Making Sense of Tantric Buddhism: History, Semiology, and Transgression in the Indian Traditions 9780231162401, 0231162405

Christian K. Wedemeyer's systematic investigation into Buddhist Tantric traditions fundamentally rethinks the natur

1,170 265 12MB

English Pages 313 [337]

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Making Sense of Tantric Buddhism: History, Semiology, and Transgression in the Indian Traditions
 9780231162401, 0231162405

Table of contents :
Front Cover
Title Page
Publication Page
Contents
Preface and Acknowledgments
Introduction
Part I: Historiography
1. ORIGINS, RELIGION, AND THE ORIGINS OF TANTRISM
2. NARRATING TANTRIC BUDDHISM
3. GOING NATIVE: TRADITIONAL HISTORIOGRAPHY
Part II: Interpretation
4. THE SEMIOLOGY OF TRANSGRESSION
5. THE PRACTICE OF INDIAN TANTRIC BUDDHISM
6. TANTRIC BUDDHIST TRANSGRESSION IN CONTEXT
Conclusion
APPENDIX I: THE INDRABHÜTI STORY ACCORDING TO PAD MADKAR PO (CA. 1575)
APPENDIX II: CHAPTER NINE OF
Notes
Bibliography
Index

Citation preview

M A K I N G SENSE OF T A N T R I C B U D D H I S M

SOUTH ASIA ACROSS THE DISCIPLINES

S O U T H ASIA ACROSS THE DISCIPLINES ♦♦♦ E D I T E D BY D I P E S H C H A K R A B A R T Y , S H E L D O N P O L L O C K , A N D SANJAY S U B R A H M A N YAM

Funded by a grant from the Andrew W. Mellon Foundation, and jointly published by the University of California Press, the University of Chicago Press, and Columbia University Press. Extreme Poetry: The South Asian Movement of Simultaneous Narration by Yigal Bronner (Columbia) The Social Space ofLanguage: Vernacular Culture in British Colonial Punjab by Farina Mir (California) Unifying Hinduism: The Philosophy of Vijnanabhiksu in Indian Intellectual History by Andrew J. Nicholson (Columbia) Everyday Healing: Hindus and Others in an Ambiguously Islamic Place by Carla Bellamy (California) The Millennial Sovereign: Sacred Kingship and Sainthood in Islam by A. Azfar Moin (Columbia) South Asia Across the Disciplines is a series devoted to Publishing first books across a wide ränge of South Asian studies, including art, history, philology or textual studies, philosophy, religion, and the interpretive social Sci­ ences. Series authors all share the goal of opening up new archives and suggesting new methods and approaches, while demonstrating that South Asian scholarship can be at once deep in expertise and broad in appeal.

MAKING SENSE OF TANTRIC BUDDHISM

H I S T O R Y , S E M I O L O G Y , AND T R A N S G R E S S I O N IN T H E I N D I A N T R A D I T I O N S

Christian 'K.'Wedemeyer

CO LUM B IA U N I V E R S I T Y PRESS

N E W YORK

Columbia University Press Publishers Since 1893

New York Chichester, West Sussex cup.columbia.edu Copyright © 2013 Columbia University Press All rights reserved

Library of Congress Cataloging-in-Publication Data

Wedemeyer, Christian K. Making sense of Tantric Buddhism: history, semiology, and transgression in the Indian traditions / Christian K. Wedemeyer. pages cm—(South Asia across the disciplines) Includes bibliographical references and index. ISBN 978-0-231-16240-1 (cloth : alk. paper)-ISBN 978-0-231-53095-8 (electronic) 1. Tantric Buddhism—India. I. Title. BQ8912.9.I5W43 2013 294.3'925-dc23 2012006168

© Columbia University Press books are printed on permanent and durable acid-ffee paper. This book was printed on paper with recycled content. Printed in the United States of America

c 10 9 8 7 6 5 4 3 2 1

Cover image: Photographed by John C. and Susan L. Huntington, June 1994. Rajshahi Museum, Rajshahi, Bangladesh. Accession #851 Cover design: Katie Poe

References to Internet Web sites (URLs) were accurate at the time of writing. Neither the author nor Columbia University Press is responsible for URLs that may have expired or changed since the manuscript was prepared.

For Josie, who gave so m uch to so many, and Gita, who w aited so patiently for me to w rite a “real book”

♦♦♦

For the scholar ofcomparative religions, Tantrism must represent the ultimate challenge. —Michel Strickmann

CONTENTS

List ofFigures and Tables xiii Preface and Acknowledgments xv List o f Abbreviations xix

INTRODUCTION: MAKING SENSE IN AND OF THE HUMAN SCIENCES

1

PART I: HISTORIOGRAPHY 1. ORIGINS, RELIGION, AND THE ORIGINS OF TANTRISM U n d e rs ta n d in g T a n tric B u d d h ism T h ro u g h Its O rig in s

18

T h e Q uest fo r O rigins as M e th o d in th e H isto ry o f R eligions

2. NARRATING TANTRIC BUDDHISM T h e P o etics o f H isto rio g ra p h y

17

32

37

38

T a n tra as End: T h e D ecline a n d Fall o f In d ia n B u d d h ism T a n tra as B eginning: T h e P rim o rd ia l U n d e r c u r r e n t

43 51

[XJ CONTENTS

Tantra as Middle: Medieval Esotericism 58 Historical Narrative and Ideological Implication 66

3. GOING NATIVE: TRADITIONAL HISTORIOGRAPHY OF TANTRIC BUDDHISM 68 Historiography and Cosmology in Exoteric Buddhism

71

Historiography and Cosmology in Esoteric Buddhism 79 Observations on Structure, Function, and Historiography 95

PART II: INTERPRETATION 4. THE SEMIOLOGY OF TRANSGRESSION

105

The Literal and the Figurative in Tantric Hermeneutics Connotative Semiotics as Exegetical M ethod

107

113

Connotative Semiotics in Tantric Ritual 117 Connotative Semiotics in Tantric Scripture

125

5. THE PRACTICE OF INDIAN TANTRIC BUDDHISM

133

Interpreting the Practice Observance I: Irony and Inversion

144

Interpreting th e Practice Observance II: Prerequisites and Temporal Frame 149 Interpreting the Practice Observance III: Śaiva Parallels

152

6. TANTRIC BUDDHIST TRANSGRESSION IN CONTEXT 170 The Social Location of Esoteric Buddhism as an Interpretative Problem 171

CONTENTS [XI}

Contriving Marginality

173

The Common Repertoire of Buddhist Professionals Carnivalesque or Rituals of Rebellion?

175

188

B u t. . . Did They Really Do It?! 192

CONCLUSION: NO TWO “WAYS” ABOUT IT

200

APPENDIX I: THE INDRABHÜTI STORY ACCORDING TO PAD MA DKAR PO (CA. 1575) 207

APPENDIX II: CHAPTER NINE OF THE BUDDHAKAPÄLA TANTRA, THE “PRACTICE” (CARYÄPATALA) 209

Notes 211 Bibliography 267 Index 301

Figure2.1

Definitions o f th e m edieval period in India 60

Figure 4.1

S tructure of th e linguistic sign according to F. de Saussure

Figure4.2

S tructure of m etalanguage 114

Figure 4.3

S tructure of connotative sem iotics

Figure 4.4

B arthes’ first exam ple

114

Figure 4.5

B arthes’ second exam ple

Figure 4.6

Suggested semiology of antinom ian discourse in

115

th e M ahäyoga Tantras

116 121

Table2.1

Conceptual binaries o f rh etoric o f T antra as beginning 56

Table 2.2

Conceptual binaries o f rhetoric o f th e medieval

Table 5.1

Central and related term s

Table 5.2

Sites of practice

137

139

Table 5.3

D ress/accoutrem ents for practice

Table 5.4

Behaviors prescribed in practice

142

Table 5.5

Behaviors proscribed in practice

143

Table 5.6

Q ualifications/prerequisites for practice

Table 5.7

D uration of practice

153

141

151

62

113

PREFACE A N D ACKNOWLEDGMENTS

he

T

I n t e r p r e t a t i o n o f I n d ia n T a n tric B u d d h ism p r e s e n te d in

th is b o o k h a s d e v e lo p e d o v e r m a n y y e a rs (ca. 1998-2010), a n d I h a v e b e n e f itte d e n o rm o u s ly fro m th e th in k in g a n d a s sis ta n c e

o f n u m e ro u s frie n d s, fam ily, a n d co llea g u es o v e r th is tim e . P a rts o f th e a r g u m e n t m a y b e tr a c e d b ac k to m y 1999 d o c to ra l d is s e r ta tio n , w h e n I firs t b e g a n to lo o k m o re c ritic a lly a t th e r h e to r ic s s tr u c tu r in g m o d e rn a c c o u n ts o f A sian h isto ry . I n s tr u m e n ta l in s h a p in g m y a p p ro a c h to th e s e q u e s tio n s w as th e in s p ira tio n a n d g u id a n c e o f T ed R iccard i, w h o u rg e d m e to re v is it th e w o rk o f H a y d e n W h ite a n d su g g e s te d t h a t I p a y clo se a t te n t io n to “h o w p e o p le w rite a b o u t India.” I h av e e n d e a v o re d to p u t t h a t ad v ic e in to p ra c tic e h e r e a n d h av e also ta k e n a r u n a t h o w In d ia n s h av e w r itte n a b o u t a n d fo r th e m se lv e s. In m a n y w ays, th e d isc u ssio n s I h a d w ith T ed a d v a n c e d lo n g s ta n d in g q u e s tio n s o f m e th o d a n d th e o r y f irs t p o se d to m e b y m y u n d e r g r a d u a te a d v iso r in relig io n , J. H. S to n e II. A lth o u g h P ro fe sso r S to n e d id n o t a c tu ally in tr o d u c e m e to th e w o rk o f R o lan d B a rth e s ( th a t w as a f o rtu ito u s r e s u lt o f a sale a t a N ew Y ork City b o o k sto re in th e e a rly 2000s), h e h a d p la n te d th e se e d lo n g b e fo re . T h e c h a lle n g in g a p p ro a c h to th e s tu d y o f re lig io n s t h a t h e ex e m p lifie d t h e n h as im p e lle d m e in th e d e c a d e s sin c e to follow u p o n th e m a n y sim ila r p r e g n a n t h in ts h e w as so g o o d a t d ro p p in g . W ritin g a n d th in k in g , in m y ex p e rie n c e , a re h ig h ly c o n te x t-d e p e n d e n t; th is b o o k c o u ld n o t h a v e b e e n w h a t it h a s b e c o m e w e re it n o t fo r th e c lim a te o f c o n s ta n t S tim u la tio n a n d c h a lle n g e o ffe re d a t th e U n iv e rsity o f C hicago. C olleagues b o th w ith in a n d o u ts id e th e D iv in ity S chool h av e se rv e d in g e n e ra l as e x a m p le s o f th e v e ry b e s t c ritic a l s c h o la rs h ip a n d h av e also o ffe re d c o n c re te e n c o u ra g e m e n t a n d c ritic is m as I v e n tu re d th e

various essays th at coalesced in this volume. Thanks in particular are due to Wendy Doniger, M atthew Kapstein, and Bruce Lincoln of the History of Religions; Divinity faculty Dan Arnold, Catherine Breckus, Clark Gilpin, M argaret Mitchell, Willemien Otten, Richard Rosengarten, Kathryn Tanner; South Asianists Muzaffar Alam, Yigal Bronner, Steven Collins, W hit­ ney Cox, and Gary Tubb; East Asianists Paul Copp and Jim Ketelaar; and Assyriologist Seth Richardson. Outside th e University of Chicago, I have benefitted immensely from discussions and assistance from many scholars of South Asian religions and civilizations, of whom particular m ention should be made: Ashok Aklujkar, Yael Bentor, David Drewes, David Gray, Janet Gyatso, Charlie Hallisey, Shaman Hatley, Roger Jackson, Karen Lang, Rob Linrothe, Dan M artin, Robert Mayer, Poornim a Paidipaty, P eter Skilling, Jan-U lrich Sobisch, Robert Thurm an, and Vesna Wallace. Allan D. Megill and Marina Bollinger offered m uch appreciated help w ith Paul Veyne. Many others have helped w ith obtaining the textual resources required for this type of study: Jim Nye, Beth Bidlack, and the staff of the interlibrary loan office at the University of Chicago libraries; the staff of the Newberry Library; and the National Archives of Nepal. The digital Sanskrit m anuscript Col­ lection of the University of Tokyo has been absolutely invaluable for my work. Leonard van der Kuijp was most generous in sharing w ith me both his own writings and copies of rare Tibetan historical m anuscripts. Vari­ ous research assistants have been invaluable over the years: Brad Aaron, Erin Burke, Amanda Huffer, Karin Meyers, and Susan Zakin. Likewise, my secretary Judy Lawrence has also been invaluable. Justin Henry was m ost kind in offering help with the Sinhala Nikayasamgrahawa. Larry Rosansky and Ted Arnold gave very kind feedback on the draft manuscript. Very special and heartfelt thanks are due to Ronald M. Davidson, who, for alm ost two decades now, has been a rem arkably generous, tolerant, and unfailingly learned interlocutor. Although we continue to disagree on num erous issues of m ethod and Interpretation, Ron has shown rem arkable goodwill in indulging my compulsion to follow out my intuitions, however contrary they may be to his own. His 2002 Indian Esoteric Bud­ dhism fundam entally reset the Standard for scholarly works on Tantric Buddhism, raising th e bar ra th e r dramatically. Had it not been for his work, I could barely have conceived of the projects th at took shape herein. All scholars working in this field owe him an immense debt of gratitude. The argum ents th at developed into chapter 2 were first crafted into a publishable form during a fruitful research stay at New College of Florida,

for which I thank Bob Schiffman, Darilyn Avery, and the late Roland Heiser. Chuck Matthews and the anonymous reviewers of the Journal ofthe Ameńcan Academy o f Religion challenged me to refine elem ents of w hat became chapter 4. Many universities and learned societies were kind enough to invite or allow me to present parts of these argum ents in their presence. Thanks are due in this regard to the American Academy of Religion, the American Oriental Society, the Columbia University Seminar on Buddhist Studies, th e Cornell University South Asia Series, the Harvard Univer­ sity Buddhist Studies Forum, the International Association of Buddhist Studies, th e University of Chicago Divinity School, and the University of Toronto D epartm ent of Religion. I am likewise indebted (fortunately, not literally) for the financial Sup­ po rt th a t I have received over the years from a variety of sources. Uni­ versity of Chicago Divinity School deans Richard A. Rosengarten and M argaret M. Mitchell have been unfailingly supportive of my work in a variety of ways, lavish in both th eir financial support and their intellectual critique. Likewise, the Committee on Southern Asian Studies provided ongoing research support and opportunities for travel to archives and institutions in South Asia. The final stages of writing were supported by a National Endowment for the Humanities Fellowship, although it is im portant to emphasize th at any views, findings, conclusions, or recomm endations expressed in this publication do not necessarily reflect those of th e National Endowment for the Humanities. Prelim inary versions of some chapters have appeared previously in various journals, and I am grateful both for this and for the chance to revisit them here. Elements of chapter 2 appeared as “Tropes, Typologies, and Turnarounds: A Brief Genealogy of the Historiography of Tantric Buddhism,” History ofReligions, vol. 40, no. 3 (February 2001), 223-259 (© 2001 by the University of Chicago; all rights reserved). The central argum ent of chapter 4 first saw publication as “Beef, Dog, and Other Mythologies: Connotative Semiotics in Mahäyoga Tantra Ritual and Scripture,’'Journal ofthe American Academy o f Religion, vol. 75, no. 2 (June 2007), 383-417. The cen­ tral argum ent in chapter 5 was first articulated in “Locating Tantric Antinom ianism: An Essay Toward an Intellectual History of the ‘P ractices/ Practice Observance’ (cary ā/caryãvrata),”Journal ofthe International Asso­ ciation o f Buddhist Studies, vol. 34 (2011). Thanks are due to the Univer­ sity of Chicago, the American Academy of Religion, and the International Association of Buddhist Studies for the opportunity to include them here. I am grateful as well to the editorial board of the South Asia Across the

Disciplines series for believing in this project and including it among the wonderful books in this series; and to Susan, John, and Eric H untington for providing the cover image. Finally, as always, my family deserves highest praise and eternal gratitude for th eir support: my parents, Anne Wedemeyer and Phillips W edem eyer; my step m o th ers, Josephine W edem eyer and Jean n e Hanchett; my nanny, M argarete Wiener; my sisters, Hope and Willow; and all my wonderful stepsisters and brothers, cousins, aunts, uncles, nieces, and nephews. The late Gen. Albert C. Wedemeyer took an early interest in my education and travels in Asia, and I belatedly acknowledge his kindness here. My m any friends and acquaintances—a kind of family them selves—bring w arm th and cheer to my world, making this kind of thing possible. Most of all, though, my deepest gratitude goes to my wife, Gita, and my two amazing girls, Maitreya and Isolde, for giving me three unsurpassable and compelling reasons to love, live, and strive. Christian K. Wedemeyer Chicago, Illinois

ABBREVIATIONS

AVP40:

Atharvaveda Pariśistā 40

BK:

Buddhakapāla Tantra

BY/PM:

Brahmayāmala/Picumata

CMP:

Caryāmelāpakapradīpa of Āryadeva

CMT:

Caņdamabarosaņa Tantra

CPAMA:

Catuspīthākhyãta-mantrāmśa

DN:

