Buddhist Studies 10: 67–89.
Harrison, P. 1987b. ‘Buddhism: A Religion of Revelation After All’. Numen 34:
256–64.
Harrison, P. 1990. The Samādhi of Direct Encounter with the Buddhas of
the Present. An Annotated English Translation of the Tibetan Version of
the Pratyutpanna-Buddha-Sam
. mukhāvasthita-Samādhi-Sūtra. Tokyo: The
International Institute for Buddhist Studies. (Studia Philologica Buddhica.
Monograph Series, V.)
Harrison, P. 1993. ‘The Earliest Chinese Translations of Mahāyāna Buddhist Sutras:
Some Notes on the Works of Lokaks.ema’, Buddhist Studies Review 10(2): 135–
77.
Hart, G. L. 1975. The Poems of Ancient Tamil: Their Milieu and their Sanskrit
Counterparts. Berkeley, Los Angeles and London: University of California
Press.
Hart, G. L. 1979. ‘The Nature of Tamil Devotion’, in Deshpande and Hook 1979,
pp. 24–33.
Hart, G. L. 1987. ‘Early Evidence for Caste in South India’, in Dimensions of Social
Life: Essays in Honor of David G. Mandelbaum, Hockings (ed.), pp. 467–91.
Berlin: Mouton-de Gruyter.
Hartzell, J. F. 1997. Tantric Yoga: A Study of the Vedic Precursors, Historical Evolution, Literatures, Cultures, Doctrines, and Practices of the 11th Century Kaśmı̄ri
Śaivite and Buddhist Unexcelled Tantric Yogas. PhD Dissertation, Columbia
University. UMI No. 9723798.
Hawley, J. S. and Wulff, D. M. (eds.) 1982. The Divine Consort: Rādhā and the
Goddesses of India. Boston: Beacon Press.
Hawley, J. S. and Wulff, D. M. (eds.) 1996. Devı̄: Goddesses of India. Berkeley and
Los Angeles: University of California Press.
Hayes, G. A. 2003. ‘Metaphoric Worlds and Yoga in the Vais.n.ava Sahajiyā Traditions of Medieval Bengal’, in Whicher and Carpenter 2003,
pp. 162–84.
Heesterman, J. C. 1962. ‘Vrātya and Sacrifice’. Indo-Iranian Journal 6: 1–37.
Heesterman, J. C. 1985. The Inner Conflict of Tradition: Essays in Indian Ritual,
Kingship, and Society. Chicago: University of Chicago Press.
372
References
Heesterman, J. C. 1993. The Broken World of Sacrifice: an Essay in Ancient Indian
Ritual. Chicago: University of Chicago Press.
Heilijgers-Seelen, D. 1994. The System of Five Cakras in Kubjikāmatatantra 14–16.
Groningen: Egbert Forsten. (Groningen Oriental Studies, 9.)
Hein, N. 1982. ‘Radha and Erotic Community’, in Hawley and Wulff 1982, pp. 116–
24. Delhi: Motilal Banarsidass.
Hein, N. 1986a. ‘Epic Sarvabhūtahite Ratah.: A Byword of Non-Bhārgava Editors’.
Annals of the Bhandarkar Oriental Research Institute 67: 17–34.
Hein, N. 1986b. ‘A Revolution in Kr.s.n.aism: The Cult of Gopāla’. History of
Religions 25: 296–317.
Heitzman, J. 1984. ‘Early Buddhism, Trade and Empire’, in Studies in the Archaeology and Paleaoanthropology of South Asia, Kennedy and Possehl (eds.), pp. 121–
37. New Delhi: Oxford and IBH.
Heller, A. 1994. ‘Early Ninth Century Images of Vairochana from Eastern Tibet’.
Orientations 25(6): 74–9.
Herrmann-Pfandt, A. 1992. D
. ākinı̄s: zur Stellung und Symbolik des Weiblichen im
tantrischen Buddhismus. Bonn: Indica et Tibetica-Verlag.
Higham, C. F. W. 1999. ‘Recent Advances in the Prehistory of South-East Asia’,
in World Prehistory: Studies in Memory of Grahame Clark, Coles, Bewley
and Mellars (eds.), pp. 75–86. Oxford: Oxford University Press for British
Academy. (Proceedings of the British Academy, 99.)
Hiltebeitel, A. 1978. ‘The Indus Valley “Proto-Śiva” Reexamined through Reflections on the Goddess, the Buffalo, and the Symbolism of Vāhanas’. Anthropos
73: 767–97.
Hiltebeitel, A. (ed.) 1989. Criminal Gods and Demon Devotees: Essays on the
Guardians of Popular Hinduism. Albany, NY: SUNY Press.
Hiltebeitel, A. 2001. Rethinking the Mahābhārata: a Reader’s Guide to the Education
of the Dharma King. Chicago and London: University of Chicago Press.
Hiltebeitel, A. 2004. ‘More Rethinking the Mahābhārata: Toward a Politics of
Bhakti’. Indo-Iranian Journal 47: 203–27.
Hinnells, J. R. 1985. Persian Mythology. Revised edn. London: Hamlyn.
Hock, N. 1987. Buddhist Ideology and the Sculpture of Ratnagiri, Seventh through
Thirteenth Centuries. PhD thesis, University of California, Berkeley.
Hodge, S. 2003. The Mahā-Vairocana-Abhisam
. bodhi Tantra with Buddhaguhya’s
Commentary. London: RoutledgeCurzon.
Hoffman, E. 1992. Visions of Innocence: Spiritual and Inspirational Experiences of
Childhood. Boston and London: Shambhala.
Holt, J. C. 2004. The Buddhist Vis.n.u: Religious Transformation, Politics, and Culture.
New York: Columbia University Press.
Hooykaas, C. 1973a. Religion in Bali. Leiden: Brill.
Hooykaas, C. 1973b. Balinese Bauddha Brahmans. Amsterdam: N. V. NoordHollandsche Uitgevers Maatschappij (Verhandelingen der Koninklijke
Nederlandse Akademie van Wetenschappen, Afd. Letterkunde. Niewe Reeks.
Deel LXXX).
Hopkins, T. J. 1999. ‘Some Reflections on Hinduism’. Unpublished typescript.
References
373
Horsch, P. 1968. ‘Buddhismus und Upanis.aden’, in Pratidānam: Indian, Iranian,
and Indo-European Studies Presented to Franciscus Bernardus Jacobus Kuiper
on his Sixtieth Birthday, Heesterman, Schokker [and] Subramoniam (eds.),
pp. 462–77. The Hague and Paris: Mouton.
Houben, J. E. M. (ed.) 2003a. Papers from workshop, The Soma/Haoma-Cult in
Early Vedism and Zoroastrianism: Archaeology, Text, and Ritual, at Leiden
University, 3–4 July 1999. Special Issue of the Electronic Journal of Vedic Studies
9, 1. E-text at www1 shore.net/˜india/ejvs/issues.html, downloaded 6 Aug 06.
Houben, J. E. M. 2003b. ‘The Soma-Haoma Problem: Introductory
Overview and Observations on the Discussion’, in Houben 2003a. Etext at www1 shore.net/˜india/ejvs/ejvs0901/ejvs0901a.txt, downloaded 6
Aug 06.
Houtman, G. 1996. ‘From Mandala to Mandalay: Vipassana Insight Contemplation as an Instrument for Aristocratic Enlightenment and Delineation
of Domain in the Mindon Era’. Seminar at Lancaster University, 21 Oct
1996.
Houtman, G. 1999. Mental Culture in Burmese Crisis Politics: Aung San Suu Kyi and
the National League for Democracy. Tokyo: Institute for the Study of Languages
and Cultures of Asia and Africa, Tokyo University of Foreign Studies.
Howe, L. 2000. ‘Risk, Ritual and Performance’. Journal Royal Anthropological
Institute 6: 63–79.
Hsu, E. 2005. ‘Tactility and the Body in Early Chinese Medicine’. Science in Context
18: 7–34.
Huber, T. 1999. The Cult of Pure Crystal Mountain: Popular Pilgrimage and Visionary
Landscape in Southeast Tibet. New York: Oxford University Press.
Hudson, D. D. 1993. ‘Madurai: The City as Goddess’, in Spodek and Srinivasan
1993, pp. 125–44.
Hudson, D. 2002. ‘Early Evidence of the Pāñcarātra Āgama’, in Harper and Brown
2002, pp. 133–67.
Humes, C. A. 1996. ‘Vindhyavāsinı̄: Local Goddess Yet Great Goddess’. In Hawley
and Wulff 1996, pp. 49–76.
Huntingford, G. W. B. 1980. The Periplus of the Erythraean Sea. Translated and
edited by G. W. B. Huntingford. London: The Hakluyt Society.
Huntington, J. C. 1987. ‘Note on a Chinese Text Demonstrating the Earliness
of Tantra’. Journal of the International Association of Buddhist Studies 10(2):
88–98.
Huntington, J. C. and Bangdel, D. 2003. The Circle of Bliss: Buddhist Meditational Art. Columbus, OH: Columbus Museum of Art; Chicago: Serindia
Publications.
Huntington, S. L. 1990. ‘Introduction’, in Huntington and Huntington 1990,
pp. 69–71.
Huntington, S. L. and Huntington, J. C. 1990. Leaves from the Bodhi Tree: The
Art of Pāla India (8th –12th Centuries) and Its International Legacy. Seattle and
London: University of Washington Press in Association with Dayton Art
Institute.
374
References
Iltis, L. L. 1987. ‘The Jala Pyākhā: A Classical Newar Dance Drama of Harissiddhi’.
In Heritage of the Kathmandu Valley, Gutschow and Michaels (eds.), 199–214.
Sankt Augustin: VGH Wissenschaftsverlag.
Iltis, L. 2002. ‘Knowing All the Gods: Grandmothers, God Families and Women
Healers in Nepal’, in Rozario and Samuel 2002a, pp. 70–89.
Inden, R. 1992. Imagining India. Cambridge, MA and Oxford,UK: Blackwell.
Insler, S. 2004. Ancient Indian and Iranian Religion: Common Ground and Divergence. 2004 Jordan Lectures in Comparative Religion, School of Oriental and
African Studies, University of London, 23–29 April 2004.
Ingalls, D. H. H. 1962. ‘Cynics and Pasupatas: The Seeking of Dishonor’, Harvard
Theological Review 55: 281–98.
Isaacson, H. 1998. ‘Tantric Buddhism in India’, in Buddhismus in Geschichte
und Gegenwart: Weterbildendes Studium. Band II. Hamburg: Universität
Hamburg.
Jackowitz, S. ‘Ingestion, Digestion, and Regestation: The Complexities of Qi
Absorption’, in Kohn 2006a, pp. 68–90.
Jacobi, H. 1895. Gaina Sûtras. Part II. The Uttarâdhyayana Sûtra. The Sûtrakritâṅga
Sûtra. Oxford: Clarendon Press. (Sacred Books of the East, 45.)
Jacobsen, K. A. (ed.) 2005a. Theory and Practice of Yoga: Essays in Honour of Gerald
James Larson. Leiden: Brill.
Jacobsen, K. A. 2005b. ‘Introduction: Yoga Traditions’, in Jacobsen 2005a, 1–27.
Jahan, S. H. 2004. ‘Location of the Port of Tmāralipti and Condition of the
Harbour’. Journal Asiatic Society of Bangladesh (Humanities) 49: 199–220.
Jaini, P. S. 1977. ‘Jina R.s.abha as an avatāra of Vis.n.u’. Bulletin of School of Oriental
and African Studies 40: 321–37.
Jaini, P. S. 1991. ‘Is There a Popular Jainism’, in Carrithers and Humphrey 1991,
pp. 187–99.
Jairazbhoy, R. A. 1994. ‘The First Goddess of South Asia, A New Theory’, in
Jairazbhoy, The First Goddess of South Asian and Other Essays, pp. 7–26. 2nd
edn. Karachi: Menander Publications.
Jamison, S. W. and Witzel, M. 2003. ‘Vedic Hinduism’. Pdf version downloaded
from www.people.fas.harvard.edu/˜witzel/vedica.pdf, 17 Sept 2006.
Jayakar, P. 1989. The Earth Mother. New Delhi: Penguin Books. [Revised and
updated edn. of Jayakar 1980.]
Jest, C. 1976. Dolpo: communautés de langue tibétaine du Népal. Paris: Éditions du
CNRS.
Jhavery, M. B. 1944. Comparative and Critical Study of Mantrasastra, with Special
Treatment of Jain Mantravada, being the Introduction to Sri Bhairava Padmavati
Kalpa. Ahmedabad: Sarabhai Manilal Nawab.
Jootla, S. Elbaum 1997. Teacher of the Devas. Kandy: Buddhist Publication Society.
(The Wheel Publication No. 414/416.)
Jordaan, R. E. and Wessing, R. J. 1999. ‘Construction Sacrifice in India, “Seen from
the East”’, in Violence Denied: Violence, Non-Violence and the Rationalization
of Violence in South Asian Cultural History, Houben and Van Kooij (eds.),
pp. 211–48. Leiden: Brill.
References
375
Joshi, M. S. 2002. ‘Historical and Iconographic Aspects of Śākta Tantrism’, in
Harper and Brown 2002, pp. 39–55.
Joshi, N. P. 1984. ‘Early Forms of Śiva’. In Meister 1984, pp. 47–61.
Jurewicz, J. 2000. ‘Playing with Fire: The Pratı̄tyasamutpāda from the Perspective
of Vedic Thought’. Journal Pali Text Society 26: 77–103.
Kaelber, W. O. 1989. Tapta Mārga: Asceticism and Initiation in Vedic India. Albany,
NY: SUNY Press.
Kakar, S. 1990. Intimate Relations: Exploring Indian Sexuality. New Delhi;
Harmondsworth: Penguin Books.
Kakar, S. 1998. The Ascetic of Desire: A Novel. New Delhi: Penguin Books.
Kala, S. C. 1951. Bharhut Vedikā. Allahabad: Municipal Museum.
Kalff, M. M. 1979. Selected Chapters from the Abhidānottaratantra: The Union of
Male and Female Deities. PhD. Dissertation, Columbia University, New York.
Kammerer, C. A. and Tannenbaum, N. (eds.) 1996. Merit and Blessing in Mainland
Southeast Asia in Comparative Perspective. New Haven, CT: Yale University,
Southeast Asia Studies. (Monograph series / Yale University. Southeast Asia
studies; no. 45.)
Kapferer, B. 1979. ‘Mind, Self and Other in Demonic Illness: The Negation and
Reconstruction of Self ’. American Ethnologist 6: 110–33.
Kapferer, B. 1983. A Celebration of Demons: Exorcism and the Aesthetics of Healing
in Sri Lanka. Bloomington, IN: Indiana University Press.
Kaplanian, P. 1981. Les Ladakhi du Cachemire: Montagnards du Tibet Occidental.
Hachette (L’Homme Vivant.)
Kapstein, M. 1980. ‘The Shangs-pa bKa’-brgyud: An Unknown School of Tibetan
Buddhism’. In Studies in Honor of Hugh Richardson, Aris and Kyi (eds.),
pp. 138–44. Warminster: Aris and Phillips.
Kapstein, M. 1997. ‘The Journey to the Golden Mountain’, in Religions of Tibet
in Practice, Lopez, Jr. (ed.), pp. 178–87. Princeton, NJ: Princeton University
Press.
Kapstein, M. 1998. ‘A Pilgrimage of Rebirth Reborn: The 1992 Celebration of
the Drigung Powa Chenmo’, in Buddhism in Contemporary Tibet: Religious
Revival and Cultural Identity, Goldstein and Kapstein (eds.), pp. 95–119.
Berkeley, Los Angeles and London: University of California Press.
Kapstein, M. T. 2005. Chronological Conundrums in the Life of Khyung
po rnal ‘byor: Hagiography and Historical Time’. Journal International
Association of Tibetan Studies 1 (October 2005). Downloaded from
www.thdl.org/collections/journal/jiats/index.php?doc=kapstein01.xml&s=
d0e1526, 3 Oct 2006.
Karlsson, K. 1999. Face to Face with the Absent Buddha: The Formation of Buddhist
Aniconic Art. Uppsala: Acta Universitatis Upsaliensis. (PhD dissertation,
Uppsala, 2000.)
Kawamura, L. 1975. Golden Zephyr. Translated from the Tibetan and annotated by
L. Kawamura. Emeryville, CA: Dharma Pub.
Kawanami, H. 2001. ‘Can Women Be Celibate? Sexuality and Abstinence in
Theravada Buddhism’, in Sobo and Bell 2001, pp. 137–156.
376
References
Kendall, L. 1985. Shamans, Housewives and Other Restless Spirits. Honolulu: University of Hawaii Press.
Kendall, L. 1988. The Life and Hard Times of a Korean Shaman: Of Tales and the
Telling of Tales. Honolulu: University of Hawaii Press.
Kenoyer, J. M. 1995. ‘Interaction Systems, Specialised Crafts and Culture Change:
The Indus Valley Tradition and the Indo-Gangetic Tradition in South Asia’,
in Erdosy 1995a, pp. 213–57.
Kenoyer, J. M. 1998. Ancient Cities of the Indus Valley. Karachi: Oxford University
Press and American Institute of Pakistan Studies.
Keyes, C. F. 1984. ‘Mother or Mistress But Never A Monk: Buddhist Notions Of
Female Gender In Rural Thailand’. American Ethnologist 11: 223–41.
Keyes, C. F. 1986. ‘Ambiguous Gender: Male Initiation in a Northern Thai
Buddhist Society’, in Gender and Religion: On the Complexity of Symbols,
Bynum, Harrell and Richman (eds.), pp. 66–96. Boston: Beacon Press.
Keyes, C. F. 1987. ‘Theravāda Buddhism and its Worldly Transformations in
Thailand: Reflections on the Work of S. J. Tambiah’. Contributions to Indian
Sociology N.S. 21: 123–45.
Khan, D.-S. 2004. Crossing the Threshold: Understanding Religious Identities in
South Asia. London: I. B. Tauris in association with The Institute of Ismaili
Studies.
Khandelwal, M. 2001. ‘Sexual Fluids, Emotions, Morality: Notes on the Gendering
of Brahmacharya’, in Sobo and Bell 2001, pp. 157–79.
Khare, R. S. 1984. The Untouchable as Himself: Ideology, Identity and Pragmatism
Among the Lucknow Chamars. Cambridge: Cambridge University Press.
Kilambi, J. S. 1985. ‘Towards an Understanding of the Mmuggu: Threshold Drawings in Hyderabad’. Res 10: 71–102.
Kim, C. 2003. Korean Shamanism: The Cultural Paradox. London: Ashgate.
Kim, T. 1988. The Relationship between Shamanic Ritual and the Korean Masked
Dance-Drama. PhD Dissertation, New York University.
King, R. 1999. Orientalism and Religion: Postcolonial Theory, India and ‘The Mystic
East’. London and New York: Routledge.
Kinnard, J. 1997. ‘Reevaluating the Eighth-Ninth Century Pala Milieu: IconoConservatism and the Persistence of Sakyamuni’. Journal International Association of Buddhist Studies 20: 281–300.
Kirsch, A. T. 1973. Feasting and Social Oscillation. A Working Paper on Religion and
Society in Upland Southeast Asia. Ithaca, NY: Cornell University Department
of Asian Studies. (Southeast Asia Program, Data Paper No. 92.)
Kirsch, A. T. 1985. ‘Text And Context: Buddhist Sex Roles/Culture Of Gender
Revisited’. American Ethnologist 12: 302–20.
Klimburg-Salter, D. E. 1989. The Kingdom of Bāmiyān: Buddhist Art and Culture
of the Hindu Kush. Napoli: Istituto universitario orientale, Dipartimento di
studi asiatici; Roma: Istituto italiano per il Medio ed Estremo Oriente. (Series
maior (Istituto universitario orientale (Naples, Italy). Dipartimento di studi
asiatici); 5.)
References
377
Klimburg-Salter, D. E. 1995. Buddha in Indien: Die frühindische Skulptur von König
Aśoka bis zur Guptazeit. Wien: Kunsthistorisches Museum Wien; Milano:
Skira.
Kobayashi, T. 1975. Nara Buddhist Art: Todai-ji. Translated and adapted by R. L.
Gage. New York and Tokyo: Weatherhill/Heibonsha.
Kohn, L. (ed.) 2006a. Daoist Body Cultivation. Magdalena, NM: Three Pines Press.
Kohn, L. 2006b. ‘Yoga and Daoyin’, in Kohn 2006a, pp. 123–50.
Kohn, R. J. 1988. Mani Rimdu: Text and Tradition in a Tibetan Ritual. PhD dissertation, University of Wisconsin, Madison.
Kohn, R. J. 2001. Lord of the Dance: The Mani Rimdu Festival in Tibet and Nepal.
Albany, NY: State University of New York Press.
Korvald, T. 2005. ‘The Dancing Gods of Bhaktapur and Their Audience’, in Allen
2005, pp. 405–15.
Kosambi, D. D. 1960. ‘At the Cross-Roads: Mother Goddess Cult Sites in Ancient
India’. Journal of the Royal Asiatic Society: 17–31; 135–44.
Kosambi, D. D. 1965. The Culture and Civilisation of Ancient India in Historical
Outline. London: Routledge & Kegan Paul.
Kosambi, D. D. 2002a. Combined Methods in Indology and Other Essays. Edited by
B. Chattopadhyaya. New Delhi: Oxford University Press.
Kosambi, D. D. 2002b. ‘The Vedic “Five Tribes”’, in Kosambi 2002a, pp. 75–
86.
Krasser, H., Much, M. T., Steinkellner, E. and Tauscher, H. (eds.). 1997. Tibetan
Studies: Proceedings of the 7th Seminar of the International Association for
Tibetan Studies, Graz 1995. 2 vols. Wien: Österreichische Akademie der Wissenschaften. (Österreichische Akademie der Wissenschaften. PhilosophischHistorische Klasse, Denkschriften, 256. Band. Beiträge zur Kultur- und
Geistesgeschichte Asiens, Nr. 21.)
Kuan, T. 2005. ‘Clarification on Feelings in Buddhist Dhyāna/Jhāna Meditation’.
Journal Indian Philosophy 33: 285–319.
Kulke, H. and Rothermund, D. 1990. A History of India. Rev. updated edn. London:
Routledge.
Kulke, H. 2004. ‘Some Thoughts on State and State Formation under the Eastern
Vākāt.akas’, in Bakker 2004, pp. 1–9.
Kvaerne, P. 1986. An Anthology of Buddhist Tantric Songs. 2nd edn. Bangkok,
White Orchid Press. (First edn. 1977, Oslo, Universitetsforlaget. (Det Norske
Videnskaps-Akademi. II.Hist.-Filos. Klasse. Skrifter. Ny Serie, 14.)
