Articles by alphabetic order
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
 Ā Ī Ñ Ś Ū Ö Ō
1 2 3 4 5 6 7 8 9 0


The Cakrasamvara Tantra (The Discourse of Śrī Heruka): A Study and Annotated Translation.

From Tibetan Buddhist Encyclopedia
Jump to navigation Jump to search
Paramasukha Cakrasamvara Yab-Yum Luipa Mandala.jpg





David B. Gray, The Cakrasamvara Tantra (The Discourse of Śrī Heruka): A Study and Annotated Translation. The American Institute of Buddhist Studies at Columbia University in New York, co-published with Columbia University’s Center for Buddhist Studies and Tibet House US, New York 2007. ISBN 0975373463

Introduction

The Cakrasaṃvaratantra (CS) is a principal tantra of the Cakra- saṃvara scriptural cycle, which is one of the largest collections of Buddhist Yoginītantra literature from the early medieval South Asian world. The Cakrasaṃvara tradition was imported into neighboring areas such as Nepal, Tibet, and Bhutan, and has func- tioned as one of the most important sources in the formation of religio-cultural systems in these areas. Its thought and practice are also maintained “in other regions influenced by Tibetan Buddhism, including Mongolia, Russia, China, and elsewhere,” notes David B. Gray, the author of the book under review, “as Tibetan lamas have been living and teaching in diaspora.” (p. xv.) Gray’s study aims at providing the first critical translation of the CS, richly annotated and accompanied by analyses of its contents and contexts. A critical edition of the CS is not included, which some may regard as a shortcoming, but Gray’s critical edition of the CS is forthcoming as a companion volume to his study.

1 Outline

Gray’s study consists of six parts: (1) an introduction into the study of the CS (pp. 1–152); (2) an annotated critical translation of the CS (pp. 153–384); (3) Sanskrit-Tibetan-English and Tibetan-Sanskrit- English glossaries (pp. 385–392); (4) a Conspectus Siglorum listing Journal of the International Association of Buddhist Studies

the Sanskrit manuscripts (Skt mss) that Gray used for his transla- tion (pp. 405–408); (5) a bibliography (pp. 409–436); and (6) an index of Sanskrit, Tibetan, and English terms (pp. 437–447). Part one provides a general introduction into the study of the CS. Gray presents the textual materials which he used for his translation and explains his translation methodology. This part also includes analyses of several important aspects of the CS and its background, such as the classification of tantra literature, dates when the CS and some other tantras belonging to the Cakrasaṃvara scriptural cycle were compiled, the contents of the CS, and the scriptural and ideological contexts within which the CS was compiled and used. It is regrettable that Gray does not provide a full textual and contex- tual study of Śaiva-Buddhist interrelations, especially since he up- holds the idea that “the Buddhist Yoginītantras were significantly influenced by Śaiva Kāpālika practices” (p. 8, n. 19). But this omis- sion may also reflect the stance that we should avoid drawing hasty conclusions on this complex issue, for he says that “the undoubt- edly complex relationships that exist between Śaiva and Buddhist tantric textual traditions will only be determined conclusively once all of the surviving texts have been critically edited and published” (p. 9, n. 19). This position may be controversial because efforts to create critical editions and efforts to determine textual relation- ships are not separate from each other – they are to be concurrently made and reciprocally associated. But we should not ignore that this first part of Gray’s study provides much information on the contents and contexts of the CS and the Cakrasaṃvara scriptural tradition, which is no doubt useful for anyone interested in Tantric Buddhism. Part two, the annotated critical translation, is the main part of the study. Gray’s translation is based on his unpublished critical edition of the CS. This edition chiefly relies on three Skt mss of the CS, on Tibetan translations of the CS, and on eleven Indian com- mentaries along with some Tibetan commentaries. It also makes use of Kalff’s edition of selected chapters of the Abhidhānottaratantra (Kalff 1979), and of Skt mss of this work which contains many parallel passages and is therefore quite useful for recovering the text of folia that are missing from the extant Skt mss of the CS. The

text of the CS as edited by Gray is partially recorded in the annota- tion. In his footnotes, he also adduces various interpretations from the commentaries, which makes this book not only the first critical translation of the CS but also a useful guidebook for comparative studies of its commentarial literature. In short, Gray’s study is the first full translation of the CS, it serves as a guide to its commentarial literature and provides much textual and contextual information on the Indian Cakrasaṃvara tradition. This makes it a ‘must-read’ for students and scholars who research the Indian Cakrasaṃvara tradition in particular and Indian and Tibetan Tantric Buddhism in general. However, it is also beset by problems. In the following I would like to focus on problems in dating the tantras of the Cakrasaṃvara scriptural cy- cle and problems of Gray’s translation of the CS and the materials used for it. I shall then turn to the analyses of the origin myths of Heruka and his maṇḍala, and of the structure and functions of the Triple Wheel maṇḍala.

2 Dating the tantras of the Cakrasaṃvara scriptural cycle

Gray notes that a precise dating of the CS and other tantras of the Cakrasaṃvara tradition is currently a difficult task (p. 11 and 20), but nevertheless offers a hypothesis of his own.

2.1 Date of the CS

Gray argues that it is likely that the CS was compiled in the eighth century, for the following reasons: (a) The CS mentions the names of Buddhist scriptures that can be dated in the late seventh century or the first half of the eighth century, such as the Sarvatathāgatatattva- saṃgrahasūtra, Guhyasamājatantra, Vajrabhairavatantra, Śrī- paramādyatantra, and the Sarvabuddhasamāyogatantra.1 (b) It is known from Tāranātha’s History of Buddhism that Jayabhadra com- posed his Cakrasaṃvarapañjikā, the oldest commentary of the CS,

1 Throughout his study, Gray spells the title of this tantra with ‘samayo- ga’. However, ‘samāyoga’ (full title: Sarvabuddhasamāyogaḍākinījālasa ṃvaratantra) is more commonly used, and very likely also correct.

during the early- to mid-ninth century. (c) In his Nāmasaṃgītiṭīkā, Vilāsavajra (mid- to late-eighth century), quotes one pāda and one word from the CS: glang chen ko rlon gos su gyon (/ zhes pa ni dpal ’khor lo bde mchog gi rgyud las te), and keng rus (ni dpal bde mchog ’khor lo’i rgyud las so). (pp. 11–14.) Given that Tāranātha’s account is ambiguous and may be un- reliable, Vilāsavajra’s very short quotations appear to be the only evidence to support Gray’s hypothesis that the CS was already ac- tive in the eighth century. Gray correlates Vilāsavajra’s glang chen ko rlon gos su gyon and keng rus respectively with hasticarma- viruddhaṃ ca in chapter 2 of the CS (where the actual reading is hasticarmāvaruddhaṃ ca) and kaṅkāla in chapter 48. However, the former correlation is problematic. Vilāsavajra’s work is a com- mentary on the Nāmasaṃgīti, and the Sankrit of the pāda in ques- tion in the Nāmasaṃgīti is gajacarmapaṭārdradhṛk,2 which cannot be found in the CS. Vilāsavajra might have read the pāda in the CS freely and related it to the pāda in the Nāmasaṃgīti freely on this basis, but this is certainly not conclusive evidence. The Sanskrit source of keng rus in the Nāmasaṃgīti is, indeed, kaṅkāla.3 But this, too, is insufficient evidence. Clear and extensive parallel pas- sages along with a reference to the name of its source text would certainly be more decisive. There is also a problem regarding the name of the tantra to which Vilāsavajra refers. Although Vilāsavajra calls the scriptural source of the pāda and word in question dPal ’khor lo bde mchog gi rgyud or dPal bde mchog ’khor lo’i rgyud, which in Sankrit is Śrīcakrasaṃvaratantra, we should bear in mind the possibility that this tantra may have previously been named Herukābhidhāna rather than Cakrasaṃvara. This is suggested by the change of its name in its chapter 51. In chapters 1 to 50 it calls itself Herukā- bhidhāna (iti śrīherukābhidhāne ….). However, in chapter 51, that name is said to refer to the large scripture of one hundred thousand verses from which this tantra was selected, and the name of this tantra is given as Cakrasaṃvara (śrīcakrasaṃvaraṃ nāma mahā-

2 Nāmasaṃgīti: Ādarśajñānam, 3d. 3 Nāmasaṃgīti: Ādarśajñānam, 1c.

yoginītantrarāja). As I shall argue below, it is very likely that chap- ter 51, together with several verses of chapter 50, was not included in the oldest version of this tantra: it was added to the oldest version after Jayabhadra, who was active after Vilāsavajra.4 Gray’s analysis of Vilāsavajra’s references is thus somewhat prob- lematic. However, we cannot deny the possibility that some form of the CS existed in the age of Vilāsavajra because, as Gray stresses, Vilāsavajra surely mentions a tantra named ‘Cakrasaṃvaratantra’ and distinguishes it from the Saṃvaratantra, by which he refers to the Sarvabuddhasamāyogatantra. It is also possible that the Ca- krasaṃvaratantra mentioned by Vilāsavajra is a work different from the CS – a work that did not survive. One should also bear in mind the possibility that the date of compilation of the CS is not so long before (or might even be very close to) the date when Jayabhadra was active. I would here like to call attention to four remarks about the historical stages of the compilation of the CS and of Jayabhadra’s commentary that I made in 20015 and that are not sufficiently taken into consideration in Gray’s study. They may be of some help for future studies about the date of compilation of the CS. [1] There were several different versions of the CS, some of which Gray mentions. Gray overlooks, however, that these can be roughly divided into two: (a) a shorter version that contains chap- ters 1 to 49 and the first half of chapter 50 (= 37a3 of the Vadodara ms)6 of the extant CS and (b) a longer version that contains all chap- ters, from 1 to 51. [2] The shorter version is very likely to be older than the longer, and Jayabhadra is very likely to have used one of the oldest texts that belong to it. The text that Jayabhadra used does not know the

4 However, there is also the possibility that this tantra was called Cakrasaṃvara from the outset because Jayabhadra, the oldest commentator on this tantra (or at least the commentator who used the oldest version of this tantra), refers to it under that name. See also Sugiki 2001 for further discus- sion of the titles Herukābhidhāna and Cakrasaṃvara. 5 See Sugiki 2001. 6 See below p. 513 for the Sanskrit mss of the CS.

chapter divisions given in the extant CS; in fact, it gives no chapter divisions at all. Furthermore, Jayabhadra’s commentary does not mention parts that are not contained in the shorter version. It is quite unlikely that Jayabhadra intentionally skipped commenting on these parts because they provide instructions into the system of the internal Heruka maṇḍala, which is a main doctrine of the Cakrasaṃvara tradition. They also introduce other systems that are highly Buddhist Mahāyānic and make the CS more Mahāyānic.7 [3] The shorter version (and also Jayabhadra’s commentary) is devoid of a clear idea of internal Cakrasaṃvara holy sites corre- sponding to their external forms. This idea first appears in the last half of chapter 50 of the extant CS and is in the Cakrasaṃvara tradition generally accompanied by such terms as bāhyādhyātma-, sabāhyādhyātma-, or the like. It became one of the principal ele- ments in the practice of ‘the creation stage’ (utpattikrama), i.e., the visualization of the Heruka maṇḍala in the Cakrasaṃvara tradi- tion after the CS. After the addition of the last half of chapter 50 and of chapter 51 to the shorter version (i.e., after the compilation of a text that belongs to the longer version), commentators of the CS began to freely read this idea into some passages in chapters that had already been present in the shorter version, and terms like bāhyādhyātma-, sabāhyādhyātma-, or the like then came to be in- serted into the shorter version, too. [4] The addition of the last half of chapter 50 and of chapter 51 to the shorter version can be dated between Jayabhadra and Kambala because Kambala, unlike Jayabhadra, comments on the last half of chapter 50 of the CS, although very briefly. Both Jaya- bhadra and Kambala very likely lived before the compilation of the Vajraḍākatantra, which was likely composed around or after the late ninth century.