Dīghanikāya

DP:

Dhammapada

Grk.:

Greek

GS:

Guhyasiddhi of Padmavajra

GST:

Guhyasamāja Tantra

HA:

Herukābhidhãna

HT:

Hevajra Tantra

JS: KKP:

Jñānasiddhi o f Indrabhūti

,

Kālīkulapañcaśataka

KMT:

Kubjikāmatatantra

LS:

Laghusamvara

MK:

Mahãkāla Tantra

MMK:

Mañjuśrīmūlakalpa

MNS:

Mañjuśrīnāmasamgīti

MS(s):

m anuscript(s)

MVT:

Mahãvairocana Tantra

PU:

Pradīpoddyotana of Candraklrtī

Skt.:

Sanskrit

SU:

Samvarodaya Tantra

ST:

Sarņputa Tantra

STTS: SU: SYM: TD: Tib.: VĀ: YRM: YS:

Sarvatathãgatatattvasamgraha Samvarodaya Tantra Siddhayogeśvarīmata(tantra) Tattvadaśaka of Advayavajra Tibetan Vajräralli Tantra Yogaratnamãlã of Känha Yoginīsamcāra Tantra

M A K I N G SENSE OF T A N T R I C B U D D H I S M

INTRODUCTION M A K IN G S E N S E IN A N D OF T H E H U M A N S C IE N C E S

śvã-kharostra-gajãdy-asrk pītvā mãmsena bhojanam nityam || istam sarvaviśesa-rakta-vilipta-mahãmāmsam samasta-kutsita-māmsam prãnaka-śata-laksasamyuktam divyam \ vairocanenâtipütam kīta-śataih simisimāyamānam śvãnanara-cchardita-miśram māmsam vajrāmbu-marjikā-yuktam | vairocana-sammiśram bhoktavyam yoginotsāhaih \ Having drunk dog, donkey, camel, and elephant blood, one should regularly feed on

. theirflesh. Human flesh smeared with the blood o f all species ofanimals is beloved. Entirely vile meat füll ofmillions ofworms is divine. Meat rendered putrid by shit, seething with hundreds of maggots, mixed with dog and human vomit, with a coating ofpiss—mixed with shit it should be eaten by theyogin with gusto.1 —Samputa Tantra

presents to the historian of religions an interpretative conundrum . It is, to all appearances, a significant branch of a m ajor world religion—one whose scriptures and practices spread from India across Asia, capturing the minds, voices, and purses of millions in Nepal, Tibet, Bhutan, Mongolia, Sri Lanka, Bali, Cambodia, China, and Japan. More recently, these traditions have begun to take root in regions as far-flung as Russia, the United Kingdom, South Africa, and Brazil. On th e o th er hand, th eir widespread popularity seems virtually incom prehensible in light of the highly objectionable features of certain of th eir scriptures and rituals. The citation from the Samputa Tantra is a clear case in point: It is difficult to im agine a m ore disgusting and degraded vision of spiritual practice, yet this passage is found in one of

T

a n t r i c b u d d h is m

the m ost respected scriptures of the esoteric Buddhist Adamantine Way (Vajrayāna). What sense can one make of a religious tradition th at seems to advocate behaviors th a t most sane hum an beings would consider aberrant (at best), if not crim inal or pathological? This precise question was in fact raised over a Century ago as m odern studies o f Tantric Buddhism were ju st beginning. The em inent scholar of Sanskrit literature, Rajendralal Mitra, was understandably troubled by sim ilar Statem ents found in th e even more highly-revered Guhyasamäja (Esoteric Community) Tantra. Finding these “at once th e most revolting and horrible th at hum an depravity could think of,” Mitra em pathized with the many among his readers who m ight consider th at such teachings “would, doubtless, be best treated as th e ravings of madmen.” He cautioned, however, th a t following this particular interpretative avenue—as attractive and instinctual as it m ight at first appear—may be prem ature insofar as this same tex t “is reckoned to be the sacred scripture of millions of intel­ ligent hum an beings.”2 Although m uch im portant work has been done in th e field of Tantric studies since these words were first w ritten, scholars continue to struggle with and offer conflicting resolutions to w hat we m ight call “M itra’s quandary.”3 A ttem pts in this direction are com plicated by th e fact th at all would-be Interpreters of Tantric Buddhism in India m ust negotiate a fundam ental methodological challenge. Statem ents such as those quoted here from th e Samputa Tantra appear in texts th a t are historically disconnected from the cultural milieu th a t created and sustained them . Scholars of th e history of Tantric Buddhism, therefore, m ust them selves reconstruct a cultural context w ithin which to in terp ret th e discourses and practices found in its literature. W ithout an inform ing context, it is not at all clear w hat such scriptural Statements m ight mean. For instance, a stated injunction to (pardon my French) “eat shit” effectively means quite distinct things in different contexts. It m ight, for instance, be spoken by a governm ent operative engaged in to rtu rin g a prisoner, forcing feces into his m outh in an effort to dem ean and disgust him and thus break his spirit. On the oth er hand, it m ight be spoken by one starving castaway to another, helpfully, suggesting a desperate m eans to stave off hunger in the absence of other sustenance. In addition to these more literal—albeit quite different—usages, the same words m ight also be spoken m etaphorically (e.g., by one frustrated driver to another, expressing the anger and hatred o f “road rage”). Just as easily, it may be spoken by one friend to another in jest, in the course of playful m utual teasing.

When confronted with Tantric Buddhist scriptures th a t say “eat shit,” th e responsible in te rp re te r is likewise not im m ediately certain how those words are to be taken. It is necessary first to determ ine the cultural milieu in which these scriptures and th eir associated rituals were created and used. This, however, presents another challenge to scholarly m ethod. There is virtually no evidence beyond textual accounts to deter­ mine when, where, why, or by whom these discourses were used in late first-m illennium India. In order adequately to interpret these scriptures, the historical contexts of th eir articulation m ust be reconstructed. In so doing, however, scholars m ust rely prim arily on the corpus of scriptures themselves. Thus, as in all historical inquiry, the interpretation of Tantric Buddhism in India unavoidably involves a complex dialectic of content and context: The transgressive Statements of the Tantras are interpreted on th e basis of an imagined context, yet th at context is itself the result of prior interpretation of the Statements. To call it a dialectic is therefore to be ra th e r charitable, for such m ethods run great risk of com m itting the fallacy circulus in probando, or circular reasoning. Even though, ultimately, any solution will inevitably be subject to some such elem ent of circularity—bound as all interpreters are in the herm eneutic circle—there are, I believe, m ethods of analysis th at provide greater purchase on the nature of the significations th a t occur in this complex corpus of literature. In particular, I believe th at the m ethods offered by semiology have trem endous prom ise for studies in this area. This is so for two reasons. On th e one hand, semiological perspectives in general allow greater insight into the nature of discursive signs and the complex uses to which they are put in a variety of social contexts. Semiology, th at is, is devoted to —and thus excels at—dem onstrating how it is th at (as J. L. Austin put it) people “do things with words.”4 Furtherm ore, semiological analysis—insofar as it tends to focus attention on macroscopic patterns of signification across entire texts or larger corpora—allows a much more reliable and nuanced approach to interpretation than the more intuitive, microscopic avenue th at attem pts to generalize directly from interpretations of expressions taken individually. By com paring usages across texts and corpora and analyzing th eir relations, semiology allows a greater and more sensitive attention to the contexts of particular expressions than is possible with a less systematic approach.5 Before these m ethods may be fruitfully applied to th e project of advancing an alternative reconstruction of the m eanings and contexts of Indian Tantric Buddhism, however, a certain am ount of preparatory

work is required in order to clear the ground. In particular, critical atten ­ tion needs first to be paid to th e interpretative models th at have been advanced in previous scholarship, so th at a new approach may proceed free of encum bering assum ptions th a t have so strongly m arked earlier accounts. This is necessary insofar as there has developed over the course of two centuries of research a consistent, hegemonic, effectively autonomous, and self-sustaining network of scholarly discourses on Tantric Buddhism th at authorizes and reinforces certain ways of speaking about and otherw ise representing it. In this respect, m odern discourses about Buddhist Tantrism are similar to the phenom enon of “Orientalism” as discussed by Edward Said. In his work, Said was concerned to illustrate inter alia the rem arkable “internal consistency of Orientalism and its ideas about the O rie n t. . . despite or beyond any correspondence, or lack thereof, with a ‘real’ Orient.”6That is, once certain rhetorical tropes by which th e East may be represented had become familiar to European audiences, they developed a certain authority. They sounded plausible— all the more so because they drew upon and redeployed well-trodden and familiar them es in the European historical and cultural imagination. Likewise, th e discourses used by m odern scholars in reading the literature of and reconstructing the nature and history of Tantric Buddhism display much the same internal, tautological consistency. The same limited stock of ideas—often couched in nearly identical language—recurs decade after decade in new publications and is rarely subjected to critical analysis of its own genealogy. Of prim ary concern in this regard is not merely scholarly repetition or redundancy, but w hat might be called methodological solipsism: a kind of "em peror’s new clothes” effect. The discourses th at circulate in the sec­ ondary literature condition w hat people see in the prim ary sources. Schol­ ars read the secondary literature before they read the prim ary literature and, having thus been prepared by the discursive com m unity to see the em peror’s clothes, are thereafter predisposed—socialized—to see the data in light of those models and to replicate them in 'dieir writings. There is, as Said indicated, a “complex dialectic of reinforcem ent” by which prior discourses condition w hat we experience of the data presented to us, which interpretations then fu rth er reinforce our com m itm ent to those very discourses.7 Predictably, then, once the path was blazed, the road of academic representations of India and Indian Buddhism quickly devel­ oped significant discursive "ruts.” Just as ruts in real roads constrain the choices of drivers, discursive ruts confront authors with a choice: Either

follow th e ruts, and accept the sometimes problem atical interpretative ride they ofFer, or risk “breaking one’s axle” (in the form of incom prehension of the texts or, worse, loss of academic recognition and prestige) in seeking to travel outside of them . Chapters 1 and 2 are thus devoted to critically reassessing the historiography of Indian Tantric Buddhism, identifying and engaging key “ru ts” in these discursive practices. This focus on the models and intellectual structures th at inform the writing of history is necessary insofar as it is precisely via the discourses of history th at scholars have sought to resolve M itra’s quandary—to “make sense” of Tantric Buddhism. There recur thro u g h o u t th e scholarly literature stylized modes of discourse concerning th e origins and history of esoteric Buddhism. These various scholarly constructs—although cast in historical idioms—are essentially interpretative: In one way or another, they all function to account for th e transgressive discourses and practices found in the Tantras. All of the m ost common models for representing the history of Tantric Buddhism exist as attem pts to make sense of the radical, transgressive features of these traditions. They all do so, furtherm ore, by suggesting (implicitly or explicitly) a social context w ithin which these developments are thought to be com prehensible. That is, the rhetorics of the origins of Tantrism and narratives of its history are fundamentally driven by the need to provide an interpretative framework for cultural analysis and they do this by arguing indirectly for the social context w ithin which these practices are to have taken place. H istoriography of w hatever so rt—etiological or n arrativ e—always involves m aking sense of its object. As Marc Bloch has observed, “The nature of our intelligence is such th at it is stim ulated far less by the will to know th an by the will to understand.”8The production of m ere antiquarian data is not in itself a fully scholarly (or intellectually satisfying) project. Human beings seek to understand—to make sense of—their world. It is precisely the purpose of historiography to take a set of data and lend it m eaning and coherence. H istoriography is always an act of making sense insofar as it brings order and (consequently) m eaning to events.9As Roland Barthes writes: The historian is not so much a collector of facts as a collector and relater of signifiers; that is to say, he organizes them with the purpose of establishing a positive meaning and filling the vacuum of pure, meaningless series.10

It is impossible for a proper history not to interpret its subject—not to make it into som ething—to transform isolated events through an act of imaginative creation into a process that can be grasped by th e reader as a coherent object of understanding. This crucial act is accom plished through w hat historians call conceptualization—th e bringing of coherence to the historical archive by identifying an organizing principle or them e. This, of course, involves using rhetoric to juxtapose two or more things, with the aim of illum inating the subject m atter through com parison, employing, as Nietzsche put it, “a movable host of m etaphors, m etonym ies, [and] anthropom orphism s.”11 This is accom plished by th a t aspect of scholarly w riting th a t Ronald Inden has term ed the “com m entative”--th a t which, above and beyond the description of texts and their contexts, represents “its subject-m atter as something.”12 Thus, to w rite a history is always to interpret a cultural practice and, thus, always to transform th e subject, representing the unknown as or in the form of som ething known. In interpreting the Tantric traditions, m odern scholars have put forward a num ber of such conceptualizations. The form these reconstructions have generally taken is an attem pt to discern the origins a n d /o r to construct a narrative of Tantric Buddhism th a t situates it w ithin a historical process. All of the Solutions offered, however, have one cru­ cial feature in common: They all concur that, ultimately, transgression in Indian esoteric Buddhism does not in fact make sense. They do so by locating antinom ian practices and discourses (historically) in social con­ texts w herein they appear as expressions of either animal impulses, prim ­ itive m entality/superstition, or merely slavish im itation. That is, Tantric transgression is ascribed to irrational or arational impulses, for which fu rth er explanation is n eith er necessary nor possible. Such an adjudication thus serves to absolve scholars of the responsibility to confront the difficult challenges of cultural Interpretation: Once it has been made, interrogation ceases. M itra’s quandary is resolved by a half-measure—not the ravings of m adm en perhaps, but in the end those of the ignorant, the semi-civilized, or the m im ic/ape. Contrariwise, I believe it is both possible and necessary to make sense of the transgressive discourses of the Buddhist Tantras, and th at this can be accomplished by employing semiological methods. As m entioned earlier, semiology is uniquely well-placed to facilitate this advance, insofar as it is the Science dedicated to understanding linguistic signs in their social aspect;13 and, further, (in its structural form) such analysis is especially

attuned to the challenges of interpreting the use of signs across large Cor­ pora of docum ents (such as the Buddhist Tantras). Rather than interpret­ ing Statem ents considered separately—and thus confined to the level of th eir ostensive or im mediately m anifest content—structural analysis of patterns of usage discernible across a substantial corpus of literature may be of inestimable aid in reconstructing the meanings of particular signs w ithin th e larger system(s) of shared conventions w ithin which they become m eaningful. Using these m ethods in concert, the task of recon­ structing the cultural contexts of Tantric transgression may then proceed on a more secure footing. Roland Barthes, who contributed so m ightily to our understanding of rhetoric, its uses and connotations, has w ritten of structural analysis th at its “constant aim [is] to m aster the infinity of utterances [paroles] by describing th e ‘language’ [‘langue’] of which they are the products and from which they can be generated.”14That is to say, behind the diversity of individual utterances (paroles), there are discernible patterns of rhetoric th at are susceptible to analysis and th at allow us to get some purchase on the larger system of signification (langue) in which those utterances make sense.15 Once the general system is grasped, this then further informs our understanding of the complex significations th a t take place in each indi­ vidual instance of com m unication, allowing a richer analysis. Grammar is a simple example of this. By describing the patterns of linguistic usage, general rules may be abstracted. By identifying how those rules play out in individual expressions, a m ore complete analysis of th e specific instance becomes possible. Grammar is, of course, only the most basic of such Systems. Semiology takes as its object the larger and intersecting Systems of signification th at take place simultaneously w ith grammatical reference or simple denotation. In addition to the langue of language itself there are langues proper to various domains of cultural practice: types of discourses or networks of articulation th at become habituated over time which are susceptible to semiotical analysis. In a sense, the two parts of this book seek to apply such analysis to two distinct corpora of literature. The essays in part I interrogate the corpora of m odern and traditional discourses on the origins and history of Tantric Buddhism (qua paroles) in an attem pt to discern the larger contours, or langue, from which they have been generated and th at lend them coherence. They analyze the ways in which m odern scholars have attem pted to make sense of these traditions through speculations as to th eir origins and through narrating their history. This investigation will dem onstrate

what is really at stäke in such m odern historiography, by situating schol­ arship on Tantric Buddhism w ithin th e larger context of m odern discourses on cultural origins and history (chapters 1 and 2) and by further comparing these m odern practices with the historiographical practice of the Buddhist Tantric traditions themselves (chapter 3). Part II shifts attention to the semiological patterns of the ritual practices and scriptural discourses of Tantric Buddhism itself in order similarly to discern the subtending structure w ithin which these transgressive State­ m ents and practices themselves make sense. These chapters dem onstrate th at, through semiological and structural analysis, one can make sense of th a t aspect of th e Buddhist Tantras th a t has proven m ost intractable for m odern w riters—the antinom ian discourses (chapter 4) and practices (chapter 5) of th e Mahäyoga and Yoginī Tantras—ofFering new answers to th e question of how one makes sense of a religious tradition th at has seemed at least puzzling, if not dow nright nonsensical, prim itive, a n d / or irrational. Such an interpretative realignm ent invites and enables a fu rth er reassessm ent of the larger issue of the social location of these traditions (chapter 6). The double entendre of Making Sense of Tantric Buddhism thus reflects what might be called the book’s implicit methodological or (after Collingwood’s usage) philosophical project. An approach th a t “never simply thinks about an object,” but ‘‘always, while thinking about any object, thinks also about its own thought about th at object.”16The aim is to consider not m erely the object itself (Indian Tantric Buddhism), or solely the theoretical frameworks by which scholars come to understand it, but to grapple w ith these two in th eir m utual relation. By so doing, the goal is to make both a critical intervention into scholarly m ethod and a constructive contribution to the interpretation of its object. In a sense, th is m ethod of critical a tte n tio n to th e stru c tu re of rh eto ric—in term s of th e subject and th e object of scholarly research— rep resen ts a practical solution to th e challenge raised by some Con­ tem p o rary critics in th e hum an Sciences. That is, m any advocates of Contem porary critical theory, when confronted w ith systemic problem s of cultural in te rp re ta tio n —stru ctu rin g presuppositions, reading in of o n e’s own cultural expectations (what has been called isogesis17), Orientalism , and the like—recom m end as a way forward th a t scholars exercise a critical self-awareness of th eir own presuppositions. This can be ra th e r a challenging task, however. To extract and abstract one’s own subjectivity from the intentional state(s) in which it is em bedded is not