Lalou, M. 1956. ‘Four Notes on Vajrapān.i’. Adyar Library Bulletin 20: 287–93.
Lamotte, É. 1966. ‘Vajrapān.i en Inde’, in Mélanges de Sinologie offerts a Monsieur
Paul Demiéville, pp. 113–61. Paris. (Bibliotheque de l’Institute des Hautes
Études Chinoises, 20.)
Landaw, J. and Weber, A. 1993. Images of Enlightenment: Tibetan Art in Practice.
Ithaca, NY: Snow Lion.
Lariviere, R. W. 1997. ‘Power and Authority: On the Interpretation of Indian
Kingship from Sanskrit Sources’, in Lex et Litterae: Studies in Honour of
378
References
Professor Oscar Botto, Lienhard and Piovano (eds.), pp. 313–27. Alessandria:
Edizioni dell’Orso.
Lee, Y.-H. 2003. Synthesizing a Liturgical Heritage: Abhayākaragupta’s Vajrāvali
and the Kālacakraman.d.ala. PhD Dissertation, University of Wisconsin at
Madison. AAT 3089732.
Lehman, F. K. (Chit Hlaing ) ‘Monasteries, Palaces and Ambiguities: Burmese
Sacred and Secular Space’. Contributions to Indian Sociology N.S. 21: 169–
86.
Leidy, D. P. 1997. ‘Place and Process: Mandala Imagery in the Buddhist Art of
Asia’, in Leidy and Thurman 1997, pp. 17–47.
Leidy, D. P. and Thurman, R. A. F. 1997. Mandala: The Architecture of Enlightenment. Thames and Hudson in association with Asia Society Galleries and
Tibet House.
Leopold, R. S. 1983. ‘The Shaping of Men and the Making of Metaphors: The
Meaning of White Clay in Poro and Sande Initiation Society Rituals’. Anthropology 8(2): 21–42.
Lerner, M. 1984. The Flame and the Lotus: Indian and Southeast Asian Art from
the Kronos Collection. New York: Metropolitan Museum of Art and Harry
N. Abrams.
Leslie, J. (ed.) 1992. Roles and Rituals for Hindu Women. Delhi: Motilal Banarsidass.
Leslie, J. (ed.) 1996a. Myth and Mythmaking. London: Curzon.
Leslie, J. 1996b. ‘Menstruation Myths’, in Leslie 1996a, pp. 87–105.
Levy, R. I. 1987. ‘How the Navadurgā Protect Bhaktapur: The Effective Meanings
of a Symbolic Enactment’, in Heritage of the Kathmandu Valley, Gutschow
and Michaels (eds.) 105–34. Sankt Augustin: VGH Wissenschaftsverlag.
Lévi, S. 1915. ‘Le catalogue géographique des Yaks.a dans la Mahāmāyūrı̄ ’. Journal
Asiatique 11: 19–138.
Levy, R. I. with Rājopādhyāya, K. R. 1992. Mesocosm: Hinduism and the Organization of a Traditional Newar City in Nepal. Delhi: Motilal Banarsidass.
Lewis, T. T. 2000. Popular Buddhist Texts from Nepal: Narratives and Rituals of
Newar Buddhism. Translations in Collaboration with S. M. Tuladhar and
L. R. Tuladhar. Albany, NY: State University of New Work Press.
Lincoln, B. 2003. ‘À La Recherche du Paradis Perdu’. History of Religions 43: 139–
54.
Lindtner, C. 1991–93. ‘Nāgārjuna and the Problem of Precanonical Buddhism’.
Religious traditions 15–17: 112–36.
Lindtner, C. 1999. ‘From Brahmanism to Buddhism’. Asian Philosophy 9: 5–37.
Ling, T. 1973. The Buddha: Buddhist Civilization in India and Ceylon. London:
Temple Smith.
Linrothe, R. 1999. Ruthless Compassion: Wrathful Deities in Early Indo-Tibetan
Esoteric Buddhist Art. London: Serindia.
Linrothe, R. (ed.) 2006. Holy Madness: Portraits of Tantric Siddhas. New York:
Rubin Museum of Art; Chicago: Serindia Publications.
Little, K. 1965. ‘The Political Function of the Poro’. Africa 35: 349–65.
Little, K. 1966. ‘The Political Function of the Poro. 2’. Africa 36: 62–72.
References
379
Littleton, C. Scott 1982. The New Comparative Mythology: an Anthropological Assessment of the Theories of Georges Dumézil. 3rd edn. Berkeley, CA and London:
University of California Press.
Locke, J. K. 1980. Karunamaya: the Cult of Avalokitesvara-Matsyendranath in the
Valley of Nepal. Kathmandu: Sahayogi Prakashan for Research centre for Nepal
and Asian Studies, Tribhuvan University.
Locke, J. K. 1985. Buddhist Monasteries of Nepal: a Survey of the Bāhās and Bahı̄s of
the Kathmandu Valley. Kathmandu: Sahayogi Press.
Lopez Jr., D. S. (ed.) 1995. Curators of the Buddha: The Study of Buddhism under
Colonialism. Chicago: University of Chicago Press.
Lorenzen, D. N. 1991. The Kāpālikas and Kālāmukhas: Two Lost Śaivite Sects. 2nd
rev. edn. Delhi: Motilal Banarsidass.
Loseries-Leick, A. 1997. ‘Psychic Sports: A Living Tradition in Contemporary
Tibet?’, in Krasser et al. 1997, vol. II, pp. 583–93.
Lubin, T. 2001. ‘Vratá Divine and Human in the Early Veda’. Journal American
Oriental Society 121: 565–79.
Lubin, T. 2005. ‘The Transmission, Patronage and Prestige of Brahmanical Piety
from the Mauryas to the Guptas’, in Boundaries, Dynamics and Construction of Traditions in South Asia, Squarcini (ed.), pp. 77–103. Firenze: Firenze
University Press.
Lynch, O. M. 1969. The Politics of Untouchability: Social Mobility and Social Change
in a City of India. New York & London: Columbia University Press.
McAlpin, D. W. 1981. Proto-Elamo-Dravidian: the Evidence and its Implications.
Philadelphia: American Philosophical Society. (Transactions of the American
Philosophical Society, v. 71, part 3).
McBride, R. D., II. 2004. ‘The Vision-Quest Motif in Narrative Literature on the
Buddhist Traditions of Silla’. Korean Studies 27: 17–47.
McDaniel, J. 1989. The Madness of the Saints: Ecstatic Religion in Bengal. Chicago:
University of Chicago Press.
McDaniel, J. 1992. ‘The Embodiment of God Among the Bauls of Bengal’. Journal
of Feminist Studes in Religion 8: 27–39.
McDaniel, J. 2004. Offering Flowers, Feeding Skulls: Popular Goddess Worship in
West Bengal. Oxford and New York: Oxford University Press.
McDermott, R. F. 1996. ‘Popular Attitudes Towards Kālı̄ and Her Poetry Tradition:
Interviewing Śāktas in Bengal’, in Michaels, Vogelsanger and Wilke 1996,
pp. 383–415.
McDermott, R. F. and Kripal, J. J. (eds.) 2003. Encountering Kālı̄: In the Margins,
At the Center, In the West. Berkeley: University of California Press.
McEvilley, T. 2002. ‘The Spinal Serpent’, in Harper and Brown 2002, pp. 93–113.
McGee, M. 1992. ‘Desired Fruits: Motive and Intention in the Votive Rites of
Hindu Women’, in Leslie 1996a, pp. 69–88.
McGee, T. G. 1967. The Southeast Asian City. London: G. Bell and Sons.
Mabbett, I. W. 1969. ‘Devarāja’. Journal Southeast Asian History 10: 202–23.
Mabbett, I. W. 1977a. ‘The “Indianization” of Southeast Asia: Reflections on the
Prehistoric Sources’. Journal Southeast Asian Studies 8: 1–14.
380
References
Mabbett, I. W. 1977b. ‘The “Indianization” of Southeast Asia: Reflections on the
Historical Sources’. Journal Southeast Asian Studies 8: 143–61.
Mabbett, I. W. 1998. ‘The Problem of the Historical Nāgārjuna Revisited’. Journal
American Oriental Society 118: 332–46.
Macdonald, A. W. 1990. ‘Hindu-isation, Buddha-isation, Then Lama-isation or:
What Happened at La-phyi?’, in Skorupski 1990, pp. 199–208.
Mackenzie, J. M. 1995. Orientalism: History, Theory and the Arts. Manchester:
Manchester University Press.
Madan, T. N. 1985. ‘Concerning the Categories śubha and śuddha in Hindu
Culture: An Exploratory Essay’. In Carman and Marglin 1985, pp. 11–29.
Revised version (“Auspiciousness and Purity”) as ch. 2 of Madan 1987.
Madan, T. N. 1987. Non-Renunciation: Themes and Interpretations of Hindu Culture.
Delhi: Oxford University Press.
Madan, T. N. 1988. Ways of Life: King, Householder, Renouncer. (Essays in Honour
of Louis Dumont.) Delhi: Motilal Banarsidass.
Madan, T. N. 1991. ‘Auspiciousness and Purity: Some Reconsiderations’. Contributions to Indian Sociology (N.S.) 25: 287–9.
Magee, P. 2004. ‘Mind the Gap: The Chronology of Painted Gray Ware and the
Prelude to Early Historic Urbanism in Northern South Asia’. South Asian
Studies 20: 37–44.
Majupuria, T. C. 1993. Erawan Shrine and Brahma Worship in Thailand: With
Reference to India and Nepal. Bangkok: Tecpress Service.
Malandra, G. H. 1997. Unfolding a Man.d.ala: The Buddhist Cave Temples at Ellora.
Delhi: Sri Satguru Publications.
Marglin, F. A. 1985a. ‘Introduction’. In Carman and Marglin 1985, pp. 1–10.
Marglin, F. A. 1985b. Wives of the God-King: The Rituals of the Devadasis of Puri.
Delhi: Oxford University Press.
Marglin, F. A. 1994. ‘The Sacred Groves: Menstruation Rituals in Rural Orissa’.
Manushi: A Journal about Women in Society 82 (May–June 1994): 22–32.
Marglin, F. A. 1995. ‘Gender and the Unitary Self: Looking for the Subaltern in
Coastal Orissa’. South Asia Research 15: 78–130.
Marriott, M. (ed.) 1969a. Village India. Chicago: University of Chicago Press.
Marriott, M. 1969b. ‘Little Communities in an Indigenous Civilization’, in
Marriott 1969a, pp. 171–223.
Marriott, M. 1979. ‘Hindu Transactions: Diversity without Dualism’, in Transaction and Meaning: Directions in the Anthropology of Human Issues, Kapferer
(ed.), pp. 109–42. Philadelphia: Institute for the Study of Human Issues.
Marshall, Sir J. 1931. Mohenjodaro and the Indus Civilization: Being an Official
Account of Archaeological Excavations at Mohenjodaro Carried Out by the
Government of India Between the Years 1922–27. Delhi: Indological Book
House.
Masefield, P. 1986. Divine Revelation in Pali Buddhism. Colombo: Sri Lanka Institute of Traditional Studies.
Maskarinec, G. G. 1995. The Rulings of the Night: an Ethnography of Nepalese
Shaman Oral Texts. Madison, WI: University of Wisconsin Press.
References
381
Matsunaga, Y. 1977a. ‘A History of Tantric Buddhism in India with Reference to
the Chinese Translations’, in Buddhist Thought and Asian Civilization: Essays
in Honor of Herbert V. Guenther on His Sixtieth Birthday, Kawamura and Scott
(eds.), pp. 167–81. Emeryville, CA: Dharma Publishing.
Matsunaga, Y. 1977b. ‘Some Problems of the Guhyasamāja-Tantra’, in Studies in
Indo-Asian Art and Culture, Chandra and Ratnam (eds.), pp. 109–19. New
Delhi: International Academy of Indian Culture.
Mayaram, S. 2000. Resisting Regimes: Myth, Memory and the Shaping of a Muslim
Identity. Delhi: Oxford University Press.
Mayer, R. 1996. A Scripture of the Ancient Tantra Collection: The Phur-pa bcu-gnyis.
Oxford: Kiscadale Publications.
Maxwell, M. and Tschudin, V. 1996. Seeing the Invisible: Modern Religious and
Other Transcendent Experiences. Oxford: Religious Experience Research Centre, Westminster College.
Meenakshi, K. 1996. ‘The Siddhas of Tamil Nadu: A Voice of Dissent’, in Tradition,
Dissent and Ideology: Essays in Honour of Romila Thapar, Champakalakshmi
and Gopal (eds.), pp. 111–34. Delhi: Oxford University Press.
Meister, M. W. (ed.) 1984. Discourses on Śiva: Proceedings of a Symposium on the
Nature of Religious Imagery. Philadelphia: University of Pennsylvania Press.
Meister, M. W. 1986. ‘Regional Variations in Mātr.kā Conventions’, Artibus Asiae
47: 233–62.
Mencher, J. 1974. ‘The Caste System Upside Down’. Current Anthropology 15:
469–93.
Mendelson, E. M. 1961a. ‘A Messianic Buddhist Association in Upper Burma’.
Bulletin of the School of Oriental and African Studies 24: 560–80.
Mendelson, E. M. 1961b. ‘The King of the Weaving Mountain’. Journal Royal
Central Asian Society 48: 229–37.
Mendelson, E. M. 1963. ‘Observations on a Tour in the Region of Mount Popa,
Central Burma’. France-Asie 179 (Mai–Juin 1963): 786–807.
Menges, K. H. 1989. ‘Aus dem animistisch-schamanistischen Wortschatz der Altajer’, in Gedanke und Wirkung: Festschrift zum 90. Geburtstag von Nikolaus
Poppe, Heissig and Sagaster (eds.), pp. 221–51. Wiesbaden: Otto Harrassowitz.
(Asiatische Forschungen, 108.)
Merz, B. 1996. ‘Wild Goddess and Mother of Us All’, in Michaels, Vögelsanger
and Wilke 1996, pp. 343–54.
Meyer, J. J. 1903. Ks.emendra’s Samayamatrika (Das Zauberbuch der Hetären.) Ins
Deutsche übertragen. Leipzig: Lotus-Verlag.
Michaels, A., C. Vogelsanger and A. Wilke (eds.) 1996. Wild Goddesses in India
and Nepal. Proceedings of an International Symposium in Berne and Zurich,
November 1994). Bern: Peter Lang (Studia Religiosa Helvetica; 2).
Miller, J. 1985. The Vision of Cosmic Order in the Vedas. London: Routledge and
Kegan Paul.
Misra, R. N. 1981. Yaksha Cult and Iconography. Delhi: Munshiram Manoharlal.
Mitchell, G. 1989. The Penguin Guide to the Monuments of India. Vol. 1: Buddhist,
Jain, Hindu. London: Viking.
382
Reference
Mitra, D. 1984. ‘Lakulı̄śa and Early Śaiva Temples in Orissa’. In Meister 1984,
pp. 103–18.
Mitterwallner, G. von 1984. ‘Evolution of the Liṅga’. In Meister 1984, pp. 12–31.
Mitterwallner, G. von 1989. ‘Yaks.a of Ancient Mathura’. In Srinivasan 1989,
pp. 368–82.
Mohan, P. N. 2001. ‘Maitreya Cult in Early Shilla: Focusing on Hwarang as
Maitreya-incarnate’. Seoul Journal of Korean Studies 14: 149–73.
Morinis, E. A. 1984. Pilgrimage in the Hindu Tradition: A Case Study of West Bengal.
Delhi and Oxford: Oxford University Press.
Morrison, B. M. 1970. Political Centers and Cultural Regions in Early Bengal. Tucson,
AZ: University of Arizona Press. (Association for Asian Studies, Monographs
and Papers, XXV.)
Movik, K. 2000. Bruk av Cannabis i Shivaittisk Tradisjon. Hovedfagsoppgave
i religionshistorie. Universitetet i Oslo. Downloaded from http://www.
normal.no/txt/shiva/info.html, Sept. 2006.
Mulder, N. 1992. Inside Thai Society: An Interpretation of Everyday Life. 3rd rev.
edn. Bangkok: Duang Kamol.
Muller-Ortega, P. E. 2002. ‘Becoming Bhairava: Meditative Vision in Abhinavagupta’s Parātrı̄śikā-Laghuvr.tti’, in Harper and Brown 2002, pp. 213–
30.
Mullin, G. H. 2006. The Practice of the Six Yogas of Naropa. Ithaca, NY and Boulder,
CO: Snow Lion.
Mumford, S. R. 1989. Himalayan Dialogue: Tibetan Lamas and Gurung Shamans
in Nepal. Madison, WI: University of Wisconsin Press.
Murti, T. R. V. 1960. The Central Philosophy of Buddhism: A Study of the
Mādhyamika System. London: George Allen and Unwin.
Mus, P. 1975. India Seen from the East: Indian and Indigenous Cults in Champa.
Monash University, Melbourne: Centre for Southeast Asian Studies. (Monash
Papers on Southeast Asia, 3.)
Mussat, M. 1978. Sou Nü King: La sexualité taoı̈ste de la Chine ancienne. Translated
by Leung Kwok Po. Paris: Seghers.
Nagarajan, V. R. 1997. ‘Inviting the Goddess into the Household’. Whole Earth
90: 49–53.
Ñān.amoli T. 1964. Mindfulness of Breathing (Ānāpānasati). Buddhist Texts from
the Pali Canon and from the Pali Commentaries. 2nd edn. Kandy: Buddhist
Publication Society.
Ñān.amoli, B. 1991. The Path of Purification (Visuddhimagga) by Bhadantācariya
Buddhaghosa. 5th edn. Kandy: Buddhist Publication Society.
Ñān.amoli, B. and Bodhi, B. 1995. The Middle Length Discourses of the Buddha: A
New Translation of the Majjhima Nikāya. Boston: Wisdom Publications.
Nanda, S. 1990. Neither Man nor Woman: The Hijras of India. Belmont, CA:
Wadsworth.
Nattier, J. 1988. ‘The Meanings of the Maitreya Myth’, in Maitreya, the Future
Buddha, Sponberg and Hardacre (eds.), pp. 23–50. Cambridge and New
York: Cambridge University Press.
Reference
383
Nepali, G. S. 1965. The Newars; An Ethno-Sociological Study of a Himalayan Community. Bombay: United Asia Publications.
Newman, J. R. 1987. The Outer Wheel of Time: Vajrayāna Buddhist Cosmology in
the Kālacakra Tantra (India). PhD Dissertation, University of Wisconsin at
Madison. UMI AAT 8723348.
Newman, J. 1998. ‘The Epoch of the Kālacakra Tantra’. Indo-Iranian Journal 41:
319–49.
Ngorchen K. L. 1987. The Beautiful Ornament of the Three Visions: An Exposition of
the Preliminary Practices of the Path which Extensively Explains the Instructions
of the “Path Including its Result” in Accordance with the Root Treatise of the
Vajra Verses of Virūpa. Singapore: Golden Vase Publications.
Nichter, M. 1977. ‘The Joga and Maya of the Tuluva Buta’. Eastern Anthropologist
30 (2):139–55.
Nyanaponika, T. 1969. The Heart of Buddhist Meditation. London: Rider and Co.
Oberoi, H. 1994. The Construction of Religious Boundaries: Culture, Identity and
Diversity in the Sikh Tradition. Delhi: Oxford University Press and Chicago:
University of Chicago Press.
O’Connor, R. 1989. ‘Cultural Notes on Trade and the Tai’, in Russell 1989, pp. 27–
65.
O’Flaherty, W. Doniger (=Doniger, W.) 1981. The Rig Veda: An Anthology. One Hundred and Eight Hymns, Selected, Translated and Annotated. London: Penguin
Books.
Olivelle, P. 1992. Sam
. nyāsa Upanis.ads: Hindu Scriptures on Asceticism and Renunciation. New York and Oxford: Oxford University Press.
Olivelle, P. 1993. The Āśrama System: The History and Hermeneutics of A Religious
Institution. Oxford: Oxford University Press.
Olivelle, P. 1997. ‘Orgasmic Rapture and Divine Ecstasy: The Semantic History of
Ānanda’. Journal Indian Philosophy 25: 153–80.
Olivelle, P. 1998. Upanis.ads. Translated from the Original Sanskrit. Oxford and
New York: Oxford University Press. (Oxford World’s Classics.)
Olivelle, P. 2003. ‘The Renouncer Tradition’, in Flood 2003, pp. 271–87.
Openshaw, J. 1997. ‘The Web of Deceit: Challenges to Hindu and Muslim
“Orthodoxies” by “Bāuls” of Bengal’. Religion 27: 297–309.
Openshaw, J. 2002. Seeking Bāuls of Bengal. Cambridge and New York: Cambridge
University Press.
Oppitz, M. 1978–80. Schamanen im blinden Land (Shamans of the Blind Country).
Documentary film. Wieland Schulz Keil Produktion/WDR.
Oppitz, M. 1981. Schamanen im Blinden Land. Ein Bilderbuch aus dem Himalaya.
Frankfurt am Main: Syndikat Verlag.
Orofino, G. 1997. ‘Apropos of Some Foreign Elements in the Kālacakratantra’, in
Krasser et al. 1997, Vol. II, pp. 717–24.
Orofino, G. 1998. [Review of Benard 1994.] Tibet Journal 23(3): 114–18.
Orofino, G. 2001. ‘Notes on the Early Phases of Indo-Tibetan Buddhism’, in Le
Parole e I Marmi: Studi in Onore di Raniero Gnoli nel suo 70◦ Compleanno,
Torella (ed.), pp. 541–64. Roma: Istituto Italiano per l’Africa e l’Oriente.
384
Reference
Orzech, C. D. 1998. Politics and Transcendent Wisdom: the Scripture for Humane
Kings in the Creation of Chinese Buddhism. University Park, PA: Pennsylvania
State University Press.
Owens, B. McCoy 1995. ‘Human Agency and Divine Power: Transforming Images
and Recreating Gods among the Newar’. History of Religions 34: 201–40.
Padoux, A. 1990. Vāc: the Concept of the Word in Selected Hindu Tantras. translated
by Jacques Gontier. Albany, NY: State University of New York Press.
Pal, P. 1986–88. Indian Sculpture: a Catalogue of the Los Angeles County Museum of
Art Collection. 2 vols. Los Angeles, CA: Los Angeles County Museum of Art
in association with University of California Press, Berkeley.
Pal, P. 1988. ‘The Fifty-one Śākta Pı̄t.has’, in Gnoli and Lanciotti 1988, pp. 1039–60.
Pal, P. 1994. The Peaceful Liberators: Jain Art from India. Los Angeles: Los Angeles
County Museum of Art; London: Thames and Hudson.
Pallath, J. J. 1995. Theyyam: An Analytical Study of the Folk Culture Wisdom and
Personality. New Delhi: Indian Social Institute.