7 One could object that the commentary on the CS by Bhavyakīrti, who is clearly one of the later commentators, also does not mention parts that are not contained in the shorter version. However, this does not invalidate my hypothesis because Bhavyakīrti’s commentary very closely follows Jayabhadra’s, as Gray also points out (p. 22).

2.2 Dates of the tantras of the Cakrasaṃvara tradition that were compiled after the CS

Gray’s approach to datingthetantras belonging to the Cakrasaṃvara tradition that were compiled after the CS can be summarized as follows: The Abhidhānottaratantra preserves Śaiva readings dat- ing to the ninth century, some of which are older than readings found in the CS. However, the compilation of the final form of the Abhidhānottaratantra cannot predate the CS because the CS mentions the Abhidhānottaratantra under the title cakrasaṃvara. Furthermore, neither the Abhidhānottaratantra nor the CS contain technical Buddhist terminology relating to the perfection stage (niṣpannakrama), which became popular in and after the ninth century.8 By contrast, other tantras of the Cakrasaṃvara tradition such as the Saṃvarodayatantra, Vajraḍākatantra, and Ḍākārṇa- vatantra contain Buddhist terminology relating to the perfection stage. For these reasons, the Abhidhānottaratantra may be dated to the eighth century, but definitely not to before the CS. Many of the tantras of the Cakrasaṃvara tradition appear to have been composed after the Abhidhānottaratantra (p. 20.) This analysis of the Abhidhānottaratantra is, however, highly problematic. It is indeed true that the Abhidhānottaratantra often preserves Śaiva readings of the early medieval age, but the same can also be said of other tantras of the Cakrasaṃvara tradition. Moreover, Gray’s idea that the Abhidhānottaratantra does not con- tain technical terminology relating to the perfection stage is hard to accept. As I argued in 1999, the Abhidhānottaratantra (like the Vajraḍākatantra, the Saṃpuṭodbhavatantra, the Ḍākārṇavatantra, and other texts of the Cakrasaṃvara tradition that can be dated around or after the late ninth century) clearly introduces the sub- tle-body system centered on the inner channels and inner circles connected with the doctrine of Four Blisses (caturānanda), evi- dently under the influence of the Hevajra subtle-body system (from around the ninth century). The Abhidhānottaratantra also gives

8 By “technical Buddhist terminology relating to the perfection stage,” Gray appears to refer to the psychosomatic subtle-body system centered on the inner channels (nāḍī), inner circles (cakra), and the like.

instructions on psychosomatic meditation based on the subtle-body system, in which the inner fire or light and the inner ambrosia are visualized to move inside and outside of the practitioner’s body. Furthermore, the Abhidhānottaratantra contains instructions of psychosomatic meditation that are closely related to Lūyīpāda’s Mahāyoga system and Kṛṣṇācārya’s Olicatuṣṭaya or Ālicatuṣṭaya system, which were regarded as instructions to the perfection stage in Lūyīpāda’s and Kṛṣṇācārya’s schools, respectively.9 It should also be noted that the Abhidhānottaratantra gives instructions on the internal Heruka maṇḍala. One of these can be regarded as a de- veloped version of instructions given in Lūyīpāda’s Cakrasaṃvarā- bhisamaya (presumably last half of the ninth century), the Saṃ- puṭodbhavatantra, the Vajraḍākatantra (both probably late ninth to tenth century), and several other ritual or meditational texts (i.e., vidhi or sādhana texts) belonging to the Cakrasaṃvara tradition. (The final section of this paper contains a discussion on the internal Heruka maṇḍala taught in these scriptures.)10 It is therefore unlikely that the date of the Abhidhānottaratantra is as early as Gray believes, and it is impossible to maintain his clear dividing line between the date and contents of the Abhidhā- nottaratantra and those of the Saṃpuṭodbhavatantra, the Vajraḍā- katantra, and other tantras of the Cakrasaṃvara tradition. Gray’s intention, although not clearly stated, may be to say that the old- est parts of the Abhidhānottaratantra were compiled in the eighth century, and that these refer to the passages that have parallels in the CS, which does not contain technical Buddhist terminol-

9 For details on the teachings of the Abhidhānottaratantra in relation to Kṛṣṇācārya’s and Lūyīpāda’s systems, see Sugiki 1999 and 2007. The con- tents of Lūyīpāda’s Mahāyoga system were already analyzed by Munenobu Sakurai in an earlier paper, although he did not mention the textual rela- tionship between Lūyīpāda’s works that teach the Mahāyoga system and the Abhidhānottaratantra (Sakurai 1997). Draft editions of two passages that explain the subtle body system and psychosomatic meditation based on it from the Abhidhānottaratantra are provided in Sugiki 2007. Since this book may be difficult to access from outside Japan, these passages are pre- sented in an appendix to the present paper. 10 For details on the historical development of the internal Heruka maṇḍala, see Sugiki 2003b and 2007.

ogy concerning the perfection stage. Even if that is so, a similar problem occurs as in his suggestion to date the CS in the eighth century. There also remains the question whether all parallel passages of the CS found in the other tantras of the same tradi- tion (such as the Saṃpuṭodbhavatantra, the Vajraḍākatantra and others) can be determined as being later than those found in the Abhidhānottaratantra. (As I mentioned above, readings that can be considered to be old and early-medieval Śaivic are also found in those tantras.) There is currently simply no conclusive evidence that proves the Abhidhānottaratantra existed in the eighth century. As in the case of the CS, there appear to have been several stag- es in the compilation of the other tantras of the Cakrasaṃvara tra- dition. There also appear to have been complex mutual references to texts between the compilers of those tantras. Finally, in order to carefully develop a plausible hypothesis on the dates of the tantras belonging to the Cakrasaṃvara traditions, we must also take the relationship between Śaiva and Buddhist tantras into consideration – and this, as mentioned above, is a point that Gray unfortunately neglects.

3 Gray’s translation of the CS and the materials used for it

3.1 Problems in the selection of materials

A Sanskrit edition of the CS, together with Bhavabhaṭṭa’s commen- tary, was published by Pandey in 2002 (henceforth CS-P). Although this edition should be respected as a pioneering achievement, it is very problematic, as many scholars in this field have pointed out; Gray also accurately shows problems in Pandey’s edition. For his own critical edition of the CS (that awaits publication), Gray used three Skt mss of the CS, which Pandey also used: an old palm- leaf ms owned by the Oriental Institute in Vadodara (accession no. 13290), and two recent copies of it. Correctly recognizing that the latter are copies of the palm-leaf ms., Gray uses the palm-leaf ms as the main basis for his edition and translation. In addition to these manuscripts, Gray also made use of other texts and supporting materials: Tibetan translations of the CS, Skt

mss and eds, as well as Tibetan translations of eleven Indian com- mentaries, some indigenous Tibetan commentaries, as well as Skt mss and the Skt ed of the Abhidhānottaratantra. Among the com- mentaries, he frequently favors Jayabhadra’s Cakrasaṃvarapañjikā and Bhavabhaṭṭa’s Cakrasaṃvaravivṛti, and also, in some pas- sages, Vajrapāṇi’s Laghutantraṭīkā. He also attaches much impor- tance to Kambala’s Herukasādhananidhi and Vīravajra’s Padār- thaprakāśikā. These commentaries are favored or considered to be important for the following reasons: (a) Skt mss or Skt eds of Jayabhadra, Bhavabhaṭṭa, and Vajrapāṇi’s commentaries are avail- able. (b) Jayabhadra’s commentary is the oldest among the surviv- ing commentaries of the CS. (c) Kambala’s commentary is also relatively early. (d) Many later commentators of the CS rely on Jayabhadra’s or Kambala’s commentaries. (e) Bhavabhaṭṭa’s com- mentary quotes many words and phrases of the CS, although he sometimes emends these in the act of quoting. (f) Kambala’s and Vīravajra’s commentaries give detailed explanations of rituals that are described only briefly in the CS. And, as I mentioned earlier, the reason for using the Abhidhānottaratantra is that it preserves old readings and contains many passages that have parallels in the CS and is hence quite useful for recovering material that is missing from the extant Skt mss of the CS. These materials, however, do not suffice for a fully critical edi- tion and translation of the CS. Most of the supporting materials are new paper mss or Tibetan translations. Gray did not use two older palm-leaf mss of Jayabhadra’s commentary, which preserve older and better readings than the new paper mss he used.11 Neither did he use a Skt ms of Kambala’s commentary, which is also an old palm-leaf ms,12 but only used a Tibetan translation of this commen- tary. Finally, Gray did not make effective use of Skt mss or Skt eds of Buddhist and Śaiva texts that have parallel or similar passages,

11 A draft-version of the Skt ed of whole text of Jayabhadra’s commentary based on these two palm-leaf mss has been published in Sugiki 2001. I plan to publish the finalized edition in the near future. 12 I have prepared an as yet unpublished Skt ed of whole text of Kambala’s commentary based on this palm-leaf ms.

such as the Vajraḍākatantra,13 the Saṃpuṭodbhavatantra, the Śaiva tantras which Alexis Sanderson mentions in his series of papers that analyze textual relationships between Śaiva Vidyāpīṭha tantras and Buddhist Yoginītantras (such as the Siddhayogeśvarīmata, the Jayadrathayāmala, the Brahmayāmala, and the Tantrasadbhāva),14 the Śaiva Vīṇāśikhatantra,15 and other related texts.16 These mate- rials are not only truly helpful in creating a critical edition and translation of the CS; they are actually indispensable for recovering missing passages of the CS that cannot be reconstructed, or only in an unsatisfactory manner, from the Abhidhānottaratantra and Jayabhadra’s, Bhavabhaṭṭa’s, and Vajrapāṇi’s commentaries. Let us look at some relevant cases.