and has never been an easy task. Were it so, we would all be objective, or at least much m ore so. In place of the somewhat impractical and imprecise openness or critical self-consciousness often advocated,181 would suggest th at there is a clear, concrete step th at scholars can take to cultivate a more critical per­ spective: One th a t is implied in prior studies, but which has not (to my knowledge) itself been abstracted as such. Bound as we are in networks of rhetoric, we cannot easily inventory our own presuppositions and ideological blinders, casting them off and purifying our perspective.19We may, however, through structural reading of the discourses in which we participate, inventory our predecessors’ presuppositions and so come to have critical distance on our own through comparison. That is to say, by understanding the patterns of scholarly discourse (langue), out of which our individual research (paroles) is generated, we can b etter understand the histories we compose and the interpretations we advance. Structural semiological analysis is thus a critical tool for both poles of the interpretative dialectic. On the one hand, it allows one to discern the consistency and perpetual replication of m odern scholarly discourses on Tantric Buddhism and so to problem atize and denaturalize them , recognizing th eir distortions and aporia. On the other, these m ethods also may be used to dem onstrate a clear semiotical structure organizing the transgressive elem ents in the Buddhist Tantras. Working from this basis, fu rth er analysis opens up novel approaches to the vexed question of the actual practice of these transgressions. The results of these inquiries, in turn, suggest a fundam ental revision of certain m odern idées fixes concerning th e social contexts of Tantric Buddhism in India. Before proceeding, it may be helpful to say a word or two about the nom enclature th a t will be used in this book. As the term will be used in w hat follows, Tantric Buddhism com prehends those forms of esoteric Buddhism th a t are nondualist in th eir conceptualization of ritual purity and pollution and are, accordingly, antinom ian or transgressive in th eir ritual praxis and scriptural discourses. What, then, is esoteric Buddhism? This expression distinguishes those forms of Buddhism th a t require a spe­ cial initiation ritual to authorize th eir central practice (excluding those not so initiated from access to ritual and doctrine) and are thus esoteric: “Designed for, or appropriate to, an inner circle of advanced or privileged disciples; com m unicated to, or intelligible by, th e initiated exclusively.”20 The esoteric Buddhist traditions encom passed a variety of different subm ovem ents and doctrinal and ritual innovations w ithin (prim arily

Mahāyāna, or bodhisattva-oriented) Buddhism, beginning in the earlymid first millennium.21 W ithin esoteric Buddhism, there are discernible, clear divergences regarding at least two forms of dualism: th a t between the divine and the practitioner and th a t betw een the pure and the im pure. Some forms of esoteric Buddhism m aintain a doubly dualistic ritual and doctrinal platform in which, on the one hand, substantial attention is paid to the maintenance of ritual purity regarding conduct within and w ithout the rites, and, on the other, the practitioner relates to the divine as petitioner, interlocutor, or invoker, but is never identified w ith the divine itself. Other forms of esoteric Buddhism break down the dualism between the practi­ tioner and the divine—allowing the practitioner to adopt the identity of a divinity or divinities in ritual/m editative practice—yet nonetheless lay relatively marked stress upon the m aintenance of rules of ritual purity. Finally, o ther forms of esoteric Buddhism—those th at form the prim ary subject m atter of this book—are thoroughly nondualistic insofar as they ritually collapse th e distinction betw een the divine and the hum an and th at between the pure and the impure. The sacred scriptures of these latte r traditions come to be called the Mahäyoga (“Great Yoga”) or Yoginī (“Female Yoga P ractitioner”) Tantras; and (as is evident in our epigram from the Samputa Tantra) advocate the deliberate transgression not only of the entire gam ut of purity strictures found in the dualistic traditions of esoteric Buddhism, but of all the most central dictates of Buddhism in its entirety. In thus inciting violation of established rules (Greek, nomoi), these traditions are accordingly described as antinom ian. Though other forms of esoteric Buddhism also feature texts called Tantras among their scriptures, because Contemporary usage has already conspired to stress th e transgressive aspects of these traditions, they will herein be designated Tantric Buddhism. Over th e course of six chapters, this book will address five major topics on which scholars have advanced divergent views concerning Tantric Buddhism: its origins, history, textual interprecation, religious practice, and social context(s). P art I, (Historiography) analyzes the discourses used to reco n stru ct th e origins and history of Tantric Buddhism, in m odernity and in traditional India. Chapter 1 argues th at there has been a consistent and overweening concern in m odern scholarship with locating th e origins of Buddhist Tantrism , and th a t this putative originary m om ent has th en been m ade to serve as th e key to subsequent interpretatio n of these traditions and th eir transgressive discourses. Three

variant origin tales are abstracted: That Tantric Buddhism is the product of a degeneration of Buddhist morality, th a t it preserves prim ordial or prim itive religious practices, or th at it is entirely borrow ed from contem poraneous Śaivism. Each of these models will be tested against the available evidence and th e inadequacy of each account will be dem onstrated. More tellingly, it will be argued th at all are essentially attem pts to explain transgression by constructing a social context w ithin which antinom ian Statem ents make sense. Etiological historiography th a t seeks explanatory purchase through appeal to origins, however, is not the only m ethod of historical explanation used in this regard. Narrative historiography is another mode of historical explanation, and we tu rn to these discourses in chapter 2. This chapter addresses Tantric Buddhism in its historical development, taking a m etahistorical perspective on m odern historiography, its rhetoric, and its ideology. Drawing on insights of Hayden White on th e role of stock story forms in th e construction of historical accounts, this chapter analyzes th e th ree predom inant modes of narrating the history of Tantric Buddhism—as the conclusion of Indian Buddhism, as a prim ordial beginning of Indian religions, and as a medieval m idpoint in a larger histori­ cal process. All such narratives are, I suggest, underdeterm ined by the historical data they serve to organize: They are fictive, contingent, and inescapably ideological products of the intellectual practice of m odern scholarship th a t appeal to commonly available models from the Euro­ pean literary im agination to offer an interpretative context for Tantric transgression. The structure of these rhetorics will be analyzed, along with th eir relationship to the modes of ideological implication th at they encode: th at is, th eir implicit “prescriptions for taking a position in the present world of social praxis and acting upon it.”22 C hapter 3 takes th e historiographical discussion further, reorienting th e critique by analyzing the historical models used by indigenous authors in interpreting the history of their own traditions. The historical discourses of Tantric Buddhists will be shown to be m uch m ore complex th an typically understood, encompassing a num ber of stylized historical motifs and distinctive narratives far beyond a simple assertion th at the Buddha taught them . These various indigenous historiographical forms are likewise given a m etahistorical analysis, situating them in larger pattern s of discursive practice w ithin Buddhism and th e broader context of contem poraneous Indian culture. Such a perspective dem onstrates that, like m odern historiography, traditional historiography of esoteric

Buddhism aims to negotiate interpretative quandaries through historical explanation. Part II advances over the course of three chapters a new interpretation of Indian Tantric Buddhism and its transgressive antinom ianism, bringing novel modes of analysis and fresh philological spade work to bear. Chap­ ter 4 offers a reading of the antinom ian discourses of Tantric Buddhist ritual and scripture, dem onstrating th at these are neither fully literal nor truly figurative, as has previously been claimed. Applying m ethods outlined by Roland Barthes to a com prehensive analysis of the Guhyasamäja Tantra, I argue th at the transgressions of Tantric ritual and scripture constitute a form of connotative semiotics. In this mode of comm unication, a complete sign in natural language (in this case, the signifier-signified complex of transgression and ritual pollution) functions as a signifier in a higher order cultural discourse. At this higher, perform ative level, the signified is the attainm ent of the goal of advanced Tantric practice: non­ dual gnosis (advayajñãna). This approach provides a new lens for appreciating the sophisticated Systems of signification functioning in esoteric scripture and ritual, and a rath er different framework for representing Tantric antinomianism. Chapter 5 advances the semiotical discussion through another mode of textual criticism: a quantitative, structural analysis of the contexts of transgression across a broad ränge of antinom ian Buddhist literature. This chapter indexes and analyzes the dress and accoutrem ents, prescribed and proscribed behaviors, sites, and prerequisites of a central rit­ ual observance as presented in over tw enty Tantric scriptures, Buddhist and non-Buddhist. The results dem onstrate a remarkable fact about the term caryä (practice) in Tantric literature, in which context the vast proportion of the antinom ianism of the Tantric scriptures occurs. Far from representing esoteric practice broadly construed—as it has uniform ly been taken by m odern interpreters—structural analysis reveals th at caryä is in fact a term of art in Tantric Buddhism, referring to a very specific and highly m anaged and contextualized ritual observance (vrata), a rarified discipline reserved for a sacerdotal elite. Chapter 6 broadens the issue of interpretation and practice further, raising the question of the social context(s) within which Tantric Buddhism flourished. This chapter argues that, although the rhetorical divergence of dualist (nontransgressive, “institutional”) and nondualist (transgres­ sive, “siddha”) Tantric traditions has been explained by postulating two independent communities, there is no good reason to infer a sociological

cleavage based on a rhetorical one. In fact, the available evidence clearly suggests to the contrary th at it was precisely the cultural milieu of the conservative m onastic institutions from which these radical movements emerged. That is, the transgressive practices were by and large the m étier not of marginal figures, but of professional Buddhists w ith m ainstream concerns, educations, and institutional locations. The social context and function of these transgressions require, therefore, to be reinterpreted, for which th e work of Max Gluckman on rituals of rebellion suggests a suitable framework. In conclusion, it will be recalled th a t—although the shift in interpre­ tation suggested herein runs against the grain of most m odern scholar­ ship on Tantric Buddhism—a very similar reorientation has, in fact, also taken place in Contemporary scholarship on the exoteric Buddhist tra ­ ditions of th e Mahāyāna. These, too, were originally imagined as sepa­ rate, breakaway com m unities of lay people outside the hegemony of the m onastic institutions. New, more critical studies, however, suggest that the radical new Mahāyāna teachings similarly developed in conservative Buddhist centers. Five points of com parison between these two models will be articulated and a revised appraisal of Tantric Buddhism suggested accordingly.

HISTORIOGRAPH Y

[1] ORIGINS, RELIGION, AND THE ORIGINS OF TANTRISM

Science deals with relations, not with origins and essences. —E. E. Evans-Pritchard

The mode in which the genesis o f a thing is explained is the candid expression o f opinion, o f sentiment respecting it. —L. Feuerbach

Les origines sont rarement helles. —Paul Veyne

N S e e k i n g to make sense of Tantric Buddhism, scholars have often looked to its origins as a way of explaining its most basic nature and causes, identifying it, and thus accounting for it. This m ethod has been common throughout hum anistic and historical studies of culture, and the study of Buddhism is no exception. Like other scholarly discourses, this one has its own coherency and structure in which a lim ited num ber of rhetorical modes recur consistently. In fact, the various accounts of the origination of Tantric Buddhism comprise a highly delim ited set of possible representations. This chapter will explore the rhetoric of the origins of Tantrism and its m odern history, observing its basic modes and its ränge of variants. Having surveyed th e discursive terrain, the discussion will then tu rn to the cogency of these various models taken on their own term s, and argue th at each of them contains fatal flaws of logic and evidence. Particular a tten tio n will be paid to two currently populär accounts th a t ascribe the origins of Tantric Buddhism to eith er tribal religions or Śaivism.

I

The discussion will conclude w ith a m ore fundam ental critique of th e search for origins as a m ethod in the hum an Sciences, and of the approach to historical interpretation of cultural forms th at it subtends and enables.

UND ERSTAN DING TANTRIC BUDDHISM TH RO U G H ITS ORIGINS Readers of historical literature will be quite familiar w ith discussions of origins as a mode of locating and interpreting figures or movements. This approach has, in fact, a very respectable pedigree in the hum an Sciences. No less a thinker th an Emile Dürkheim stressed the need to trace one’s scholarly subjects from th eir origins and only then through the course of th eir subsequent existence. Historical objects, Dürkheim argued, are shaped by and irrevocably linked with the circumstances of their birth: Every time that we undertake to explain something human, taken at a given moment in history—be it a religious belief, a moral precept, a legal principle, an aesthetic style or an economic system—it is necessary to commence by going back to its most primitive and simple form, to try to account for the characteristics by which it was marked at that time, and then to show how it developed and became complicated little by little.1 In the study of Tantric Buddhism, this m ethod has been extrem ely populär. T hroughout th e scholarly literature, one finds chapters and subchapters w ith titles such as “Tantric Buddhism: Its Characteristics and Origins”2 (L. de La Vallée Poussin 1898), “Origin of Buddhist Magic: Rise of Vajrayāna” (B. Bhattacharyya 1931), “The Genesis of Vajrayäna”3 (H. von Glasenapp, 1936), “Origin and Development of Tantric Buddhism” (S. Dasgupta 1946), “Origins” (D. Snellgrove 1959), “The Seventh-Century Beginning” (R. Davidson 2002), and so on. An entire volume on The Origins ofYoga and Tantra (G. Samuel) appeared as recently as 2008. Even among those authors who do not explicitly use the language of origins, recourse to this mode of explanation is frequent. How th en have we m oderns discussed the origins of Tantric Buddhism? There seem to be three prim ary modalities.4According to some, Buddhist Tantrism em erged as an outlet for transgressive or degenerate impulses by monks. Others discern its roots deep in th e religious prim ordium of India. Yet others refer the rise of Tantric Buddhism to a Wholesale borrowing from the traditions of its Śaiva com patriots. In a certain sense,

each of these im puted origins is keyed to a particular narrative account of th e history of esoteric Buddhism. That issue will be set aside for the moment, as it will form a part of the discussion in chapter 2. For now, let us get a sense of the variety of rhetoric th at has been employed in discussing the origins of Tantric Buddhism and the interpretative work that these conceptions perform. One of the original causes to which scholars have turned in accounting for the rise of Buddhist Tantrism is a lack of moral rigor in the flourishing Buddhist m onasteries of the first millennium. One of the most influential advocates of this etiology is Benoytosh Bhattacharyya. This influential Bengali scholar rightly observes th at “it is very doubtful w hether we will ever be in a position to trace the origin of the Tantra in the m ost precise m anner possible.”5 He nonetheless feit th at an adequate account of its principal causes could be produced. In several of his published works, B hattacharyya attrib u tes authorship of the Tantric m ovem ent to the influence of degenerate monks and a need to accommodate these tendencies. Observing th at sources relate th at many monks either left on their own or were expelled from the Buddhist m onastic com m unity because they were unable to practice the strict morality required of them by the Discipline (Vinaya), Bhattacharyya speculates that there were many others who were not bold enough to proclaim a war against the rules imposed on them, but violated them in secret. It is thus very natural to expect that there arose secret conclaves of Buddhists who, though professing to be monks, violated all the rules of morality and secretly practiced things that were considered by others to be revolting. After the death of the Buddha, such secret conclaves must have grown in number in every province, until they formed into a big Organization. If we add to this the yoga practices and the practice of mantras, we get a picture of the Täntrika cult at its early stage.6 Here, th e inspiration for the developm ent of th e Tantric traditions is attributed to a moral turpitude alleged to have been widespread even in the tim e of th e Buddha himself, to which period Bhattacharyya locates th e earliest esoteric com m unities. There are some unexplained gaps in this account. For one, the argum ent is largely based upon speculative assum ptions (“it is very natural to e x p e c t. . . ”), which premises are implicit ra th e r th an acknowledged. Furtherm ore, while one can certainly understand th at monks may desire pleasures prohibited them , it is more