Palmié, S. 2006. ‘A View from Itia Ororó Kande’. Social Anthropology 14: 99–118.
Pandey, R. 1969. Hindu Sam
. skāras. 2nd rev. edn. Delhi: Motilal Banarsidass.
Pandian, J. 1982. ‘The Goddess Kannagi: A Dominant Symbol of South Indian
Tamil Society’, in Mother Worship: Theme and Variations, Preston (ed.),
pp. 177–91. Chapel Hill: University of North Carolina Press.
Parpola, A. 1988. ‘The Coming of the Aryans to Iran and India and the Cultural
and Ethnic Identity of the Dāsas’. Studia Orientalia (Helsinki) 64: 195–302.
Parpola, A. 1994. Deciphering the Indus Script. Cambridge and New York:
Cambridge University Press.
Parpola, A. 1995. ‘The Problem of the Aryans and the Soma: Textual-Linguistic
and Archaeological Evidence’, in Erdosy 1995a, pp. 353–81.
Parpola, A. 1999a. ‘The Formation of the Aryan Branch of Indo-European’, in
Blench and Spriggs 1999, pp. 180–207.
Parpola, A. 1999b. ‘Sāvitrı̄ and Resurrection: The Ideal of Devoted Wife, Her
Forehead Mark, Satı̄, and Human Sacrifice in Epic-Purān.ic, Vedic, HarappanDravidian and Near Eastern Perspectives’, in Parpola and Tenhunen 1999,
pp. 167–312.
Parpola, A. 2002a. ‘Pre-Proto-Iranians of Afghanistan as Initiators of Śākta
Tantrism: On the Scythian/Saka Affliiation of the Dāsas, Nuristanis and
Magadhans’. Iranica Antiqua 37: 233–324.
Parpola, A. 2002b. ‘sndaih and Sı̄tā: On the Historical Background of the
Sanskrit Epics’. Journal American Oriental Society 122: 361–73.
Parpola, A. 2005. ‘The Nāsatyas, the Chariot and Proto-Aryan Religion’. Journal
of Indological Studies (Kyoto) 16–17: 1–63.
Parpola, A. and Christian 2005. ‘The Cultural Counterparts to Proto-IndoEuropean, Proto-Uralic and Proto-Aryan: Matching the Dispersal and
Contact Patterns in the Linguistic and Archaeological Record’, in Bryant
and Patton 2005, pp. 107–41.
Parpola, A. and Tenhunen, S. (eds.) 1999. Changing Patterns of Family and Kinship
in South Asia. Proceedings of an International Symposium on the Occasion of
Reference
385
the 50th Anniversary of India’s Independence Held at the University of Helsinki
6 May 1998. Helsinki: Finnish Oriental Society. (Studia Orientalia, 84.)
Parry, J. P. 1991. ‘The Hindu Lexicographer? A Note on Auspiciousness and Purity’.
Contributions to Indian Sociology (N.S.) 25: 267–86.
Parry, J. P. 1994. Death in Banaras. Cambridge: Cambridge University Press.
Pejros, I. and Shnirelman, V. 1998. ‘Rice in Southeast Asia: A Regional Interdisciplinary Approach’, in Blench and Spriggs 1998, pp. 379–89.
Pflueger, L. W. 2003. ‘Dueling with Dualism: Revisioning the Paradox of Purus.a
and Prakr.ti’, in Whicher and Carpenter 2003, pp. 70–82.
Pirart, É. 1998. ‘Historicité des forces du mal dans la R.gvedasam
. hitā’. Journal
Asiatique 286: 521–69.
Pollock, S. 2007. The Language of the Gods in the World of Men: Sanskrit, Culture,
and Power in Premodern India. Delhi: Permanent Black.
Possehl, G. L. and Rissman, P. C. 1992. ‘The Chronology of Prehistoric India: From
Earliest Times to the Iron Age’, in Ehrich 1992, I, pp. 465–90; II, pp. 447–
74.
Possehl, G. L. 1998. ‘Sociocultural Complexity Without the State’, in Archaic
States, Feinman and Marcus (eds.), pp. 261–91. Santa Fe, School of American
Research Press.
Poster, A. G. 1986. From Indian Earth: 4,000 Years of Terracotta Art. Brooklyn, New
York: The Brooklyn Museum.
Potts, D. T. 2005. ‘Cyrus the Great and the Kingdom of Anshan’, in Curtis and
Stewart 2005, pp. 7–28.
Prebish, C. A. 1995. ‘Ideal Types in Indian Buddhism: A New Paradigm’. Journal
American Oriental Society 115: 651–66.
Prebish, C. A. 1996. ‘Śaiks.a-dharmas Revisited: Further Considerations of
Mahāsām
. ghika Origins’. History of Religions 35: 258–70.
Quigley, D. 1995. The Interpretation of Caste. Oxford: Clarendon Press. (Oxford
India Paperback Edition.)
Quintanilla, S. R. 2000. ‘Āyāgapat.as: Characteristics, Symbolism, and Chronology’. Artibus Asiae 60: 79–137.
Qvarnström, O. 1998. ‘Stability and Adaptability: A Jain Strategy for Survival and
Growth’. Indo-Iranian Journal 41: 33–55.
Qvarnström, O. 2000. ‘Jain Tantra: Divinatory and Meditative Practices in the
Twelfth-Century Yogaśāstra of Hemacandra’. Tantra in Practice, White (ed.),
pp. 595–604. Princeton, NJ: Princeton University Press.
Qvarnström, O. 2002. The Yogaśāstra of Hemacandra: A Twelfth Century Handbook on Śvetāmbara Jainism. Translated by O. Qvarnström. Cambridge, MA:
Harvard University, Department of Sanskrit and Indian Studies. Distributed
by Harvard University Press.
Qvarnström, O. 2003. ‘Losing One’s Mind and Becoming Enlightened: Some
Remarks on the Concept of Yoga in Śvetāmbara Jainism and Its Relation to
the Nāth Siddha Tradition’, in Whicher and Carpenter 2003, pp. 130–42.
Rabe, M. D. 1996. ‘Sexual Imagery on the Phantasmagorical Castles at Khajuraho’.
International Journal of Tantric Studies 2(2) (Nov. 1996).
386
Reference
Rabe, M. D. 1999. ‘“Not-Self” Consciousness and the Aniconic in Early
Buddhism’. In Modeling Consciousness Across the Disciplines Symposium, Jodan
(ed.), pp. 269–80. Lanham, MD: University Press of America. On-line
version downloaded from www.sxu.edu/˜rabe/asia/bodhgaya/index.html,
25/12/2001.
Radhakrishnan, S. 1963. The Bhagavadgı̄tā: With an Introductory Essay, Sanskrit
Text and Notes. 2nd edn, 7th impression. London: George Allen and Unwin.
Rajaram, N. S. and Frawley, D. 1995. Vedic ‘Aryans’ and the Origins of Civilization:
A Literary and Scientific Perspective. Quebec: WH Press.
Rao, S. R. 1973. Lothal and the Indus Civilization. Bombay: Asia Publishing House.
Rawson, P. 1981. Oriental Erotic Art. New York: Gallery Books.
Ray, R. A. 1997. ‘Nāgārjuna’s Longevity’, in Sacred Biography in the Buddhist
Traditions of South and South Asia, Schober and Woodward (eds.), pp. 129–59.
Honolulu: University of Hawai’i Press.
Ray, R. A. 1999. Buddhist Saints in India: A Study in Buddhist Values and Orientations. New York: Oxford University Press.
Reid, A. 1988. ‘Female Roles in Pre-Colonial Southeast Asia’. Modern Asian Studies
22: 629–45.
Rhie, M. M. ‘Mahakala: Some Tangkas and Sculptures from the Rubin Museum
of Art’, in Demonic Divine: Himalayan Art and Beyond, Linrothe and Watt
(eds.). pp. 44–98. Rubin Museum of Art; Chicago: Serindia Publications.
Rhum, M. R. 1987. ‘The Cosmology of Power in Lanna’, Journal of Siam Society
75: 91–107.
Rhum, M. R. 1994. The Ancestral Lords: Gender, Descent, and Spirits in a Northern
Thai Village. DeKalb: IL: Northern Illinois University. (Center for Southeast
Asian Studies. Monograph Series on Southeast Asia, Special Report 29.)
Robinson, E. 1996. The Original Vision: A Study of the Religious Experience of Childhood. Oxford: Religious Experience Research Centre, Westminster College.
Robinson, J. B. 1979. Buddha’s Lions: The Lives of the Eighty-Four Siddhas.
Caturaśı̄ti-siddha-pravr.tti by Abhayadatta. Translated into Tibetan as Grub
tho brgyad cu rtsa bzhi’i lo rgyus by sMon-grub Shes-rab. Translated into
English by J. B. Robinson. Berkeley, CA: Dharma Publishing.
Robinson, S. P. 1985. ‘Hindu Paradigms of Women: Images and Values’, in Women,
Religion and Social Change, Haddad and Findly (eds.), pp. 181–215. Albany,
NY: SUNY Press.
Rocher, L. 1985. ‘The Kāmasūtra: Vātsyāyana’s Attitude Toward Dharma and
Dharmaśāstra’. Journal American Oriental Society 105: 521–9.
Roerich, G. N. 1976. The Blue Annals. 2nd edn. Delhi: Motilal Banarsidass.
Rozario, S. 2002. ‘The Healer on the Margins: The Dai in Rural Bangladesh’, in
Rozario and Samuel 2002a, pp. 130–46.
Rozario, S. and Samuel, G. (eds.) 2002a. The Daughters of Hariti: Childbirth and
Female Healers in South and Southeast Asia. London and New York: Routledge.
Rozario, S. and Samuel, G. 2002b. ‘Tibetan and Indian Ideas of Birth Pollution:
Similarities and Contrasts’, in Rozario and Samuel 2002a, pp. 182–208.
Reference
387
Ruegg, D. S. 1964. ‘Sur les rapports entre le bouddhisme et le “substrat religioux”
indien et tibétain’. Journal Asiatique 252: 77–95.
Russell, S. D. (ed.) 1989. Ritual, Power, and Economy: Upland-Lowland Contrasts in
Mainland Southeast Asia. DeKalb, IL: Northern Illinois University Center for
Southeast Asian Studies. (Monograph Series on Southeast Asia. Occasional
Paper No. 14.)
Saheb, S. A. A. 1998. ‘A “Festival of Flags”: Hindu-Muslim Devotion and the
Sacralising of Localism at the Shrine of Nagore-e-Sharif in Tamil Nadu’,
in Embodying Charisma: Modernity, Locality and the Performance of Emotion
in Sufi Cults, Werbner and Basu (eds.), pp. 55–76. London and New York:
Routledge.
Said, E. W. 1978. Orientalism. New York: Pantheon.
Salomon, C. 1991. ‘The Cosmogonic Riddles of Lalan Fakir’, in Gender, Genre,
and Power in South Asian Expressive Traditions, Appadurai, Korom and Mills
(eds.), pp. 267–304. Philadelphia: University of Pennsylvania Press.
Samanta, S. 1992. ‘Maṅgalmayı̄mā, Sumaṅgalı̄, Maṅgal: Bengali Perceptions Of
The Divine Feminine, Motherhood and “Auspiciousness”’. Contributions to
Indian Sociology (N.S.) 26: 51–75.
Sampurn.ānand, Ś. 1956. The Atharva Veda, Vrātyakān.d.a With Srutiprabha commentary in English by Sri Sampurn.ānand. Madras: Ganesh.
Samuel, G. 1982. ‘Tibet as a Stateless Society and Some Islamic Parallels’. Journal
of Asian Studies, 41: 215–29. (Revised version in Samuel 2005, pp. 27–51.)
Samuel, G. 1989. ‘The Body in Buddhist and Hindu Tantra: Some Notes’. Religion
19: 197–210.
Samuel, G. 1990. Mind, Body and Culture: Anthropology and the Biological Interface.
Cambridge and New York: Cambridge University Press.
Samuel, G. 1992. ‘Gesar of Ling: the Origins and Meaning of the East Tibetan Epic’,
in Tibetan Studies: Proceedings of the 5th Seminar of the International Association
for Tibetan Studies, Narita, 1989, Ihara and Yamaguchi (eds.), pp. 711–22.
Narita: Naritasan Shinshoji. (Reprinted in Samuel 2005a, pp. 165–91.)
Samuel, G. 1993. Civilized Shamans: Buddhism in Tibetan Societies. Washington
and London: Smithsonian Institution Press.
Samuel, G. 1997. ‘Women, Goddesses and Auspiciousness in South Asia’. Journal
of Interdisciplinary Gender Studies 4: 1–23. (Reprinted in Samuel 2005a,
pp. 256–87.)
Samuel, G. 2000. ‘The Indus Valley Civilization and Early Tibet’, in New Horizons
in Bon Studies, Karmay and Nagano (eds.), pp. 651–70. Osaka: National
Museum of Ethnology. (Bon Studies 2) (Reprinted in Samuel 2005a, pp. 138–
64.)
Samuel, G. 2001a. ‘The Religious Meaning of Space and Time: South and Southeast
Asia and Modern Paganism’. International Review of Sociology vol. 11 no. 3
(Nov. 2001), pp. 395–418.
Samuel, G. 2001b. ‘The Effectiveness of Goddesses, or, How Ritual Works’.
Anthropological Forum 11: 73–91. (Reprinted in Samuel 2005, pp. 229–55.)
388
Reference
Samuel, G. 2001c. ‘Tibetan Medicine in Contemporary India: Theory and
Practice’, in Healing Powers and Modernity in Asian Societies, Connor and
Samuel (eds.), pp. 247–68. Westport, CT: Bergin and Garvey (Greenwood
Publishing).
Samuel, G. 2002a. ‘The Epic and Nationalism in Tibet’, in Religion and Biography
in China and Tibet, Benjamin Penny (ed.), pp. 178–88. Richmond, Surrey:
Curzon Press.
Samuel, G. 2002b. ‘Introduction’, in Rozario and Samuel 2002a, pp. 1–33.
Samuel, G. 2002c. ‘Ritual Technologies and the State: The Mandala-Form
Buddhist Temples of Bangladesh’. Journal Bengal Art 7: 39–56.
Samuel, G. 2002d. ‘Buddhism and the State in Eighth Century Tibet’. In Religion
and Secular Culture in Tibet: Tibetan Studies II (PIATS 2000: Tibetan Studies:
Proceedings of the 9th Seminar of the International Association of Tibetan Studies,
Leiden 2000) ed. H. Blezer with the assistance of A. Zadoks, pp. 1–19. Leiden:
Brill. (Reprinted in Samuel 2005a, pp. 94–116.)
Samuel, G. 2005a. Tantric Revisionings. New Delhi: Motilal Banarsidass; London:
Ashgate.
Samuel, G. 2005b. ‘Subtle Bodies in Indian and Tibetan Yoga: Scientific and
Spiritual Meanings’. Paper for Second International Conference on Religions
and Cultures in the Indic Civilisation, Delhi 17–20 December 2005.
Samuel, G. 2006a. ‘Healing and the Mind-Body Complex: Childbirth and Medical
Pluralism in South Asia’, in Multiple Medical Realities: Patients and Healers
in Biomedical. Alternative and Traditional Medicine, Johannessen and Lázár
(eds.), pp. 121–35. New York and London: Berghahn Books.
Samuel, G. 2006b. ‘A Short History of Indo-Tibetan Alchemy’. Paper for the
Medicine, Religion and History Panel at the 11th Seminar of the International Association for Tibetan Studies, Königswinter, German, 27 August to
2 September 2006.
Samuel, G. 2006c. ‘Tibetan Medicine and Biomedicine: Epistemological Conflicts,
Practical Solutions’. Asian Medicine 2: 72–85.
Sanderson, A. 1985. ‘Purity and Power among the Brahmans of Kashmir’, in
Category of the Person, Carrithers (ed.), pp. 190–216. Cambridge and New
York: Cambridge University Press.
Sanderson, A. 1988. ‘Śaivism and the Tantric Traditions’, in The World’s
Religions, Sutherland, Houlden, Clarke and Hardy (eds.), 660–704. London:
Routledge.
Sanderson, A. 1994. ‘Vajrayāna: Origin and Function’, in Buddhism into the Year
2000: International Conference Proceedings, Bhikkhu et al. (eds.), pp. 87–102.
Bangkok and Los Angeles, CA: Dhammakaya Foundation.
Sanderson, A. 1995. ‘Meaning in Tantric Ritual’, in Essais sur le rituel III: Colloque
du centenaire de la section des sciences religieuses de l’école pratique des hautes
etudes, Blondeau and Schipper (eds.), pp. 15–95. Louvain and Paris: Peeters.
Sanderson, A. 2001. ‘History through Textual Criticism in the Study of Śaivism,
the Pañcarātra and the Buddhist Yoginı̄tantras’, in Grimal 2001, pp. 1–47.
Sanderson, A. 2003–4. ‘The Śaiva Religion Among the Khmers’. Bulletin de l’École
Française d’Extrême-Orient, 90–1: 349–462.
Reference
389
Sanderson, A. 2004. ‘Religion and the State: Śaiva Officiants in the Territory of the King’s Brahmanical Chaplain’. Indo-Iranian Journal 47: 229–
300.
Saraswati, S. S. 1985. Kundalini Tantra. 1st Australian edn. Gosford, NSW:
Satyananda Ashram.
Sardar, H. n.d. ‘Trance-Dancers of the Goddess Durga’. Downloaded from
www.asianart.com/articles/hamid/ on 10 Oct. 2001.
Satyananda P. (Swami Satyananda Saraswati) 1980. Four Chapters on Freedom: Commentary on Yoga Sutras of Patanjali. 2nd Australian edn. Mangrove Mountain,
NSW: Satyananda Ashram.
Schaeffer, K. R. 2002. ‘The Attainment of Immortality: From Nāthas in India to
Buddhists in Tibet’. Journal Indian Philosophy 30: 515–33.
Schalk, P. 1994. ‘The Controversy about the Arrival of Buddhism in Tamilakam’.
Temenos 30: 197–232.
Schipper, K. 1994. The Taoist Body. Translated by K. C. Duval. Taipei: SMC
Publishing Inc. (First published in 1982 as Le corps taoı̈ste by Librairie Arthème
Fayard, Paris.)
Schlingloff, D. 1964. Ein buddhistisches Yogalehrbuch. Ed. and trans. Berlin:
Akademie verlag. (Sanskrittexte aus den Turfanfunde, 7.)
Schmithausen, L. 1997. Maitri and Magic: Aspects of the Buddhist Attitude Toward
the Dangerous in Nature. Wien: Verlag der Österreichischen Akademie der
Wissenschaften.
Schopen, G. 1996. ‘Immigrant Monks and the Proto-Historical Dead: The
Buddhist Occupation of Early Burial Sites in India’, in Festschrift für Dieter
Schlingloff, Wilhelm (ed.), pp. 215–38. Munich: Reinbek.
Schopen, G. 1997a. Bones, Stones, and Buddhist Monks: Collected Papers on the
Archaeology, Epigraphy, and Texts of Monastic Buddhism in India. Honolulu:
University of Hawai’i Press.
Schopen, G. 1997b. ‘Archeology and Protestant Presuppositions in the Study of
Indian Buddhism’, in Schopen 1997a, pp. 1–22.
Schrempf, M. 1999. ‘Taming the Earth, Controlling the Cosmos: Transformation
of space in Tibetan Buddhist and Bon-po Ritual Dances’, in Sacred Spaces
and Powerful Places in Tibetan Culture: A Collection of Essays, Huber (ed.).
Dharamsala, The Library of Tibetan Works and Archives.
Schrempf, M. 2001. Ethnisch-Religiöse Revitalisierung und rituelle Praxis einer osttibetischen Glaubensgemainschaft im heutigen China am beispiel ritueller Maskentanzenaufführungen der Bönpo-Klosterföderation von Gamel Gingka in Amdo
Sharkhog in der Zeit von 1947 bis 1996. Inaugural-Dissertation zur Erlangung
des Doktorgrades am Fachbereich Politik- und Sozialwissenschaften der
Freien Universität Berlin.
Schubring, W. 1942–52. Isibhāsāiyam: Ein Jaina Text der Frühzeit. Göttingen:
Vandenhoeck & Ruprecht. (Nachrichten der Akademie der Wissenschaften
in Göttingen. I, Philologisch-historische Klasse; Jahrg. 1942, Nr. 6, Jahrg.
1952, Nr. 2.)
Schwartzberg, J. 1992. A Historical Atlas of South Asia. 2nd impression, with additional material. New York: Oxford University Press.
390
Reference
Scott, D. 1993. ‘Śiva and the East Iranians of Bactria: Hindu cross-cultural expansion – 1’. International Journal of Indian Studies 3: 87–106.
Searle-Chatterjee, M. and Sharma, U. (eds.) 1994. Contextualising Caste: PostDumontian Approaches. Oxford: Blackwell Publishers/The Sociological
Review. (Sociological review monograph; 41.)
Sen, S. 1953. Vipradāsa’s Manasā-Vijaya: A Fifteenth Century Bengali Text [. . .]
Calcutta: Asiatic Society. (Bibliotheca Indica, 277.)
Seneviratne, H. L. 1978. Rituals of the Kandyan State. Cambridge and New York:
Cambridge University Press.
Shaffer, J. G. 1992. ‘The Indus Valley, Baluchistan, and Helmand Traditions:
Neolithic through Bronze Age’, in Ehrich 1992a, I, pp. 441–64; II, pp. 425–46.
Shaffer, J. G. and Lichtenstein, D. A. 1995. ‘The Concepts of “Cultural Tradition”
and “Palaeoethnicity” in South Asian Archaeology’. In Erdosy 1995a, pp. 126–
54.
Shah, U. P. 1984. ‘Lakulı̄śa: Śaivite Saint’. In Meister 1984, pp. 92–102.
Shahbazi, A. Shapur 2005. ‘The History of the Idea of Iran’, in Curtis and Stewart
2005, pp. 100–11.
Shakabpa, W. D. 1967. Tibet: A Political History. New Haven: Yale University Press.
Sharma, A. 1995. ‘The Aryan Question: Some General Considerations’, in Erdosy
1995a, pp. 177–91.
Sharma, J. P. 1968. Republics in Ancient India, c. 1500B.C.–500B.C. Leiden: Brill.
Sharma, J. P. 1989. Jaina Yakshas. Meerut: Kusumanjali Prakashan.
Shaw, J. 2000. ‘Sanchi and its Archaeological Landscape: Buddhist Monasteries,
Settlements and Irrigation Works in Central India’. Antiquity 74: 775–76.
Shaw, J. and Sutcliffe, J. V. 2001. ‘Ancient Irrigation Works in the Sanchi Area:
An Archaeological and Hydrological Investigation’. South Asian Studies 17:
55–75.