13 Draft versions of the Skt eds of chapters 1, 7, 8, 14, 18, 22, 36, 38, 42, 44, 48 of the Sanskrit Vajraḍākatantra have been published in Sugiki 2002, Sugiki 2003a, and Sugiki 2008. I have also prepared as yet unpublished draft versions of other chapters of this tantra. 14 Sanderson 1995, Sanderson 2001. 15 The Vīṇāśikhatantra’s instruction on the vetālasādhana (Skt ed, 190cd–193) contains a passage that is very similar to or identical with those of the CS, the Vajraḍākatantra, and the Herukābhyudayatantra. See Sugiki 2008 for details and references. I express my heartfelt thanks to the reviewer of this article that was published in the journal Tantric Studies (The Center for Tantric Studies, University of Hamburg) who suggested that I check care- fully the Vīṇāśikhatantra’s passage in question before submitting the final version of paper. I would like to add here that the verse mahāśaṅkhamayaṃ kuryād athavā kacchapasya tu of the Vīṇāśikhatantra (Skt ed, 113cd) is also a parallel of the CS’s mahāśaṅkhamayaṃ kuryād abhedyaṃ kaccha- pasya tu (Skt ms, 25a3–a4). 16 For example, the Catuṣpīṭhatantra, the Saṃvarodayatantra, the Ḍākārṇavatantra, Lūyīpāda’s Cakrasaṃvarābhisamaya, and Kṛṣṇācārya’s Cakrasaṃvarasādhana. Lūyīpāda’s Cakrasaṃvarābhisamaya (= Bhaga- vadabhisamaya) have been published in Sakurai 1998. The Skt ed of the whole text of Kṛṣṇācārya’s Cakrasaṃvarasādhana has been published in Sugiki 2000, which also contains a list of parallel passages found in the CS and Kṛṣṇācārya’s Cakrasaṃvarasādhana.


3.2 Textual problems and problems of translation


Gray recovers pāda 3a of chapter 26 (whose folia are missing from the Skt mss of the CS) as ‘taṃ dūtīṃ sarvasiddhidaṃ’ from one of the two paper mss of Jayabhadra’s commentary (p. 265, note 4).17 He mentions that the text in question is improperly declined, as Bhavabhaṭṭa notes in his commentary (tam iti tāḥ, dūtī dūtayaḥ, CS-P: 483) (p. 265, note 4), and translates “These messengers bestow all powers” (p. 265). However, the text can be recovered from the two palm-leaf mss of Jayabhadra’s Cakrasaṃvarapañjikā and the two palm-leaf mss of the Vajraḍākatantra as ‘tāṃ dūtīṃ sarvasiddhidāṃ,’ ‘that female messenger bestowing all supernatu- ral effects [or accomplishments].’ Jayabhadra comments that ‘tāṃ dūtīṃ sarvasiddhidāṃ’ should be read as tā dūtyaḥ [[[sarvasiddhi]]- dāḥ, TS], which means ‘those female messengers bestow all super- natural effects [or accomplishments].’18 Though not optimal, this is at least clearer and more natural than Gray’s taṃ dūtīṃ sarva- siddhidaṃ, and is likely to be the older version because the sources are older than those used by Gray. The whole verse 3 may be re- covered from the Brahmayāmala, Jayabhadra’s commentary, Kam- bala’s commentary, the Vajraḍākatantra, and Bhavabhaṭṭa’s com- mentary as follows: tāṃ dūtīṃ sarvasiddhidāṃ darśanāt sparśanāt tathā / cumbanāvagūhanān nityaṃ (metrically bad) yogapīṭhe viśeṣataḥ //.19

17 Pandey’s reconstruction of the verse in question is ‘taṃ dūtī tu sattvārthasiddhidaṃ,’ which except for ‘tu’ is based on the text quoted in Bhavabhaṭṭa’s commentary. 18 Jayabhadra’s Cakrasaṃvarapañjikā, Skt ed, 26.1. The Vajraḍākatantra, my unpublished Skt ed, 34.4a.

19 tāṃ dūtīṃ sarvasiddhidāṃ ] J; taddravyaṃ sarvadā siddhaṃ – BY. tāṃ dūtīṃ sarvārthasiddhidāṃ – VḌT. taṃ dūtī sattvārthasiddhidam – Bh. taṃ dūtī tu sattvārthasiddhidaṃ – CS-P. : darśanāt sparśanāt tathā ] em.; darśanāt sparśabhakṣaṇāt – BY. darśanāt sparśanāt – J. darśanaṃ sparśanaṃ tathā – VḌT, Bh, CS-P. : cumbanāvagūhanān nityaṃ ] J, VḌT; cumbanād gūhanāc caiva – BY. cumbanāvagūhanām (ityādi) – K. cumbane(-tyādi) – Bh. cumbanaṃ gūhanaṃ nityaṃ – CS-P. : yogapīṭhe viśeṣataḥ ] J; śivapīṭhe viśeṣataḥ – BY. yogapīṭhaviśeṣataḥ – K and VḌT (very likely a corruption of yogapīṭhe viśeṣataḥ). yogapīṭham and viśeṣata

As for verse 2ab of chapter 27, whose folia are also lost, Gray fa- vors the reading given in one of the two paper mss of Jayabhadra’s commentary and reconstructs ‘grāme grāme vrajanti ca dūtayo (rūpalakṣaṇam, TS),’ translated as “The messengers travel from town to town. [As for their] physical characteristic[s], …” (p. 271, n. 4). Pandey, on the other hand, recovers this verse as ‘grāme grāme vrajan tasya dūtayo rūpalakṣaṇam.’ Indeed, one of the two palm- leaf mss of Jayabhadra’s Cakrasaṃvarapañjikā20 supports Gray’s reconstruction. However, Pandey’s reconstruction is better than Gray’s because it is confirmed by older sources. It is fully confirmed by the other palm-leaf ms of Jayabhadra’s Cakrasaṃvarapañjikā, the palm-leaf ms of the Abhidhānottaratantra, and Bhavabhaṭṭa’s Cakrasaṃvaravivṛti.21 The corresponding verse in the Skt ms of the Brahmayāmala reads grāme grāme vratan tasya devatārū- palakṣaṇaṃ,22 which is closer to Pandey’s reconstruction than to Gray’s grāme grāme vrajanti ca. The text Pandey reconstructed means: ‘He (= the practitioner) travels (vrajan tasya: vrajan is vrajaṃ) from village to village. [In these villages, the] female mes- sengers [show their] physical characteristic[s to him].’ In this con- text, the one who travels is not a messenger but a practitioner. The comments on this verse by Jayabhadra, Kambala, and Bhavabhaṭṭa also support this interpretation.

Gray translates pāda 6a of chapter 41, whose folia are also lost, as “[They are:] in Kulutā (better: Kulatā, TS)23 and [[[Maru]]], …’.24

(iti) – Bh. yogapīṭhaṃ viśeṣataḥ – CS-P.

20 Jayabhadra’s Cakrasaṃvarapañjikā, Skt ed 27.1 and the footnote there.

21 Jayabhadra’s Cakrasaṃvarapañjikā, Skt ed, 27.1. Bhavabhaṭṭa’s Ca- krasaṃvaravivṛti, Skt ed, 488.

22 The Brahmayāmala, Skt ms (NGMPP A42/6). 326b3. Skt eds, San- derson 2006: 22 (grāme grāme vrataṃ tasya devatārūpalakṣaṇam); Hatley 2007: 180 (grāme grāme vrataṃ tasya devatārūpalakṣaṇam).

23 ‘Kulatā’ is more common in Buddhist Cakrasaṃvara scriptures than Gray’s ‘Kulutā.’ See also Bhavabhaṭṭa’s comment on this pāda, kulatāyām ityādinā (CS-P, 547).

24 Pandey’s reconstruction of this text is as follows: kulatāyāṃ vivikte ca. However, vivikte is not attested in any surviving Sanskrit sources that are closely related to this pāda.

He put ‘Maru’ in brackets because “the Sanskrit text here is not preserved, and the Tibetan translations list the variant mgon pa (PM 239a, SL 130b), which is unattested elsewhere.” (p. 330, n. 10.)25 However, for the word in question, we find ‘araṇya’ in Jaya- bhadra’s commentary, which is very likely derived from the word ‘araṇyeśe’ that appears in a parallel passage found in the Śaiva Tantrasadbhāva.26 The passage that includes pāda 6a provides the archaic list of Cakrasaṃvara holy sites that is derived from a cor- responding Śaiva list, such as the one found in the Tantrasadbhāva. We may thus recover pāda 6a from the Tantrasadbhāva and from Jayabhadra’s commentary as ‘kulatāyām araṇye ca,’ and the trans- lation should be ‘[They are:] in Kulatā, Araṇya, …’

Gray translates verse 10ab of the same chapter as “The six yoginīs are in Kulutā (better: Kulatā, TS), and the six mothers are in the land of Maru” (p. 331–332). He appears to have followed Pandey’s reconstruction ‘ṣaḍ yoginyaḥ kulatāyāṃ marudeśe ṣaḍ mātarāḥ.’ However, the last pāda must be ‘marudeśe ca mātaraḥ,’ (not ṣaḍ but ca,) which can be recovered from the palm-leaf mss of the Tantrasadbhāva (araṇyeśe ca mātaraḥ), Jayabhadra’s com- mentary (mātaraḥ [: no ṣaḍ]), Kambala’s commentary (marudeśe ca mātaraḥ), the Vajraḍākatantra (marudeśe ca yā mātarāḥ), and Bhavabhaṭṭa’s commentary (mātarā iti [: no ṣaḍ]).27 The Tibetan translations of the CS (mya ngam yul na ma mo rnams) also sup- port this reconstruction, and no old sources support Pandey’s and Gray’s ‘marudeśe “ṣaḍ” mātarāḥ.’ The translation should there- fore be corrected to: ‘the mothers are in the land of Maru’. It is