difficult to understand why they would necessarily gravitate tow ard the revolting. (Presumably, Bhattacharyya has in m ind here the ritual consum ption of polluting substances, w ith which this book began—a topic th at will be discussed in detail in chapter 4.) The rise of Tantric Buddhist literature is described in sim ilar term s in B hattacharyya’s works. “Those monks who saw salvation only in leading a natural life w ent on devising plans and probably by w riting w hat we call the original Tantras which were secretly handed down through th eir tru sted disciples who could practice th e rites only in secret.”7 The notion invoked h ere—of a “natural life”—is central to this hypothesis. For B hattacharyya, the m onastic regulations were com posed largely of “unnatural rules of discipline,”8which he also calls “u nnatural and strict rules.”9 Here, “natural life” m eans enjoying pleasures,10 a habit th a t is attrib u ted to th e Buddha him self who, th e reader is told, “took food and n ourishm ent in a natural way.”11 As th e m onastic discipline asked Buddhists to go against th e ir m ost basic hum an nature, it could only be expected th a t th ere would arise an im pulse to circum vent the rules th a t fru strated th e realization of such a “n atu ral life.” This im pulse, B hat­ tacharyya insisted, found expression in th e Tantric scriptures, w herein “everyw here any casual reader can detect a desire on the p art of th e au th o rs to th w art all un n atu ral rules and regulation forcibly chained on to th e followers of Buddhism.”12 In short, Bhattacharyya m aintained th a t th e Tantras were com posed in o rd er th a t u n reg en erate m onks m ight enjoy the pleasures of life w ith Buddha’s im prim atur, thus giving rise to a long trad itio n of Buddhist th o u g h t and practice dedicated to th e realization of this goal.13 A nother influential account of the origins of Tantric Buddhism m ain­ tained th at the Tantras could be traced back to the most rem ote antiquity in India. Tantrism , th at is, has no discernible origins per se but ra th e r represents the oldest indigenous religious tradition of the Indian subcontinent. These traditions are here regarded as “pre-Äryan”—belonging to th e culture of India th at preceded the alleged advent of Central Asian im m igrants who brought the Vedic revelations southeast into new territories. The great art historian Stella Kramrisch wrote in 1929 th at Śāktism (the worship of the female element, often considered the sine qua non of “T antra”) had “its roots in the m ost rem ote antiquity.”14 Indologist E. J. Thomas made much the same claim in 1933, com m enting th at “Tantrism as a form of religion is of unknown origin, and may possibly have arisen am ong some indigenous and non-Aryan people.”15 This view too has

continued to have its proponents. In 1962, R. 0. Meisezahl w rote th at “the Tantra, w h eth er Hindu or B u d d h ist. . . consists essentially of religious methods and practices which were current in India from times immemorial.”16The French Indologist Andre Bareau similarly claimed in 1966 th at “the origins of the Tantric movement go back rath er far in tim e and seem allied w ith ancient magical and religious beliefs th at rem ain as alive in India as elsewhere.”17 Sinologist Robert van Gulik, in a m onograph on the esoteric Buddhist divinity Hayagriva, provides a more thorough elaboration of this idea: The roots of this curious System may be traced back to very old, probably even pre-Indo-Aryan days. The belief in the power of the magic formulae... seems to be particularly rooted in the propensity towards magic existing among the ancient aboriginal tribes of India. Many of these ancient conceptions were adopted by the Indo-Aryan conquerors and made an integrant part of their own conceptions. Van Gulik’s account incorporates a related but distinct claim th a t characterizes m uch of Contemporary thought about the origin of the Tantras. This is th e notion th a t—even after the ascendency of the Vedic cultural model—prim ordial, pre-Äryan religious currents continued to be practiced am ong “m arginal” com m unities th a t included prim itive groups such as India’s so-called “tribal” population. Van Gulik continues: In different parts of India, however, situated outside the centra of IndoAryan culture, where the aboriginal population was better able to preserve its own character, the native usages of magic and witchcraft maintained themselves in a form more closely resembling the pristine.18 Vedic culture, it is claimed, adopted some of th e magical rites of the aboriginal population of India in the ancient period in th e form of the Atharvavedic rituals. The original rites, however, lived on pristine among prim itive aboriginal com m unities, from which they were subsequently incorporated into the later Tantric literature. This same link betw een pri­ m ordial Indian religion and the timeless tribal cultures is m ade by Cintaharan Chakravarty in his article “The Antiquity of Tantrism,” although he characterizes the ancient inheritance differently. Rather th an merely magic, for Chakravarty Tantra represents th e “sex-magic” of prim itive peoples th a t “seem[s] to have come down from prim itive tim es and [is]

known to be prevalent even in the present days am ong people w ith a primitive culture ,”19 This “aboriginal/tribal” theory of origins continued to flourish in the late tw entieth Century. Miranda Shaw, in her 1994 book Passionate Enlightenmenty claims th a t “practices th a t had great antiquity in India’s forests, m ountains, and rural areas, among tribal peoples, villagers, and the lower classes, were em braced and redirected to Buddhist ends.”20Geoffrey Sam­ uel too has claimed th a t “much of both Tantric vocabulary and Tantric techniques . . . seems to have derived from ‘tribal’ or folk shamans.”21This view has continued to attract advocates into the tw enty-first Century, in which Ronald Davidson has argued for the tribal origins of m any Tantric Buddhist practices and divinities.22 Others argue th at Buddhist Tantrism came entirely by way of borrowing from contem poraneous Śaiva literature and practices. Occasionally, this appears as a variant on the “pre-Āryan” them e, since Śiva has frequently (if very tenuously) been associated with the ithyphallic, horned god on the famous Harappan seal 420.23 However, the theory th at the ori­ gins of Tantra are specifically Śaiva and th a t it was from these traditions th at some Buddhists ultim ately adopted it, has its own rhetorical density as a theory of origins and deserves separate treatm ent. The association, even equation, of Tantrism and Śaivism appears quite early on in the study of Buddhism. Indeed, E. Burnouf him self thought it “likely” th a t the Buddhist Tantras were propagated by “Buddhists who, while entirely preserving th eir beliefs and th eir philosophy, consent to practice certain Śaiva rites th at promise them success in this world.”24 As is frequently the case in B urnouf’s work, he is following the lead of a Brit­ ish researcher based in South Asia. In this context, he echoes the views of Horace Hayman Wilson, who had w ritten th at in the Buddhist Tantras of Nepal “the worship of s i v a , and Tantra rites, are . . . widely blended with the practices and notions of the Bauddhists.”25 This conception of the origins of Tantric Buddhism has likewise had a long career and has recently experienced a vigorous resurgence. In 1911, Louis de La Vallée Poussin categorically claimed th at “Buddhist täntrism is practically B uddhist Hinduism , Hinduism or Śaivism in B uddhist garb.”26 In 1987, David Snellgrove replicated exactly Burnouf’s model of two kinds of Buddhist Tantra: those th at share more features in common w ith the Mahāyāna sūtras and those "with Non-Buddhist Associations,”27 by which he means Śaiva associations. More recently, a specialist in Śaiva

traditions, Alexis Sanderson, has made an expansive case, arguing th at the rise of Tantric Buddhism “had been achieved by absorbing and adapting non-Buddhist practices,” specifically Śaiva practices.28 Thus, over the course of decades of research on Tantric Buddhism, the best scholars of several generations have provided a num ber of competing (and sometimes intersecting) accounts by which they claim to locate the origins of these traditions. Nearly all theories offered, however, replicate one or th e o th er of a rath er limited stock of origin concepts. One thing all of these scholars have in common, however, is the presupposition th at it is in fact im portant th a t historical studies locate the origins of religious traditions, insofar as doing so is thought to provide a privileged perspec­ tive on th e trad itio n ’s m ost essential nature, providing a key by which the m eaning of its discourses and practices may be best interpreted. The conclusion of this chapter will be devoted to evaluating the cogency of this shared presupposition. Before turning to th e larger interpretative context o f these claims, however, it may be of some use first to evaluate these models on th eir m erits. Bracketing for the m om ent the larger explanatory argum ents of which they provide key propositions, or the more subtle question of w hether or not it is in fact explanatorily useful or im portant to identify th e origins of religious traditions, it is w orth considering at least briefly w hether or not the scholarly accounts commonly advanced are actually supported by the available evidence. The notion th at Tantric Buddhism was created to accommodate degenerate tendencies in th e Buddhist m onastic community, whose members sought pleasures of the flesh prohibited them by the disciplinary rules of the order, is perhaps the easiest to dismiss. It is, after all, an example of the clum siest sort of origin tale: a “ju st so” story. Much like the account of how th e elephant got his nose (or the tiger his stripes) or other such folk etiologies, there is no real evidence for this view at all. It is entirely speculative, based upon a series of assumptions about men, th eir desires, and th eir behavior.29Indeed, insofar as some have sought to found it more solidly upon literary evidence, this is limited to farcical, bawdy representations in Sanskrit satires, such as the Mattaviläsaprahasana, or tendentious allegories, such as the Prabodhacandrodaya.30 W hatever the desires of first-m illennium Indian Buddhist monks may have been, th e cogency of this account also hangs critically on the ques­ tion of w hether Tantric practice was, in fact, prim arily pleasurable. Alex Wayman questioned this basic presum ption when he noted th at

to be practical, it is passing stränge that anyone would bother with the Tantra to justify his “degenerate” practice, for who so bent among worldly persons would divert his energies by muttering a mantra a hundred thousand times at dawn, noon, sunset, and midnight, with fasting and other inhibitions, to engage in a “degenerate” practice, when, as we know so well, people at large engage in degenerate practices without bothering to mortify themselves at dawn, noon, sunset, and midnight!31 Wayman suspects th at the contrived sexuality of Tantric ritual would not have been much fun—or, at least, not so much th at it would be w orth executing all the elaborate rites accompanying it. Of course, both Waym an’s view and the one he rejects are entirely speculative, based merely upon the au thors’ intuitions as to w hat would or would not have been pleasurable to a late first-m illennium Indian Buddhist. W hether or not it was due to an awareness of this fact, this model has largely been abandoned. Yet som ething of its influence continues to operate at a more subtle level. In a widely praised 2002 work, the composition of Tantric scriptures is attributed to a desire by Buddhist monks to have scriptural w arrant for their lusts for “drinking wine and making love to nubile women.”32 While this account is quite easily discounted, th e o th er two models have been more durable and thus call for more detailed treatm ent. There do remain a few retrogradë w riters who continue to m aintain that, for instance, the presence of large-breasted female terraco tta figurines in th e prehistoric sites of the Harappan civilization are evidence of the antiquity of Śāktism and Śākta tendencies in pre-Äryan Indian society; however, virtually no serious scholar of today would advocate this view.33 The notions of Tantrism as a prim ordial tradition of m atriarchal m other worship, magic, and sex rites—or as representing a perennial “pre-Äryan” religion of India—are largely defunct.34 On the other hand, more historically and anthropologically saw y w riters continue to m aintain th at these rites and traditions m ust have come from similarly primitive tribal comm unities in India. This view consequently continues to attract considerable attention and assent in Contemporary scholarly accounts. To ascribe the origins of Tantrism to tribal groups, however, is no less problem atic th an the o th er variants on this them e; and it is clearly a variant, not an independent, account, for th e operative concept in this interp retatio n is similarly th e notion th a t it m ust be simple, prim itive societies, in which “magic” held sway, from which the Tantras derive.35 Underpinning this view is th e notion th at India is host to tribal groups

th at have continued to practice prim itive religious traditions, alongside but independent of th e Indian m ainstream . The anthropologist C. von Fürer-Haimendorf, for instance, has w ritten th at these tribes “persist in an economic and social Organization th at elsewhere feil into desuetude at the end of the neolithic era.”36 Thus, via the medium of India’s tribes, the primordial, prim itive rituals made their way into the religious m ain­ stream in the form o f the Tantric traditions. Geoffrey Samuel and Ronald Davidson—am ong the m ore influential scholars of esoteric Buddhism in recent decades—m aintain variants of this view. However, very little at all is clear concerning any communities th a t m ight be called “trib al” in the late first-m illennium . There are no prim ary source docum ents or m onum ents upon which scholars might draw in order to understand these cultures in their own right. The only sources available are inscriptions and literary works of m ainstream Indian society th a t make reference to them , and the latter do not give uniform or detailed accounts of the religions of these cultures. In general, these sources merely make derogatory remarks th at highlight their alterity, m uch like British accounts of Indians in the nineteenth Century, who similarly considered th eir subjects “backward.” The very notion of w hat m ight constitute a “trib e” is itself quite vague in m odern scholarship on the Tantras. Samuel him self notes th at “it is difficult to know quite what ‘tribaF m ight m ean at this period,” since “‘tribaP in the m odern sense . . . is the product of a long-term relationship between the populations which we now label tribal and those which we now term caste Hindu” who “may well have separated out from an initially more uniform population.”37 Thus, while there seems to be some good evidence th at there did exist com m unities in the late first millennium with distinctive nam es—Doms, Gonds, Śibis, Nisädas, and th e like—th at were considered “o th er” than the Brahmanical com m unities of the time, very little is known of th eir cultures, how they were distinctive, or even if they were truly distinc­ tive. These tribal peoples, fiirtherm ore, were not enclaves of autochthonous peoples living in a prim itive state of social organization. Sources indicate th a t at least some of these groups m igrated w ithin India, such th a t th ese “autochthonous trib es” were often centered far from th eir alleged hom elands.38 In term s of social form ation, insofar as we know from inscriptional sources, m any tribal com m unities were well-established States th a t carried on high-level cultural and diplom atic intercourse w ith neighboring Brahmanical polities.39 Nor is it clear, as it has been claimed (if not assumed), th a t their religious practices centered on

devotion to local nature deities (such as were allegedly later incorporated as Tantric divinities). In the ninth Century, for instance, a “tribal chief” Pulindaräja (king of the Pulinda “trib e”) “prevailed upon a Bhaumakara ruler in Orissa to grant land for the m aintenance of a Śaiva tem ple and Śaiva ascetics.”40 In fact, it has been observed th at in the Tantric period many of these “tribes” were “champions of neo-brähm anism ” or Hinduism.41The Nisädas, a famous tribal group considered to be “m arginal” and a subject of scorn in Brahmanical sources—and thus precisely the sort of com m unity to which the source of Tantrism is ascribed—are known from early first-m illennium literature to be a settled people, who engaged in unobjectionable rites such as offering caru (a dairy offering com m on in m ainstream Hindu ceremonies,) and even perform ed orthodox Vedic fire offerings.42 Furtherm ore, w hat very little evidence we have for the direction of transm ission of Tantric culture suggests unm istakably th at th e tribal or marginal communities were in fact the targets of Tantric transmission, not the source. Alexis Sanderson, for instance, has drawn attention to a passage in the Buddhist Guhyasiddhi th at advocates Buddhist Tantric yogins traveling among untouchable com m unities, giving initiation and teaching Tantric scriptures to them .43 Likewise, where there is clear evidence of local (and thus marginal, if not necessarily tr ib a l) deities in Tantric traditions—far from constituting the m ain focus of these cults, bringing w ith them well-formed tribal ritual program s—they are found instead incorporated into already well-established, pan-Indian esoteric Systems such as Buddhism and Śaivism. For instance, certain Kashmiri Śaiva rit­ ual manuals (paddhati), based on pan-Indian Ägamic sources such as the Netra and Svacchanda Tantrasf were at a very late date “elaborated through th e insertion of the worship of num erous subsidiary deities drawn from various sources, some of them local goddesses . . . and others drawn from m ainstream traditions.”44 More im portantly—and this point is central to the overarching thesis of this book—th e critique of the theory of the tribal origins of Tantrism does not stand or fall on w hether or not such com m unities existed or w hat they were like (although I think th e evidence to this effect rath er seriously weakens, if not destroys, the case).45Rather, th e notion th at Tan­ trism derives from tribal traditions is largely the result of a widespread tendency by m odern scholars to read the rhetoric of the Tantras hyperliterally. As a consequence of this interpretative choice, the occurrence of (allegedly) tribal term s is taken as evidence of tribal provenance and,

thus, of the social location, origins, and interpretative context proper to Buddhist Tantrism. A clear example of the general trend may be seen in the logic employed by the historian R. S. Sharma, in making the case for the tribal origins of Tantrism: [The goddess] Śakti is known as Mātańgī, which shows that originally the goddess belonged to the Mātańga tribe. She is also called Candālī, which indicates that she was a goddess of the Candälas. According to the Kulämava Tantra, a candãlī, carmakãrī, mãgadhī, pukkasī, śvapacī, khattakī, kaivartU vaiśyayositah, śastrajīvinī, kauñcikī (or kandukī), Śaundikī, rañjakī, gāyakī, rajakī, śilpinī, or kaulikī. . . is to be worshipped as Śakti. . . . Obviously the list contains the names of mostly śūdra and untouchable women, generally of tribal origin.46 One can very clearly discern the theoretical assumptions th at underlie this interpretation. The first two claims are both based upon the premise th at w hen well-known tribal ethnonym s are employed as names of divinities, this necessarily implies th at the divinity in question had its ori­ gins in th a t community. However, this interpretation is based on a fun­ damental m isunderstanding of the semiological system at play in the use of these term s, taking them as instances of prim ary reference or denotation. In an o th er example of the same error, Sharma asserts th at the Matańgapãrameśvara Tantra (an influential scripture of the Śaiva Siddhānta) “was evidently composed to serve the needs of the [tribal] Mātańgas living in eastern Madhya Pradesh and Andhra.”47 However, even the most casual perusal of this latter work will reveal that it takes its name from the legend that it was taught by Śrīkantha [Śiva] to the sage (muni) Matańga.48There is nothing in the description of this sage to suggest that he is of tribal ancestry, or th at his background is anything but m ainstream Brahmanical. He is described as outstanding in his gnosis (jñãna) and m editation (samädhi)— key term s in Brahmanical religious discourse—such th at he is described as a “preem inent lion of sages” (munīnām śãrdülah simhah).49 In this regard, comparison m ight be made to the “Śabara Com mentary” (Śabarabhãsya) on the famous Mīmãmsāsūtra of Jaimini. No one in their right mind would ever speculate th at it was w ritten for (tribal) śabaras or th at its author Śabarasvamin was a tribal. Indeed, a ritually im pure śabara would never have been allowed near the hyper-Brahmanical Mīmãmsãsütra, Yet, due to an idée fixe th at has become established from early studies of Buddhism, this logic is deemed compelling in the Tantric case.