Shaw, M. 1994. Passionate Enlightenment: Women in Tantric Buddhism. Princeton,
NJ: Princeton University Press.
Shawn, A. 2006. ‘Life Without Grains: Bigu and the Daoist Body’, in Kohn 2006a,
91–122.
Shen, W. 2006. ‘Reconstructing the History of Buddhism in Central Eurasia
(11–14 Cent.): An Interdisciplinary and Multilingual Approach to Khara
Khoto Texts’. Paper for the 11th Seminar of the International Association
of Tibetan Studies, August–September 2006, Königswinter, Germany.
Sherburne, R. 1983. A Lamp for the Path and Commentary by Atı̄śa. London: Allen
and Unwin.
Shrestha, B. G. 1999. ‘Visible and Invisible Aspects of the Devı̄ Dances in Sankhu,
Nepal’, in Tambs-Lyche 1999, pp. 100–13.
Sick, D. H. 2004. ‘Mit(h)ra(s) and the Myths of the Sun’. Numen 51: 432–67.
Silburn, L. 1988. Kun.d.alinı̄: The Energy of the Depths. A Comprehensive Study Based
on the Scriptures of Nondualistic Kaśmir Śaivism. Albany, NY: State University
of New York Press.
Silk, J. 2002. ‘What, if Anything, is Mahāyāna Buddhism? Problems of Definitions
and Classifications’. Numen 49: 355–405.
Reference
391
Sims-Williams. N. and Cribb, J. 1996. ‘A New Bactrian Inscription of Kanishka
the Great’. Silk Road Art and Archaeology 4: 75–142.
Singh, U. 1993. Kings, Brahmanas and Temples in Orissa. An Epigraphic Study AD
300–1147. Delhi: Munshiram Manoharlal.
Sircar, D. C. 1971–72. ‘Mahamayuri. List of Yaksas. Translation’. Journal of Ancient
Indian History 5: 262–328.
Sircar, D. C. 1973. The Śākta Pı̄t.has. 2nd rev. edn. Delhi: Motilal Banarsidass.
Sivaramamurti, C. 1976. Śatarudrı̄ya: Vibhūti of Śiva’s Iconography. New Delhi:
Abhinav Publications.
Skilling, P. 2004. ‘Mahayana and Bodhisattva: an Essay Towards Historical Understanding’, in Phothisatawa barami kap sangkhom thai nai sahatsawat mai
[Bodhisattvaparami and Thai Society in the New Millennium], Proceedings
of a Seminar in Celebration of the 4th Birth Cycle of HRH Princess Maha
Chakri Sirindhorn, Limpanusorn and Iampakdee (eds.), pp. 141–56. Bangkok:
Thammasat University Press. (Chinese Studies Centre, Institute of East Asia,
Thammasat University.)
Skjærvø, P. O. 1995. ‘The Avesta as Source for the Early History of the Iranians’,
in Erdosy 1995a, pp. 155–76.
Skjærvø, P. O. 1997. ‘The State of Old-Avestan Scholarship’. Journal American
Oriental Society 117(1): 103–14.
Skjærvø, P. O. 2005. ‘The Achaemenids and the Avesta’, in Curtis and Stewart
2005, pp. 52–84.
Skorupski, T. (ed.) 1990. Indo-Tibetan Studies. Tring: Institute of Tibetan Studies.
Skorupski, T. 1998. ‘An Analysis of the Kriyāsam
. graha’, in Sūryacandrāya: Essays
in Honour of Akira Yuyama on the Occasion of his 65th Birthday, Harrison and
Schopen (eds.), pp. 181–96. Swisttal-Odendorf: Indica et Tibetica Verlag.
(Indica et Tibetica, 35.)
Slusser, M. Shepherd 1982. Nepal Mandala: a Cultural Study of the Kathmandu
Valley. 2 vols. Princeton: Princeton University Press.
Smith, B. K. 1989. Reflections on Resemblance, Ritual, and Religion. New York and
Oxford: Oxford University Press.
Smith, B. K. 1994. Classifiying the Universe: The Ancient Indian Varn.a System and
the Origins of Caste. New York and Oxford: Oxford University Press.
Smith, D. 1985. Ratnākara’s Haravijaya: An Introduction to the Sanskrit Court Epic.
Delhi: Oxford University Press.
Smith, D. 2003. ‘Orientalism and Hinduism’, in Flood 2003, pp. 45–63.
Smith, Relph B. and Watson, William (eds.) 1979. Early South East Asia: Essays in
Archaeology, History, and Historical Geography. New York and Kuala Lumpur:
Oxford University Press.
Snellgrove, D. L. 1959. The Hevajra Tantra: A Critical Study. 2 vols. London: Oxford
University Press. (London Oriental Series, 6.)
Snellgrove, D. L. 1987. Indo-Tibetan Buddhism: Indian Buddhists and Their Tibetan
Successors. London: Serindia.
Snellgrove, D. L. 1988. ‘Categories of Buddhist Tantras’, in Gnoli and Lanciotti
1988, pp. 1353–84.
392
Reference
Snellgrove, D. L. 2000. Asian Commitment: Travels and Studies in the Indian SubContinent and South-East Asia. Bangkok: Orchid Press.
Sobo, E. J. and Bell, S. (eds.) 2001. Celibacy, Culture and Society: The Anthropology
of Sexual Abstinence. Madison, WI: University of Wisconsin Press.
Sopa, G. L. 1985. ‘The Kalachakra Tantra Initiation’, in Sopa, Jackson and Newman
1985, pp. 85–117.
Sopa, G. L., R. Jackson and J. Newman 1985. The Wheel of Time: The Kalachakra
in Context. Madison, WI: Deer Park Books.
Southwold, M. 1979. ‘Religious Belief’. Man 14: 628–44.
Southwold, M. 1983. Buddhism in Life: the Anthropological Study of Religion
and the Sinhalese Practice of Buddhism. Manchester: Manchester University
Press.
Spess, D. L. 2000. Soma: The Divine Hallucinogen. Rochester, VT: Park Street
Press.
Spiro, M. E. 1967. Burmese Supernaturalism: A Study in the Explanation and Reduction of Suffering. Englewood Cliffs, NJ: Prentice-Hall.
Spiro, M. E. 1971. Buddhism and Society: A Great Tradition and its Burmese Vicissitudes. London: Allen and Unwin.
Spodek, H. and Srinivasan, D. M. (eds.) 1993. Urban Form and Meaning in South
Asia: The Shaping of Cities from Prehistoric to Precolonial Times. National
Gallery of Art, Washington; Hanover and London: University Press of New
England.
Sponberg, A. 1992. ‘Attitudes Towards Women and the Feminine in Early
Buddhism’, in Buddhism, Sexuality, and Society, Cabezón (ed.), pp. 3–36.
Albany, NY: State University of New York Press.
Srinivas, M. N. 1952. Religion and Society Among the Coorgs of South India. Oxford:
Clarendon Press.
Srinivasan, D. M. 1979. ‘Early Vais.n.ava Imagery: Caturvyūha and Variant Forms’.
Archives of Asian Art 32: 39–54.
Srinivasan, D. M. 1984. ‘Significance and Scope of Pre-Kus.ān.a Śaivite Iconography’. In Meister 1984, pp. 32–46.
Srinivasan, D. M. (ed.) 1989. Mathura: The Cultural Heritage. New Delhi: AIAS.
Srinivasan, D. M. 1997. Many Heads, Arms and Eyes: Origin, Meaning and Form of
Multiplicity in Indian Art. Leiden: Brill. (Studies in Asian Art and Archaeology,
20.)
Stablein, W. G. 1976. The Mahākālatantra: A Theory of Ritual Blessings and Tantric
Medicine. PhD dissertation, Columbia University. (University Microfilms
76–29, 405.)
Stablein, W. 1991. Healing Image: The Great Black One. Berkeley and Hong Kong:
SLG Books.
Stacul, G. 1992. ‘Further Evidence for the “Inner Asia Complex” from Swat’, in
South Asian Archaeology Studies, Possehl (ed.), pp. 111–22. New Delhi: Oxford
& IBH Publishing House.
Stargardt, J. 2000. Tracing Thoughts Through Things: The Oldest Pali Texts and the
Earliest Buddhist Archaeology of India and Burma. Seventh Gonda Lecture.
Reference
393
Amsterdam, 12th November 1999. Amsterdam: Royal Netherlands Academy
of Arts and Sciences.
Stein, B. 1994. Peasant State and Society in Medieval South India. Delhi: Oxford
University Press.
Stein, M. A. 1900. Kalhana’s Rājataranginı̄: A Chronicle of the Kings of Kasmı̄r.
Translated, with an introduction, commentary, and appendices by M. A.
Stein. 2 vols. Westminster: A. Constable.
Stephen, M. 2001. ‘Barong and Rangda in the Context of Balinese Religion’. Review
of Indonesian and Malaysian Affairs 35: 137–94.
Stone, E. R. 1994. The Buddhist Art of Nāgārjunakon.d.a. Delhi: Motilal Banarsidass.
Strong, J. S. 1992. The Legend and Cult of Upagupta: Sanskrit Buddhism in North
India and Southeast Asia. Princeton, NJ: Princeton University Press.
Studholme, A. The Origins of Om
. Man.ipadme Hūm
. : A Study of the Kāran.d.avyūha
Sūtra. Albany, NY: State University of New York Press.
Stutley, M. 1980. Ancient Indian Magic and Folklore. Boulder, CO: Great Eastern.
Sugiyama, J. 1982. Classic Buddhist Sculpture: The Tempyō Period. Translated and
adapted by S. Crowell Morse. Tokyo, New York and San Francisco: Kodansha
International Ltd and Shibundo.
Sukthankar, V. S. 1936. ‘Epic Studies VI. The Bhr.gus and the Bhārata: A TextHistorical Study’. Annals of the Bhandarkar Oriental Research Institute 18(1):
1–76.
Sullivan, H. P. 1964. ‘A Re-Examination of the Religion of the IndusCivilization’.
History of Religions 4: 115–25.
Sullivan, H. P. 1971. ‘The Prehistory of Indian Religion’. [Review of Allchin 1963.]
History of Religions 11: 140–6.
Sutherland, G. H. 1991. The Disguises of the Demon: The Development of the Yaks.a
in Hinduism and Buddhism. Albany, NY: State University of New York Press.
Svoboda, R. E. 1986. Aghora: At the Left Hand of God. New Delhi: Rupa & Co.
Svoboda, R. E. 1993. Aghora II: Kundalini. New Delhi: Rupa & Co.
Swearer, D. 1976. Wat Haripuñjaya. A Study of the Royal Temple of the Buddha’s
Relic, Lamphun, Thailand. Missoula, MT: University of Montana. (Published
by Scholars Press for the American Academy of Religion.)
Tambiah, S. J. 1970. Buddhism and the Spirit Cults in North-east Thailand. Cambridge: Cambridge University Press.
Tambiah, S. J. 1973. ‘From Varna to Caste through Mixed Unions’, in The Character of Kinship, Goody (ed.), pp. 191–229. Cambridge: Cambridge University
Press.
Tambiah, S. J. 1984. The Buddhist Saints of the Forest and the Cult of Amulets:
A Study in Charisma, Hagiography, Sectarianism and Millennial Buddhism.
Cambridge: Cambridge University Press.
Tambiah, S. J. 1985. ‘The Galactic Polity in Southeast Asia’, in Culture, Thought,
and Social Action: An Anthropological Perspective, pp. 252–86. Cambridge, MA:
Harvard University Press.
Tambiah, S. J. 1987. ‘At the Confluence of Anthropology, History, and Indology’.
Contributions to Indian Sociology N.S. 21: 187–216.
394
Reference
Tambs-Lyche, H. (ed.) 1999. The Feminine Sacred in South Asia (Le sacré au feminine
en Asie du Sud). Delhi: Manohar.
Tannenbaum, N. 1987. ‘Tattoos, Invulnerability and Power in Shan Religion’.
American Ethnologist 14: 693–71.
Tannenbaum, N. 1989. ‘Power and its Shan Transformations’, in Russell 1989,
pp. 67–88.
Tarling, N. (ed.) 1999. The Cambridge History of Southeast Asia: Volume One. From
Early Times to c.1500. Paperback edn. Cambridge: Cambridge University Press.
Tatelman, J. 2005. The Heavenly Exploits: Buddhist Biographies from the
Divyāvadāna. Vol. I. New York: New York University Press (JJC Foundation). (Clay Sanskrit Library.)
Taylor, J. 1993. Forest Monks and the Nation-State: An Anthropological and Historical Study in Northeastern Thailand. Singapore: Institute of Southeast Asian
Studies.
Taylor, K. W. 1999. ‘The Early Kingdoms’, in Tarling 1999, pp. 137–82.
Templeman, D. 1983. Tāranātha’s bKa’.babs.bdun.ldan. The Seven Instruction
Lineages by Jo.nang Tāranātha. Dharamsala, Library of Tibetan Works and
Archives.
Templeman, D. 1989. Tārānatha’s Life of Kr..sn.ācārya/Kān.ha. Dharamsala: Library
of Tibetan Works and Archives.
Templeman, D. 1997. ‘Buddhaguptanātha: A Late Indian Siddha in Tibet’, in
Proceedings of the 7th Seminar of the International Association for Tibetan Studies,
Graz 1995, Eimer et al. (eds.), pp. 955–65. Wien: Verlag der Österreichischen
Akademie der Wissenschaften.
Terwiel, B. J. 1975. Monks and Magic: An Analysis of Religious Ceremonies in Central
Thailand. Studentlitteratur, Lund. (Scandinavian Institute of Asian Studies
Monograph Series No. 24.
Thakur, U. n.d. Madhubani Painting. New Delhi: Abhinav Publications, c.1981 or
1982.
Thapan, A. R. 1997. Understanding Gan.apati: Insights into the Dynamics of a Cult.
New Delhi: Manohar.
Thapar, R. 2003. The Penguin History of Early India: From the Origins to AD 1300.
New Delhi: Penguin Books.
Thieme, P. 1960. ‘The “Aryan” Gods of the Mitanni Treatise’. Journal American
Oriental Society 80: 301–17.
Thompson, G. 2003. ‘Soma and Ecstasy in the Rgveda’, in Houben 2003a. E-text at
www1.shore.net/˜india/ejvs/ejvs0901/ejvs0901e.txt, downloaded 6 Aug 2006.
Thurman, R. A. F. 1997. ‘Mandala: The Architecture of Enlightenment’, in Leoidy
and Thurman 1997, pp. 127–45.
Thurman, R. A. F. 2004. The Universal Vehicle Discourse Literature (Mahāyānasūtrālam
. kāra) by Maitreyanātha/Āryāsaṅga Together With Its Commentary
(Bhās.ya) by Vasubandhu. Translated from the Sanskrit, Tibetan, and Chinese
by L. Jamspal, R. Clark, J. Wilson, L. Zwilling, M. Sweet, R. Thurman,
editor-in-chief R. A. F. Thurman. New York: American Institute of Buddhist
Reference
395
Studies, co-published with Columbia University’s Center for Buddhist Studies
and Tibet House US.
Tinti, P. 1998. Between Two Civilisations: History and Self-Representation of
Bangladeshi Buddhism. D.Phil. dissertation, University of Oxford.
Tiyavanich, K. 1997. Forest Recollections: Wandering Monks in Twentieth-Century
Thailand. Honolulu: University of Hawaii Press.
Tiyavanich, K. 2004. The Buddha in the Jungle. Seattle: University of Washington
Press.
Toffin, G. 1984. Société et religion chez les Néwar du Népal. Paris: Éditions du
C.N.R.S.
Toffin, G. 1996. ‘A Wild Goddess Cult in Nepal: The Navadurgā of Theco Village
(Kathmandu Valley)’, in Michaels, Vogelsanger and Wilke 1996, pp. 217–
51.
Toffin, G. 1999. ‘Possession, danses masquées et corps divin dans la vallée de
Katmandou (Népal)’. Purus.ārtha 21: 237–61.
Trainor, K. 1997. Relics, Ritual and Representation in Buddhism: Rematerialising
the Sri Lankan Theravada Tradition. Cambridge and New York: Cambridge
University Press.
Trautmann, T. R. 1979. ‘The Study of Dravidian Kinship’, in Deshpande and
Hook 1979, pp. 153–74.
Trungpa, C. 1982. The Life of Marpa the Translator, by Tsang Nyön Heruka, translated
from the Tibetan by the N. Translation Committee under the direction of
Chögyam Trungpa. Boulder, CO: Prajñā.
Tsuchida, R. 1991. ‘Two Categories of Brahmins in the Early Buddhist Period’.
Memoirs of the Toyo Bunko 49: 51–95.
Tucci, G. 1940. Travels of Tibetan Pilgrims in the Swat Valley. [Reprinted 1971 in
Tucci, Opera Minora, 2 vols., Rome: G. Bardi.]
Tucci, G. 1980. The Religions of Tibet, translated by Geoffrey Samuel. London:
Routledge and Kegan Paul; Berkeley, CA: University of California Press.
Tuladhar-Douglas, W. 2005. ‘On Why it is Good to Have Many Names: the Many
Identities of a Nepalese God’. Contemporary South Asia, 14: 55–74.
Turner, V. W. 1957. Schism and Continuity in an African Society: A Study of Ndembu
Village Life. Manchester: Published on behalf of the Rhodes-Livingstone Institute by Manchester University Press.
Turner, V. W. 1968. The Drums of Affliction: a Study of Religious Processes among the
Ndembu of Zambia. Oxford: Clarendon Press; London: International African
Institute.
Turner, V. W. 1970. The Forest of Symbols: Aspects of Ndembu Ritual. Ithaca, London:
Cornell University Press.
Uchiyamada, Y. 2000. ‘Passions in the Landscape: Ancestor Spirits and Land
Reforms in Kerala, India’. South Asia Research 20: 63–84.
Upadhyaya, S. C. 1961. Kama Sutra of Vatsyayana: Complete Translation from the
Original Sanskrit. Bombay: Taraporevala’s.
Urban, H. 1999. ‘The Extreme Orient: The Construction of “Tantrism” as a
Category in the Orientalist Imagination’. Religion 29: 123–46.
396
Reference
Urban, H. 2001. ‘The Path of Power: Impurity, Kingship, and Sacrifice in Assamese
Tantra’. Journal of the American Academy of Religion 69: 777–816.
van Buitenen, J. A. B. 1975. The Mahābhārata. 2. The Book of the Assembly Hall;
3. The Book of the Forest. Translated and edited by J. A. B. van Buitenen.
Chicago and London: University of Chicago Press.
van Buitenen, J. A. B. 1983. The Mahābhārata. 1. The Book of the Beginning. Translated and edited by J. A. B. van Buitenen. Chicago and London: University
of Chicago Press.
van den Hoek, A. W. (Bert) 2005. ‘The Death of the Divine Dancers: The Conclusion of the Bhadrakali Pyakham in Kathmandu’, in Allen 2005, pp. 374–404.
van Skyhawk, H. 1993. ‘Nası̄ruddı̄n and Ādināth, Nizāmuddı̄n and Kāniphnāth:
Hindu-Muslim Religious Syncretism in the Folk Literature of the Deccan’,
in Brückner, Lutze, and Malik 1993, pp. 445–67.
Venkatraman, R. 1990. A History of the Tamil Siddha Cult. Madurai: Ennes
Publications.
Vergati-Stahl, A. 1979. ‘Taleju, Sovereign Deity of Bhaktapur’, in Asie du Sud,
Traditions et Changements: [actes du VIe colloque européan sur les études
modernes concernant l’Asie du Sud], Sèvres, 8–13 juillet 1978 / [organisé par
M. Gaborieau et A. Thorner], pp. 163–7. Paris: Éditions du CNRS.
Vergati, A. 1986. ‘Les associations religieuses (gut.hi) des temples de la vallée de
Kathmandou’. Purus.ārtha 10: 97–123.
Vergati, A. 2000. Gods and Masks of the Kāthmān.d.u Valley. New Delhi: D. K.
Printworld.
Vetter, T. 1990. ‘Some Remarks on Older Parts of the Suttanipāta’, in Earliest
Buddhism and Madhyamaka, Ruegg and Schmithausen (eds.), pp. 36–56.
Leiden: E. J. Brill.
Victoria, B. 2005. Zen at War. New York and Tokyo: Weatherfield.
Viswanathan, G. 2003. ‘Colonialism and the Construction of Hinduism’, in Flood
2003, pp. 23–44.
Vogt, B. 1999. Skill and Trust: The Tovil Healing Ritual of Sri Lanka as Cultur-Specific
Psychotherapy. Translated by M. H. Kohn. Amsterdam: VU University Press.
Volwahsen, A. 1969. Living Architecture: Indian. Calcutta: Oxford and IBH Publishing Company.
Walshe, M. 1995. The Long Discourses of the Buddha. A Translation of the Dı̄gha
Nikāya. Boston: Wisdom Publications.
Wadley, S. S. 1980. ‘Hindu Women’s Family and Household Rites in a North
Indian Village’, in Unspoken Worlds, Falk and Gross (eds.), pp. 94–109. New
York: Harper and Row.
Wallace, V. A. 1995. ‘The Buddhist Tantric Medicine in the Kālacakratantra’. Pacific
World: J. of the Institute of Buddhist Studies (N.S.) 10–11: 155–74.
Wallace, V. A. 2001. The Inner Kālacakratantra: A Buddhist Tantric View of the
Individual. New York and Oxford: Oxford University Press.
Wallace, V. A. 2004. The Kālacakratantra: The Chapter on the Individual Together
with the Vimalaprabhā. Translated from Sanskrit, Tbetan, and Mongolian by
V. A. Wallace. New York: American Institute of Buddhist Studies at Columbia
Reference
397
University, New York. Co-Published with Columbia University’s Center for
Buddhist Studies and Tibet House US.
Walshe, M. 1995. The Long Discourses of the Buddha: A Translation of the Dı̄gha
Nikāya. Boston: Wisdom Publications.
Walter, M. L. 1979. ‘Preliminary Results from a Study of Two Rasayana Systems in
Indo-Tibetan Esotericism’, in Tibetan Studies, Aris and Kyi (eds.), pp. 319–24.
Warminster: Aris and Phillips.
Walter, M. L. 1980. The Role of Alchemy and Medicine in Indo-Tibetan Tantrism.
PhD Dissertation, Indiana University. UMI 8024583.
Walter, M. 1992. ‘Jabir, The Buddhist Yogi. Part One’. Journal Indian Philosophy
20: 425–38.
Walter, M. 1996. ‘Jabir, The Buddhist Yogi. Part Two: “Winds” and Immortality’.
Journal, Indian Philosophy 24: 145–64.
Walter, M. L. 2000. ‘Cakravartins in Tibet’. Paper for the 9th Seminar of the
International Association for Tibetan Studies, 24–30 June 2000, Leiden
University.