25 Gray goes on to state that “several verses down, however, marudeśe is attested by Bhavabhaṭṭa (CS-P, 548).” However, this ‘marudeśe’ is a quota- tion not from the passage in question but from another passage of the CS (41.10b). 26 Jayabhadra’s Cakrasaṃvarapañjikā, Skt ed, 41.2 (araṇyaṃ marubhū- miḥ). The Tantrasadbhāva, Skt ed, Sanderson 1995: 100, n. 20 (kulūtāyām araṇyeṣe). 27 The Tantrasadbhāva, Sanderson 1995: 100, n. 20. Jayabhadra’s Cakra- saṃvarapañjikā, Skt ed, 41.3. Kambala’s Herukasādhananidhi, my unpub- lished Skt ed (: Skt ms, 70a4). The Vajraḍākatantra, Skt ed, 18.4. Bhava- bhaṭṭa’s Cakrasaṃvaravivṛti, Skt ms, 127a4 (: Skt ed, 488).

very likely that Gray, as well as Pandey, misread Bhavabhaṭṭa’s ‘ṣaḍ yoginya ityādi / … vajravārāhīyāminyādayaḥ ṣaṭ / mātarā iti saptamātṛrūpāḥ marudeśe /’ (Skt ms, 127a2–a4).28 But the word ṣaṭ in this passage is not the number modifying mātarāḥ; it refers to ṣaḍ yoginyaḥ, i.e., the six yoginīs beginning with Vajravārāhī and Yāminī.29

There are also cases where Gray creates unnatural translations, some of which appear to have been caused by unnecessary or in- correct emendations of manuscript readings. Some examples fol- low. The Vadodara ms of the CS 4a5 (: Skt ed, 3.17ab) reads sarvā kiṅkarī tasya sādhakasya na saṃśayaḥ,30 ‘all [the ḍākinīs] are fe- male servant[s] of that adept; no doubt.’ Gray translates: “There is no doubt regarding anything done by that adept.” (p. 175.) He does not explain how he emended the Sanskrit text.

A further example is CS 4b1–b2, where the Skt ms reads:

eṣa yogavaraḥ śreṣṭhaḥ sarvayogeṣu cottamaḥ / yaḥ kāṅkṣiṣyate kaścit sa devāsuramānuṣān / abhibhūya gamiṣyaty atra maṇḍale yo ’bhiṣiktaḥ / sarvatantroktasādhakaḥ /31 ‘This is the supreme yoga, the most excellent, and it is the highest among all yogas. Anyone who wishes [this supreme yoga] will go, conquering gods, titans, and men. [The one] who was initiated in this maṇḍala is the adept of what is taught in all tantras.’

Gray translates: “This yoga is the most excellent, the highest among all yogas, which can kill anyone, gods, titans or men. The adept who has been taught all tantras, and who has been initiated in the

28 Pandey’s edition of this line reads ṣaḍ yoginya ityādi / … / vajravārāhī yāminyādayaḥ ṣaḍ mātarā iti / saptamātṛrūpāḥ marudeśe / (CS-P, 548). 29 See also Jayabhadra’s comment on ṣaḍ yoginyaḥ and mātaraḥ: ṣaḍ yo- ginyo vajravāhyādicaṇḍikāntāḥ // mātaraḥ kākāsyādyāḥ // [Skt ed, 41.3] 30 Pandey’s edition reads sarvāḥ kiṅkarīs tasya sādhakasya na saṃ-śayaḥ.

31 For yaḥ kāṅkṣiṣyate kaścit, which is supported by Bhavabhaṭṭa’s com- mentary, Kambala’scommentaryreads yaḥkāṅkṣiṣyatinityaṃ (Sktms, 11b6– b7). For atra maṇḍale yo ’bhiṣiktaḥ sarvatantroktasādhakaḥ, Kambala’s commentary reads atra maṇḍalābhiṣiktaḥ sarvatantroktasādhanaḥ (Skt ms, 11b7). These variant readings are also acceptable.

maṇḍala, will go forth, conquering.” (p. 176.) Again, he does not adduce his version of the text.

The Skt ms of CS 4b7–5a1 reads tato jñātvā bhāvayen nityaṃ siddhis tathāgatavaco yathā, ‘therefore, should he know and al- ways visualize [the maṇḍala], [there will be] accomplishment (or supernatural effect), as taught by the Tathāgata.’ Gray translates, again without adducing the text: “Knowing thus, one should always meditate on the powers taught by the Tathāgata.” (p. 180.) The Skt ms of the CS 26a3–a4 (: Skt ed, 34.7) reads eṣate cakrodbhāsaṃ kuryād yathākramaṃ sarvasiddhi<ḥ>prasādha- kaḥ,32 ‘he seeks the radiance of the wheel. Should he practice [this wheel] in due succession, [he] accomplishes all supernatural effects (or accomplishments).’ Gray reads eṣate cakrodbhāsaṃ as eṣa te cakrodbhāsaṃ and emends to eṣa tricakrodbhāsaṃ by misread- ing Bhavabhaṭṭa’s comment on the word eṣate33 and by favoring the reading of one of the paper mss of Jayabhadra’s commentary, which is not attested in other materials; he then translates as “He should successively make the Three Wheels radiant. This is the ac- complishment of all powers.” (p. 311 and n. 15 on that page.)


4 Origin myths of Heruka and his maṇḍala Heruka is the highest deity of the Cakrasaṃvara tradition. Hence, researching the origin myths of Heruka and his maṇḍala has been a main concern of scholars studying this tradition. Gray unpacks the history of Indian versions of this myth mainly on the basis of the Sarvabuddhasamāyogatantra, the Sarvatathāgatatattvasaṃgraha, and Indrabhūti’s commentary on the CS.

32 For cakrodbhāsaṃ, Kambala reads cakranirdiṣṭaṃ. For the whole line, Bhavabhaṭṭa reads eṣate cakranirdiṣṭaṃ sarvasiddhipradāyakam. (CS-P, 528. I corrected Pandey’s eṣa te into eṣate.) Pandey’s edition reads eṣa te cakranirdiṣṭaṃ sarvasiddhiprasādhakam / cakrodbhāsaṃ tathā kuryād yathākarmānurūpataḥ //


33 Following Pandey’s edition (CS-P, 528), he reads Bhavabhaṭṭa’s eṣate mṛgayate, an explanation of the meaning of the word eṣate, as eṣa te mṛgayate, which makes less sense.


According to Gray, the myth in the Sarvabuddhasamāyogatantra described the birth of Heruka as a generation through yogic heat via controlled breathing. Heruka burns the triple world and Hindu deities such as Rudra, Mahādeva, Viṣṇu, Brahmā, etc., reduces them to ashes, and restores or reanimates them. Although these Hindu deities are roasted in Heruka’s process of cosmic cleansing, this does not mean that they are vilified. They are rather portrayed as victims of a cosmic disorder in which Māras (the traditional Buddhist villains) are active and which is the result of the inevi- table process of karmic conditioning. However, the version of the myth that eventually came to predominate portrays Śaiva deities as the perpetrators of cosmic disorder. The myth in the Sarvatathā- gatatattvasaṃgraha (which is a story of Vajrapāṇi’s subjugation of Mahādeva and Mahādeva’s conversion to Buddhism) and the myth found in Indrabhūti’s commentary on the CS are examples of this version. The myth in Indrabhūti’s commentary is especially impor- tant because many Tibetan versions of Heruka’s origin myth con- tain the story of the origin of the Cakrasaṃvara Heruka maṇḍala, and Indrabhūti’s commentary on the CS is the only known Indian text that presents a complete version of the myth. It is a likely source of the Tibetan versions.

Gray provides a translation of the whole text of the myth as it is introduced in Indrabhūti’s work and analyzes its content, using the Tibetan versions in support for his analysis. He then argues that the myth represents the adoption of non-Buddhist elements and that these elements are at the same time subordinated within a Buddhist cosmic hierarchy. The myth is therefore clearly a reaction to Hindu tripurāntaka myths.

Gray’s portrayal of the history of Heruka myths is acceptable, and it is beneficial to those who are interested in this topic. His discovery that the myth in question occurs in Indrabhūti’s work is no doubt a great contribution to the study of the Cakrasaṃvara tra- dition; I myself had completely overlooked it. However, he would have been able to paint a fuller picture by also considering the ver- sion of this myth that is found in Nāropāda’s ’Khor lo bde mchog gi rnam par ’phrul pa (*Cakrasaṃvaravikurvaṇa);34 Nāropāda’s version is as likely a source of the Tibetan representations of this myth as Indrabhūti’s.

Although some descriptions differ in the two versions, Gray’s analysis of Indrabhūti’s version can also be applied to Nāropa’s: both versions represent the adoption of non- Buddhist elements and subordination of these elements within a Buddhist cosmic hierarchy. However, Nāropāda’s version should also be considered because its explanations of the origination of the Heruka maṇḍala and of the subjugation of non-Buddhist di- vinities are more detailed than those given in Indrabhūti’s version. In comparison to the latter, Nāropāda’s version lends itself more naturally to the interpretation that the myth represents the adoption of non-Buddhist elements, and their subordination within Buddhist cosmic hierarchy, and need not be complemented with information taken from Tibetan versions.

Let us examine the contents of Nāropāda’s version briefly. 35 The beginning scene of the myth in Nāropāda’s version can be sum- marized as follows:

During the era of Kali, (1) a deity from the Thirty-three Heaven, (2) Gandharva, (3) the chief Yakṣa and (4) his attendant (g’yog), (5) the chief Rākṣasa and (6) his attendant, (7) the chief Nāga and (8) his at- tendant, and (9) the chief Asura and (10) his attendant, transforming themselves into twenty-four Bhairavas or ‘awful divinities’ (drag po), each took a consort; they then captured twenty-four sites located on the Jambū continent: (1’) four sites classified as pīṭha, (2’) four sites classified as upapīṭha, (3’) two sites classified as kṣetra, (4’) two sites classified as upakṣetra, (5’) two sites classified as chandoha, (6’) two sites classified as upacchandoha, (7’) two sites classified as melāpaka, (8’) two sites classified as upamelāpaka, (9’) two sites classified as śmaśāna, and (10’) two sites classified as upaśmaśāna, respectively. The four-bodied, four-natured, and four-faced Mahādeva, who resides on the summit of Mt. Meru with his four goddesses and his four secret goddesses, became the lord of these Bhairavas at their request. They

34 This work is preserved in the Peking edition of Tibetan Tripiṭaka, Otani University catalogue 4628. 35 For further details, see Sugiki 2006, 2007, and 2009.

terrorized the people living on the Jambū continent and brought this world to a state of degeneration.