The reference to Śakti as M ātańgī and Candālī—both of which are untouchable caste nam es—is clearly an outgrow th of th e semiological System th a t inform s th e list given in the latter p art of Sharm a’s argument. Contrary to Sharma’s view, these women are not at all “generally of tribal origin.” In fact, of the sixteen types of women advocated as śakti in this passage from the Kulämava, there is only one th at is arguably an ethnonym: Māgadhī (“woman from Magadha”). Of th e others, twelve of them are occupations despised within the Brahmanical System: leatherw orker (carmakäri), two types of dyer (pukkasī, rañjakī), butcher (kattiki), fisherwoman (kaivartī), m ercenary (śastrajīvinī), barber (kanduki), liquor-dealer (śaundikī), singer (gāyakī), washerwom an (rajaki), craftswom an (śilpinī), and weaver (kauliki). The three rem aining are caste names: one a cliché for the lowest of untouchables (candālī), another a stylized, derogatory term for the alien subaltern (śvapacī, or “dog-cooker”), and the last, not tribal, but twice-born Hindu women of the “m erchant” caste (vaiśya-yositah). Even th e one ethnonym , Māgadhī, does not hold up as a reference to a woman of tribal origin. Magadha is a toponym th at refers to the “central land” of Buddhism, around the older royal Capital Räjagrha and the later imperial Capital of Aśoka, Pätaliputra (m odern Patna). It is listed by the Äpastamba Śrautasūtra as an outlying region, others being Kaliñga (Orissa) and Gandhära50—the latter the location of the stereotypical center of let­ tered learning in Pāli literature, Taksaśila (Taxila). Thus, Māgadhīs are not m entioned as examples of primitive, tribal womanhood. Rather, the relevance of the term here derives from the fact th at the Indian Dharma lit­ erature considers people from this region to be of mixed caste and, thus, impure.51 This list becomes even m ore coherent w hen one notes th a t th e term “m erchant w om en” (vaiśya-yositah) is a very unlikely reading. The edito r o f th e Kulārnava reports vaiśya- as a tte ste d in two o f five m anuscripts b ut relegates it to th e textual apparatus. The preferred reading of th e editor, T ārānātha V idyāratna, which fits th e context perfectly, was n o t vaiśya-yositah, but viśva-yositah: “everyone’s w om en” or prostitu tes.52 Indeed, it is strongly to be suspected th a t th e tru e reading of th e v arian t tex ts is n o t vaiśya-yositah, but veśya-yositah: quite literally “p ro stitu te women.”53 The structure th at lends coherence to this otherw ise heterogeneous list may be found in th e fact th at these term s all refer to occupations or castes th at are considered low a n d /o r polluting by th e orthodox Indian socioreligious order. The women suggested as objects of Śākta worship in this

Tantra (and in other, sim ilar lists elsewhere in the Tantric literature) are uniformly those who are considered to be the m ost ritually im pure persons. Although my full argum ent for this position will not be presented until chapter 4, it is sufficient to say here th at this list has nothing to do with tribals and should not be construed as evidence for tribal origins of the traditions o f th e Kulâmava. This is merely the ostensive content of the discourse. Structural analysis reveals th at it is the expression of a pervasive discourse o f ritual purity and pollution in dialog with m ainstream Brahmanical Hinduism, not an extem al reference to m arginal communities or to tribal societies. This same herm eneutical principle also accounts for Śakti being given the nam es M ātańgī and Candālī: N either of these names occurs here as an ethnonym ; they appear rath er as m etonym s for ritual impurity, which Śakti (it is thereby communicated) transcends. The sam e in terp retatio n recurs throughout m odern analyses. References to (allegedly) tribal groups are read as directly referring to those groups, w hereas a close, critical attention to those term s indicates th at they occur in clusters of semiotically related words th a t reference concepts of purity and pollution. It is as if one were to read Contemporary advertising and conclude th a t Ivory Soap was derived from elephantine com m unities or, at least, th eir byproducts. Of course, anyone cognizant of the sem iotical com m unity th a t produced these ads would know th at “Ivory” refers to the whiteness of th e soap (and, thus, its [99.44 percent] purity), not its derivation or ingredients. To read references to tribal or outcaste ethnonym s as evidence for th e origins of th e Tantric traditions within these com m unities is similarly and fundam entally to m isconstrue the m eaning of th e term s in context. This notion o f th e tribal origins of Tantrism has, however, developed trem endous au th o rity th ro u g h continual re p etitio n in th e scholarly literature. So stro ng has this presum ption becom e th a t scholars have begun to see tribalism where there is none. For instance, in a section on “[Tantric] Siddhas in th e Tribal Landscape,” Ronald Davidson makes the strong claim—central to his argum ent—th a t “we find . . . canonical and exegetical references to trib a l. . . peoples almost at every turn.”54 Consult­ ing the footnote th a t will presumably detail these abundant references to tribals, th e reader is directed to two passages: one from the Guhyasamäja Tantra (and its com m entary, th e Pradipoddyotana) and an o th er from the Krsnayamäri Tantra.ss These passages in th e Tantras both employ th e term mahātavīpradeśa, which Davidson translates as “dom ains of th e great for­ est tribes.”56 However, atavī does not mean “forest tribe.” It m erely refers

to a forest.57 The Krsnayamãń verse is the first of a passage th at describes the special practice (caryä—the subject of chapter 5) th at typically takes place in liminal or rem ote places such as forests. The Guhyasamäja pas­ sage speaks of religious practice taking place uon a rem ote m ountain, in great forest regions provided with fruits, flowers, and the like.”58 The commentary, which Davidson also cites as evidence of tribal provenance, glosses mahãtavīpradeśesu as mahãtavyãh pradeśesu “in regions of great for­ ests,” which, the com m entator further notes, means “in excellent places” (prakrstesu deśesu), their excellence due to the fact th at they are “delightful [on account of having] ponds and so fo rth ” (jalãśayãdimanoharatvãt). Thus, there is nothing in any of these passages th at suggests anything about “forest tribes.”59The “references . . . at every tu rn ” tu rn out to be a chim era of a long-standing scholarly rhetoric and testify to the rem arkable durability of habitual ways of thinking and speaking about Tantric Buddhism inherited from (and consecrated by) more th an a Century of scholarly repetition. This is a clear example of the m anner in which theo­ ries concerning the social context of Tantric Buddhism m arkedly structu re and reinforce the interpretations Contemporary scholars advance concerning Tantric Buddhist literature. The list is “about” tribals and thus was probably composed by tribals, which in tu rn accounts for why the list and the transgressive practices it describes “appear to us so meaningless and puerile.”60 This assum ption operates in concert w ith another im portant, structu rin g scholarly axiom, one th a t undergirds the second major account still active in Contemporary scholarship. This is the notion th at Tantric Buddhism is not in fact Buddhism at all and, thus, m ust have originated somewhere eise. There is an extensive academic literature arguing for the foreignness of the Tantric movements, locating them either to the west of India among the “Magi priests of the Scythians,”61 or to the east in Chi­ nese Daoist circles.62 More recently, this foreign source has been identified w ith Śaiva groups th a t were the neighbors and com patriots of the esoteric Buddhist communities. This notion continues to be very influential in shaping scholarly discussions, although Louis de La Vallée Poussin critiqued it as early as 1898: One commonly regards idolatrous and superstitious Tantrism as “no longer Buddhism;’' one forgets that Buddhism is not separable from Buddhists, and that the Indian Buddhists (les Hindous bouddhistes) were willingly idolatrous, superstitious or metaphysicians.63

I do not m ean to suggest th at there is no trace of interaction with eso­ teric Śaivism in the Buddhist Tantric traditions. To adequately address the issue of Śaiva influence on esoteric Buddhism would require a monograph in itself. (In fact, Alexis Sanderson has recently devoted a m onographlength essay to exactly this topic.64) Suffice it to say that, unlike theories of tribal origins, there is substantial evidence of sustained and intense interaction between contem poraneous esoteric Śaiva and Buddhist communities. That said, it seems equally clear th at the influence was mutual, with each tradition leaving significant traces of th eir own thought and practice on currents in the other.65 Consequently, it is not cogent to speak of esoteric Buddhism having “originated” in esoteric Śaivism. Just as Śaivism and Buddhism interacted in the preceding m illennium (ca. 450 b . c . - a .d . 550)—sharing the same social spaces, political structures (that generally patronized religious Orders eclectically), and economy, as well as a common spatial and tem po­ ral framework for religious activity, those communities centered around various Śivas, Visnus, Buddhas, and (jaina) Tīrthańkaras in the late first millennium ( a .D.) participated mutually in a pan-Indian religious culture, most of whose structuring assum ptions were th e same and in which a variety of ritual forms were shared and developed across traditions.66The only way in which esoteric elem ents tout court can be considered to have originated in Śaivism is if one reifies these traditions as having been not only institutionally but also intellectually, socially, and ritually isolated from one another at the beginning of the Tantric period. That is to say, the only way th at it is cogent to speak of esoteric Buddhism having originated from Śaivism is if one begins with an already fully-formed notion of Bud­ dhism th at does not include Tantric elements. Yet the only way to do so is to adopt a norm ative position on w hat “real” Buddhism is. We will return to this issue in a moment. Before moving on to the larger methodological question, lest there be any doubt th a t these same old chestnuts about the origins of Tantra are alive and well today, consider this passage from a recent (2006) undergraduate textbook issued by the British academic publisher Routledge: Tantra derives from a wide constellation of beliefs and practices that mostly belonged to the non-Vedic religious traditions of the Indian subcontinent. Tantra may date back to the period of the Indus Valley Civilization which, together with the religions of chthonic tribes, form part of its source [Later,] beliefs and practices of these once marginalized groups

began to bleed into the Sanskritic tradition.. . . The content of the Bud­ dhist Tantras appears to have Hindu, especially Śaiva, origins.67 This sum m ary may unfortunately be said to represent th e current, m ain­ stream consensus on the origins of the Tantric traditions: The theory of decline has dropped out, leaving a synthetic account—a bricolage of the prehistoric, tribal, and Śaiva models.

THE QUEST FOR ORIGINS AS M ETHOD IN THE HISTORY OF RELIGIONS As noted previously, the m ethod of locating the origins of a historical object and tracing its developm ent thereafter is a widely practiced and well-pedigreed m ethod in th e hum an and social Sciences. In the nineteen th Century, and well into th e tw entieth, it was considered a funda­ m ental approach to historical scholarship. As Eric Sharpe has observed of late nineteenth-century intellectual currents: Virtually every philologist, every historian, every archaeologist would at this time have been able to subscribe to what Jane Harrison had written in 1885 in her Introductory Studies in Greek Art, that “The historical instinct is wide awake among us now. We seek with a new-won earnestness to know the genesis, the origines of whatever we study.. ”68 That is, th e quest for origins was considered in this period to be not merely a, but the, prim ary tool of history (and an epitom e of fashionable French theory69). This is im portant to grasp; w ithout this insight, the m any claims made by, for example, phenom enologists of religion to study th eir subject “historically,” are incomprehensible. What is very often m eant by historical scholarship in religious studies well into the tw entieth Century is essentially to identify the identity or core nature of a tradition by reference to its origins and to trace the fortunes of this (self-consistent) object over time.70 As Paul Harrison observed in 1995, this approach is very much alive in Contemporary studies on Buddhism: The fascination with origins, beginnings or sources does appear to be a kind of scholarly universal. Part of this—and this much is clear enough—is the idea that if we can understand the beginnings of something, we are better

placed to understand the whole thing, as if its essential character were somehow fixed and readable in the genetic encoding of its conception.71 This notion th at the origins of a thing determ ine its fundam ental nature has a longer history but was perhaps most clearly articulated in the early eighteenth Century by the Italian scholar Giambattista Vico. In his Scienza Nuova (New Science), Vico distilled the essence of this method, summed up in the equation of nature (natura) and origin (nascimento), This appears as the fifteenth axiom of his historical method: The inseparable properties of things must be due to the mode or fashion in which they are born. By these properties we may therefore teil that the nature or birth (natura o nascimento) was thus and not otherwise.72 That is, historical things have essential (or inseparable) qualities th at are fundamentally conditioned by the circumstances of their origin—th at the birth of a historical thing brings forth a “quiddity” th at makes the thing what it is and no other. This is an elem ent of th e “new science” th at is stubbornly entrenched in academic research. Historian Marc Bloch describes an obsession with origins as “th e idol of th e historian trib e” and even as an intellectual “hypnosis”73 More recently, historian of religions Daniel Dubuisson has rem arked on w hat he sees as a pervasive “m ythic im agination” in the hum an Sciences “where origins are considered by m any—and often—in an alm ost spontaneous way as the locus of perfection, of initial fullness and simplicity.”74 This is perhaps especially true in scholarship on reli­ gions. No m atter how m uch training is given in historical m ethod (and in most religious studies departm ents, this is practically none), a seemingly basic hum an impulse to grasp things through th eir origins continues to function.75 There is, furtherm ore, an observable and intriguing codependence between th e acts of defining a historical entity and identifying its ori­ gins. For example, it is only through the intellectual Operation of defin­ ing esoteric Buddhism as one thing and not another th at an origin can be constructed in the first place. One can only identify th e origins of a thing if one has already defined w hat th at thing m ost essentially is; and, conversely, identifying the most essential nature of a historical thing is generally made w ith reference to its origins. It is only, for instance, by identifying Tantrism as essentially th e worship of women th a t one can

claim to locate its origins in th e prehistoric world of buxom terraco tta figurines; by identifying it as pursuit of pleasure th a t one can point to derivation from m onastic unrest; or by identifying it as essentially Śaiva (or un-Buddhist) th at one finds its origins in Śaivism. A consistent appeal to origins would (and, in fact, practically does) make th e study of Tantric Buddhism impossible. This is so because the various definitions involved (and the structuring assumptions th at undergird them ) render the very term “Tantric Buddhism” an oxymoron. Insofar as any religious phenom enon is truly judged “Tantric,” and thus fundamentally tribal/m arginal/Śaiva, it is thereby rendered non-Buddhist—a foreign grow th grafted onto the Buddhist tree. Conversely, insofar as texts, artifacts, or practices are considered “Buddhist,” for much of the schol­ arly com m unity they become ipso facto non-Tantric, since “real” Tant­ rism (we all “know”) is not the "sem anticized” and “bowdlerized” version practiced by the Buddhists but rather the tribal, marginal, Śaiva variety. In a bizarre fashion, Tantric Buddhism finds itself an awkward stepchild of studies of both Tantrism and Buddhism. Accounts of origins, then, are fundam entally the product of the histo rian (of religion)’s constructive activity in identifying a m ost central aspect of th e tradition.76 Dubuisson, in fact, argues th at a scholarly focus on th e origins of things is fundam entally m ythical, not scientific, for “th e typical and almost exclusive question posed by m yths is th at of origins” and “the fascination th a t the hum an Sciences have for such simple, ‘theological’ explanations [that reduce the infinite diversity of reality to a principle or to a unique, ontologically homogenous cause] probably represents the greatest obstacle th at they have to overturn and overcome.”77 Analogies can easily be found in the hum an Sciences. Clothing, for instance, is likewise not susceptible to a monocausal account of origins th at serves as a totalizing interpretative key to its historical meanings. It very well may have been originally crafted for the purpose of warm th, but if clothing as a hum an phenom enon is consistently interpreted in light of this it will introduce major distortions. A trem endous am ount of the cul­ tural life of societies is devoted to clothing w ithout much, if any, reference to its actual pragmatic value. If a scholar were to define its origins and to attu n e her cultural interpretations accordingly, she could not in fact be said to understand clothing in any but the m ost superficial and ahistorical sense. Much the same can be said of the Tantric Buddhist traditions. Were one to find a satisfactory account of its origins—and, as I have endeavored already to show, this has not successfully been accomplished to date—one

will certainly “make sense,” but one would still be no closer to understanding Tantric Buddhism. Scholars, consequently, would be best served by getting out of the origins business. More im portantly for our purposes, however, in identifying its ori­ gins (and thus, essential nature), the scholar sim ultaneously (if surreptitiously) constructs an ideal social context for Tantric Buddhism th a t then serves as an interpretative fram e w ithin which to make sense of transgressive discourses and practices. W hat Hayden White has said of historical works in a narrative m ode (a topic to which we will proceed shortly) is as tru e of historical models in an etiological mode, insofar as both pu rp o rt to be “a model, or icon, of past structures and processes in the interest of explaining what they were by representing them!’78 By locating an origin, these approaches seek to “make sense” of Tantric transgression by ascribing them to an historical agent or agency. That is, by depicting transgressive discourses and practices as being expressions of a particular kind of actor, they thereby become assim ilated to known types of transgression, and thus meaningful (or, at least, com prehensibly meaningless). Thus, for instance, w hen the Tantric traditions become owned by rebellious, pleasure-seeking (ex-)monks, th e motives and m eaning of th eir transgressions are thereby also com m unicated. Why do they advocate breaking all th e rules? Obviously, since they are boys subject to “unnatural rules of restrain t” and they “naturally” seek to transgress them . By attributing them to prim ordial or Contemporary primitive tribal peoples, their motives may be less clear, but the explanation is nonetheless transparent: these are stränge rites of unfamiliar people with “m arginal” ideas about the power of sex and natural forces they do not fully understand. No furth er explanation is needed, since Tantric transgression thereby becomes an expression of the “prim itive mind: superstitious, childlike, incapable of either critical or sustained thought.”79 The popularity of this mode of explanation may perhaps be attributable to such longstanding m odern associations as, for instance (as Bruce Lincoln has indicated), th e association of “episodes of incest and cannibalism ” with “irrationalities th at reveal th e childhood of hum an thought.”80 By ascribing Tantric practices to im itation of the Śaiva traditions, the meaningfulness is displaced, and thus deferred. The rites do not need to make sense (in and of them selves they can be perfectly m eaningless)—one only needs to account for why the Buddhists im itated them , for which patronage jealousy is a readyrciade and common attribution.