Walter, M. 2003. ‘Jabir, The Buddhist Yogi, Part III: Considerations on an International Yoga of Transformation’. Lungta 16: 21–36.
Warder, A. K. 1970. Indian Buddhism. Delhi: Motilal Banarsidass.
Wayman, A. 1961. ‘Totemic Beliefs in the Buddhist Tantras’. History of Religions
1(1): 81–94.
Wayman, A. 1977. Yoga of the Guhyasamājatantra: The Arcane Lore of Forty Verses. A
Buddhist Tantra Commentary. Delhi: Motilal Banarsidass; New York: Samuel
Weiser.
Wayman, A. 1978. Calming the Mind and Discerning the Real. New York: Columbia
University Press.
Wayman, A. and Tajima, R. 1992. The Enlightenment of Vairocana. (Book I. Study
of the Vairocanābhisam
. bodhitantra by Alex Wayman, Book II. Study of the
Mahāvairocana-Sūtra by R.Tajima.) Delhi: Motilal Banarsidass.
Weber, M. 1976. The Protestant Ethic and the Spirit of Capitalism. Translated by
Talcott Parsons. Introduction by A. Giddens. London: Allen and Unwin.
Wedemeyer, C. K. 1999. Vajrayana and its Doubles: a Critical Historiography, Exposition, and Translation of the Tantric works of Āryadeva. PhD dissertation,
Columbia University.
Wedemeyer, C. K. 2001. ‘Tropes, Typologies and Turnarounds: A Brief Genealogy
of the Historiography of Tantric Buddhism’. History of Religions 40: 223–59.
Wedemeyer, C. K. 2002. ‘Antinomianism and Gradualism: The Contextualization
of the Practices of Sensual Enjoyment (Caryā) in the Guhyasamāja Ārya
Tradition’. Indian International Journal of Buddhist Studies 3: 181–95.
Wedemeyer, C. K. 2006. ‘Visions and Treasures: Jo-nang Tāranātha and the Historiography of the Nāgārjunian Tantric Corpus’. Paper for the XIth Seminar
of the International Association for Tibetan Studies, Königswinter, August
2006.
Wedemeyer, C. K. in press. Āryadeva’s Lamp that Integrates the Practices
(Caryāmelāpakapradı̄pa): The Gradual Path of Vajrayāna Buddhism according
398
Reference
to the Esoteric Communion Noble Tradition. New York: Columbia University
Press.
Weinstein, S. 1987. Buddhism under the T’ang. Cambridge University Press.
Werner, K. 1989. ‘The Longhaired Sage of RV 10,136: A Shaman, a Mystic or a
Yogi?’, in The Yogi and the Mystic. Studies in Indian and Comparative Mysticism,
Werner (ed.), pp. 33–53. London: Curzon Press.
Wheatley, P. 1979. ‘Urban Genesis in Mainland South East Asia’, in Smith and
Watson 1979, pp. 288–303.
Wheatley, P. 1983. Nāgara and Commandery: Origins of the Southeast Asian
Urban Traditions. Chicago, IL: University of Chicago, Dept. of Geography. (Research paper / University of Chicago, Department of Geography;
nos. 207–8.)
Whicher, I. 1997. ‘Nirodha, Yoga Praxis and the Transformation of the Mind’.
Journal Indian Philosophy 25: 1–67.
Whicher, I. 1998. The Integrity of the Yoga Darśana: A Reconsideration of Classical
Yoga. Albany, NY: State University of New York Press.
Whicher, I. 2002. ‘The Integration of Spirit (Purus.a) and Matter (Prakr.ti) in the
Yoga Sūtra’, in Whicher and Carpenter 2002, pp. 51–69.
Whicher, I. 2002–3. ‘The World-Affirming and Integrative Dimension of Classical
Yoga’. Cracow Indological Studies 4–5: 619–36.
Whicher, I. and Carpenter, D. (eds.) 2003. Yoga: The Indian Tradition. London
and New York: RoutledgeCurzon.
White, D. G. 1986. ‘Dakkhin.a and Agnicayana: An Extended Application of Paul
Mus’s Typology’. History of Religions 26: 188–211.
White, D. G. 1991. Myths of the Dog-Man. Chicago and London: University of
Chicago Press.
White, D. G. 1996. The Alchemical Body: Siddha Traditions in Medieval India.
Chicago and London: University of Chicago Press.
White, D. G. 1998. ‘Transformations in the Art of Love: Kāmakalā Practices in
Hindu Tantric and Kaula Traditions’. History of Religions 38(4): 172–98.
White, D. G. 2003. Kiss of the Yoginı̄: ‘Tantric Sex’ In Its South Asian Contexts.
University of Chicago Press.
White, D. G. 2005. ‘Review of Indian Esoteric Buddhism, by R. M. Davidson’.
Journal International Association of Tibetan Studies 1: 1–11.
White, D. G. 2006. ‘“Open” and “Closed” Models of the Human Body in Indian
Medical and Yogic Traditions’. Asian Medicine: Tradition and Modernity 2:
1–13.
Whitney, W. D. 1905. Atharva-Veda Sam
. hitā Translated with a Critical and Extegetical Commentary. Cambridge, MA: Harvard University Press.
Wiener, M. J. 1995. Visible and Invisible Realms: Power, Magic, and Colonial
Conquest in Bali. Chicago and London: University of Chicago Press.
Wijeyewardene, G. 1986. Place and Emotion in Northern Thai Ritual Behaviour.
Bangkok: Pandora.
Wile, D. 1992. The Art of the Bed Chamber: Chinese Sexual Yoga. Albany, NY: State
University of New York Press.
Reference
399
Wilke, A. 1996. ‘Śaṅkara and the Taming of Wild Goddesses’, in Michaels,
Vogelsanger and Wilke 1996, pp. 123–78.
Willemen, C. 1982. The Chinese Hevajratantra. Leuven: Uitgeverij Peeters.
(Rijksuniversiteit te Gent, Orientalia Gandensia, VIII.)
Williams, P. 1989. Mahāyāna Buddhism: The Doctrinal Foundations. London and
New York: Routledge.
Williams, R. 1966. ‘Before Mahāvı̄ra’. Journal Royal Asiatic Society 2–6.
Willis, M. 2001. ‘Inscriptions from Udayagiri: Locating Dimensions of Devotion,
Patronage and Power in the Eleventh Century’. South Asian Studies 17: 41–
53.
Willis, M. 2004. ‘The Archaeology and Politics of Time’, in Bakker 2004, pp. 33–58.
Wilson, H. H. 1832. ‘Sketch of the Religious Sects of the Hindus’ 17: 169–313,
Asiatic Researches (Calcutta).
Wilson, I. D. 1999. ‘Reog Ponorogo: Spirituality, Sexuality, and Power in a
Javanese Performance Tradition’. Intersections: Gender, History & Culture in the
Asian Context 2 (May 1999). Downloaded from www.sshe.murdoch.edu.au/
intersections/issue2 contents.html, 10 Aug. 2001.
Wiltshire, M. 1990. Ascetic Figures Before and in Early Buddhism. Berlin and
New York: Mouton de Gruyter.
Winn, M. 2006. ‘Transforming Sexual Energy with Water-and-Fire Alchemy’, in
Kohn 2006a, pp. 151–78.
Witzel, M. 1989. ‘Tracing the Vedic Dialects’, in Dialectes dans les littératures Indoaryennes, Caillat, (ed.), pp. 97–264. Paris: de Boccard.
Witzel, M. 1995a. ‘Early Indian History: Linguistic and Textual Parametres [sic]’,
in Erdosy 1995a, pp. 85–125.
Witzel, M. 1995b. ‘R.gvedic History: Poets, Chieftains and Politics’, in Erdosy 1995a.
pp. 307–52.
Witzel, M. 1995c. ‘Early Sanskritization’. Electronic Journal of Vedic Studies 1(4),
Dec. 1995, www1 shore.net/˜india/ejvs.
Witzel, M. 1997. ‘The Development of the Vedic Canon and its Schools: The
Social and Political Milieu. (Materials on Vedic Sakhas 8)’, in Inside the Texts,
Beyond the Texts: New Approaches to the Study of the Vedas, Witzel (ed.), pp. 257–
345. Cambridge, MA: Department of Sanskrit and Indian Studies, Harvard
University. (Harvard Oriental Series. Opera Minora, vol. 2.)
Witzel, M. 2001. ‘Autochthonous Aryans?: The Evidence from Old Indian and
Iranian Texts’. Electronic Journal of Vedic Studies 7(3): 1–118.
Witzel, M. 2003. ‘Vedas and Upanis.ads’, in Flood 2003, pp. 68–101.
Witzel, M. 2005. ‘Indocentrism: Autochthonous Visions of Ancient India’. In
Bryant and Patton 2005, pp. 341–404.
Wojtilla, G. 1984. ‘Notes on Popular Śaivism and Tantra in Eleventh Century
Kashmir (A Study on Ks.emendra’s Samayamātr.kā)’, in Tibetan and Buddhist
Studies Commemorating the 200th Anniversary of the Birth of Alexander Csoma
de Körös, Ligeti (ed.), vol. 2, pp. 381–9. Budapest: Akadémiai Kiadó.
Wojtilla, G. 1990. ‘Vaśı̄karan.a Texts in Sanskrit Kāmaśāstra Literature’, in The Sanskrit Tradition and Tantrism (Panels of the 7th World Sanskrit Conference,
400
Reference
Kern Institute, Leiden: Augustus 23–29, 1987, vol. 1.), Goudriaan (ed.),
pp. 109–16. Leiden: E. J. Brill.
Wolters, O. W. 1979. ‘Khmer “Hinduism” in the Seventh Century’, in Smith and
Watson 1979, pp. 427–42.
Woodroffe, Sir J. 1974. The Serpent Power: Being the S.at-Cakra-Nirūpana and
Pādukā-Pañcaka. Two Works on Laya-Yoga, Translated from the Sanskrit,
with Introduction and Commentary by Arthur Avalon (Sir John Woodroffe).
New York: Dover.
Wright, J. C. 1966. Non-Classical Sanskrit Literature: An Inaugural Lecture Delivered
on 24 Nov. 1965 by J. Clifford Wright, Professor of Sanskrit in the University
of London. London: University of London, School of Oriental and African
Studies.
Wujastyk, D. 1997. ‘Medical Demonology in the Kāśyapasam
. hitā’. Unpublished
MS.
Wujastyk, D. 1999. ‘Miscarriages of Justice: Demonic Vengeance in Classical
Indian Medicine’, in Religion, Health and Suffering, Hinnells and Porter (eds.),
pp. 256–75. London and New York: Kegan Paul International.
Yamamoto, C. 1990. Mahāvairocana-sūtra Translated into English from Ta-p’ilu-che-na ch’eng-fo shen-pien chai-ch’ih ching, the Chinese version of
Śubhākarasim
. ha and I-hsing (A.D. 725). New Delhi: International Academy
of Indian Culture and Aditya Prakashan. (Śata-Pit.aka Series, 359.)
Yamasaki, T. 1988. Shingon: Japanese Esoteric Buddhism. Boston and London:
Shambhala.
Yokochi, Y. 1999. ‘The Warrior Goddess in the Devı̄māhātmya’, in Living with
Śakti: Gender, Sexuality and Religion in South Asia, Tanaka and Tachikawa
(eds.), pp. 71–113. Osaka: National Museum of Ethnology. (Senri Ethnological
Studies, 50.)
Yokochi, Y. 2004. The Rise of the Warrior Goddess in Ancient India: A
Study of the Myth Cycle of Kauśikı̄-Vindhyavāsinı̄ in the Skandapurān.a’.
PhD dissertation, University of Groningen. Available on-line at dissertations.ub.rug.nl/faculties/theology/2005/y.yokochi/ (downloaded 2 October
2005).
Yuthok, L. Choedak Thubten 1990. A Study on the Origin of Lam ’Bras Tradition in
India between 630–940 A.D. B.Litt. thesis, Faculty of Asian Studies, Australian
National University.
Yuthok, L. Choedak Thubten 1997. The Triple Tantra by Panchen Ngawang
Choedak.Translated and annotated by Lama Choedak Yuthok. Canberra:
Gorum Publications.
Zanen, S. M. 1986. ‘The Goddess Vajrayoginı̄ and the Kingdom of Sankhu’.
Purus.ārtha 10: 125–66. (L’Espace du Temple II, ed. J-C Galey).
Zigmund-Cerbu, A. 1963. ‘The S.adaṅgayoga’. History of Religions 3: 128–34.
Zvelebil, K. V. 1973. The Poets of the Powers. London: Rider and Co.
Zvelebil, K. V. 1988. The Irulas of the Blue Mountains. Syracuse, NY: Maxwell
School of Citizenship and Public Affairs, Syracuse University.
Reference
401
Zvelebil, K. V. 1991. Tamil Traditions on Subrahman.ya-Murugan. Madras: Institute
of Asian Studies.
Zvelebil, K. V. 1992. Companion Studies to the History of Tamil Literature. Leiden:
Brill. (Handbuch der Orientalistik. 2. Abt., Indien. Ergänzungsband; 5.)
Zvelebil, K. V. 2001. Hippalos. Oxford: Mandrake.
Zvelebil, K. V. 1996. The Siddha Quest for Immortality. Oxford: Mandrake.
Zydenbos, R. J. 1992. ‘The Jaina Goddess Padmavati’, in Contacts Between Cultures:
South Asia vol. 2, Koppedrayer (ed.), pp. 257–92. Lewiston, Queenston and
Lampeter: Edwin Mellon Press.
Index
Abhayadatta 329
Abhayākaragupta 292, 329
Abhidharma Pit.aka 31, 216
Abhidharmakośa 220
Abhinavagupta 224, 235, 253, 270, 283, 325, 326,
327, 341, 349
aborigines 52
Acala 313
Achaemenid 26, 47, 50, 60, 63, 74, 75, 79–80
see also Iran; Persian
Ādi-Buddha 218
Ādityas 96, 117
Advaita Vedanta 217
Afghanistan 54, 199, 212
Africa 43, 66, 82, 98, 116
age-sets, East African 116
cult-associations, West African 82, 180, 346
Agastya 157
Agathocles 203
aggression, cultural approaches to 189, 190
aghori 130
Agni 67, 76, 156, 249
Agni Purān.a 305
agricultural societies 74, 98, 107, 344
agriculture 42–5, 47, 49
Ahura Mazdā 74–5, 80, 97
Ajanta 142
Ajātaśatru 56, 60
Ājivikas 26, 60, 75, 82, 120–1, 124, 130, 135, 174,
197
see also śraman.a
Aks.obhya 220, 225, 226
alchemy 331
Chinese 280
inner see subtle body
Islamic 278
Jaina 332
medical 275
origins of 278–85
outer see rasāyana
physical 306 see also rasāyana
Alexander 33, 60, 63, 79
Ali, Daud 272–3, 275, 347
Allchin, B. 47
Allchin, B. and F. R. 3
Allchin, F. R. 55
Allen, N. 87, 162
alpana 96 see also floor-painting, ritual
Alter, J. 181, 336
Amarāvatı̄ 197, 330
Ambat..tha Sutta 68
Ambedkar Buddhism 344–5
Am
. bika 112, 205
Amitabha 216, 220, 225–6, 227
Amitāyurdhyāna Sūtra 219
amman goddesses 248
Amoghavajra 310, 312
Ānāpānasati Sutta 139
Andaya, B. W. 186
Andhra Pradesh 83, 84
Aṅga 58–60
Angkor 200
Aṅgulimāla 245
Aṅgulimāla Sutta 131
aniconic 112
animal sacrifice 97
Aniruddha (Burmese king) 322
anthropology, social and cultural 13, 19, 66, 149
Anuttarayoga see Tantra
Āpastamba Dharmasūtra 122
apsarās 6, 8, 63, 74
and ascetics 181
Arakan 201, 309, 344–5
Āran.yakas 24, 25, 203
archaeology 22–3, 36–7, 42
Archer, W. 273, 303
Ardoxsho 109, 205, 248
arhat 162
ari monks of Burma 322
aripan 96 see also floor-painting, ritual
Arthaśāstra 29, 237, 347
ārya 48, 55, 97, 164 see also Aryan
402
Index
Arya Samaj 16
Aryan 48, 49, 55–6, 58, 68, 77, 80, 100 see also
ārya
Aryan invasion model 61, 343
Aryan origins 17, 19, 20
as.a 163
Asaṅga 199, 217, 220, 274–5, 287, 288, 341
ascetic
Brahmanical 120, 122–3, 153–64, 173, 174, 205
culture, early shared 124, 265
female 125, 126, 182–3
male 182–3
movements 48, 75, 88, 113, 119, 194 see also
śraman.a
non-Brahmanical see Ājivikas; Buddhism;
Jainas; śraman.a
Śaiva ascetics placing themselves outside
society 252
asceticism 119, 161, 164, 173, 189
and militarism 127
Asia
Central 199, 295, 343
East 194, 306–7, 309–13, 334
Southeast 47, 169, 174, 194, 197, 200–1, 240,
295, 343, 345
highland populations 92
Āśmaka 60
Aśoka 33, 60, 72, 127, 174
Aśokan inscriptions 80, 145
āśramas, four 123, 153, 158, 160
āsrava 136–7
Assaka see Āśmaka
Assam 199
As.t.amātr.kā see Mātr.kā goddesses
astrology 81
astronomy, in Kālacakra Tantra 331
asura 74, 97, 216
Aśvaghos.a 197
aśvamedha 73, 98, 113, 117, 209
Āt.ānāt.iya Sutta 142, 145
Atharvaveda 24–5, 51, 64, 161, 232, 237–8, 249,
251, 297
hymns to battle-drums 236
transmission of 237
Vrātyakān.d.a 237, 238, 242
Atiśa 264, 292, 295, 329, 334
Atre, S. 3, 6–7
auspiciousness 91, 93, 107–9, 110, 147, 159,
166–9, 344, 345
Austroasiatic 54
avadhūti 246
Avalokiteśvara 212
Avanti 49–50, 58
Avatam
. saka Sūtra 199, 217
Avestan 26–7, 80, 95, 97
403
āyāgapat.a 108
Ayodhyā 13, 50, 64, 68, 69, 71, 74, 215 see also
Sāketa
Āyurveda 275, 278–9, 282
Ayutthaya 215
Ayyapan 185
Aziz, B. 235
Babylon 80
Bactria 97, 195, 199, 203
Bāhubali 63
Bailey, G. and Mabbett, I. 175–6, 179, 180,
239–40, 347
Bailey, H. 239–40
Bakker, H. 24, 30, 71, 204, 226
Balarāma 101, 202–3 see also Sam
. kars.an.a
Bali 152, 194, 233, 313, 314, 320–1, 322, 324, 337
Bāmiyān 199, 212, 215
Bān.a 233, 293
Bangladesh
Buddhist monasteries 193–4
Barong 320
Barth, F. 346
Barua, B. 144
Bāul 232, 246, 253, 285, 335
Bayly, S. 85, 166
bcud len 278
Bechert, H. 34
Beck, G. 254
belief, identification of religion with 17
Belo, J. 320
Benard, E. 232
Bengal 246, 253, 293, 329
Delta 50, 63, 76, 79, 81–2, 184, 185, 343 see also
Ganges Delta
East 199, 344–5
Bernbaum, E. 330
Besnagar 193–4, 203 see also Vidisa
Beyer, S. 277, 287, 319
Bhagavadgı̄tā 67, 202, 203, 210–11
Bhāgavata 36, 202–3, 256
Bhāgavata Purān.a 63, 91, 162
bhaginı̄tantra 260 see also Tantra, yoginı̄ class
Bhairava (deity) 114, 152, 243–6, 266, 298,
315–17, 319, 321
becoming Bhairava 270
Tantras of Bhairava 254
Bhairavas (ascetics) 243
Bhais.ajyaguru 306–7
Bhaktapur 297, 315
bhakti 15, 210, 325, 344
Bhāradvāja 123, 150
Bharata 63, 90
Bharati, Agehananda 281
Bhārgavas 67, 72, 209
404
Index
Bhārhut 141, 142–5
bhikkhu 186, 188, 189
bhiks.u 130
bhrūmadhya 284
Bhubaneswar 240
bhuta 319
Bhutan 346
Biardeau, M. 7, 160, 257
Bihar 293
Bimbisāra 60
bindu 284
Black Slipped Ware 45
Blezer, H. 260
Bodhgayā 112
bodhi 1, 119
bodhicitta 246
identified with sexual substances 287, 350
Bodhipathapradı̄pa 329
bodhisattva
ideal 350
vow 217
body
channels within 221 see also sus.umn.ā
open model of 221
transfer of consciousness into another 221,
223, 340, 342
Bogar 276, 281
Bollée, W. 115, 124, 127, 128, 183, 184, 340
Bollywood 233
Bon religion 194
Bouillier, V. 127–8, 343
Brahmā 56, 99, 140, 150, 243–5
brahmacārin 116–17, 122, 158–9, 160–1, 185,
189
as cultural ideal 185, 343, 348
brahmacarya 132
brahman 163–4, 217
Brāhman.as 24, 25, 58, 65, 97, 114, 156, 161, 162,
202, 203, 232, 237, 281 see also Jaiminı̄ya
Brāhman.a, Śatapatha Brāhman.a
Brahmanical