The cosmic disorder was primarily a result of the inevitable proc- ess of karmic conditioning (i.e., the arrival of the Kali era). In this era of cosmic disorder, Śaiva divinities in Bhairava forms changed the Jambū continent – with the twenty-four sites in its center – into a Śaiva maṇḍala, took control of it, and caused it to be in an un- wholesome situation. The myth subsequently explains how the Heruka maṇḍala originated and how Śaiva divinities were subju- gated through enjoyment (longs spyod pa), dissolution (thim pa), and control (dbang du byas pa), which are only briefly mentioned in Indrabhūti’s version: Unhappy about this unwholesome situation, the Samyaksaṃbuddha came down from the Akaniṣṭha Heaven to the summit of Mt. Meru in order to subjugate these awful divinities. The Samyaksaṃbuddha manifested himself as an experiential-body divinity (longs sku) with one face, two arms, a white complexion, and the nature of Vajradhara, and he took Samantabhadrī (kun tu bzang mo) as his consort. He then transformed himself into the divinity named Heruka, who had a dark complexion, four faces, and twelve arms, and who took Vajravārāhī as his consort. Subsequently, Heruka and Vajravārāhī created twen- ty-four pairs of male and female heroic divinities who came to be those of the triple wheels (i.e., the origination of the Cakrasaṃvara Heruka maṇḍala). Each stage in the entire process of the manifesta- tion of the Heruka maṇḍala as described above reflected a particular characteristic of each of five Tathāgatas (i.e., Vairocana, Amitābha, Ratnasaṃbhava, Amoghasiddhi, and Akṣobhya).36

These Buddhist divinities conquered Mahādeva and his retainers, and, subjugating them, (1) made them objects of enjoyment through sexual assemblage and by making ornaments of their bones (= enjoy- ment), (2) effected the disintegration and incorporation of their con- sciousnesses (= dissolution), and (3) took control over their bodies, words, and minds (= control). In these steps of the process, the male and female Śaiva divinities were subjugated along the paths of anger and sexual passion respectively. Assimilating the essence of the Śaiva

36 The text does not expound the particular characteristic of each of the five Tathāgatas. They commonly symbolize the five kinds of gnosis and the five aggregates.

divinities by the incorporation of their consciousnesses and the orna- mentation of their bones, the Buddhist divinities then took over the twenty-four sites on the Jambū continent. Heruka then created four female divinities as gate-keepers, and created four other female di- vinities in addition (i.e., the eight ḍākinīs of the samayacakra portion of the Cakrasaṃvara Heruka maṇḍala). These eight female divini- ties conquered and subjugated Kinnaras of both sexes found at sites located in eight directions surrounding the above twenty-four sites formerly controlled by the Śaiva divinities.

Although the twenty-four pairs of Buddhist divinities established themselves at the twenty-four sites, they had not yet attained Buddhist enlightenment. Hence, they ascended Mt. Meru. Asked by them to give instructions on Buddhist truth, and receiving their various offer- ings and hymns, the Saṃyaksaṃbuddha at the summit of Mt. Meru produced the various tantras of the Cakrasaṃvara tradition.

As described in this myth, the summit of Mt. Meru and the Jambū continent were taken over by Buddhist divinities, and the good Buddhist dharma (i.e., tantras of the Cakrasaṃvara tradition) was brought to this world. The Samyaksaṃbuddha’s act of cos- mic cleansing was completed. It should be noted that the Buddhist divinities did not sweep away the Śaiva elements in this process, but instead assimilated the Śaiva essence into their cosmos. This is demonstrated by Buddhist divinities’ taking over of the Śaiva maṇḍala consisting of Mt. Meru and the twenty-four sites, which resulted in its change into Buddhist Heruka maṇḍala, and by the enjoyment, dissolution, and control process, through which the bodies, words, and minds of the Śaiva divinities became constitu- ents of their Buddhist counterparts. These processes therefore rep- resent the adoption of non-Buddhist elements and their subordina- tion within a Buddhist cosmic hierarchy.

5 The structure and functions of the Triple Wheel maṇḍala

Tantras belonging to the Cakrasaṃvara tradition describe many varieties of maṇḍalas. Among them, the most popular and widely used for practice is the Heruka maṇḍala consisting of five con-

centric wheels, i.e., the Great Bliss Wheel (mahāsukhacakra),37 the Mind Wheel (cittacakra), the Speech wheel (vākcakra), the Body Wheel (kāyacakra), and the Pledge Wheel (samayacakra). The Great Bliss Wheel, on which Heruka, his consort Vajravārāhī, four ḍākinīs, and four skull-bowls are depicted, is located at the center of this maṇḍala.38 The Great Bliss Wheel is surrounded by three concentric wheels, the Mind, Speech, and Body Wheels. These three are collectively called ‘the Triple Wheel’ (tricakra), and twen- ty-four holy sites and twenty-four coupled deities (i.e., twenty-four pairs of ḍākinī and vīra) assigned to these holy sites are depicted on them (i.e., eight holy sites with eight couples on each wheel ☓ 3 = twenty-four holy sites with twenty-four couples.) They are sur- rounded by the Pledge Wheel, on which eight ḍākinīs reside.39 This Heruka maṇḍala can be roughly described as having two forms, external and internal, and the deities and holy sites that constitute this maṇḍala symbolize traditional Mahāyānic or Indian concepts such as the triple world (sky, earth, underground), the three bodies of the Buddha (trikāya), the five elements (pañcabhūta), the ten spiritual levels (daśabhūmi), the ten perfections (daśapāramitā), the ten kinds of gnosis (daśajñāna), the eight vows (aṣṭasamaya), the thirty-seven conditions that contribute to awakening (sapta- triṃśadbodhipākṣikadharma), and so forth.40 The CS per se does not introduce the fully developed form of the Heruka maṇḍala, but explains its prototypical form,41 as Gray duly notes (p. 55, 58).

37 Gray names this circle ‘gnosis wheel’ ( jñānacakra) (p. 55), but it is more commonly called ‘great bliss wheel’ (mahāsukhacakra) in the Buddhist Cakrasaṃvara tradition. 38 Heruka and Vajravārāhī are situated at the center of the Great Bliss Wheel. They are surrounded by four ḍākinīs (i.e., Ḍākinī, Lāmā, Khaṇḍarohā, and Rūpiṇī) and four skull bowls in the cardinal directions and quarters, re- spectively. The four skull bowls are not explicitly mentioned in the CS.

39 The four gate-keeper ḍākinīs Kākāsyā, Ulūkāsyā, Śvānāsyā, and Śūka- rāsyā reside in the four directions and four other ḍākinīs, Yamadāḍhī (Ya- madāḍhī is more common than Gray’s Yamadāhī, p. 55 n. 169), Yamadūtī, Yamadaṃṣriṇī, and Yamamathanī, are in the four quarters.

40 For details of the structure and symbolism of this Heruka maṇḍala, see also Sugiki 2003b, 2007, and 2009.

41 For details of the prototypical form of the Heruka maṇḍala in the CS,

After describing the structure of the Heruka maṇḍala as above, Gray focuses his analysis on the structure and function of the Triple-wheel part of the Heruka maṇḍala and its doctrinal contexts. Since the CS does not explain every detail in full, Gray further re- lies on other sources, in particular on the Abhidhānottaratantra, the Yoginīsaṃcāratantra, the Saṃvarodayatantra, Umāpatideva’s Vajravārāhīsādhana, Lūyīpāda’s Bhagavadabhisamaya (= Cakra- saṃvarābhisamaya), Atiśa’s Abhisamayavibhaṅga, Abhayāka- ragupta’s Āmnāyamañjarī, and Bu-ston’s bDe mchog nyung ngu rgyud kyi spyi rnam don gsal.

I will now examine Gray’s portrayal of the Triple Wheel, i.e., the twenty-four Cakrasaṃvara holy sites beginning with Pullīramalaya and ending with Kulatā, and the coupled deities assigned to the twenty-four sites. The examination will focus on two points: (1) the mapping of the twenty-four Cakrasaṃvara holy sites and (2) the development of systems of the twenty-four internal holy sites.

5.1 The mapping of the twenty-four holy sites

Gray explains the geographical locations of the twenty-four Cakra- saṃvara holy sites on the Indian continent on the basis of Bu- ston’s bDe mchog nyung ngu rgyud kyi spyi rnam don gsal (notes on pp. 329–333), and their remapping over Kathmandu Valley and Tibetan and Mongolian areas on the basis of Abhayākaragupta’s Āmnāyamañjarī and some earlier studies on the topic (pp. 70– 71). The twenty-four Cakrasaṃvara sites, which originally re- ferred to the twenty-four sites on the Indian subcontinent, were remapped over areas outside India – such as Kathmandu Valley, Tibet, and Mongolia – during the process of transmission of the Cakrasaṃvara tradition from India to those outside areas. This in- terpretive flexibility was an essential factor in the transformation that the tradition had to undergo as it crossed regional boundaries. Abhayākaragupta’s definition of the nature of the Cakrasaṃvara holy sites – any sites, including Tibet and China, where living hu-


see also Sugiki 2003b, 2007, and 2009.

man female ḍākinīs resided could be regarded as Cakrasaṃvara holy sites – functioned to legitimate their remapping.

Gray’s account manages to capture an important aspect of the expansion of the Cakrasaṃvara tradition, but it deserves to be sup- plemented by a consideration of how Indian texts of the Cakra- saṃvara tradition prior to Abhayākaragupta discuss the mapping of the Cakrasaṃvara holy sites in India.42 While the instructions given in most of these texts are fragmentary, Nāropāda’s Yul nyi bcu bshi’i rgyu mtshan bstan pa43 gives detailed instructions and is very likely the most important Indian source for Tibetan versions like Bu-ston’s bDe mchog nyung ngu rgyud kyi spyi rnam don gsal, which Gray used.

Nāropāda identifies geographical locations of sites that are giv- en unnatural or obscure names by the Cakrasaṃvara scriptures: Himālaya is Mt. Kailāsa, Pretapurī (also called Pretādhivāsinī) re- fers to the valleys located on the border between India and Tibet, Gṛhadevatā (which, as Sanderson argued, was originally a name of the deity of the site Saurāṣra in the Śaiva Tantrasadbhāva) is Li yul, which may refer to Khotan. Suvarṇadvīpa is located of the coast of west India,44 but some say that it is in east China, and Nagara refers to Laṅkāpura, the land of rākṣasa, but some say that it is an area around a monastery standing on the border of Kaśmīra and northwest India. However, other Indian sources give different information on the geographical locations and features of the Cakrasaṃvara sites, which indicates that there were different maps of them. For example, Arbuda has been identified with Mt. Abu in modern Rajasthan since it was mentioned in the Mahābhārata, but it is identified with Takṣaśilā by Nāropāda. There are three different descriptions of the geographical location of Nagara according to

42 For details of the following analyses, see Sugiki 2006, 2007, and 2009.

43 This text is preserved in Peking edition of Tibetan Tripiṭaka, Otani University catalogue 4628 (the same catalogue number as Nāropāda’s ’Khor lo bde mchog gi rnam par ’phrul pa mentioned above). 44 Generally, Suvarṇadvīpa refers to the island in the ocean off the south tip of India, often Sri Lanka.