Therefore, it should be clear th at th e construction of etiological histories of Tantric Buddhism are driven fundam entally by—and their most palpable effect is in potentiating—th e project of making sense of the transgressive aspects of these traditions. As Mitra made clear, we cannot merely w rite them off as the “ravings of m admen;” but, through imaging th eir origins (and, correlatively, the social location in which these prac­ tices paradigmatically take place), they can be accounted for adequately, if patronizingly. These are practices, one concludes, driven either by animal impulses of the sexual drive, primitive superstitions, or slavish im itation. All are, ultimately, explanatory paradigms, which is precisely w hat the etiological mode of historiography is m eant to provide. By identifying its “tru e ” m eaning in and by a scenario of origination, interpretation may proceed and be fixed, w hereby—w onder of wonders—Tantric Buddhism “makes sense.”

{2] NARRATING TANTRIC BUDD HIS M

[The] goal [ofhistorical research] is less to teil new stories than to retell familiar ones. —Philippe Carrard

I

N A D D R ESSiN G

issu es o f c u ltu ra l u n d e r s ta n d in g a n d I n te r p r e ta tio n ,

th e r e is p e r h a p s n o d isc ip lin e m o re c ru c ia l th a n h isto ry . In th e p a s t tw o c e n tu r ie s , fo r b e t te r o r w o rse , h is to ry h a s b e c o m e a d o m in a n t

(p e rh a p s th e d o m in a n t) m o d e o f u n d e r s ta n d in g th e w o rld a n d o u rselv e s. It is th e p riv ile g e d m e d iu m fo r e x p re ssin g id eas a n d valu es, a n d fo r sign ify in g m e a n in g in th e h u m a n S ciences. T h e re a re o th e r d isc o u rs e s, o f c o u rse , b y w h ic h p e rs o n s , id e as, a n d in s titu tio n s m a y b e re p r e s e n te d , asso c ia te d , a n d ev a lu a te d ; b u t w h e n it co m es to u n d e r s ta n d in g d y n am ic p ro c e sse s o f c h a n g e , th e la n g u a g e o f p o w e r—th e effec tiv e la n g u a g e —is histo ry . It is th u s to be e x p e c te d th a t m o d e rn sc h o lars w o u ld e n d e a v o r to m ake se n se o f T a n tric B u d d h ism th r o u g h th e m e d iu m o f n a r ra tiv e h is to ­ rio g rap h y . T elling th e s to ry o f a relig io u s tr a d itio n allow s th e s c h o la r (an d h e r re a d e r) so m e p u rc h a se o n its d e v e lo p m e n t—h o w it c h a n g e d o v er tim e a n d w as in flu e n c e d b y v a rio u s c irc u m sta n c e s. In allo w in g a d ia c h ro n ic p e rsp e c tiv e , th e w ritin g o f n a rra tiv e s is g e n e ra lly th o u g h t to d a y to b e a m o re tr u ly h isto ric a l a p p ro a c h to th e s tu d y o f relig io n s th a n t h a t allo w ed by th e se a rc h fo r o rig in s. It is to th e m o d e rn h is to rio g ra p h y o f th e s e tr a d itio n s, its rh e to ric , a n d th e s tr u c tu re th e r e o f th a t w e sh all n o w tu r n o u r a tte n tio n . N a rra tiv e s tr u c tu re is fu n d a m e n ta l to h is to ry as it is u n d e r s to o d today. T his is w h a t d is tin g u is h e s h is to rie s fro m m e re a n n a ls o r c h r o n ic le s .1 H isto ric a l n a r r a tiv e s m a k e se n se o f a n o th e rw is e m e a n in g le s s se rie s o f ev e n ts, a n d a re a fu n d a m e n ta l m o d e o f h u m a n u n d e rs ta n d in g . N a rra tiv e s

com m unicate a logic of change, a connection between events, and they situate historical actors in com prehensible contexts w ith discernible lines of development. It is, first and foremost, through casting events and actors in narratives th at the otherwise nonsignifying (if not insignificant) flow of events is rendered sensible. In w hat follows, I will consider briefly th e poetics of narrative rep resen tatio n . I will touch on th e stru ctu re of th e historical im agination and the role of narrative form in making historical sense. Subsequently, I will outline the three m ajor modes in which the history of Indian Tantric Buddhism has been narrativized and the subtending structures (langues) th a t inform these various articulations (paroles).

THE POETICS OF HISTORIOGRAPHY

In coming to understand the rhetoric employed in the historiography of Indian Buddhism, it is necessary first to consider the nature of histori­ ography itself. Hayden White, one of the leading lights of m odern historiographical thinking, has observed, “It is often said th a t history is a m ixture of Science and art. But, while recent analytical philosophers have succeeded in clarifying the extent to which history may be regarded as a kind of science, very little attention has been given to its artistic components.”2It is in this latter area th at W hite’s own work has made an invaluable contribution. In his research, White draws attention to th e fictive, contingent nature of the rhetorics th a t structure and inform historical writing. Following his lead, historians have begun to take more seriously th e fundam ental fact th at narrative forms are in principle independent of th e evidence they serve to organize.3 Rather, they are the result of an imaginative process by which the historian constructs them as a story—a set of events with a narrative arc. Historical accounts thus consist of at least two elem ents—a factive aspect and a fictive one. That is, histories consist of certain factual ele­ m ents or data (which them selves may be m ore or less independent of a subtending in terpretative fram ework) th a t are organized and given m eaning by a fundam entally fictive,4narrative structure. Once a phenom enon has been constituted as an object of historical discourse—itself an act of imaginative construction—a ränge of rhetorical moves are potentiated. The phenom enon in question can th e n be conceived as having an origin, a development, and a resolution—th at is, it now can become, in the Aristotelian sense, a story to be told.5

Louis Mink, whose m arvelous essays on the Historical Understanding were m uch adm ired by White, has described a widespread (albeit naive) attitude in historiography th at claims (implicitly) th a t “the historian . . . finds th e story already hidden in w hat his data are evidence for; he is Creative in th e invention of research techniques to expose it, not in the art of narrative construction.”6 Mink, quite rightly, finds this view highly problematical. It is th e historian, after all, who im parts identity, m eaning, and narrative function to the data at her disposal. The narrative role and, thus, th e historical m eaning of any historical fact are in themselves indeterm inate. Any event may be cast in a variety of narrative contexts and serve a variety of narrative fimctions, while rem aining entirely faithful to the historical record. At the most basic level, events may be cast as either a beginning, a middle, or an end: the three fundam ental elem ents of narrative according to Aristotle. However, while Aristotle seems to have believed th at events were naturally and necessarily so structured,7 such is dem onstrably not the case. One example from (relatively) recent history should suffice: the independence of India. One can easily see th at its narrative role is underdeterm ined. At m idnight on the 15th of August 1947, certain events seem definitely to have taken place—new pieces of colored cloth were raised on poles at Lai Qila and elsewhere, words were spoken, new authorities were vested, festivities undertaken, while, at the same time, a cer­ tain M ohandas Karamchand Gandhi slept (perhaps uneasily) in Calcutta. Yet, in w hat way do these disparate events cohere in a unified narrative? Did these events constitute, for instance, the end of British rule of th e subcontinent? Were they instead a m edian point in larger processes of social and political change taking place in South Asia? Or were they the beginning—th e dawning—of a new age and a new order? It is all, I would say, and none, and m ore than these. To use a Buddhist idiom, these are samvrtisatyay not paramärthasatya: Each is a reality conjured forth by the consensual agreem ent of a signifying community, not realities th a t exist in and of themselves. Each are possible ways of interpreting those events; each serves th e aims of a certain set of embodied interests; but there is no independent, epistemic criterion by which we may privilege one over the others as a more true presentation of the realities they represent. The appropriate question in discussing narrative em plotm ents of his­ torical events is therefore not which is true. Rather, the key issues are (a) w hat is th e sem iotical logic or structure th a t inform s each n arra­ tive? And, (b) w hat ideological ends does each serve? W hite’s Metahistory

advanced this discussion by highlighting th e mechanics of the modes of em plotm ent, explanation, and ideological im plication th at structure m odern historiography. Drawing on the work of N orthrop Frye, White explored th e m anner in which identical series of events could be rhetorically cast in either comedic, tragic, rom antic, or satiric modes. For instance, th e h istory of any given phenom enon could be told as an instance of the trium ph of good over evil (romance), a transitory trium ph (comedy), a m om entary defeat (tragedy), or as a failure to m aster a world th at is captive to death and the specter of meaninglessness (satire).8What is im portant to note about these choices is th e irreducibly im aginative elem ent in them . The narrative form is nowhere found in the data itself. Indeed, both Mink and White are concerned to elucidate the extent to which “histories” are not ultim ately th e product of the facts th at inspire them , but of th e poetical im agination of th e historian who “em plots” th em —an im agination which, in short, situates these facts within one of several conventional narrative structures. The availability of historical narratives is thus largely independent of the data th at they emplot. Moreover, the variety of narratives available to a historian is (like theories of origins) limited. There are only a handful of narrative forms available in any cultural idiom, with a limited ränge of distinctive varieties of story structures based upon them , and these latter are by no means universal. They are rooted in the narrative traditions of specific cultures. As White noted, the historian brings uto his consideration of th e historical record .. . general notions of the kinds of stories th at m ight be found there.” These notions are provided by the culture(s) into which the historian has been socialized. For instance, with regard to nineteenth-century historiography, W hite observed th a t “the norm ally educated historian of the nineteenth Century would have been raised on a staple of classical and Christian literature. The mythoi contained in this literature would have provided him w ith a fund of story forms on which he could have drawn for narrative purposes.”9 A skillful historian draws on th e stories best-known, best-loved, by her audience. By retelling these stories using data drawn from the period she is describing, a historian is able to teil a persuasive tale. These stories are those th at seem “natural” and obvious w ithin the culture. They are familiar and, thus, seemingly self-evident. The ränge of interpretative models available and persuasive w ithin a culture in p art predeterm ines the interpretations found by the historian. One sees what one knows—w hat one has been accustomed to seeing. The

m anner in which the historical imagination shapes our understandings of history and th e ränge of possibilities w ithin the historical field is another of the im portant im plications of Contemporary, critical historiography. Not only are historical narratives fundamentally fictive and based on precritical choices, but the ränge of historical imagination w ithin a culturallinguistic group also limits the types of em plotm ent available and, thus, the data deem ed relevant to th at group. Cultural habits of historiography can serve as m ethodological blinders or, as suggested in the introduction, “ruts” in th e avenues of scholarly research. A th ird im portant contribution of this type of analysis is th e m anner in which it highlights th e ideological implications th at the various story forms entail. Ideology, in this context, is deflned by White as “a set of presuppositions for taking a position in the present world of social praxis and acting upon it.”10 By em plotting events in a certain way, one implies a valuation of those events and, concurrently, consequences th a t under­ standing has for current actions and attitudes. History is no t merely, as some have suggested, “w ritten by the winners.” The Services of a skillful historian, or historical m ythologist, are a necessary precondition for being a winner. Effective leaders or visionaries are those who can craft (or Commission) compelling narratives, such th at others find it sensible to in terp ret th eir own experience and activity within the same narrative ffamework.11 To re tu rn to our example of Indian independence, it may be em plotted in a variety of ways w ith very different ideological implications.12 For instance, independence as a beginning seems to correspond to a rom antic or comedic mode of em plotm ent. Beginnings carry a load of imagina­ tive baggage: dawning, freshness, light, and promise. Here, independence means th e nascence of a new political and social order, w herein “young India”13 em barks on fulfilling its destiny and taking its rightful place among th e family of nations. Yet, for others, it may be em plotted as an ending—a sunset, portending darkness and dismay. This mode of em plot­ m ent foregrounds the demise of British rule in India, which had brought such glories and achievem ents as political unification of th e subcontinent, th e railway System, the Delhi Golf Club, punch, and th e like. Here, one constructs this event as a tragedy: a death to be m ourned, as some— British and Indian alike—still do. For others, it is a fulcrum point—an ethically neutral shift of power between two equally loathsom e, equally corrupt legions of bureaucrats—one foreign, one native, yet cut of the same cloth. In this case, one sees an ironic or satiric mode of em plotm ent.

None of th ese n arratives, it should be clear, is unqualifiedly true. They all shape (and by shaping distort) th e limitless com plexity of the event, w ith its m anifold actors, am bivalent or polyvalent m otivations, and interacting social, political, cultural, religious, linguistic, and eco­ nom ic cu rrents. Likewise, th ey all have palpable ideological implications. Such valuations are not some extrinsic elem ent of bias th a t may be system atically elim inated by th e scientific h isto rian . B enedetto Croce, for instance, observed th a t “historical affirm ation is th e quintessence of judgm ent, indeed is th e only tru e ju d g m e n t;. . . historical works are a web of narrative appraisals.”14 Histories make sense, not tru th ; and sense is always for som eone or som eones.15 One is rem inded of Alex W ayman’s observation th at th e m easure of an adequate study of esoteric Buddhism is precisely th a t it enables th e reader to come to a ju d g m en t about it.16 Croce was a m ore subtle thinker th a n Wayman; he was well aware th a t good historiography does not traffic in petty tribunals by which one may become (as Wayman put it) “genuinely for [or] against it.” Rather, by staking out an interpretative context (plot, characters, etc.) and locating h er subject therein, th e historian inevitably and unavoidably com m unicates a relationship betw een the reader and th e subject(s) of th e history, sim ultaneously indicating th e appropriate norm ative stance w ith regard to th e subject(s). Historical judgm ent sheds light and “opens the way,” w rites Croce, on “th e struggle of good against bad, useful against harm ful, beautiful against ugly, tru e against false, in a word, value against non-value.”17 With these considerations in mind, let us tu rn our attention to the various narratives th a t have inform ed th e w riting of the history of Tantric Buddhism in India. What modes of em plotm ent have typically been used in representing these traditions? What implicit explanatory logics or cul­ tu ral associations inform th e rhetorics used in each? And w hat do they indicate to the reader concerning the proper interpretative stance to take w ith regard to Tantric Buddhism in “the struggle of good against bad, use­ ful against harmful, beautiful against ugly, true against false, in a word, value against non-value?” What sense do these narratives make of Indian esoteric Buddhism? Like Indian independence, the narratives th at have been used to structure histories of Tantric Buddhism follow three basic models, each of which corresponds to one of the three m ost fundam ental narrative term ini. That is to say, one may read of Tantric Buddhism as the end of a prior process (the history of Indian Buddhism as a whole), or as the ancient beginnings of Indian religion, or as a medieval waypoint.