asceticism see ascetic, Brahmanical
cultural pattern 151, 165–72, 193
law-codes 207 see also Dharma Sūtras
religion and culture 1, 9, 21–2, 24–30, 34, 48,
58, 61, 79, 83, 84
expansion of 76–9
ritual, in Buddhist states 215
scholarship 163
texts 56, 98
tradition, reasons for survival 345
brahmavihāra 135, 174, 222, 223
Brāhmı̄ script 80
Brahmin-killing 169
Brahminisation 171
Brahmins 55, 56, 67, 75, 76, 83, 86–7, 88, 99, 100
and purity 165, 345
as authorities for domestic ritual 164–5, 344
jat.ila Brahmins 122, 125–6, 153–4, 177
land grants to Brahmins 77
semi-ascetic style, evolution of 164, 165, 184,
188–9
sponsorship of Brahmins by kings 78, 165–6
varieties of 153–4, 177
Brahmo Samaj 16
breathing, mindfulness of 139
Brereton, J. 95
Br.hadāran.yaka Upanis.ad 69, 114, 117, 123, 225,
282
sexual rituals and practices 282
Briggs, G. 326, 332
Brockington, J. 27, 29–30, 68, 208, 221
Bronkhorst, J. 120–1, 122, 123, 125, 126, 134–5,
137, 216, 221, 222
Brooks, D. 254, 326, 335
Brown, C. H. 167
Brown, M. F. 233–4
Bryant, E. and Patton, L. 54
bSam-yas 307
BSP see Black Slipped Ware
Buchanan, F. 336
Buddha
accessing power of 228
and Brahmins 87
as term for enlightened being 70
bodies of 286
colossal statues of 215
historical (Śākyamuni) 31, 32, 33, 54, 58, 60,
68, 76, 81, 100, 101, 120–2, 151, 154, 173,
225–6, 308
original teachings of 133–40
predecessors of historical Buddha 347
revised dating for 32, 34, 50, 60, 75, 125, 126
Buddhaghosa 218–19
Buddhas
of four directions 225–6 see also man.d.ala of
five tathāgatas
Buddhism 1, 9, 10, 15–18, 25, 50, 60, 64, 68, 70,
75, 82–3, 84, 100–1, 119–26, 127–31, 169,
174, 211–16 see also śraman.a
and Brahmanical ascetics 154, 163–4
as a state religion 146, 194, 215
attraction of Tantric techniques for 270
Hinayāna 212
historical relation to Jainas 125, 126
Mahāyāna 15, 199, 212–14, 215
modernist versions of 10
monasteries and temples in early Southeast
Asia 193–4
monasteries and temples in South Asia 193–4
Index
not a protest against Brahmanical caste
system 100
origins of Buddhist saṅgha 175–80
persecution of, in China 201
philosophical schools 216–18
and meditation 216
reason for disappearance in India 344–5
religion of the countryside 194, 213
Tantric 104 see also Vajrayāna
Theravāda 15, 16–17, 31, 138, 212, 309, 314, 323,
337
Buddhist societies of Southeast Asia 146,
344
esoteric traditions 325
modernist forms of 337–8
trade, links to see trade
Vajrayāna 16, 181, 182, 194, 199, 215, 279, 314
see also Tantra, Buddhist
Buddhist
caryāgı̄ti songs 245 see also caryāgı̄ti
charismatic forest-monks 214
dialectic between urban and forest traditions
214–15
forest monk tradition 213–14, 220
monk 350
monks, and death 129
‘Pure Land’ Buddhism 216
sūtras 56, 77, 101, 123, 162, 174, 248
sūtras, Mahāyāna 211, 220, 222, 223, 227–
8, 291, 340 see also names of individual sūtras
Tantra see Tantra, Buddhist; Vajrayāna above
teachings, synthetic schemes 329–30
texts 30–2, 56, 87, 99 see also jātaka; Pali
Canon
Chinese translations of 31, 195, 309, 311,
312, 313
early 56
references to Vedas and Upanis.ads 100–1
Buddhist and Jain cultural pattern 146–51, 193–4
Burghart, R. 69
Burma 146–7, 170, 201, 309, 321, 322, 324–5, 337
C14 dating 42, 45, 125, 126 and see radiocarbon
dating
Caitanya 343
caitya 141, 145
cakra (literally, ‘wheel’)
locations within subtle body 255, 271, 277,
280, 283, 285, 289, 291, 305, 330–1, 333
Tantric ritual gathering 230
Cakrasam
. vara Tantra 152, 260, 262, 265, 286,
289, 308, 328–9, 331, 341
myth 266
cakravartin 73, 215, 304
Cambodia 337 see also Khmer
405
Cambyses 80
Campā (Northeast India) 43, 58, 126
Cāmun.d.ā 246, 247, 251–2, 319
can.d.āla 88
Can.d.ālı̄ (goddess) 251, 286
can.d.ālı̄ (low-caste woman) 245–6
Candragupta II 251
Candragupta Maurya 33, 60
Candravam
. śa see dynasties, Lunar and Solar
Cardona, G. 34–6
Carrithers, M. 213, 337
Carrithers, M. and Humphrey, C. 169
caryāgı̄ti songs 245
caryātantra see Tantra
caste system 82, 83–9, 165–6, 195 see also
Dumont; jāti; low-caste groups, with
ritual functions; varn.a
cattle-centred culture 95 see also pastoralist
societies
Cedi see Cetiya
celibacy 116–17, 125, 126, 181–2, 183, 185, 189, 350
and Buddhist Tantric practitioners 264, 292,
330, 334
and purity 182
in Brahmanical society 187–9
in Northern Thai Buddhism 185–6
Central Gangetic region 45, 47–8, 49–51, 58, 60,
61, 68, 71, 72, 75, 78, 79, 80, 81, 87, 100,
104, 177, 184, 185, 205, 207, 209, 210,
343–4
centralisation, political 43, 77, 81
Cetiya 58
Chakrabarti, D. 43, 45, 81, 111, 126
’cham 319
Champa (Southeast Asia) 201, 295
Chāndogya Upanis.ad 122, 123, 202, 204, 282, 284
Chandra, Lokesh 312
Chandraketugarh 81, 107, 110, 112, 144, 184, 205
Chanhujo-daro 7
Chatterji, B. C. 127–8
Chawla, J. 248
Chiang Mai 146–7, 150, 171
childbirth 90, 110, 131, 169
China 47, 197, 200, 215, 227, 313
Buddhism in 310, 311–12
Chinese
alchemical practices 278, 279, 341
sexology 278–9
chödpa 130
choice, individual, in early urban environment
178
cholera, goddess of 248
Choong, M. 31
Christianity 115
chronological issues 12, 27–8, 32–6, 113, 195
406
Index
chronological limits of book 9–10
Cı̄nacāra practices 281
Cittamatrā see Yogācāra
Cittar 276–8, 280, 281
city states 56
clan see kula
classical synthesis of Indian culture 193 see also
Guptas
Claus, P. 319
Coburn, T. 243
Coedès, G. 200
Cohen, R. 142, 212
coinage 80
Col.a 293, 299
Collins, S. 212
colonial period 10
Comfort, A. 273
communalism 14
Connolly, P. 284–5
Connor, L. 320
conversion experience 173
Conze, E. 216, 220
Coomaraswamy, A. 102
Cort, J. 140, 169, 232, 268
cosmology, in Kālacakra Tantra 331
Courtright, P. 322
Cousins, L. 31, 34, 137, 139, 174, 325, 338
cow-dung images 96
Crangle, E. 117
cremation grounds 257, 289, 291, 325, 326
Cribb, J. 81, 204
cult-associations, early Indian 83, 126–7, 180
cultivation, mental and physical 1, 2, 11
cults, Hellenistic initiatory 82, 84, 180, 346
Cynics 241–2
Cyrus 80
daeva 74, 97
dāı̄ 235, 249
d.āin 247
Dainichi-kyô 259
d.ākinı̄ 246–7, 249
Daks.ināmūrti 185
dakśin.āpatha 58
dalit 166 see also low-caste groups
Dalton, J. 264, 286–8, 289
Daoism 278–80
daoyin exercises 279
Darius 80
Das, R. 335
Das, V. 167, 168
Davidson, R. 199, 224, 227, 250, 256, 257, 264,
266, 292–3, 295, 302, 304, 328, 330, 348
de Michelis, E. 336
de Sales, A. 238, 249
de Silva, L. 131
de Visser, M 313
death
domain of Buddhist clergy and other
śraman.as 128–9, 187, 239–40
holy 134, 223, 340
religious meaning of 142 see also transgression
transfer of consciousness at death 342 see also
death, holy; rebirth
DeCaroli, R. 101, 102, 129, 140, 142, 254
Deccan 49, 50, 60, 61, 80, 205, 308, 335, 343
see also India, Central
Deeg, M. 157
Dehejia, V. 254–5, 300, 301
deities see also goddesses
as divine reflection of king 207
dangerous female 110, 246, 247–9, 291, 341
and death of children 248
deities in animal form, Indus Valley 95
divine patrons of aggressive rulers 207
Indo-Aryan 61, 113 see also deities, Vedic
laukika 101, 140, 142
local 101–2, 140, 165, 205, 207 see also yaks.a
of rivers and lakes 101
protective, of towns and states 101
supreme 207, 209–10, 247
Vedic 27, 28, 55, 80, 96, 97, 158, 201, 205
warrior 101, 205
worldly 101
deity
new conceptions of 201–7
demon devotee 266
Śiva as demon devotee of Buddha 266
Deshpande, M. 48, 54
devadāsı̄ 323
Devadatta 214
devas 80, 97, 102, 143
Devı̄māhātmya 30, 243, 251
dhāran.ı̄ 220
Dhargyey, N. Geshe 332
dharma 160, 210
Dharma Sūtras 15, 25, 243–5, 347
Dharmaguptaka 31
dharmarāja 73, 210, 212, 345
dhutanga practices 131
dhyāna states see meditation, dhyāna states
Dhyansky, Y. 4
dialectic between institutionalised and less
regulated components of spiritual
traditions 348
dichotomy, of good and evil 115
Diehl, C. 325
Dı̄ghanikāya 31, 68, 77, 139, 145, 202 see also
individual suttas̀
Dı̄pavam
. sa 215
Index
Dı̄rghāgama 31
disorder, religious meaning of 114–15 see also
misfortune; transgression
disrepute, deliberate courting of 240, 242
dissimulation, religion of 326, 332
Divakaran, O. 247
dividuals 207
Divyāvadāna 104
Doi Suthep 146–7
d.ombı̄ 245–6, 286
domestic ritual see ritual, domestic
Donaldson, T. 300, 301
Doniger, W. see O’Flaherty, W. D.
Dorje, G. 260, 273
Dowman, K. 328
Dravidian 19, 52, 54, 83, 90
dress 92, 303
dualists 253
dukkha 177–8
Dumézil, G. 87
Dumont, L. 85, 87, 131, 160, 165–7, 181, 182
Dundas, P. 32, 127–8, 134, 183, 224, 340
Dunhuang 287, 295
Dupuche, J. 254, 325, 349
Dūrga 5–6, 151, 204, 243, 247, 249, 250, 257,
297, 315, 321
Dut.t.hagāman.i 127
dvārapālas 102
Dvāravatı̄ 200
Dyczkowski, M. 254, 325
dynasties see also names of individual dynasties
Lunar 67, 71, 74
Lunar and Solar 63–9, 208
Solar 63–9
Dzogchen see rDzogs-chen
earth, goddess of 70, 73, 310
Eastern Gangetic 50
ecology 49 see also agriculture
ecstatic 113, 154, 158 see also shamanic
Egypt 83
ejaculation of semen
as offering to goddess 288–9
as Vedic offering 282
control over 181, 182, 274, 331 see also seminal
continence and loss; sexual, energy, uses
of
Ekānam
. śa 202, 256
Elamite 54, 80
elephants 110 see also Gan.eśa
Eliade, M. 158, 160, 340
embryology, in Kālacakra Tantra 331
Emigh, J. 320
Enamul Haque 110
English, E. 259
407
epics 102 see also Mahābharata; Rāmāyan.a
epigraphic material 23, 195
equanimity 135
Erdosy, G. 45, 56, 126
erotic images, on temples 109, 184, 299, 300
see also Khajuraho; Konarak; mithuna
and Tantra 301
erotics 22 see also kāmaśāstra
esoteric practitioners 1
ethical component, of early Buddhism and
Jainism 176, 180
external influences on South Asian cultures 81
Evans-Pritchard, E. 66, 163
Falk, H. 51, 115–16, 127, 155, 183
family life, as opposed to renunciation 181
see also auspiciousness
Farmer, S., Sproat, R. and Witzel, M. 6, 7, 54
fasting, to death 134 see also sallekhanā
female bodily functions as polluting 90–1, 182
Ferguson, J. and Mendelson, E. M. 322, 325, 337
Ferreira-Jardim, A. 137
fertility 107–9, 111–12, 169, 183, 184
feudalisation see militarisation; sāmantisation
fig-tree deity 5, 95
Filliozat, J. 278
Finn, L. 250, 321
fire, ritual, Vedic 68, 98, 122, 129
Fitzgerald, J. 30, 64, 67, 72, 208, 209
Flood, G. 222, 232, 267, 319, 336
floor-painting, ritual 96
Fluegel, P. 125
forest 123
Foucault, M. 273, 339, 345
Frawley, D. 33
Freeman, R. 319
Fruzzetti, L. 167
Fūdô 313
Fynes, R. 83–4
Gadon, E. 248
Gajalaks.mı̄ 107
galactic polity 227
Galvão, A. 89
gan.a 128
gāna-sam
. gha 58, 75
Gān.apatyas 232
Gandhāra 33, 48, 50–1, 60, 63, 104, 109, 199,
205–6, 217, 248 see also India, Northwest
Gandhārı̄ 32
Gandhi, Mahatma 182
Gan.eśa 110, 112, 298, 315–17
Ganga-Yamuna Doab 49
Ganges Delta 60, 81, 110–12 see also Bengal
Delta
408
Index
Gangetic, Central see Central Gangetic region
Gaud.apāda 217
Gaurı̄ 297
Gayā 142
gdung brgyud 235
Geertz, C. 320
Gellner, D. 314, 337
gender issues 88, 89–93, 181, 182–3, 189, 348–50
see also auspiciousness; celibacy; women,
status of
and Islam 92–3
genealogy 66 see also dynasties, Lunar and Solar
geography, religious 211
George, C. 273
Germano, D. 264
Gesar 73
Gibson, T. 199
Gitomer, D. 209
goddesses
Brahmanical 249–52
disease-causing 248–9, 250, 256
fierce 289
goddess cult 107, 184
goddess figurines 95–6 see also Mehrgarh
figurines
gradual introduction of goddess cult into
Brahmanical mainstream 256
Indus Valley 5–6, 7, 95
Warrior Goddess 247
gods see deities
Gold, D. and Gold, A. 335
Goldman, R. 67
Gombrich, R. 32, 100, 121, 123, 131, 178, 179, 245,
247, 249
Gombrich, R. and Obeyesekere, G. 338
Gonda, J. 51, 204
Gorakh (Gorakhnath) 276, 328
gotra 155
Goudriaan, T. 305–7
graha 249
Grayson, J. 311
Greece 56, 81
gr.hastha 122
gr.hya ritual 97
Gr.hya Sūtras 15, 25, 164–5
gtum-mo 286
Gud.imallam 204, 205
Guenther, H. 191, 273, 281, 286
guhyābhis.eka 287
Guhyagarbha Tantra 260, 264, 287,
288
Guhyasamāja Tantra 222, 259–60, 262,
264, 287, 288–9, 328–9, 331
Gujarat 125, 185, 346
Gupta, S. 167, 267
Gupta, S. and Gombrich, R. 297, 324
Guptas 33, 64, 67, 71, 74, 165, 195, 197, 203, 207,
208, 211, 217, 237, 247, 251, 269, 292, 346,
347
Gurjara-Pratı̄hāra 293
Gutschow, N. 298, 315
guthi 315
Haaland, G. and Haaland, R. 96
Hallisey, C. 212
Hamilton, S. 139, 216, 223
Hanseatic League 82
Hanssen, K. 253, 335
haoma 33
Harappa 3 see also Indus Valley, cultural
tradition
Haravijaya 277, 299
Hardy, F. 210
Harikeśa 151
Hāritı̄ 109, 112, 205–6, 314
Harivam
. śa 29, 67, 256
Harlan, L. 303
Harman, W. 172
Harner, M. 234
Harper, D. 278
Harper, K. 250, 251
Harper, K. and Brown, R. 229
Harrison, P. 137–8, 219
Hars.a (Kashmiri king) 301, 306, 334
Hars.a (North Indian king) 293
Hars.acarita 233
Hart, G. 83, 86, 88, 89–93, 134, 235–7, 321
Hartzell, J. 68, 259, 284, 331
hat.ha yoga 279, 336
Hat.hayogapradı̄pikā 336
Hayagrı̄va 262
healing deities 305
health, long life and immortality, practices to
obtain see sexual, practices; Tantra
Heesterman, J. 25, 51–2, 115, 117, 155, 160
Heilijgers-Seelen, D. 277
Hein, N. 67, 72, 210–11, 335
Heitzman, J. 347
Heliodorus 203
Hellenistic mystery cults see cults, Hellenistic
initiatory
Hemacandra 221, 224, 273, 289, 295, 332–3
Hephthalite 292
Herakles 206, 241 see also Hercules
Hercules 241 see also Herakles
Herrmann-Pfandt, A. 249
Hevajra Tantra 250, 260, 262, 264, 285, 286, 289,
308, 329, 330, 341
in Cambodia 304, 307
man.d.ala of 251
Index
hijra 169
Hiltebeitel, A. 3, 29–30, 64, 208, 209, 266
Hindu 17
Hinduism 1, 10, 15–18, 24, 26 see also
Brahmanical, religion and culture
as problematic term 17, 195
devotional styles 335, 344 see also bhakti
establishment in modern form 195, 198
modern Hinduism and local deities 151
modernist versions of 10, 20, 335, 336
Hiṅglāj 257, 295
Hinnells, J.
Hodge, S. 259
Hoek, van den, A. 314, 315
Hoernle, A. 55
Hooykaas, C. 324, 337
Hopkins, T. 48, 50, 51, 60, 61, 73, 343
Horsch, P. 123
Houben, J. 97
Houtman, G. 174, 321
Howe, L. 321
Hsu, E. 282
Huangting Jing 279
Huber, T. 266, 286
Hudson, D. 203
human sacrifice 317, 321
Hūn.a 292
Huns 292
Huntington, J. 109, 226
Huntington, S. 308
hymns, Vedic 155
dialogue hymns 157
philosophical and speculative 162–4
iconographic material 24, 42
iconography 206
Id.ā 68
Iks.vaku 64, 67, 68–9
historical dynasty 195
Ilā 67–8
Iltis, L. 315
India
Central 49 see also Deccan
Eastern 205 see also Assam; Bengal; Orissa
North 205 see also Kuru-Pañcāla Region
Northeast see Bengal Delta; Central Gangetic
region
Northwest 50, 61, 79–81, 295, 343 see also
Gandhāra
South 47, 63, 79, 170, 171, 199–200, 212,
235–6, 249, 343
links to China 281
Western 335
Indianisation 199, 201
Indic religions 1
409
individual
self-conscious, development of 120, 177
subordinated to greater whole 207
Indo-Aryan
cultural heritage 61, 73–5, 76, 83, 99, 113, 166
see also Vedic, deities; Vedic, religion
deities see deities, Indo-Aryan
languages see languages, Indo-Aryan
religion, early 96–7 see also deities,
Indo-Aryan; deities, Vedic
Indo-European aspects of caste system 83
Indo-Greek 35, 63, 80, 81, 195, 202 see also
Seleucid
Indo-Parthian 195
Indra 56, 61, 74, 96–7, 99, 140, 150–1, 169, 203,
209, 238
Indrabhūti (King of Zahor) 302, 334
Indus Valley 60
cultural tradition 2–8, 19, 23, 41, 43, 49, 94–5
integration era 41, 54, 95
localisation era 41, 55, 95–6
seals 54, 95
Ingalls, D. 162, 241–2, 252
initiation 252, 346
in early śraman.a cults 346–7
sexual form 252, 327, 349
two systems of initiatory knowledge 346
inscriptions, see epigraphic material
insight 1
liberating 136–7, 164, 174, 252–3, 259, 264,
269–70, 271, 277, 287, 288, 289, 291–2,
339, 341, 342, 351–2 see also prajñā
Insler, S. 74, 96, 97
Iran 26–7, 33, 47, 48, 50, 54, 60, 74, 97, 109, 113,
127–8, 163, 199, 203 see also Achaemenid;
Avestan; Parthian; Persian
iron 43, 45–7, 49–50, 81, 200
Isaacson, H. 30, 259
Isibhāsāiyam 124
Isis 84, 180
Islam 9, 115
modernist versions of 20, 335
Ismailis 335
Jagannāth 151, 336
Jaiminı̄ya Brāhman.a 158, 282
Jainas, Jainism 1, 10, 15–17, 18, 25, 34, 50, 60, 63,
64, 68, 70, 75, 82–3, 84, 99, 100–1, 112,
119–25, 126, 127–31, 137, 140, 146, 164,
169, 174, 175, 180, 185, 211, 216, 291, 295
see also śraman.a
and Nāth Siddhas 332, 333
fierce goddess rituals 232
internal subtle-body practices 289
model of society 344
410
Index
Jainas, Jainism (cont.)