Nāropāda as mentioned in the previous paragraph; but according to the Yoginījālatantra (and the two commentaries on the Hevajra- tantra by Kāṅhapāda and Ratnākaraśānti), Nagara refers to Pāṭa- liputra (east India). Finally, Nāropāda describes many of the twen- ty-four sites as sites whose center is formed by sacral stones such as stone liṅgas of various shapes and stone dharmodayas. (The stone liṅgas conform in shape to the body parts which the Cakrasaṃvara scriptures equate with external holy sites.) These stone liṅgas and dharmodayas are very likely to be a Buddhist recasting of Śaiva śivaliṅgas and yonis. But many other authors, including Abhayāka- ragupta, regard goddesses or living human ḍākinīs as sacral cent- ers of the holy sites.

While the Indian compilers of texts belonging to the Cakrasaṃ- vara tradition attempted to pinpoint a specific geographical loca- tion and to define a specific feature for each individual site, they did not always reach a consensus. It is therefore likely that the locations and features of these sites were flexible rather than fixed. This sug- gests that the list of names of the twenty-four Cakrasaṃvara holy sites in India was rather idealized, serving as a symbolic frame- work along which individual sites were arranged, to a certain ex- tent, according to the respective compiler’s preference. This is also supported by other facts. First, as Sanderson pointed out, the list of twenty-four Cakrasaṃvara sites was produced in the process of the Buddhist redaction of the Śaiva list of holy sites. Second, although the CS provides a list of the sites in question, it does not per se give any clear descriptions of their actual geographical locations, or, for that matter, of the practice of actual pilgrimage to them. In terms of practice the CS rather focuses on the visualization or contemplation of the holy sites in the form of a maṇḍala. Only later scriptures, such as Nāropāda’s work, consider them in terms of geographical locations.

Attention should also be paid to the change of descriptions from the CS to the Saṃvarodayatantra with regard to the travel of the Cakrasaṃvara practitioner. In the CS, the places where the practitioner travels in search for ḍākinīs are described as ‘villages’

(grāma),45 and these are not yet defined as the twenty-four Cakra- saṃvara sites. The Saṃvarodayatantra, on the other hand, defines these locations to be the twenty-four sites.46 This change of descrip- tion could result from an attempt to interpret these holy sites as a symbolic framework, mapped to sets of villages or towns in areas that were actually controlled by Cakrasaṃvara Buddhists, or at least accessible to them. Gray argues that Indian Cakrasaṃvara Buddhists emphasized the internal practice of the twenty-four sites (i.e., meditational practice of the body maṇḍala, in which all the sites are visualized in one’s body), and that this may have reflected the political reality that Buddhists did not have control over many, or any, of them (pp. 68–70). But his explanation covers only half of the history of the theology of these holy sites because it ignores that Indian Cakrasaṃvara Buddhists eagerly attempted to map them to the human body and to map and remap them over the Indian conti- nent already before Abhayākaragupta.

Interpretive flexibility regarding the mapping of the twenty-four Cakrasaṃvara holy sites was already, and often, the hermeneutic stance of Indian Cakrasaṃvara Buddhists prior to Abhayāka- ragupta; it is not exclusively linked to the tradition’s subsequent transmission to areas such as Kathmandu Valley, Tibet, and Mongolia. Abhayākaragupta’s statement that any sites where living ḍākinīs reside can be regarded as Cakrasaṃvara holy sites should be understood in this hermeneutic context of Indian Cakrasaṃvara Buddhism, as well as in the context of the tradition’s transmission from India to its outlying areas. Finally, I would like to make a small suggestion concerning Abhayākaragupta’s mention of Tibet and China. Gray states that “the mention of Tibet and China is surely not accidental, as these were major destinations for its (= the Cakrasaṃvara tradition’s, TS) transmission, of which erudite Indian Buddhists such as Abhayākaragupta were certainly aware (p. 70).” This may be correct, but it is also possible that Abhayā- karagupta merely followed Nāropāda, who had mentioned Tibetan

45 See again the passage grāme grāme vrajan tasya dūtayas rūpalakṣa-ṇam discussed in section 3.2 of this paper.

46 The Saṃvarodayatantra, 9.

and Chinese Cakrasaṃvara sites in his Yul nyi bcu bshi’i rgyu mt- shan bstan pa, and the Vajraḍākatantra, which, together with the Ḍākārṇavatantra, defined Bhoṭa (i.e., Tibet) as one of the Cakra- saṃvara holy sites.

5.2 The development of systems of twenty-four internal holy sites

The twenty-four Cakrasaṃvara holy sites have both external and internal forms. In the internal practice of the twenty-four holy sites, these sites are visualized at various places in the practitioner’s own body. The collective body of these twenty-four holy sites is noth- ing other than the highest deity, Heruka. Hence, the practitioner, through the visualization of the internal holy sites, experiences an identification with Heruka as his or her innate Buddha nature.

As Gray mentions, the meditational practices of internal holy sites are often seen as the highest form of practices centered on holy sites in the Cakrasaṃvara tradition (pp. 68–70). The Cakrasaṃvara Buddhists were very eager to develop a system of internal holy sites and created many varieties of such a system. However, Gray’s por- trayal of the system of internal Cakrasaṃvara holy sites covers only half of the tradition (which may well have been his intention).

As I argued in 2003,47 a more comprehensive view suggests that the theories regarding the internal Heruka maṇḍala comprised of the twenty-four sites developed in two stages, with Gray’s portrayal being limited to the first: (1) the stage of the internalization of the twenty-four holy sites (i.e., the Triple Wheel), which symbolize the daśabhūmi and daśapāramitā, and (2) the stage of the internaliza- tion of the entire Heruka maṇḍala including the twenty-four holy sites (i.e., the Great Bliss, Triple, and Pledge Wheels), which sym- bolizes the saptatriṃśadbodhipākṣikadharma as well as the daśa- bhūmi and daśapāramitā.48


47 Sugiki 2003b; see also Sugiki 2007.

48 Note that in some texts, the trikāya, the daśajñāna, the trayodaśabhūmi, and some other concepts traditionally taught in Mahāyāna Buddhism are in- ternalized along with the daśabhūmi, daśapāramitā, and saptatriṃśadbod hipākṣikadharma.

The Cakrasaṃvara literature describes a variety of forms of the internal Heruka maṇḍala. They can be classified into five types. The first type appears in the last half of chapter 50 of the CS. The second type is introduced in the Abhidhānottarottaratantra, the Vajraḍākatantra, the Saṃvarodayatantra, the Saṃpuṭodbhava- tantra, Lūyīpāda’s Cakrasaṃvarābhisamaya, Jayabhadra’s Cakra- saṃvarasādhana, and so forth. The third type can likewise be found in the Abhidhānottaratantra, as well as in Dhīmat’s Cakra- saṃvarodayamaṇḍalopāyikā and Kumārakalahaṃsapāda’s Saṃ- vararahasyanāmasādhana. The fourth type appears in Atiśa’s Abhisamayavibhaṅga, Prajñārakṣita’s Abhisamayapañjikā, Tathā- gatavajra’s Abhisamayavṛtti, Abhayākaragupta’s Cakrasaṃvarā- bhisamaya, and Śubhākaragupta’s Abhisamayamañjarī, which, except for the last two, are commentaries on Lūyīpāda’s Cakra- saṃvarābhisamaya. The internal Heruka maṇḍala given in the Yoginīsaṃcāratantra can also be considered as of this fourth type. The fifth type is described in Dārikapāda’s Cakrasaṃvarasādhana, Ghaṇāpāda’s Cakrasaṃvarasādhana and Kāyamaṇḍalābhisamaya, Kṛṣṇācārya’s Cakrasaṃvarasādhana and Vasantatilakā, and in the Jñānodayatantra. The versions of the first and second types of the internal Heruka maṇḍala emerged in the first stage of development, whereas the third, fourth, and fifth types developed in the second.

Let us see the five types of the internal Heruka maṇḍala in de- tail. The following elements constitute instructions of this maṇḍala:

(0) Basic philosophy:

A somatic philosophy that enlightenment can be obtained through one’s own body: one’s body is a means for attaining enlightenment.

(i) Internalized objects:

(i-1) Twenty-four holy sites and twenty-four coupled deities (i.e., the Triple wheel), which are equivalent to the daśabhūmi and the daśapāramitā.

(i-2) Thirty-seven coupled and single deities (i.e., the whole maṇḍala including the Triple wheel), which are equivalent to the saptatriṃśadbodhipākṣikadharma as well as the daśa-bhūmi and the daśapāramitā.

(ii) Body counterparts:

(ii-1) The channels (nāḍī) together with their corresponding body ingredients (dhātu), and body sites (sthāna etc.) where the channels are seated.

(ii-2) Heruka’s supernatural form: Heruka’s four faces and the objects in Heruka’s twelve hands, and the external Vārāhī. (ii-3) The four principal circles (cakra): the mahāsukhacakra in the head, the saṃbhogacakra in the throat, the dharma- cakra in the heart, and the nirmāṇacakra in the abdomen; and the eight gates of the body (i.e., the eight orifices: right and left ears, right and left eyes, right and left nostrils, mouth, and anus).

(iii) Methods for actual practice (i.e., meditation):

Meditational process for visualization of the internal Heruka maṇḍala.

All five types of the internal Heruka manḍala share the somatic philosophy (the factor (0) above) which legitimizes the internal practice of the Heruka maṇḍala. But the five types are distin- guished from each other by the elements (i), (ii) and (iii) as shown in the following table.


(i) (ii) (iii)

First type (i-1) (unclear) (unclear)

Second type (i-1) (ii-1) Described

Third type (i-2) (ii-1) and (ii-2) Described

Fourth type (i-2) (ii-1) and (ii-3) Described

Fifth type (i-2) (ii-1) Described


The first and second types internalize the twenty-four holy sites and the twenty-four coupled deities (i-1) and therefore can be said to aim at the somatic application of the somatic philosophy focuss-

ing on the daśabhūmi and daśapāramitā. Unlike the second type the first type gives no explanation for (ii) and (iii) even though it ar- gues that the twenty-four holy sites should be practiced internally. Put in another way, the system of the first type remains idealized. This idealized system, however, functions as a foundation for the second type. The systems of the third, fourth, and fifth types are in turn based on the second type with regard to the Triple-wheel part of the maṇḍala. For this reason, the idealized system of the first type can be defined as the prototypical form of the internal Heruka maṇḍala. The second type attaches the elements (ii-1) and (iii) to this prototype; in the second type channels, body ingredients, and body sites are equated with the twenty-four ḍākinīs, the twenty-four vīras, and the twenty-four sites on the Triple Wheel, respectively. The third, fourth and fifth types internalize the element (i-2).