T A N T R A AS END: THE DECLINE A ND FALL OF IND IAN BU D D H ISM

Anyone who has read even a modicum of the scholarly literature on the history of esoteric Buddhism cannot help but be struck by how frequently narratives link these traditions inseparably w ith the decline and disappearance of Buddhism in India. The narrative of decline was undeniably the single m ost populär m otif used to structure the history of Buddhism in the years 1820-1930. Even as late as 1975, Per Kværne could w rite that “to regard tantricism as a ‘degeneration’ of earlier Buddhism has been— and in many circles still is—extremely widespread.”18Again, titles of book chapters clearly attest to how prevalent this model has been. The relevant chapter in William Theodore de Bary’s sourcebook The Buddhist Tradition (1969), penned by em inent Indologist A. L. Basham, is frankly titled “Tan­ tricism and th e Decline of Buddhism in India.”19The conceptual linkage of Tantric Buddhism and the end of Indian Buddhism was well-established considerably before this and was so taken for granted in th e early tw en­ tieth Century th at Louis de La Vallée Poussin, in “Notes de Bibliographie Bouddhique”—a running series of bibliographical notices—treated Tant­ ric Buddhism and the disappearance of Buddhism as one unified rubric: “44. Tantrisme, disparition du bouddhisme.”20 Seen from a certain perspective, Tantric Buddhism as a historical endpoint seems to make a good deal of intuitive sense, insofar as the flourishing of these traditions coincided with the last centuries (the end of the timeline) of th e flourishing of Indian Buddhism itself. This observation, however, hardly begins to account for the prevalence of this narrative in the m odern historiography of Buddhist Tantrism. As in all three of the narrative forms we will explore, there is a clear “semio-logic” th at structures these discourses: a deep structure of the m odern historical imagination. U nderpinning the various individual articulations of narratives of decline is one of the m ost populär and recurrent poetic models, both East and West: th e m etaphor of organic development. To stru ctu re narratives according to stages of organic life has been extrem ely com m on not only in the historiography of Buddhism, but equally so in historiography generally. Its use can be traced from hoary antiquity through the present, having been the model of choice among discerning authors from the very advent of W estern historiography. It has been utilized by w riters such as Homer, Hesiod, Plato, Vico, Hegel, and Marx, to name only a few. In brief, this archetype conceives that, just

as plants and animals are seen to go through a process of birth, growth, maturity, decline, and death, so other (even all) phenom ena can be traced across this same trajectory. Thus, cities, nations, schools of thought, political parties, and even religions, have been conceptualized in these term s, and the events of th eir histories interpreted accordingly. We m ust insist, nevertheless, on the metaphorical nature of this model. While we may, for instance, quite genuinely speak of the childhood, adulthood, decline, and death of individual m en (although even here th ere is frequently an elem ent of m etaphorical comparison), we are speaking in a poetic mode when we talk of th e childhood of Man. This m etaphorical em plotm ent becam e codified and objectified by Vico, w hen his New Science posited universal cycles of organic develop­ m ent in hum an history.21 In Vico’s historiography, we see a model of historical developm ent in which civilizations follow a regulär cycle of eras—a divine period, a heroic period, and a hum an period portending a decline into barbarism. R. G. Collingwood describes a fiirther analysis into six periods: First, the guiding principle of history is brüte strength; then valiant or heroic strength; then valiant justice; then brilliant originality; then constructive reflection; and lastly a kind of spendthrift and wasteful opulence which destroys what has been constructed.22 Although it becam e the foundation for m uch of the m odern practice of history, this vision of a determ inate and regulär succession of eras—eras th a t end in decline—was not a novel creation of m odernity. It is m erely a refinem ent of th e ancient m ythopoeic vision of th e successive ages of civilization: th e Golden, Silver, Bronze, and Iron Ages, in which th e natu re of hum anity progressively declines. This trope is operative too in th e sim ilar theory of the four ages in India: th e Krta, Dvāpara, Tretä, and Kali Yugas, w herein living beings becom e by stages less and less intel­ ligent, ethical, and vital. We find a sim ilar series of four stages, ending in decadence, in th e sociohistorical theories of Ibn Khaldün.23 In m ore recent memory, one finds Rousseau, in a strangely Buddhistic m om ent, com m enting th a t “the body politic, like th e hum an body, begins to die from the very m om ent of its birth, and carries w ithin itself th e causes of its destruction.”24 Clearly, this m etaphorical reading of historical processes as conforming to the pattern of the individual organic life cycle has been endemic

to historiographical practice throughout its history. The early nineteenth Century, in which the historiography of Buddhism was initiated, marked the zenith of popularity for this vision. History became a quest to find the stories waiting “out th e re ” in th e data. Of these stories, at least one thing was certain: They would follow, with law-like regularity, a cycle of organic developm ent. “Hegel,” says White, “broke down th e history of any given civilization and civilization as a whole into four phases: the period of b irth and original growth, th at of maturity, th at o f ‘old age,’ and that of dissolution and death.”25 For Hegel, not only the total structure of civilizational developm ent, but all the microcosmic histories w ithin it (in fractal fashion), traverse the selfsame four historical mom ents—mom ents th at correlate to his vision of the successive transform ations of hum an consciousness. In light of this narrative structure, so characteristic of European his­ toriographical practice, consider the following com m ent m ade by Cecil Bendall (professor of Sanskrit at Cambridge) w herein, w ith acute clarity, the model of organic developm ent is used to structure the history of Buddhism: Much... has been written about the glorious and vigorous youth of Indian Buddhism; something about its middle age of scholasticism and philoso­ phy; but next to nothing about its decay, decrepitude and dotage, as shown in the Tantra-literature.26 In line w ith this model, the following com m on version of Buddhist history is constructed. First th ere was Śākyamuni Buddha, th e original propounder of Buddhism, (of whom most reputable scholars will adm it that we really have no reliable data). The first period of Buddhism per se, then, is said to be th at of the so-called Hīnayāna/Theravãda. Here we see the traditions and the literature of Theraväda Buddhism, th e currentlydom inant school of Buddhism in m ost of Southeast Asia, defined as functionally equivalent to original Buddhism. This Buddhism, while not quite as “p u re” as th at taught by Śākyamuni (and certainly not in its Contem­ porary form in colonial Ceylon), is fairly faithful to the source. Then, the story goes, the literature of the Mahāyãna began to emerge. At this point, after th e pure ethical teachings of the early Buddhist schools (which, one is cautioned, were a philosophy or a way of life, not a religion), Indians were no longer able to follow the dictates of such a lofty path. They began to rationalize th eir instinctive, plebian bowing and scraping to idols as

orthodox Buddhist practice. At the end of this process, Buddhism finally goes off the deep end. After being continually eroded by the slothful, sensual tendencies natural to Indians (and o th er natives of warm climes), the Buddhist tradition finally decided to give free license to do w hatever one w anted and to call it Buddhist practice. To this end, however, it was thought necessary to fabricate apocryphal scriptures (Tantras) in which such sensual indulgences could be passed off as orthodox practice, sanctioned by th e Buddha. This is clearly the view ascribed to by Monier Williams, Boden Professor of Sanskrit at Oxford, in his volume on Buddhism. All of the foregoing mod­ els are brought together in this influential work. “The tendency of every religious movement,” claims Williams, “is towards deterioration and disintegration.”27 After the Buddha’s death, he claims, “the eternal instincts of hum anity . . . insisted on making them selves feit notw ithstanding the unnatural restraint to which the Buddha had subjected them ”28 and Buddhists quickly began to give up the celibacy, ethics, and other teachings enjoined by the Buddha. Then, he claims: The Protean system called Mahā-yāna arose, and grew, by the Operation of the usual laws,. . . into a congeries of heterogeneous doctrines, including the worship of Bodhi-sattvas, deified saints, and personal gods.29 Yet, “far worse th an this, Buddhism ultim ately allied itself with Täntrism or th e worship of the female principle (śakti), and under its sanction encouraged the grossest violations of decency and the worst forms of profligacy.”30 Substantially the same narrative is found in Louis de La Vallée Poussin’s later work: Criticism can admit this tripartite division: a Buddhism undevotional and exclusively monastic, or the Little Vehicle, which goes back without doubt [!!] to the founder; a Buddhism much more composite, monastic and secular, devotional, polytheistic, at times monotheistic, highly commingled with pure philosophy and gnosticism (gnose): this is the Great Vehicle. . . ; finally, the degraded and denatured Buddhism of the Tantras, attested since the Vllth Christian Century.31 Repeatedly, the same story appears in the Standard works on the history of Buddhism. There is no need to m ultiply exam ples—anyone who has

read works on Buddhist history has come across this story or one very much like it. The question this poses for the critical historiographer is how, w ith a variety of narrative forms available, did this one so quickly become dom inant? The ans wer may be found by attending to patterns observable in the use of historical narrative and historical explanation in European literature. The narrative of civilizational decline following upon moral (especially sexual) degeneracy was well-established in the classical historical tradition—and was thus readily available to the historical imagination of early scholars of Buddhism, whose education was founded in large part on the study o f classical literature.32 Perhaps the paradigm atic example is the tale of th e Etruscan decline. Here, in a significant and populär his­ torical episode of Roman history, the fall of Etruria—a powerful neighbor of early Rome (subsequently incorporated into the empire)—is attributed to th eir m oral degeneracy.33 R. A. L. Fell States in his work on Etrwria and Rome: The decline of the Etruscan people is often ascribed to the nature of their religion, and the depravation of their morals. Greek writers have much to teil us of the luxury and the vices of the Etruscans, of their elaborate feasts and flowery coverlets, silver vessels and numerous attendants, and the Roman poets echo the taunt.34 It is w orth noting th a t this trope is later co-opted by Christian historians—developing from th e Roman intellectual tradition—to explain the fall of Rome itself. The decrepit civilization of paganism w ith its Neros and Caligulas, phallic cults and games, they claimed, m ust necessarily give way to the vigorous, youthful moral power of Christianity.35 It is clear here from whence Vico derived his final phase of “spendthrift and wasteful opulence.” It was precisely this historical archetype th at inform ed th e fashioning of the history of Tantric Buddhism. Given the basic datum so strikingly evi­ dent to w riters of British In d ia -th e absence of a Buddhist presence and, hence, th e ostensible disappearance of Indian Buddhism—one needed to account for this fact historically. For many, Tantrism fit the exigencies of narrative quite nicely, providing a familiar and easily-digestible account. The idea most commonly associated with Tantra from the outset (and still widespread today) was sex;36and sex, of course, was associated with decadence. Inevitably, this conception of the Tantric traditions suggested to

the narrative im agination of the nineteenth Century the classical arche­ type of the decline and fall. The resulting tale, it should be apparent, is a familiar one, recapitulating th at of Etruria: A once strong and vital culture becomes seduced by pleasure and renounces its earlier com m itm ent to purity and virtue. In particular, th e Iure of the pleasures of the flesh— so difficult to keep in check—overcomes the people and society becomes decadent. The ultim ate outcom e is the death of the once-great society. Of course, this choice of fictive em plotm ent is predicated upon two prior interpretative choices: For one, it foregrounds a theory of civilizational decline due to m oral degeneracy and, second, it identifies the Tantric traditions prim arily w ith moral failings. The arbitrary and essentially fictive elem ent of these choices becomes strikingly apparent when one considers alternative narratives based on alternative interpretative choices. Alexander Cunningham, for instance, although well aware of the existence of Tantric Buddhism, gives the following variant account of the Buddhist decline: Buddhism had in fact become an old and worn-out creed, whose mendicant monks no longer begged their bread, but were supported by lands long since appropriated to the monasteries. The Srämanas and Bhikshus were not like those of ancient days, the learned and the wise, whose bodily abstinence and contemplative devotion, combined with practical exhortations and holy example, excited the wonder of the people. The modern Buddhists had relapsed into an indolent and corrupt body, who were content to spend a passive existence in the monotonous routine of monastic life there were still the same outward signs of religion; but there was no fervent enthusiasm in the lifeless performance of such monotonous routine.37 Cunningham invokes another populär archetype of the nineteenth Cen­ tu ry historian’s arsenal. In this account we hear—not the echoes of the classical tale of the Etrurian debauches—but rather the strains of the (neoclassical) tale of th e Reformation (and Enlightenment). Here, the relevant connection is not sex, but ritual. Late Buddhism is hom ologized w ith Romish religion, as opposed to the pure serm ons of the Son of God. We see yet another clergy th at has become pam pered and luxurious, content to defraud the populace w ith their “priestly mummery.”38 The invocation of this narrative model bears witness to Cunningham ’s place among the heirs of the Renaissance, the Reformation, and the Enlightenm ent. It is

not, however, convincing witness to actual events in India. This emplotm ent too is a fundam entally fictive account th at crafts a unified understanding of a complex process. While making sense of the same evidence, it Stands in direct com petition w ith those who would account for the putative decline of Indian Buddhism in term s sexual and moral, rath er than ritual and ecclesiastical.39 In this regard, we m ay note the following, extrem ely illum inating, Statement of T. W. Rhys Davids th a t reveals how the exigencies of plot structure can far outweigh (and even supplant) the testim ony of concrete evidence. Starting from the prem ise of the putative decline and fall of Buddhism, Rhys Davids leaves the reader of his Buddhist India w ith the following considerations: Gibbon has shown us, in his great masterpiece, how interesting and instructive the story of such a decline and fall can be made. And it is not unreasonable to hope that, when the authorities, especially the Buddhist Sanskrit texts, shall have been made accessible, and the sites shall have been explored, the materials will be available from which some historian of the future will be able to piece together a story, equally interesting and equally instructive, of the decline and fall of Buddhism in India.40 In case th e re had been any doubt about th e fundam ental, form ative influence of precritical, fictive, theoretical models on the construction of Indian Buddhist history, here th ere can be no question. Rhys Davids indicates in essence th at, before scholars have even collected the evi­ dence available from literary and archaeological remains, they can a pri­ ori assum e a narrative structure along the lines of Gibbon’s Decline and Pall ofthe Roman Empire. Gibbon’s m asterw ork had allowed a new way of making sense of th e fall of a hugely successful enterprise (Rome); some such account was seemingly needed to understand India’s loss of Bud­ dhism as well. At times, then, narratives of the fall of Buddhism were crafted on the model of th e classical story of decline through sexual degeneracy. At others, th e decline is attributed to an alleged disconnection of scholastic Buddhism and its ritually-oriented priesthood from the needs of the laity—reflecting th e populär E nlightenm ent/P rotestant narrative of the decline of Catholicism. In very recent work too one finds echoes of this rhetoric of degeneration, of the decline and loss experienced by Buddhist

communities in the Tantric period.41 One may be excused a sense of déjà vu when one reads in such works th at yet another culprit was responsible, since skeptical Centrist (Madhyamaka) thought “constructed the ideal justification for the morally indolent to buttress th eir unwillingness to adhere to the precepts. Such indolence was ever lurking in th e backg ro u n d . .. the history of Buddhist monasticism is a narrative about the extended testing ofpreceptorial boundaries by the morally challenged .” Even the media are called to account in this Contemporary rendition of the decline and fall of Buddhism, insofar as they featured “erotized com positions of the Sanskrit and Prakrit poets.”42 W hat is m ost striking among these various discourses is th a t regardless of w hether an author blames sex, or scholasticism, skepticism, indo­ lence, titillation by the literary media, or a com bination of all of these, the fundam ental narrative remains the same. The ideological implications of these narratives—their “prescriptions for taking a position in the present world of social praxis and acting upon it”43—are abundantly clear. Narra­ tives of decline never entail a positive assessment. They are m eant, rather, as object lessons in what to avoid. In casting Tantric Buddhism in the role of a conclusion, these traditions are made to appear as destructive and dangerous—causes for the erosion of quality and goodness. The same is tru e of the inverse of this m odel—th e sim ilarly teleological tro p e of trium ph. Such narratives m erely represent th e ideological inverse of th e supercilious scolding of th e decline model, serving to valorize and advocate for the traditions. For those w ith an opposite ideo­ logical agenda—drawing on an alternative theory of cultural history—a n arrative of Buddhist Tantrism as an end may be cast in a progressive mode, w herein th e Buddhist traditions would have been cut down at th e ir height w ithin the flourishing artistic, intellectual, and political culture of Päla period Bengal. In practice, such narratives are ra th e r rare in scholarly literature; however, an interesting example of this may be found in M iranda Shaw's Passionate Enlightenment, w herein Tantric Buddhism is depicted as “th e crow ning cultural achievem ent of Päla period India.”44 It is notew orthy that, although her narrative prom otes an opposing norm ative agenda, the various data Shaw cites to argue for h er n arrative of progress (e.g., various em olum ents offered to successful scholar-m onks) are precisely those cited by Davidson to justify his narrative of decline. The two differ only in th eir choice of interpretative model.

TA N T R A AS BEGINNING: THE PRIM ORDIAL UNDERCURRENT

In th e early tw en tieth Century, w ith the narrative of decline firmly established in scholarly discourses, a second historical model em erged that told a very different story of the history of Tantric Buddhism. This new rhetoric seems originally to have been the product largely of Indian (chiefly Bengali) scholars but was quickly taken up and amplified by the more Romantically-inclined interpreters of the West. The core idea of this narrative is th a t th e Tantric traditions represent the prim ordial religion of the Indian subcontinent th a t was driven Underground by the invading Äryans whose patriarchal, Vedic religion established itself on top of this earlier, m atriarchal tradition, which was dom inated thereafter but never entirely extinguished. As in the case of the trope of decline, what is most w orthy of notice is the consistent structure of the rhetoric across its many iterations. In this case, throughout the many works th at employ this narrative, there appears a set of hierarchical, binary oppositions that lend m eaning and coherence (and ideological valency) to the discourse. These core binaries are below/above and past/present—the Tantric tradi­ tions are a cultural undercurrent th at derives from the most ancient past. Yet these basic dualities ramify throughout the literature, articulating in turn not merely a spatial and tem poral difference, but an entire network of such hierarchical binaries, including th at of gender. Both of th e valences of this discourse—the spatial and the tem poral— appear quite clearly in Giuseppe Tucci’s influential 1949 essay on Vajrayäna Buddhism. Tantric Buddhist rites show the gradual ascent to the surface, the invasion and the spread, of older intuitions, which dig their roots deep into India’s spiritual and religious bedrock and hand down ffom it primitive, sometimes barbarous, ideas.45 The historical em ergence of the Tantric Buddhist traditions in the midlate first millennium , then, may be explained as an occasion in which— like Jed C lam pett’s black gold, bubbling up from th e ground—“the old bedrock again came to the surface.” Tucci considered this bedrock to consist in “the inexhaustible fund of Indian folklore.”46 The language of spatial hierarchy is pervasive in this narrative. George Eider, w riting in 1978, echoes Tucci, claiming th at “the [Tantric] cults may