modernist versions of Jainism 10, 20
original teachings of 133–5
survival in India 344–5
texts 32, 56, 76
Jaini, P. 63, 140
Jairazbhoy, R. A. 8
Jalandhara 257
Janaka 56, 69, 70, 71, 80
Janakpur 69, 71
janapada 47 see also mahājanapadas
Japan 127–8, 312–13, 337
jātaka 69, 74, 143, 248
jāti 85, 87
Java (Indonesian island) 215, 241, 295, 313, 337
Javā (kingdom) 304
Jayakar, P. 6, 7, 8
Jayavarman II 304
Jayavarman VII 304
jhāna states see meditation, dhyāna states
Jhavery, M. 268
jñāna-kān.d.a 159, 164
Jogjakarta 74, 215
Jordaan, R. and Wessing, R. 321
Jorwe culture 49
Joshi, M. 30
Jurewicz, J. 2000
Jvālāmālinı̄ 268
Jwālamukhı̄ 257
Kaelber, W. 156, 158–60, 163–4, 181
kaivalya 1, 119, 223
Kak, S. 33
Kakar, S. 39
Kālacakra Tantra 218, 224, 260, 295, 329, 330–2,
334–5, 341
Kālamukhas 243, 254
Kalhan.a 301, 322, 328
Kālı̄ 115, 204, 246, 247, 249, 319
Tantras of 254
Kālidāsa 74
Kalighat 257
Kālikā Purān.a 6
Kāmāks.ı̄ 151
Kāmarūpa 257, 280
kāmaśāstra 272–3, 347
Kāmasūtra 272–3, 275, 289
Kamboja (Northwest India) 60, 63
Kambuja (Southeast Asia) 295, 304 see also
Khmer
Kanchi 151
Kān.ha 245–6, 328 see also Kr.s.n.ācārya
Kannada (Karnataka) 319
Kannaki 303
Kanyakumari 151
kāpālika-style practitioners 220, 233, 243, 245,
246, 252, 257, 265, 268–9, 270, 275, 276,
289, 291, 326, 327, 341, 348, 349–50
Kāpālikas 243 see also kāpālika-style practitioners
Kapferer, B. 319
Kapila 216
Kapstein, M. 303
Karakhoto 295
Kāran.d.avyūha Sūtra 212
Karlsson, K. 107, 219
karma 124, 242, 252
karma-kān.d.a 159, 163–4
Karnataka 83
Kārttikeya 150, 206, 250 see also Skanda
Kashmir 298, 299, 301, 308, 322, 328, 334
Kāśı̄ 56, 58, 61, 76, 280 see also Vārān.ası̄
Kat.ha Upanis.ad 114
Kathmandu 297, 315
Kathmandu Valley 297, 314, 317, 337
Kātyāyana 34–5
kaula 229, 235, 257, 276, 277, 291, 297, 299, 301,
326–7, 328, 333, 341, 349–50
Kaulajñānanirn.aya 277, 283, 285
Kauśāmbı̄ 43, 45, 58
Kaut.ilya 29, 213
Kendall, L. 234
Kenoyer, J. M. 43
Kerala 83, 319, 322
keśin 113, 157
kevala 1
Keyes, C. 170, 186, 193–4
Khajuraho 300, 302, 344
Khan, D. 335
khandroma 246
Khāravela 195
Khare, R. 269
Kharos.t.hı̄ script 79, 80
khecarı̄ mudrā 279
Khmer 241, 295, 304–7, 334
Khubilai Khan 334
Khuddakanikāya 31
Kim, C. 129, 234, 322
king
and Tantric theology 297
as converter of power, in South India 236
Buddhist model of kingship 212 see also
dharmarāja
eroticisation of kingship and warfare 293,
300, 348
Four Great Kings (Lokapālas) 96, 144–5, 146,
309–10
kingship, models of 208, 209 see also warrior
king, wisdom king below
kingship, nature of 66–7, 69, 72, 74
man.d.ala as model of kingship 226–7
Index
renunciate king 70, 72, 75, 209 see also rājar..si
sympathies with śraman.a traditions 177, 180
warrior king 69, 73–4, 113 see also cakravartin
wisdom king 69–71, 72, 73, 80, 82, 175, 177,
205, 209, 293, 345 see also dharmarāja
King, R. 336, 338
Kinnard, J. 308
Kinnarı̄ Jātaka 74
kinship
bilateral 170
Dravidian 170
in Central Gangetic region 90
matrilineal 170
matrilocal residence 186
patrilineal 63, 92, 170
Kirsch, A. T. 170
Klimburg-Salter, D. 101, 194, 199, 212, 215
Kōbō Daishi 313
Kohn, L. 278, 279
Kohn, R. 319
kolam 96, 167 see also floor-painting, ritual
Konarak 300, 302, 344
Kongōchō-gyō 259
Korea 127–8, 234, 310, 311, 312, 313, 322
Korvald, T. 315
kośa 284
Kosala 50, 58, 61–3, 68, 70, 71, 75, 76
Kosambi, D. D. 7–8, 79, 81, 248
Kot.avı̄ 256
kriyā practices see Tantra
Kriyāsam
. graha 329
kriyātantra see Tantra
krodha-vighnāntaka 262, 329
Kr.s.n.a 64, 67, 151, 202–3, 210, 226
as child and lover 211
Kr.s.n.ācarya 328 see also Kan.ha
Kr.ttikā sisters 249–50
ks.atriya 56, 60, 67, 76, 82, 83, 86–7, 99, 127, 164,
166, 177, 187, 208–9
Ks.emendra 301, 306, 322, 328
Ks.udrakāgama 31
Kubera 102, 104, 112, 144, 205 see also Vaiśravan.a
Kubjikā 265
Kubjikātantra 277, 285
Kukkuravatika Sutta 162
kula 231, 235, 269, 276, 291, 341
essence of, transmitted sexually 252, 287, 291
Kulacūd.āman.i Tantra 250, 321
Kulke, H. and Rothermund, D. 50
Kumāra 206
Kumārajı̄va 310, 311–12
Kumārı̄ 151
Kumbhakāra Jātaka 70
Kun.d.alinı̄ 255, 277, 283, 286
Kuntāpa rituals 52
411
Kuru state see Kuru-Pañcāla Region
Kuru-Pañcāla Region 45, 48, 50, 55–6, 58, 60,
61–4, 71, 73, 74, 75–8, 79, 87, 98, 99–102,
113, 115–17, 155, 163, 207, 209, 281, 343
brahmins of 155, 225, 226
Kuruks.etra 51, 52–3, 64, 71, 75, 100, 102
Kus.an.as 33, 165, 195–7, 203, 205–7, 217, 226,
247
Pre-Kus.ān.a 205–7
Kūt.adanta Sutta 77
Kuvera see Kubera
Kvaerne, P. 246
Ladakh 237
Lai, W. 309
Lajjāgaurı̄ 108
Laks.mı̄ 107, 110, 144, 167, 169, 204, 205, 248, 297
Lakula 254
Lakulı̄śa 241
Lakulı̄śas 243 see also Lakula; Pāśupatas
Lalon Fakir 285
Lalou, M. 205
lam ’bras teachings 330
lam rim teachings 329
Lamotte, E. 205
Lan Na 146, 147–9
languages 42, 48, 53–5
Indo-Aryan 25, 27, 28, 42, 52, 53, 54, 55, 61,
78, 83, 99, 155
Laṅkā 71
Laṅkāvatāra Sūtra 217
Lariviere, R. 155
laukika 145, 146, 149, 170, 187 see also deities,
laukika
lay practitioners 333, 334–5
Leslie, J. 91
Levy, R. 298, 314, 315
liberating insight see insight, liberating
Licchavi 67
limits of our knowledge 11–12
Lindtner, C. 213
lineages of practice 253, 260
divergent lineages 260
liṅga 203, 280
worship of 204
Linrothe, R. 24, 260, 328, 340
periodisation of Buddhist Tantra 262, 329
living without grains 278
Locke, J. 314
lokottara 145, 146, 149, 187
Lopāmudrā 157
Lorenzen, D. 132, 233, 241–2, 245, 322
Loseries-Leick, A. 279
Lotus Sūtra 216 see also Saddharmapun.d.arı̄ka
Sūtra
412
Index
low-caste groups
myths of 269
trance-dancers 298, 319
with ritual functions 83, 84, 86, 88, 166,
235–6, 252, 258, 268–9, 291
without ritual functions 88
Lubin, T. 132, 161, 164–5, 210, 344
Lungmen 215
Lynch, O. 269
Mabbett, I. 200–1, 264
Maccha 58
Macdonald, A. W. 266
Madan, T. N. 166–8
Madhva 218
Madhya Pradesh 254–5
Madhyamāgama 31
Mādhyamika 216, 220
Madurai 151, 171
Magadha 47, 50, 58–60, 73, 75, 76, 77, 81, 104
Magar 238, 249
magical power 162, 174 see also ritual, pragmatic
Mahābharata 15, 27, 29–30, 63, 64–8, 69, 72, 90,
91, 114, 122, 123, 162, 204, 207–8, 209,
215, 221, 240, 249, 250, 340
Mahabodhi Society 16
mahājanapadas 42, 47, 48, 56–60, 75, 79, 81,
310, 311
as city states 56
oligarchies 58
republics 58
Mahākāla 151, 152, 266
Mahākāśyapa 124
Mahāmāyūrı̄ 102, 104, 150, 151, 247
Mahanadi 81
Mahāsaccaka Sutta 135, 136
Mahāsamaya Sutta 145
mahāsiddhas 328–9
Mahāvairocanābhisam
. bodhi Sūtra (or Tantra)
227, 259, 262, 313
Mahāvam
. sa 144, 145, 215
Mahāvı̄ra 33, 54, 68, 100, 120–2, 124, 125, 126,
130, 151, 173
predecessors 347 see also Nami; Nimi; Parśva
mahāvrata
of Brahmin-killers 243–5
of Jainas 132
of Kālāmukhas 132
of kāpālika-style practitioners 243–5
of vrātya 52, 116, 132, 158
Mahāyāna see Buddhism; Buddhist
Mahāyānasūtrālam
. kāra 274–5, 287
mahāyogatantra see Tantra
Mahis.amardinı̄ 247, 250, 257
Maithili 336 see also Mithilā
Maitreya 199, 220
maitrı̄ see mettā
Maitri Upanis.ad 221, 222, 284
Majjhimanikāya 31, 70, 139, 162 see also
individual suttas
Makhādeva Sutta 70
Makkhali Gosala 120, 124, 173
Malabar 83–4
Malwa culture 49
Manasā 112, 205
man.d.ala 151, 224–7, 250, 291–2, 293, 305, 340
and vyuha scheme 226
as representing galactic polity 226–7
as social model 224
Australian Aboriginal 224
goddesses incorporated in 297
gradual incorporation of goddesses and
Bhairava deities in Buddhist man.d.alas
262, 340
in Br.hadāran.yaka Upanis.ad 225
in Buddhist architecture 307, 309
in Suvarn.aprabhāsa Sūtra 225–6
in Vrātyakān.d.a of Atharvaveda 225, 238
incorporating fierce goddesses 341 see also
Mātr.kā goddesses
of Bhais.ajyaguru 307
of eight bodhisattvas 227
of five tathāgatas 226, 262
of Hevajra Tantra 251
of Kālacakra Tantra 331
of Peaceful and Fierce Deities 260
visualised 259
mangala 168
Man.ibhadra 205
Mañjuśrı̄ 199
Mañjuśrı̄mūlakalpa 305
mantra 220, 254
Manu, Laws of see Manusmr.ti
Manusmr.ti 88, 115, 116, 122, 210, 237, 303
Mar-pa 334
Marglin, F. A. 167–8, 323, 336
Mariamman 248
maritime contacts
India and China 280
India and Egypt 83–4
marriage
in Buddhist societies 149, 187
in Vedic religion 98
marriage system 63, 89–90
Marriott, M. 171, 207, 256
Marshall, Sir John 3, 19
martial imagery 127–8
Masakado, Tairo no 313
Masefield, P. 88, 137–8, 174
Maskarinec, G. 249
Index
masked ritual dances see ritual
Mataram 313
Mathurā 50–1, 58, 61, 109, 112, 205–6, 248
Mātr.kā goddesses 298, 305 see also Navadurgā
As.t.amātr.kā (eight Mātr.kā) 298, 315
Saptamatr.ka (seven Matr.kā) 250–2, 269, 297
Matsya see Maccha
Matsya Purān.a 151
Matsyendra (Matsyendranāth) 276–7, 326, 328,
332
Mauryas 33, 47, 56, 60, 72, 95–6, 107, 199, 347
post-Mauryan period 165
Mawangdui texts 278, 280, 281
Mayer, R. 220, 266
McAlpin, D. 52, 54
McDaniel, J. 265, 303, 325, 335, 336
McEvilley, T. 4, 284
medicine 22, 221, 248–9, 275, 281–2
meditation 1, 2, 156 see also Tantra; yoga
and contemporary society 350–3
as technology of the self 347
Brahmanical 220–3
buddhānusmr.ti 219, 220, 259
Buddhist 218–20
and philosophy 216
early 135–7
Mahāyāna 213, 218–19
Theravāda 218–19
definition of 342
dhyāna states 135–7, 139, 174, 219, 221–2, 223,
333, 337, 339, 340
Jaina 134, 221, 223–4
mindfulness of breathing 219
reality of effects 351
sexual practices 220, 231, 232, 264, 274, 341
see also Tantra, sexual practices
transgressive and non-transgressive 275,
341–2
techniques associated with fierce deities 267
using earth-kasin.a 219
visualisation of Buddha 220
Meenakshi, K. 327
Meghavāhana 83, 195
Mehrgarh figurines 94–5
Mekong 47
Melanesia 346
Melusine 74
men, young, and aggression 189–90
Mendelson, E. M. 325, 337
Menges, K. 239–40
menstrual blood
of goddess 278
ritual use of 253, 327
menstruation 90–1, 168–9
mercury 276, 277–8, 280
413
Merz, B. 315
Mesopotamia 79
mettā 135
Meyer, J. 328
microcosm and macrocosm 221
militarisation of Indian society 92 see also
samantisation
Mināks.ı̄ 151, 171, 172
mind–body processes 2, 173
misfortune, religious meaning of 114–15, 129
see also disorder; transgression
Mitanni 27, 32, 96, 97
Mithilā 56, 61, 64, 69, 70, 167, 175, 347 see also
Videha
Mithras 127–8
mithuna
as cultural ideal 184, 344, 348
images 107, 112, 184, 300
Mitra 96–7
Miyāvutta Sutta 74
mleccha 203
Mohan, P. 343
Mohenjo-Daro 3, 4–7 see also Indus Valley,
cultural tradition
moks.a 1, 119, 159–60, 273, 289
Mon 201
monasteries
monastic universities 307
of northeast India, as Tantric centres 264
rituals for construction and consecration 329
monasticism 119
Christian 183
Mongols 194, 295, 312, 337
monism 207, 218
monotheism 207
Moon 280
morality 181–2 see also ethical component
mudang 322
muggu 96 see also floor-painting, ritual
Mulder, N. 325
Müller, M. 27, 33
Muller-Ortega, P. 270, 349
Mumford, S. 177
Munda 54
muni 113, 157, 158
Murti, T. R. V. 218
Murugan 172, 185, 235, 281
Mus, P. 255, 256
Muslim expansion into South Asia 92–3, 295,
303, 342
Myongnang
mystery cults 180
nād.ı̄ 255, 271, 285, 289, 305, 330–2
nāga 101, 102, 143
414
Index
Nagarajan, V. 167
Nāgārjuna 22, 27, 34, 197, 211, 217, 220, 264, 328
Nāgārjunakon.d.a 112, 184, 197
nakedness 130, 158, 303
Nālandā 302
Nami 69–70, 124, 140 see also Nimi
Nandas 60, 81
Nārada 124
Nārāyan.a 203
Naritasan Shinshōji 313
Nāropā 328, 329
Nāth Siddhas 224, 265, 276–8, 279, 291, 326, 333
and Jainas 332, 333
householder Nāths 335
Nattier, J. 199
Navadurgā 298, 315, 321
navarātrı̄ 297
Nayaka 171
NBPW see Northern Black Polished Ware
Ndembu 98
neidan 280
Nemi 125, 185, 332
Nepal 194, 237, 252, 308, 344–5 see also
Kathmandu Valley; Newar
Nepali, G. S. 314
Netra Tantra 298–9, 325
Newar 152, 179–80, 292, 297, 314, 315, 317, 323
see also Kathmandu valley; Nepal
Newman, J. 330–1
Nichter, M. 319
Nikāyas 212
Nimi 69–70, 124 see also Nami
Ninno-kyō see Renwang Jing
nirgun.i devotional traditions 344
nirvān.a 1, 119
non-dualism 253
Northern Black Polished Ware 45, 49–50, 81
numismatic material 23–4, 42
O’Connor, R. 1989
O’Flaherty, W. D. 157–8, 162, 166, 204
Oc Eo 47, 200
Odantapurı̄ 308
Od.d.iyāna 295
Odin 116
Oesho 203, 204, 206
Okkāka 68
oligarchies 58
Olivelle, P. 20, 27, 33, 51, 123, 124, 153–4, 158,
160, 162, 177, 221, 283, 284
chronology for Upanis.ads 284
Openshaw, J. 265, 335
Oppitz, M. 249
ordination, temporary 147, 149, 185–6
as rite of passage 186
orgasm and divine bliss 283
Orientalism 10
Orissa 240, 241, 254–5, 308, 320, 336
Orofino, G. 259, 331
orthogenetic 18–22, 98, 160, 204, 255, 280–1
outside, viewing society from 175, 178
Owens, B. 314
Padmāvatı̄ 112, 205, 268
Padoux, A. 254
Paesi 124
Pagan 295
Paharpur 307–8
Painted Gray Ware 45, 49–50
Pāla 293, 302, 307, 308, 328
importance of external patronage of
Buddhism in Pāla state 308
Palani 281
Pali Canon, Pali textual tradition 12, 16–17,
31–2, 87, 88, 212, 245, 248
Kālı̄ in 247
Pali literalism 121
Pallava 83, 199, 203
pan-Indian religious substrate 255
Pañcāla 52, 58 see also Kuru-Pañcāla
Pāñcarātra 203, 226, 227, 232, 267, 291
pañcāvars.ika rituals 215
Pāñcika 104, 109, 205–6
Pān.d.avas 209
Pān.d.ya 171
Pān.ini 29, 34–6, 56, 71, 202, 213
parda 170
paritta 131
Parpola, Asko 4, 6, 7, 8, 27, 48, 52, 53, 54, 55, 97
Parry, J. 168
Parsva 124, 125, 126, 173, 185, 332
Parthian 197
Parvatı̄ 151, 171, 182, 204, 249, 297
pastoralist societies 74, 97, 98, 209
values of 98, 161, 163, 183, 343
Pāśupata Sūtra 242
Pāśupatas 162, 240, 241, 242, 243, 252, 254,
268–9, 270, 304, 348, 349
Pāśupati 162
Pāt.aliputra 58, 104
Patan 297, 315
Patañjali 29, 35–6, 101, 132, 137, 221–2, 336
Pattinı̄ 84, 150
Pāyāsi 124
Perfection Stage of anuttarayoga Tantra 287, 288
Periplus of the Erythrean Sea 83
Persian 29, 33 see also Achaemenid; Iran
perspectives, internal and external 234–5, 238–9,
349–50
Peru 233
Index
PGW see Painted Gray Ware
Pharro 109, 206
philosophical understandings of Indian religion
350, 351
’pho ba practices in Tibet 221, 342
’phrul ’khor 279
physical exercises 279
Pirart, E. 29
pı̄t.has (ritual gathering sites) see Śākta, pı̄t.has
Plato 284
Pleiades 6, 249
pluralism, religious 197, 211, 218
politics, contemporary, and historical writing 84
Pollock, S. 36
pollution see purity
polytheism 207, 218
Ponorogo 320
Popper, K. 176
Porus 79
Possehl, G. L. 43
post-colonial scholarship 12 see also
representation, issues of
pottery 45 see also Northern Black Polished
Ware; Painted Gray Ware
power
civilised and uncivilised forms of 204
legitimate exercise of 204
royal 205
Prajāpati 150, 156, 238
prajñā 136, 174
prajñā Tantras 260 see also Tantra, yoginı̄ class
Prajñāpāramitā (goddess) 304
Prajñāpāramitā Sūtras 217, 309
prakr.ti 222, 223
prān.a 238, 271, 284–5, 289, 333, 341
pratyekabuddha 70
Pratyutpannabuddhasam
. mukhāvasthitasamādhi
Sūtra 219, 220, 223, 226
Prebish, C. 212
precepts, Buddhist 132
prosperity, as religious value see auspiciousness
Protestant presuppositions 16–18
Protestantism, rise of 179
Proto-Elamite 4
Proto-Elamo-Dravidian 54
proto-śraman.a movements 125, 126, 175,
346
psychic economy 187, 189, 317–18
psychoactive substances 98 see also soma
psychoanalysis 115
pulse-reading 282
Pun.d.ra 60
Purān.as 15, 27, 30, 65–6, 102, 211, 212,
243–5
Puri 151, 323, 336
415
purification of the saṅgha 214
purity 166, 188, 345
purohita, royal 237, 251, 297, 298–9
Puru 63, 64, 67, 79
purus.a 222, 223
Purus.a-Sūkta 166
purus.ārtha 273, 289
Pus.karavatı̄ 60
Pyu 201
qi 278–9, 282, 341
qigong 279
Quigley, D. 166, 314, 315
Qvarnström, O. 224, 232, 268, 332–3
Rabe, M. 300, 301
Rādha 60
radiocarbon dating 19, 23, 45, 50 see also C14
dating
raiding expeditions, for cattle 115
Rājagr.ha 43, 58, 104, 125, 126, 205
rājaguru 298
Rajaram, N. S. 33
rājar..si 70–1 see also king, renunciate
rājasūya 98, 113
Rājatarangin.ı̄ 301
rājayaks.man 275, 278–9
Rajput 67, 303
raks.ası̄s 248
Rāma 64, 68, 69, 70–1, 73, 205, 208–9
Rāma Jāmadagnya 209
Rāmānuja 218, 343
Rāmāyan.a 15, 27, 29–30, 50–1, 64–5, 69, 70–1,
74, 91, 114, 207–9, 215, 293
Rangda 320–1
rasāyana 275, 276, 278
Ras.t.rakūt.a 293, 308
Ratnagiri 308
Ratu Kidul 74
Ravidas 269
Ray, R. 213–14, 220, 264, 348
rDzogs-chen 264, 281, 288
rebirth
in heaven 122, 223, 340 see also death, holy;
yoga
liberation from 119–20, 122, 124, 134, 174, 175,
179
Wheel of 142
regional issues 28
Reid, A. 89, 170
religion
everyday see auspiciousness; ritual, pragmatic
pragmatic see ritual, pragmatic
social and political grounding of 170, 176,
179, 228, 342–3
416
Index
renunciate traditions 1, 119, 125, 126, 131, 181 see
also ascetic, movements; śraman.a
origins of 175–80
renunciation 131, 160, 173, 178–9 see also
renunciate traditions
Renwang Jing 309, 310, 313, 324
representation, issues of 10–11, 12–14
republics 58
reversal, of everyday life 273–4
revitalisation movement 176
R.gveda 15, 20, 24–9, 33–4, 41, 51–3, 55, 58, 63–5,
74, 92, 95, 96, 97, 113–14, 115, 123, 132,
154, 156–8, 162–3, 203, 204
rgyud 229
Rhie, M. 266
rice 42, 43, 200
ring stones 107
ritual
abandonment of village ritual 177
actions, sets of four or six 268
aggressive and destructive rituals, market for
349
ascetics taking over functions of hereditary
practitioners 239
Brahmanical, for householders 177
domestic ritual 97, 118
kingship, Brahmanical rituals of 98, 101,
194
kingship, Buddhist rituals of 101, 194
kingship, low-caste rituals to support 235–6
magical see pragmatic ritual below
masked ritual dances 298, 315, 320, 321
pragmatic ritual 155, 174, 188, 189, 194, 232,
233, 242–3
psychic effectiveness 317–18
rituals of everyday life 171
sexual 281
śrauta ritual 98, 164
Vedic sacrificial 155, 166
village ritual 194, 347
women’s ritual 10, 132, 167, 182
rNying-ma-pa 287
Robinson, J. 328–9
Robinson, S. 167
Roman 197
trade with Malabar coast 74, 84
trade with Southeast Asia and China 200
Roy, Ram Mohan 17
Rozario, S. 235
R.s.abha 63
.r.si 113, 123, 124, 147, 154–5, 156–7, 158, 163, 209,
249
.rta 163
Rudra 3, 6, 114, 116, 158, 204, 243
Ruegg, D. S. 255
.sadaṅga 222
Saddharmapun.d.arı̄ka Sūtra 216, 223, 309, 310
sādhaka 306
sādhana 161, 287, 336
sages 147
Said, E. 10
Śailendra 295
Śaiva 101
dualist 253
myths 256, 267
non-dualist 253, 254
‘orthodox’ 322, 336
religion in Southeast Asia 304–7
Siddhānta 253, 267
systems of thought and practice in mediaeval
Kashmir 253
Tantra see Tantra, Śaiva
Śaivism 207
Śaivite
ascetics 240–5
Śaka 195–7
Sāketa 45, 50, 58, 71 see also Ayodhyā
Śākta 245, 254, 255, 257, 265, 289, 336, 341
disaggregating 257, 336
pı̄t.has 151, 254, 257–8, 265–6, 280, 341
Brahmanical mythology of 257
proto-Śākta 255
śakti 277, 297
Śakuntalā 63
Śākyamuni see Buddha, historical
Śākyas 58, 68
sallekhanā 134
Salomon, C. 265, 285, 335
Śālvas 76
samādhi 174, 219, 221–2, 223, 339, 340
sāmanta 293, 295
samantisation 92, 170, 293, 296, 303, 348
Samatat.a 307
samatha 338 see also śamatha
śamatha 137, 337
Samaveda 113
samaya 288–9
Samayamātr.kā 306
Śambhala 295
Sam
. dhinirmocana Sūtra 217
Sam
. kars.an.a 101, 114, 202, 205 see also Balarāma
Sām
. khya 216, 223
Sam
. nyāsa Upanis.ads 221
sam
. prajñāta samādhi 222
Sampurn.ānand, Hon. Sri 238
sam
. skāra 137, 171, 188, 211, 283
Samuel, G.