Their aim can be described as the somatic application of the somat- ic philosophy focussing on the saptatriṃśadbodhipākṣikadharma along with the daśabhūmi and daśapāramitā. This shift from (i-1) to (i-2) seems to have some relation to the development of an exter- nal five-wheeled Heruka maṇḍala in the scriptures of the Cakra- saṃvara tradition composed after the CS. The third type applies the concept of (ii-1) for the internaliza- tion of the Triple Wheel, and applies (ii-2) for the internalization of the Great Bliss Wheel and the Pledge Wheel. The fourth type, on the other hand, introduces the concept of (ii-3) for the internali- zation of the Great Bliss Wheel and the Pledge Wheel. The fifth type applies the concept of (ii-1) not only for the internalization of the Triple Wheel but also for the internalization of the Great Bliss Wheel and the Pledge Wheel.

As noted above, the historical development of internal forms of the Heruka maṇḍala began in its first stage with the internaliza- tion of the twenty-four holy sites/the daśabhūmi and daśapāramitā (i.e., the Triple Wheel). This was followed by the internalization of the whole maṇḍala/saptatriṃśadbodhipākṣikadharma along with the daśabhūmi and daśapāramitā (i.e., the Great Bliss, the Triple, and the Pledge Wheels).49 The significance of the instruction in the

49 However, it should be noted that the arrival of the the third, fourth, and

internal Triple-wheel Heruka maṇḍala given in the CS (i.e. the first of the five types), is that it provides the basic conceptual frame- work for Cakrasaṃvara systems of practice of the internal Heruka maṇḍala, around which the later and more elaborate types were developed.

Conclusion

Criticizing the work of others is a relatively easy task, whereas producing original studies is difficult. Although Gray’s pioneering study on the CS has some problems, it is nevertheless a significant contribution to the study of Indian Buddhist Yoginītantra litera- ture, for the reasons I mentioned in the first part of this article: This is the first full translation of the CS; with its copious annotation, Gray’s study can serve as a guidebook to the commentaries on the CS, and it provides much textual and contextual information on the Indian Cakrasaṃvara tradition in general. Together with the book under review, Gray’s critical edition of the Sanskrit and Tibetan CS, announced as a companion volume, will hopefully further pro- mote the study of Tantric Buddhism.

Acknowledgments

I am responsible for any mistakes found in this paper. However, I would like to express my heartfelt thanks to Mr. Joseph M. Logan for having assisted me with the English and to Dr. Birgit Kellner for suggestions and editorial contributions. When I was writing this review article, Péter-Dániel Szántó also published a short re- view [[[Tantric]] Studies 1 (2008) 215–219]. When I was finalizing this article, Alexis Sanderson’s paper “The Śaiva Age – The Rise and Dominance of Śaivism during the Early Medieval Period”

fifth types does not signal the extinction of the tradition of the second type. Versions of the internal Heruka maṇḍala that can be classified as belong- ing to the second type continued to be practiced even after the third, fourth, and fifth types appeared, presumably because the second type was taught in the classic Cakrasaṃvara canons such as the Abhidhānottaratantra, the Vajraḍākatantra, the Saṃpuṭodbhavatantra, etc. and therefore often seen as authoritative.

appeared [In: Genesis and Development of Tantrism, ed. Shingo Einoo. Tokyo 2009: Institute of Oriental Culture, University of Tokyo, 41–349]. Sanderson’s paper also includes his criticism of Gray’s work. Some of the arguments that I made here overlap with arguments by Szántó and Sanderson, which, I hope, the readers will kindly pardon.

(Research for this review article was partially supported by the Ministry of Education, Science, Sports and Culture, Grant-in-Aid for Scientific Research (C), 20520047, 2008.)

Appendix: draft editions of two passages from the Abhidhā- nottaratanta that explain the subtle body system and psycho- somatic meditation based on it

These passages were originally presented in Sugiki 2007 and are reproduced here because this paper may be difficult to access from outside Japan. These are not the only passages that explain the sys- tem in question in the Abhidhānottaratantra, but they are sufficient to validate my analysis in section 2.2 above. Two special conventions are used in the text:

  • virajaṃ [→ viramaṃ]: all manuscripts read virajaṃ, but this is likely to be a later emendation or corruption; the reading viramaṃ is more likely to be original.
  • ekaṃ [= prathamaṃ]: while the reading outside the brackets (here: ekaṃ) is irregular, or its meaning is obscure, it is nevertheless deemed to be the original reading (and, hence, acceptable in the context of the edited work). However, it should be regarded as equivalent to the term given in brackets.

The Abhidhānottaratantra. Skt mss: IASWR I-100 149b5–150a3, Matsunami 10 158a4–b2. Matsunami 12 186a2–b1. catuḥsandhyānuṣṭheyaṃ1 caturānandanandanaṃ2 / nābhihṛtpadmas ta- nau madhye3 jihvāmūle śiropari // ānandaṃ paramam4 caiva *virajaṃ [→ viramaṃ]5 sahajaṃ6 tathā / catuḥṣaṣṭidalam7 *ekaṃ [= prathamaṃ]8

1 catuḥsandhyā- ] IASWR; catusandhyā Matsunami 10 and 12. 2 -nandanaṃ ] em.; nandanāṃ IASWR and Matsunami 12. nandanā Matsu- nami 10. 3 tanau madhye ] IASWR; tanūmadhya Matsunami 10. tanumadhye Matsu- nami 12. 4 paramam ] IASWR; paramām Matsunami 10 and 12. 5 The third of the Four Blisses (caturānanda) is generally named virama. (The Four Blisses are: ānanda, parama-ānanda, virama-ānanda, and sahaja- ānanda.) 6 sahajaṃ ] Matsunami 10 and 12; om. IASWR. 7 catuḥṣaṣṭi- ] Matsunami 10 and 12; catuṣaṣṭi IASWR. 8 The intention of this phrase is that the first inner circle connected with ānanda, the first Bliss, is of the shape of a lotus with sixty-four petals. Hence, I

dvitīyam9 aṣṭadalam uttamam // tṛtīyaṃ10 ṣoḍaśadalaṃ caturthaṃ dvātriṃśaddalam11 / vārāhī nābhimūlasthaṃ12 sahajaṃ13 herukotta- mam // * caturāryasatyatāṃ bhāvyaṃ [→ caturāryasatyatā bhāvyā]14 sandhyākāleṣu15 *yoginām [→ yoginā]16 / duḥkhaṃ nirmāṇacakraṃ tu samudayo dharmacakrayoḥ // nirodhaṃ saṃbhogacakraṃ17 mārgaṃ cai- va mahāsukham / evaṃ sandhyā18 anuṣṭhānaṃ kṛtaṃ yogasuniścitam //

The Abhidhānottaratantra. Skt mss: IASWR I-100 83b1–b5, Matsu- nami 10 78a1–5, Matsunami 12 89b3–90a2. This passage is closely related to Kṛṣṇācārya’s Olicatuṣṭaya or Ālicatuṣṭaya system. There are two versions of the Olicatuṣṭaya or Ālicatuṣṭaya system: the ver- sion taught in Kṛṣṇācārya’s Vasantatilakā and the version taught in the same author’s Olicatuṣṭaya or Ālicatuṣṭaya. The former version is closely related to instructions of psychosomatic meditation given in the Saṃpuṭodbhavatantra and Vajraḍākatantra, and the latter version is to the passage edited below. (For details, see Sugiki 1999 and 2007.) A similar passage also appears in Vanaratna’s Rahasya- dīpikā Skt ed: pp. 87–88.

note that the word ekaṃ means prathamaṃ or first in this context. 9 dvitīyam ] IASWR and Matsunami 12; dvitiyam Matsunami 10. 10 tṛtīyaṃ ] IASWR; tṛtīya Mastunami 10 and12. 11 dvātriṃśad- ] Matsunami 12; dvātriśad IASWR. dvātriṃśata Matsunami 10. 12 In Sugiki 2007 I emended -sthaṃ to -sthā because it is vārāhī who resides at the base of the navel region. However, -sthaṃ is acceptable because it is possible to read this line as: ‘Vārāhī, [who is] the Innate (sahajaṃ), resides at the base of the navel region. Heruka [, who is also the Innate, resides at] the upper place (i.e., the head).’ 13 sahajaṃ ] Matsunami 10 and 12; saha IASWR. 14 caturāryasatyatā bhāvyā ... yoginā is grammatically better and makes bet- ter sense in this context. (A Yogin should conceive the nature of the Four Noble Truths in all the times [i.e., the four sandhi connected with the four inner circles connected with the Four Blisses].) 15 -kāleṣu ] IASWR and Matsunami 10; kāla Matsunami 12. 16 See note 14. 17 saṃbhogacakraṃ ] em.; saṃbhogikacakraṃ IASWR. saṃbhogacakre Matsunami 10 and 12. 18 sandhyā ] IASWR; sadhyām Matsunami 10. sandhyām Matsunami 12. (In Sugiki 2007 I edited as sandhyām.)