Civilized Shamans 337
Mind, Body and Culture 224
Sam
. vara Tantra see Cakrasam
. vara Tantra
Index
samyama 222, 223
Samye see bSam-yas
Sanatkumāra 202
Sāñcı̄ 107, 110, 141, 302
Sanderson, A. 245, 251, 252–3, 255, 256, 264,
267, 276–7, 292, 298–9, 304, 325–6,
327–8, 332, 349
saṅgha 75, 128, 162, 212
admission to 154
purification of 337
Śaṅkara 27, 30, 185, 223, 240, 271, 284, 289, 322,
343
Śaṅkaravijaya 322–3
Sankhu 315
sannyāsin 123, 129, 160, 188–9, 210, 347, 350
Sanskrit 26, 27, 165, 200
Sanskritisation 53, 78, 171, 199
Saptamātr.kā see Mātr.kā goddesses
Śāradātilaka 305
Sarāha 328
Sarasvatı̄ 41, 53, 268
Śāriputra 124
Sarvāstivādin 31
Sarvatathāgatatattvasam
. graha 259, 262
S.as.t.hı̄ 110, 167, 248
Śatapatha Brāhman.a 68, 69, 74, 76, 114, 116, 284
Śatarudrı̄ya 114, 204
Sātavāhana 83, 195–7
S.atcakranirūpana 279
satı̄ 303
Satı̄ (consort of Śiva) 257
Satipat..thāna Sutta 139
S.at.karmadı̄pika 305
Schaeffer, K. 232
Schipper, K. 278–9
Schlingloff, D. 219
Schmithausen, L. 135
Schopen, G. 142
Schrempf. M. 319
Scott, D. 199
sculpture 202–3, 205 see also iconography
Scythians 165
Sdok kak Thom inscription 299, 304–5
Searle-Chatterjee, M. and Sharma, U. 166
Seleucid 60 see also Indo-Greek
semen 280
identified with bodhicitta 287, 288–9
seminal continence and loss 181, 275–6, 277,
282–3, 288 see also ejaculation of semen,
control over
Sena 302
sex and violence 302
sexual
energy, uses of 156–7, 158–9, 161, 181
excitation, and movements in subtle body 283
417
fluids, as fuel for yoginı̄s 252–3
fluids, as polluting and ritually powerful 231,
265, 269, 326, 327, 349
intercourse, as symbolically equivalent to
Vedic sacrifice 282
practices, aestheticisation of 276
practices, origins of 278–85
practices in Gupta period 271–6
practices in meditation see meditation
practices in Tantra see Tantra
practices in Vedic and Upanis.adic material
282
practices to obtain health and long life 299
seminal fluid, importance of 277
substances, identified with bodhicitta 350
yoga see meditation, sexual practices; Tantra,
sexual practices
sexuality
and sam
. sāra 275
celebration of 161
normal 117
transgressive 117, 275
Shaffer, J. G. 43
shaman
possible derivation of term from śraman.a
239–40
shamanic 113, 115, 139, 154–5, 157, 158,
340
shamanic-type practitioners
distrust of 233–4
incorporation of animal spirits into body for
healing 238
internal and external perspectives on see
perspectives, internal and external
Sharma, A. 27–8
Sharma, J. 58, 140, 268
Sharrock, P. 304
Shaw, J. 193–4, 213
Shaw, M. 302–3
Shen, W. 295
Sherburne, R. 329
Shingon 259, 313
Shrestha, B. 315
siddha 242, 264
Tamil see Cittar
Siddha medical tradition 282
Siddhi 304
Silburn, L. 255, 283, 325, 349
Silk, J. 212
Sindhu see Indus
Sinhalese 54
Sircar, D. C. 257
Śiśunāga 60
Sı̄tā 64, 69, 70, 74, 208
Śı̄talā 248, 315
418
Index
Śiva 3–4, 5, 6, 19, 114, 115, 116, 150, 151, 158, 171,
172, 195, 199, 201–2, 203–5, 206–7, 210,
238, 242, 243–5, 251, 321 see also Śaiva
identification with 238–9, 242–3, 245, 253, 270
king as earthly representative of 297
Skanda 6, 101, 114, 172, 202, 203–4, 206–7, 249,
251 see also Karttikeya
Skanda Purān.a 30, 91, 256
Skilling, P. 212
Skjærvø, P. O. 80
Skorupski, T. 329
Slusser, M. 315
smallpox, goddess of 248
Smith, B. 163
Smith, D. 277, 299
Snellgrove, D. 250, 274–5, 304, 329
Soma (deity) 150
soma (ritual substance; also name for Moon) 33,
97, 98, 113, 277, 280, 340
as entheogen 340
Somavam
. śı̄ 299–300
Sopa, Geshe L. 330
Sopa, Geshe L., Jackson, R. and Newman, J. 330
sound, theology of 254
Southwold, M. 17
spirit cults, in Southeast Asia 194
spirit-deity cults see deities, laukika; deities, local
Spiro, M. 149, 194, 324, 337
śrama 155, 156–7, 161
śraman.a 8–9, 15, 25, 60, 70–1, 73, 79, 117–18,
120–31, 153, 157, 160, 164, 165, 174, 197
see also Ājivikas; ascetic, movements;
Buddhism; Jainas; proto-śraman.a
movements
and Brahmanical deities 149
and local deities 140
and the dead 128–9, 141
as intermediaries with ‘tribal’ populations
180, 239–40
attraction to lay followers 179–80
attraction to rulers 180
early 339 see also proto-śraman.a movements
ethical component 176
origin of śraman.a orders 175–80
patrons and supporters, early 176
Sraoša 206
śrauta ritual see ritual, śrauta
Śrauta Sūtras 15, 25, 156, 158, 160
Śrāvastı̄ 50, 58, 71, 126, 310, 311
Śrı̄ (Śrı̄-devı̄) 102, 111, 144
Sri Lanka 47, 54, 71, 84, 149, 193, 194, 212, 213,
299, 319, 344–5
Śrı̄ Siṅgha 281
Srinivas, M. 53, 78
Srinivasan, D. 24, 61, 204, 226, 247
Śrı̄vidyā 322, 326, 335
Śrı̄vijaya 295, 304
Stablein, W. 232, 337
Stargardt, J. 201
state
models of 79, 80, 82, 83, 228
nature of 67 see also kingship; centralisation,
political
rituals of state 270
State Protecting Sūtras 309, 312
Stein, B. 171
Stein, M. A. 328
Stephen, M. 321
strı̄-ācār 10, 167
Studholme, A. 212
stūpas, Buddhist 107, 141, 142, 184, 213, 329
śubha 168
substrate languages 53, 54
subtle body 271, 281, 289, 291
and sexual practices 271
in Brahmanical sources 284–5
internal practices 289, 291, 327, 341
internal structure 281 see also cakra; nād.ı̄
knots 283
practices 264, 271, 278–9, 281
śūdra 60, 83, 86–7, 166, 203
Sudyumna 68
suffering see dukkha
Sufi 180, 258, 265, 335, 342
Suhr.llekha 211
Sukhasiddhi 303
Sukhāvatı̄vyūha Sūtra 219
śukladhyāna 224, 333 see also meditation, dhyāna
states
sūks.ma-śarı̄ra 271, 289
Sukthankar, V. 67, 72
Sullivan, Herbert 3–4
sulphur 278
Śuṅgas 67, 195
śūnyatā 264, 334
Śūrasena 58, 61
Sūrya 102, 114
Sūryavam
. śa see dynasties, Lunar and Solar
sus.umn.ā 221 see also body, channels within
Sutherland, G. 74, 248
Sūtra Pit.aka 31
sūtras
Brahmanical see Dharma Sūtras, Gr.hya
Sūtras, Śrauta Sūtras, etc.
Buddhist see Buddhist, sūtras
Suvarn.abhāsottama Sūtra see Suvarn.aprabhāsa
Sūtra
Suvarn.aprabhāsa Sūtra 211, 225–6, 259, 309, 324,
340
Svacchandabhairava 254
Index
Śvetāmbara 32, 332–3
Śvetāśvara Upanis.ad 202
Swat Valley 295
Taittirı̄ya Upanis.ad 114, 284
Taittirı̄yasam
. hitā 91, 114
Taks.ası̄lā 60, 79, 199, 213
Taleju 315
Tambiah, S. 88, 89, 129, 131, 138, 167, 194, 213,
224, 227, 323, 325, 337
Tamil
goddesses 248
kingship 84, 86
literature, early 84
people 127
Tamilnadu 83, 199, 293
Tamluk 81 see also Tamralipti
Tamralipti 81, 110
Tangut 295
Tannenbaum, N. 194
Tantra 1, 2, 9, 15, 22, 88, 89, 114, 130, 151, 152,
161–2, 204, 205, 220
aggressive and violent uses of Tantric ritual
251–2
and contemporary society 231, 350–3
and human psychology 351–2
and self-realisation 325
anuttarayoga, correct Sanskrit form 259, 329
anuttarayoga class of Buddhist Tantras 224,
257, 286–7, 304, 308, 312, 329, 330, 342
attitudes of early Western scholars 230
Buddhist 161, 250–1, 254 see also Vajrayāna
Buddhist appropriation of Śaiva Tantra 264–5
caryā class of Buddhist Tantras 220, 227, 258,
259, 286, 304, 340
classificatory schemes of Buddhist Tantra 258
components of Śaiva and Buddhist Tantra
291–2
definition of 229, 232, 342 see also Tantra,
disaggregating below
disaggregating 257, 291–2
distinction between sūtra and tantra 227
dual support of Buddhist and Śaiva Tantra
313, 324
female deities in Tantra 249, 255
fierce goddess practices, origin of 255–7, 297
health, long life and immortality, practices to
obtain 271, 276–7, 278, 301, 306, 331, 349
historiography 229
impersonation of Tantric deities 302 see also
ritual, masked ritual dances
identification with deity in Buddhist Tantra
259, 286
identification with deity in Śaiva Tantra 270
see also Śiva
419
in East Asia 310
in frontier areas 307
interiorisation of ritual 286–7, 327–8
internal and external perspectives 234–5
internal yogic practices see subtle body
issue of legitimacy 322–3
Jaina Tantra 232, 267–8, 332–3, 342
problematic nature of term 229
kriyā class of Buddhist Tantras 220, 227, 258,
286, 304, 340
kriyā practices in Śaiva tradition 279
mahāyoga class of Buddhist Tantras 258, 259,
276, 287–9, 329, 341, 348
modern Western 231
mutual borrowings between Śaivas and
Buddhists 232
negative stereotyping, in India 233
passion as liberating from passion
possible borrowings from China 281, 327
priests, in Kathmandu Valley 314, 315
purposes of Tantric practice 306
rDzogs-chen and Tantra 264
removal of transgressive elements 325–6, 349
respectable and transgressive versions 227, 228
right-hand and left-hand styles 233
ritual dance 250, 267
Śaiva 30, 161, 224, 238, 250, 252–8 see also
kaula
Śaiva Tantra, problematic nature of term 229
Śaiva Tantra and the yoginı̄ cults 252–8
sexual and transgressive elements, meaning of
348–50
sexual practices 230, 231, 262–4, 291–2, 327–8,
341 see also meditation, sexual practices
sexual practices, reduction of importance in
Tibet 330
sexual practices in Kālacakra Tantra 331–2
sexuality and gender relations 348–50
spiritualised understandings of 349–50
state patronage 296–302, 304–23, 325
Tantric studies and Western fantasy 229
‘Tantric’ style practice in Burma and Thailand
transgressive styles of 232–3, 341
Vais.n.ava 232 see also Pāñcarātra
Vais.n.ava Tantra, problematic nature of term
229
yoga class of Buddhist Tantras 258, 286–7, 307
yoginı̄ class of Buddhist Tantras 258, 260, 264,
289, 329, 341
tāntrika 322, 342 see also Tantra
tapas 154, 155–6, 157, 158–9, 161
Taxila see Taks.aśı̄lā
Taylor, J. 213
technologies of the self 273, 339–42, 345–8
Chinese 278–9
420
temples, Buddhist 107
Terwiel, B. 325
texts 24–32, 42
privileging of 18–22, 37
teyyam 319
Thailand 146–9, 170, 324–5, 337
Thapan, A. 205
Thapar, R. 47, 58, 77
Theism, early 201–7, 210
Theravāda see Buddhism, Theravāda
theriomorphic deities, Indus Valley 95
thig-le 284
Thurman, R. 217, 274–5, 289
Tibet 10, 130, 152, 182, 194, 221, 227, 233,
237, 246, 259, 307, 323, 329, 334,
337, 345
medicine 282
ritual dance
Tibetan lamas in contemporary world 338
Tinti, P. 309, 344–5
tı̄rthaṅkara 63, 68, 70, 146, 185, 267–8, 332
Tirumantiram 285
Tirumūlar 276, 285
Tiyavanich, K. 213, 337
Toffin, G. 314–17
trade 47, 61, 82, 185, 295, 308
and origins of śraman.a orders 126, 175,
179–80, 346
transfer of consciousness into another body
see body, transfer of consciousness into
another
transgression
Guhyasamāja Tantra 260
immorality 231
impurity 231
pollution 231, 291
religious meaning of 114–15, 117, 130 see also
disorder; misfortune; sexuality,
transgressive
state uses of transgressive power 235–7,
324
Tantric schools avoiding transgressive
practices 291
transgressive practices 253, 291, 327, 341
uses of 220
Trautmann, T. 90
trees, and local deities 141
tribal 78, 91–2
Trika 253, 254, 283, 325, 326
Truths, Four Noble 136–7
Tsong-kha-pa 264
Tsuchida, R. 122, 123
Tuladhar-Douglas, W. 102, 314
Tumburu 255, 304–6
Turner, V. 98
Index
Uchiyamada, Y. 322
Udayagiri 251, 297, 327, 334
Ujjayinı̄ 43, 49, 60, 126
Umā 204
untouchables see dalit; low-caste groups
Upagupta 147
Upanis.ads 15, 17, 25, 33, 56, 78, 100–1, 114, 117,
123, 164, 177, 202, 217, 221, 232, 281, 340
see also names of individual Upanis.ads
Urban, H. 229
urban life, problems of 197
urbanisation
First, of South Asia 43 see also Indus Valley
Cultural Tradition
Second, of south Asia 41–51, 98, 175–80
Urvaśı̄ 63, 74, 89–90
Uttarādhyāyana Sūtra 70, 125, 140, 216
Uttarapatha 60
Vāgı̄śvarı̄ 268
Vaidika 322
vaikhānasa 122
Vairocana 216, 227, 259
Vaiśālı̄ 58
Vaiśes.ika 216
Vais.n.ava 36, 101, 202, 210, 226, 256, 291, 322,
336
ascetics 336
Sahajiyā 232
Vais.n.avism 207
Vaiśravan.a 104–5 see also Kubera
vaiśya 83, 86–7, 99, 164, 166
vājikaran.a 275
Vajjian confederacy 58–60
vajra 104
Vajrabhairava 266
Vajrakı̄laya 266
Vajrapān.i 104, 146, 205–6, 241, 250, 262
Vajrasattva 220, 262, 277
Vajraśekhara 259
Vajrayāna see Buddhism, Vajrayāna; Tantra,
Buddhist
Vajrayoginı̄ 315
Vākāt.aka 195, 197, 292, 347
Vālmı̄ki 69
value systems in Indic societies 343–5
Vam
. sa see Vatsa
van Skyhawk, H. 335
vānaprastha 122–3
Vaṅga 60, 111
Vārān.ası̄ 43, 126, 151 see also Kāśı̄
varn.a 60, 83, 85, 86–8, 166, 187
Varun.a 74, 75, 96–7, 150
vāsanā 137
Vasis.t.ha 71, 123, 155, 208, 209, 281
Index
Vasubandhu 199, 217
Vāsudeva 101, 114, 202, 203, 205, 226
Vatsa 58
Vātsyāyana 272–3
Vedānta 15, 217, 322, 336
Vedic
civilisation 19
deities see deities, Vedic
non-Vedic components of Indian culture 20,
49–51, 61, 175
philosophy 162–4
religion 97–8, 119
sages 56 see also .r.si
texts 17, 19, 20, 21–2, 24–9, 33, 42, 48, 51–3,
255–6 see also Atharvaveda; R.gveda
periodisation of 51, 52–3
Vedic-Brahmanical
culture 45, 49, 50–1, 73, 75–6, 96, 99 see also
Brahmanical religion, and culture;
Brahmanical, tradition
religion, early 99
ritual complex, early 99
Vergati-Stahl, A. 315
Victoria, B. 343
Videha 58, 61–3, 69–71, 73, 76 see also Mithilā
Vidiśā 49–50, 193–4, 213 see also Besnagar
vidyādevı̄s 268, 333, 334–5
Vidyottama Mahātantra 250
Vijñānavāda see Yogācāra
Vikramorvaśı̄ 74
Vimalaprabhā 330–1
Vı̄n.āśikha Tantra 304–6, 325
Vinaya 31, 154, 174, 212, 273, 347
Vindhyavāsinı̄ 256
vipassanā 338 see also vipaśyanā
vipaśyanā 137, 174
Virūpa 330
vision quests 119
visionary techniques 213
and Mahāyāna Sūtras 220, 340
Vis.n.u 114, 150, 151, 171–2, 195, 201–2, 203,
204–5, 209
avatars of 203
king as earthly representative of 297
Vis.n.udharmottara Purān.a 305
Visuddhimagga 218–19
Viśvāmitra 63, 123, 155, 208–9
Vivekacūd.āman.i 284
Vivekananda 221, 336
Vogt, B. 319
Volkmann, R. 257
vow see vrata
vrat 167
vrata 131–3, 161–2, 164
imitating animals 162, 242
421
vrātya 113, 115, 117, 127–8, 132, 142, 158, 183–5,
186, 237–9, 242, 340, 346
rituals of 115–16 see also mahāvrata
self-identification as deity in ritual 238–9
Vrātyakān.d.a see Atharvaveda
Vr.s.n.i
deities 114, 202
people 202, 226
vyūha 203, 226
wall-painting, ritual 96
Wallace, V. 330–2
Walter, M. 101, 215, 232
Ward, W. 230
Warder, A. 217
warrior
brotherhoods 183
imagery, as reworked in Brahmanical tradition
343
imagery, in Buddhism and Jainism 187, 189,
190
Wayman, A. 235, 260, 329
Weber, M., and elective affinities 170, 176, 179,
228
weddings, in Buddhist societies 171
Wedemeyer, C. 229, 260–2, 264, 287
Weinstein, S. 310–11
Werner, K. 157
Western followers of yoga and Tantra 230, 231
Whicher, I. 222
White, D. 115–16, 129, 209, 221, 232, 235, 241,
243, 252, 269, 275, 276–9, 280–1, 299,
301, 325, 332, 335–6, 340, 341, 349
periodisation of Śaiva Tantra 326–7, 329
Wiener, M. 322
Wilke, A. 322
Williams, P. 199, 220
Williams, R. 125, 185, 346
Willis, M. 29, 74, 237, 251, 297
Wilson, H. 230–1
Wilson, I. 320
Wiltshire, M. 69–70
wisdom literature, shared among early ascetics
124
Witzel, M. 25, 27, 28, 29, 42, 48, 51–3, 58, 63–4,
68, 71, 76, 78, 87, 98–9, 113–17, 127, 128,
155, 163, 237
Wojtilla, G. 306, 328
women
as Buddhist practitioners 170, 186, 302–3
as kāpālika-style practitioners 242
as Pāśupata practitioners 242
as Tantric practitioners 302–3
in śraman.a movements 125, 126
status of 88, 89–93, 189, 302–3
422
women’s ritual see ritual, women’s
world-renouncer see renunciate traditions
Wright, J. C. 29, 74, 80
Wu, Empress 215
Xixia 295
Xuanzang 293, 310–11
Yājñavalkya 124, 225–6
Yajurveda 24–5, 113
yak tovil 319
yaks.a 74, 101–12, 140, 141, 143–5, 146, 147, 151,
161, 169, 198, 201, 202, 205, 207, 247,
250, 267–8, 309–10 see also yaks.ı̄
generals 104–5, 307, 309
yaks.ı̄ 6, 102, 146, 151, 267–8 see also yaks.in.ı̄
yaks.in.ı̄ 74, 248 see also yaks.ı̄
Yamāntaka 262
Yamasaki, T. 337
yantra 224
yavana 35–6 see also Indo-Greek
Yayāti (king in Mahābharata) 209
Yayāti II (king of Orissa) 300, 302
Yijing 109, 212
yoga 1, 2, 114 see also meditation
and contemporary society 350–3
and Sām
. khya 216
as technology of the self 347
definition of 342
‘internal’ yogic practices 255, 291,
341
isolation from world 221
modernist reconstruction of 336
origins 8–9
physical aspects 2
Index
postures 4
pragmatic uses of 223
reality of effects 351
sexual, in Buddhist Tantra 261
sexual, reconceptualisation of 255
shamanic origin 340
transfer of consciousness into another body
see body, transfer of consciousness into
another
warrior’s transport to heaven 128, 221
yoga, Tantra and meditation, growth of global
interest in 1–2
Yogācāra 216, 217, 220
Yogaratnamālā 286
Yogaśāstra 221, 224, 332–3
Yogasūtra 35, 132, 137, 221–3, 336, 340
dating of 222
engagement with world 222–3
yogatantra see Tantra
Yogavāsis..tha 305
yoginı̄ 246, 254
temples 254–5, 299, 301
yoginı̄tantra see Tantra
Yokochi, Y. 30, 243, 251, 256
Yon.i Tantra 332
Yudhis.t.hira 72, 209
Yungang 215
Yuthok, C. 328
Zanen, S. 315
Zarathustra 26–7, 74, 97
Zigmund-Cerbu, A. 222
Zoroastrian 26, 80, 166 see also Zarathustra
Zvelebil, K. 84, 285
Zydenbos, R. 140