  • mantha[→ manthya]manthānayogena19 jñāna*raśmir[→ vahnir]20 iha karmaṇā mārutena prerito nābhimaṇḍale dhūmāyati jvalati21 dīp- tibhiḥ22 / < samayacakre gatān sugatān *dagdhā[→ dagdhvā]23,>24 ta-

thāgatānāṃ25 saṃbhogacakra*gatān[→ gatam]26 upāyaṃ triḥprada- kṣiṇīkṛtya, ūrṇākośagatena27 marmodghāṭanadvāreṇa28 niḥsṛtya29, daśadiglokadhātusthitānāṃ tathāgatānāṃ jñānāmṛtaṃ30 gṛhītvā, śikhārandhragatena kanakadvāreṇa31 jālandharasaṃjñakena32 praviś- ya, dantasīmottaragatarandhreṇa33 saṃbhogacakre viśramya34, dag- dhānāṃ35 tathāgatānām ānandaṃ janayantī, nābhimaṇḍale36 sthirī-

19 Both the phrases manthamanthāna and manthyamanthāna can be found in Buddhist esoteric scriptures, but the latter is better. 20 Generally the psychosomatic fire of gnosis is named jñānavahni, jñānāgni, or jñānānala, but jñānaraśmi appears to be acceptable. 21 jvalati ] IASWR and Matsunami 10; jvaranti (or -ra- is cancelled?) Matsunami 12. 22 dīptibhiḥ ] Matsunami 10 and 12; jva(five letters blurred) IASWR. 23 This line explains the process of the jñānaraśmi or jñānavahni’s upward movement from the navel circle to the heart circle. The meaning of this line is: ‘Having burnt the Sugatas residing on the samayacakra (i.e., the dharmacakra in the heart), …’ Hence, dagdhvā is better. 24 From samayacakre to *dagdhā[→ dagdhvā] (inside the brackets) ] blurred in IASWR. 25 tathāgatānāṃ ] Matsunami 12; tānāṃ Matsunami 10. 26 What resides on the saṃbhogacakra (i.e., the cakra in the throat) is the sound OṂ, which is here referred to with upāyaṃ. 27 -gatena ] Matsunami 10; gate IASWR and Matsunami 12. 28 marmodghāṭana-] em.; ma(five or six letters blurred) IASWR. rmodghāṭana Matsunami 10. mamodgheṭana Matsunami 12. 29 niḥsṛtya ] em.; blurred (nisṛtya or nisṛjya?). IASWR. niśṛtya Matsunami10 and 12. The jñānaraśmi or jñānavahni goes out of the practitioner’s body through his or her marmodghāṭanadvāra. Hence, niḥsṛtya is better than niśritya, another possible emendation of niśṛtya. 30 -mṛtaṃ ] Matsunami 10 and 12; mṛtā LASWR. 31 kanakadvāreṇa ] Matsunami 10 and 12; kanakakalajalena IASWR. 32 jālandhara- ] IASWR and Matsunami 12; jāraṃdhara Matsunami 10. 33 -randhreṇa ] IASWR and Matsunami 12; caṃdreṇa (or readable as raṃ- dhreṇa?) Matsunami 10. 34 viśramya ] IASWR and Matsunami 10; viśamā Matsunami 12. 35 dagdhānāṃ ] Matsunami 10 and 12; dagdhānā IASWR. 36 -maṇḍale ] Matsunami 10; maṇḍala (or readable as maṇḍale?) Matsu-

bhavati37 //

Abbreviations

Skt ms(s). Sanskrit manuscript(s).

Skt ed(s). Sanskrit text(s) critically edited.

IASWR. Mss on microfilm copies kept at the [recently defunct] Institute for the Advanced Studies of World Religions, Stony Brook, NY. Catalogue numbers according to: Christopher S. George and Mānabajra Bajrācārya, Buddhist Sanskrit Manuscripts. A Title List of the Microfilm Collection of The Institute for Advanced Studies of World Religions. Stony Brook 1975. Matsunami. Mss on microfilm copies kept at the library of the University of Tokyo. Catalogue numbers according to: Seiren Matsunami, A Catalogue of the Sanskrit Manuscripts in the Tokyo University Library. Tokyo 1965: Suzuki Research Foundation. NGMPP. Mss on microfilm copies kept at National Archives in Kathmandu. Reel numbers according to Nepal-German Manuscript Preservation Project.

Primary sources

Abhidhānottaratantra. Skt ms: IASWR I-100, Matsunami 10 and 12. Skt ed (selected chapters): Kalff 1979. Cakrasaṃvaratantra, or CS. Cakrasaṃvaramahāyoginītantrarāja. Skt ms: Oriental institute Vadodara, accession no 13290. CS-P. CS edited by Janardan Shastri Pandey. See Pandey 2002. Brahmayāmala, or BY. Brahmayāmala. Skt ms: NGMPP A42/6. Skt ed (se- lected paragraphs or chapters): Sanderson 2006 and Hatley 2007. Cakrasaṃvarapañjikā, or J. Cakrasaṃvarapañjikā, Jayabhadra’s commen- tary on the CS. Skt ed: Sugiki 2001. Herukasādhananidhi, or K. Herukasādhananidhipañjikā, Kambala’s com- mentary on the CS. Skt ed: Unpublished edition by Sugiki. Vajraḍākatantra, or VḌT. Vajraḍākamahāyoginītantrarāja. Skt ed: Unpublished edition by Sugiki. Cakrasaṃvaravivṛti, or Bh. Cakrasaṃvaravivṛti, Bhavabhaṭṭa’s commentary on the CS. Skt ed: Pandey 2002.

nami 12.

37 -bhavati ] Matsunami 10 and 12; blurred IASWR.

Nāmasaṅgīti. Āryamañjuśrīnāmasaṅgīti. Skt ed: Lal 1994. Rahasyadīpikā. Vanaratna’s commentary on Kṣṇācārya’s Vasantatilakā. Skt ed: Rinpoche and Dwivedi 1990.

Secondary sources

Hatley 2007. Shaman Hatley: The Brahmayāmalatantra and Early Śaiva Cult of Yoginīs. PhDdiss. Philadelphia 2007: University of Pennsylvania. Kalff 1979. Martin M. Kalff: Selected Chapters from the Abhidhānottara- Tantra: The Union of Female and Male Deities. PhD diss. 2 vols. New York 1979: Columbia University. Lal 1994. Banarsi Lal: Āryamañjuśrīnāmasaṅgīti with Amtakaṇikā-tippaṇī by Bhikṣu Raviśrījñāna and Amtakaṇikodyota-nibandha of Vibhūtican- dra. Sarnath 1994: Central Institute of Higher Tibetan Studies. Pandey 2002. Janardan Shastri Pandey: Śrīherukābhidhānam Cakrasaṃ- varatantram with the Vivṛti Commentary of Bhavabhaṭṭa, Vol. I and II. Sarnath 2002: Central Institute of Higher Tibetan Studies. Rinpoche and Dwivedi 1990. Samdhong Rinpoche and Vrajvallabh Dwivedi: Vasantatilakā by Caryāvratī ŚrīKṣṇācārya with Commentary Rahasyadīpikā by Vanaratna. Sarnath 1990: Central Institute of Higher Tibetan Studies. Sakurai 1997. Munenobu Sakurai: Cakrasaṃvarābhisamaya-no daiyuga-wo megutte (On Mahāyoga in the Cakrasaṃvarābhisamaya). Indo-gaku- bukkyō-gaku-kenkyū (Journal of Indian and Buddhist Studies) 46 (1997) 124–129. Sakurai 1998. Munenobu Sakurai: Cakrasaṃvarābhisamaya-no genten- kenkyū – bonbun-kōtei-tekusuto (A Critical Study of the Cakrasaṃvarā- bhisamaya). Chisan-gakuhō (Journal of Chisan Studies) 47 (1998) 1–32. Sanderson 1995. Alexis Sanderson: Vajrayāna: Origin and Function. In: Buddhism into the year 2000. Bangkok and Los Angeles 1995: Dhamma- kāya Foundation, 87–102. Sanderson 2001. Alexis Sanderson: History through Textual Criticism in the Study of Śaivism, the Pañcarātra and the Buddhist Yoginītantras. In: Les Sources et le temps. Sources and Time: A Colloquium, Pondicherry, 11–13 January 1997, ed. François Grimal. Publications du département d’Indologie 91. Pondicherry 2001: Institut Français de Pondichéry/École Française d’Extrême-Orient, 1–47. Sanderson 2006. Alexis Sanderson: The Śaiva Model of Post-initiatory Caryā in the Yoginītantras. Handout accompanying a lecture in the Graduate School of Humanities, University of Tokyo, 14 September 2006.

Sugiki 1999. Tsunehiko Sugiki: Kṛṣṇācārya-no shi-shidai (Kṣṇācārya’s Four-staged Meditational Process). Indo-gaku-bukkyō-gaku-kenkyū (Journal of Indian and Buddhist Studies) 47/2 (1999) 880–883. Sugiki 2000. Tsunehiko Sugiki: Kṛṣṇācārya’s Śrīcakrasaṃvarasādhana – Critical Edition with Notes. Chisan-gakuhō (Journal of Chisan Studies) 49 (2000) 45–62. Sugiki 2001. Tsunehiko Sugiki: Cakrasaṃvaratantra-no seiritsu-dankai- nitsuite – oyobi Jayabhadra-saku Śrīcakrasaṃvarapañjikā-kōtei-Bonpon (On the Making of the Cakrasaṃvaratantra and a Critical Study of Jaya- bhadra’s Śrīcakrasaṃvarapañjikā). Chisan-gakuhō (Journal of Chisan Studies) 50 (2001), 91–141. Sugiki 2002. Tsunehiko Sugiki: A Critical Study of the Vajraḍākamahā- tantrarāja (I) – Chapters 1 and 42. Chisan-gakuhō (Journal of Chisan Studies) 51 (2002) 81–115. Sugiki 2003a. Tsunehiko Sugiki: A Critical Study of the Vajraḍākamahātan- trarāja (II) – Sacred Districts and Practices Concerned. Chisan-gakuhō (Journal of Chisan Studies) 52 (2003) 53–106. Sugiki 2003b. Tsunehiko Sugiki: Five Types of Internal Maṇḍala Described in the Cakrasaṃvara Buddhist Literature – Somatic Representations of One’s Innate Sacredness. Tōyō-bunka-kenkyū-sho-kiyō (The Memoirs of the Institute of Oriental Culture) 144 (2003) 157–231.

Sugiki 2006. Tsunehiko Sugiki: Haiburid-na seichi-no tajigensei: bukkyō- saṃvarakei-niokeru ikai-no chikara. In: Ikai-no kōsaku vol.2, ed. Ayako Hosoda and Kazuko Watanabe. Tokyo 2006: Lithon, 55–85. Sugiki 2007. Tsunehiko Sugiki: Saṃvara-kei mikkyō-no shosō (Aspects of Saṃvara Esoteric Buddhism). Tokyo 2007: Tōshindō. Sugiki 2008. Tsunehiko Sugiki: The Homa System of the Vajraḍākatantra: A Critical Edition and a Preliminary Analysis of its Homa System. Tantric Studies 1 (2008) 131–154. Sugiki 2009. Tsunehiko Sugiki: The Structure and Traditions of the Systems of Holy Sites in Buddhist Saṃvara Cycle and its Related Scriptural Cycles in Early Medieval South Asia: The Geography of Esoteric Buddhism in the Eyes of the Compilers of the Scriptures. In: Genesis and Development of Tantrism, ed. Shingo Einoo. Institute of Oriental Culture Special Series 23. Tokyo 2009: Institute of Oriental Culture, University of Tokyo, 515–562.


Tsunehiko SUGIKI