have been a hidden force w ithin India from pre-Aryan tim es” and th at the rise of Buddhist esoterism in the first m illennium represented “a sudden and even violent eruption of pre-Aryan religious life never really conquered by chariot or fire sacrifice but seething more or less Underground for many centuries.”47 That is, the Tantras come down from ancient (preÄryan) times, were suppressed (but not conquered) by the culture of the Äryans, and later erupted to th e surface from Underground. It is also notew orthy th at the ambiguity of the word primitive (meaning both first an d /o r crude) is very much in play in this discourse. For his part, Eider sees these pre-Aryan religions as existing in prim itive fertility rites of ancient days. As was not uncom m on in scholarship of this period, th ree very dif­ ferent constituencies were incorporated w ithin a unified academic dis­ course about ancient Tantrism: early peoples, Contemporary primitives, and Contemporary lower classes. The cultural forms of the lower social Orders were considered to be of ancient provenance (folklore), providing a conceptual (and historical) link between the past and the provincial.48 Similarly, as Johannes Fabian has so clearly dem onstrated, there was a pervasive anthropological conflation of ancient societies and Contempo­ rary cultures possessed of simple (primitive) technology.49Thus, in Tucci’s analysis, th e bedrock of folklore may inevitably be found among provin­ cial peoples insofar as “in the literary descriptions of the Vajrayäna pantheon late though it be, prehistorical Indian religion survives, w ith its old deities of th e tribes and villages.”50 Shashibhusan Dasgupta uses much the same rhetoric when he asserts th at “Täntricism seems to be a religious under-current, originally inde­ pendent of any abstruse m etaphysical speculation, flowing on from an obscure point of tim e in the religious history of India.”51 Here again, one sees a rhetoric in which tem poral priority and spatial and intellectual/ cultural inferiority are unified. Dasgupta stresses the im portant role of the laity as a force in leading Buddhist com m unities to move toward Tan­ tric ideas, insofar as he considers Tantricism both as reflecting the massm ind and as being aboriginal (again conflating early a n d /o r prim itive peoples and Contemporary com m oners).52A. L. Basham likewise describes a historical scenario in which ancient Tantricism was practiced among the “lower social Orders” before (presumably) rising to appear in the lit­ erature of th e elite.53 In th e writings of Edward Conze one can see a further elaboration of this discourse. The beginnings of Buddhist Tantrism, he writes, “go back

to the dawn of hum an history, when an agricultural society was pervaded by magic and w itchcraft, hum an sacrifice and the cult of the m other god­ dess, fertility rites and chthonic deities. The Tantra was not really a new creation, but the result of an absorption of prim itive beliefs by the literary tradition.”54 As in Tucci’s work, these ancient peoples are identified as those indigenous (autochthonous) populations th at inhabited India before the coming of th e Äryans—here identified with the Dravidian peo­ ples of the South: The erotic mysticism and the stress on the female principle owed much to the Dravidian stratum of Indian culture which, in the cult of the Village Goddess had kept alive the matriarchal traditions about the Mother God­ dess to a greater extent than the Vedic religion had done.55 Note th a t another, related hierarchy is here articulated: th a t of gender. Shimchi Tsuda, in an extrem ely influential 1978 article, likewise invokes a hierarchy associating the female and the provincial when he describes the Tantric practices of the charnel ground as “prevalent among the lowest strata of rural, m atriarchal community of the time.”56 H istorian of religions Mircea Eliade brought practically all of the various p aram eters o f this discourse of binary oppositions to g eth er w hen he rep resen ted T antrism as a “great Underground cu rren t of autochtho­ nous and populär spirituality.” “Tantrism ,” Eliade teils us, “developed in the provinces . . . w here the spiritual counteroffensive of th e aboriginal inhabitants was in full force.” And, further, th a t “here we recognize the ‘religion o f th e Mother’ th a t in ancient tim es reigned over an immense Aegeo-Afrasiatic te rrito ry and which was always th e chief form of devotion am ong th e autochthonous peoples of India. In this sense, the irresistible ta n tric advance also im plies a new victory for th e pre-Äryan populär strata.”57 It should be noted th a t this discourse continues to circulate in Contem porary scholarly accounts, as in The World’s Religionst published in 1988: Historically speaking, it seems possible to explain the literary emergence of esoteric material by reference to the changed social and cultural circumstances in northern India after the collapse of the Gupta Empire. A “normative” and intellectual superstructure collapsed which made it possible for religious undercurrents to rise to the surface and find literary expression.58

Likewise, in 2002: The religions of the Äryans, from the very beginning of their expansion, had been compelled to tolerate and even to assimilate the populär cults of the masses, including primitive erotic fertility rites and animistic beliefs. These folk cults in their canonized forms lived on through the ages under the shadow of the Äryan religions, to become overwhelmingly powerful again in the medieval period of India when Äryan culture began to wane.59 In th e light of the evident density of this discourse, its rem arkable consistency, and its u tte r lack of any but the m ost tenuous evidentiary foundation, th e critical reader will be bound to wonder: From where does this extraordinary univocality derive? And, why was (and is) this histori­ cal vision considered so com pelling by so m any m odern interpreters? I believe there are two explanations—one historical and one structural. Historically speaking, this model was—like th at of the decline of Etruria and Rome—very m uch available to m odern interpreters in the late-nineteen th and early-tw entieth-century context in which it em erged. This period witnessed a trem endous am ount of scholarly activity concerning what was called Mutterrecht und Urreligion: M other Right and Prim ordial Religion. This was the title of a 1926 collection of the most essential writings of the German classicist and scholar of ancient myth, Johann Jakob Bachofen, originally penned in the m id-nineteenth Century. The m ost influential of Bachofen’s writings was his 1861 m onograph Mother Right: An Investigation ofMatriarchy in the Ancient World in its Religious and Juridical Character.60 In this highly regarded work, Bachofen set out his theory of Mutterrecht—1“m other-pow er”—arguing th at in all societies there was a cultural stage preceding the rise of patriarchy in which social power was located in the mother, society was organized around m atrilineal descent, and the very semiotics of its cultures and its values were oriented toward w hat Bachofen considered the characteristically feminine. MutterrechU in essence, was “the law of the m aterial-corporeal, not of higher spiritual life . . . a product of the m aternal-tellurian, not of the paternal-uranian.”61 These cultures, moreover, were to be found “am ong peoples who never achieved th e level of classical culture” who followed “a more prim itive way of life.”62 Where would one find such cultural forms? They “are to be observed chiefly am ong the pre-Hellenic peoples and are an essential com ponent of this archaic culture” th at “began to decline only w ith the victorious developm ent of the paternal System.”63

If th e read er is beginning once again to experience a sense of déjà vu, she will be perfectly entitled to feel so. It is, after all, transparently obvious th a t th e elaboration of the narrative of Tantric Buddhism as an archaic and prim itive religious form ation, deriving from a m atriarchal culture associated w ith the earth and the body—a culture th at was preclassical and declined due to the victory of a later, classical, patriarchal people—is precisely predicated on this vision of the universal existence of archaic m atriarchal cultures elaborated by Bachofen and the many he influenced.64 Historically speaking, this narrative—like th at of decline— derived from one of a limited stock of story forms circulating in the contem poraneous intellectual culture of Europe. In this case, it was not a narrative derived from classical antiquity,65but one th at had its source in nineteenth-century scholarly interpretation of the classical cultures of the M editerranean basin. This historical reason for th e w idespread plausibility of this n arra­ tive as an explanation of the trajectory of Tantric Buddhism is further reinforced by th e highly structured and cohesive intellectual perspective th at underlies and finds expression in this model. As m entioned previously, all of these various narratives—in each of its distinctive articulations—are structured around a series of conceptual binaries. The ränge and coherency of these binaries have been schematized in table 2.1. As previously noted, the prim ary binaries are spatial and temporal: The Tan­ tric traditions are phenom ena of below (bedrock, undercurrent, etc.) and before (prehistorical, old, ancient). These basic binaries are supplem ented by a series of fu rth er conceptual pairs clustered around gender, ce n te r/ periphery, and oth er polarities; and all of these stand in relationship to the putative opposite of the Tantric traditions. Thus, Tantric Buddhism is not only temporally, but culturally then (primitive), not now (civilized). It is signified as subordinate in gender (m other/fem ale vs. father/m ale, etc.) and ontology (body rather than mind). It is socially exterior (villages, provinces, agricultural), rather than interior (cities, center, m ercantile). It is inferior in term s of dass (lower), intelligence (mass-mind), and religion (m agic/w itchcraft). All of these binaries correlate conceptually and are linked historically with a quasi-racial Classification (Āryan/pre-Āryan). None of these polarities will be terribly surprising: They are the classic binaries of patriarchal, literate, urban societies, such as those inhabited by most m odern scholars. They are all indisputably hierarchical; and Tantrism finds itself consistently cast as the inferior pole in this regime. This binary schematization by which the Tantric traditions are conceptualized

CATEGORY

TANTRA

NON-TANTRA

Hierarchical metaphors

D epth /Underground

Surface

Under

Over

Prehistory

History

Ancient

(Modern?)

Aboriginal

(Contemporary)

Pre-Aryan/Dravidian

Aryan

Autochthonous

(Metropolitan)

Primitive

(Advanced)

Barbarous

Civilized

Mother

Father

Matriarchy

Patriarchy

Female

Male

Body

Mind (“abstruse

Temporal metaphors

Gender metaphors

m etaphysical speculation”)

Other metaphorical hierarchies

Related (implied?) semioses

Mass-mind

Individuated mind

Lay (amateur)

Monastic (professional)

Tribes

(Developed societies)

Villages

Cities

Provinces

Center

Popular/folk

(Cosmopolitan)

Agricultural

(Trade)

M agic/witchcraft

(Religion)

(Nature)

(Culture)

(Animal)

(Human)

is precisely th at found in Bachofen’s theory of archaic m atriarchal societies. This cultural stage (note the hierarchy implicit here as well) is marked by a conceptual consistency, Bachofen insists: There was a “homogeneity of a dom inant idea” in th e ancient m atriarchal cultures, consisting of a thorough-going privileging of the mother, the female, the left, the pas­ sive, the body, earth, the moon, and so on.66 These binaries are by th eir very nature hierarchical, and encode an overarching system of valuation marked by either thoroughgoing derision or valorization. Of course, the predom inant valuation in m ost cultures (past and Contemporary) has been th at which Privileges the patriarchal pole of th e binary. These discourses function to devalue Tantric Bud­ dhism: infantilizing it, feminizing it, casting it as m arginal and obsolete. The good is th e Hellenic/Äryan rath er than the pre-Hellenic/pre-Äryan, the m ale/patriarchy ra th e r th an the fem ale/m atriarchy, religion rather than magic, th e elite ra th e r than the mass, individual rather than group, civilized rath er th an barbarous, (developed) societies rather than (primi­ tive) tribes, th e p resent ra th e r th an the past, and so on. Ultimately, of course, these articulated binaries betray a tendency th at term inates in considering one pole culture and the other nature, or even one (fully) human and th e other sub-hum an or animal. A whole cottage industry of critical theory developed in the late twentieth Century around an attem pt to erode the ideological hegemony of this network of associations—evidence itself for how hegemonic this ideology has been. Once again, then, the dom inant form of this narrative is one that devalues Tantric Buddhism as an inferior cultural form: a “survival” or obsolete vestige of a stage of cultural developm ent long since past. However, in this case, there has been more room for ideological ambivalence in the treatm en t of the narrative. That is, historically speaking, unlike the model of decline, a positive evaiuative interpretation of this conceptualization has had some play in W estern intellectual culture. In response to this (Enlightenment) conceptual structure, there arose as well a (Roman­ tic) counter-narrative. As offen happens in the case of such hegemonic ideologies, however, th e counterculture accepted the basic premises of their opponents’ perspective. They did not challenge the tem poral and cultural binaries themselves, but merely inverted th eir valuation—valorizing the ancient and the primitive (though rarely the feminine). Mircea Eliade is a prim e example of this type of thinking—an author who held up archaic thinking as a spiritual salve for a world wounded by th e excesses of m odern ideology and saw wisdom, not ignorance, in the culture of the

peasantry of the villages.67 Yet, these voices were very m uch a m inority and—given that, in making their case, they had acceded to the fundam en­ tal dualistic conceptualization of Tantrism th a t undergirded the ideology of its detractors—theirs was rather a lost cause. In the end, then, this narrative too is the product of a mode of cultural im agination populär in the late nineteenth and early tw entieth centuries, not the product of evidence and sustained argum ent. It is a projection of a culturally available narrative onto a new sphere of cultural history, based on little more than a superficial resemblance of Tantric worship of the feminine with alleged m atriarchal tendencies in ancient M editerranean societies.

TA N T R A AS MIDDLE: MEDIEVAL ESOTERICISM

Alongside these two, a third rhetorical model (representing the last of Aristotle’s m ajor plot elem ents) has em erged recently as the dom inant narrative of Tantric Buddhist history. In this mode, Tantric Buddhism is cast as a m idpoint in a larger historical m ovem ent—in a word, a "medieval” phenom enon. This discourse has had a remarkable surge in popularity over the last decade or two. One m easure of this can be taken from the relative frequency of usage of the word medieval in the treatm ents of Tantric Buddhism across th e five editions of The Buddhist Religion, originally authored by Richard Robinson and arguably the most populär introductory textbook on Buddhism in the United States since its first publication in 1970.68 The first edition contains no references to Tantra as medieval, although it does make four references to witches a n d /o r covens (terms often associated with the medieval). The second and third editions (1977 and 1982) are yet more circumspect: While they do mention wizards and medieval (each once), these are in reference to pre-Tantric Buddhist saints and Theresa of Avila, respectively. The rhetoric begins to shift by the fourth edition (1997), which features three references to the medieval or middle ages, and one each to feudatory and fiefdoms. A major swing is evident in the fifth edition (2005), however. Therein, one not only gets scattered references to wizards (once), and feudalism (thrice), but the word medieval is used no less than th irteen times in the sections on late Indian Buddhism!69 While this rhetoric of the medieval is experiencing (as it were) a major renaissance at present, its use in the historical representation of esoteric Buddhism is by no means a recent phenom enon. Monier Monier-Williams uses this term in his 1885 book on Hinduism, describing Tantricism as the

“last and worst stage of medieval development.”70 This style of representation draws in part on the association of the medieval period with magic. As Helmut Hoffman put it, “m any of the texts of the so-called Diamond Vehicle . . . [show] a stränge relationship with similar hokus-pokus pop­ ulär in Europe in the Middle Ages.”71 This usage was further reinforced by its use in David Snellgrove’s influential 1987 work, Indo-Tibetan Buddhism, in which he explicitly defends his use of this idiom. According to Snellgrove, “th e resem blance between much of th at superstition and magic [of th e Middle Ages] w ith tantric rites aiming at magical powers of a m undane kind cannot be denied.”72 In line with this thinking, Snell­ grove fu rth er advocates translating m antra as spell. However, the recent, marked increase in this type of rhetoric is doubtless attributable to the influence of Ronald M. Davidson’s Indian Esoteric Buddhism. In this work, Davidson consistently uses the term to describe the Tantric period; he further employs a constellation of related key term s (feudalism, etc.) that constitute th e essential elem ents of the im agination of the medieval in m odern cultures. Like the two previous modes of narration, here too we see a model widespread in contem poraneous W estern thought serving to structure the historical representation of Tantric Buddhism. The idea of an interm ediate period in European history appears as early as th e fifteenth Century, but the historical model of three periods—classical, middle, and m odern— is only fully attested ra th e r later, from about the late sixteenth Century.73 “Medieval” itself is prim arily a nineteenth-century usage.74 Accordingly, this Schema was im ported into the periodization of Indian history in the late n ineteenth Century. There are, as one m ight imagine, a num ber of problematical issues surrounding th e use of this periodization in tracing the history of India. Not the least of these is the remarkable inconsistency observable in its application to Indian chronology. Typically, those who employ this term to describe th e history of Tantric Buddhism m ark the beginning of the medi­ eval period by the decline of the Gupta Empire—metaphorically equating the Guptas w ith th e Romans as the keepers of the classical civilization which is alleged (by definition) to have declined in the medieval period. In a similar vein, the Annual Bibliography of Indian Archaeology for the Year 1933 conceives of th e medieval period precisely as falling between the Guptas and the Delhi Sultanate, insofar as they categorize numismatic researches thus: “a. Early Indian Coins, b. Indo-Scythian and Kushän, c. Guptas, d. Medieval, e. Moslem.”75 This usage is by no means the norm , however.

[60} NARRATING TANTRIC BUDDHISM FiGURE 2.1 D efin ition s o f th e m ed ieval p eriod in India

A.D.

200

400

600

800

1000

1200

1400

1600+

Annual Bibliography (1933) Banerjee (1940)

■ ■ ■ ■ ^ ■ ■ ■ ■ ■ ■ -1 8 0 0

Dagupta (1940

■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ -1 7 0 0

Basham (1954)

■■■■■■■■■■■■■■■■■■■■■■I

Chanana (1968) Indian History

■ ■ ■ ■ ■ ■ ^ ■ -1 7 0 7

Congress National Museum Choudhary (1974) Jha (1974) Silk (1992) Harrison (1995)

(No medieval period: “ancient” until ca. 1300, then Muslim period) I 2

Davidson (2002) Ali (2004)

1“Early medieval" ends in 1300. 2 Mahāyãna “entered its medieval period” in the second Century. 3“Early medieval” ends in 1200. 4 “Early medieval” ends in 1200.

As can be seen in figure 2.1, some authors use this term to refer to periods as early as A .D . 200, while num erous others (perhaps most historians of India) consider th e medieval period to have begun in A .D . 1200 or even later. That is, while some consider th e medieval (or early medieval) to end with establishm ent of Islamic hegemony, others consider it to begin therew ith. In this latter, very common usage, medieval refers essentially to the period of Islamic rule of India.76The Indian History Congress in fact

“formally adopted 1206 c.E. as the date when ‘ancient India’ ended and ‘medieval India’ began.”77 Other authors see no utility in distinguishing a medieval period at all. Jonathan Silk, for example, uses the term “Ancient India” to refer to the entire period before the Muslim hegemony.78 Thus, there is little uniform ity observable in the application of this periodization to Indian history. W hat there is tends on the whole to define the period as n ot beginning until after the era in which Tantric Buddhism flourished in India. In fact, th e use of this periodization seems little more consistent with regard to th e European histories it was developed to account for. There one likewise finds considerable disagreem ent among professional historians concerning the proper application of this term (if any).79 In a presidential address to the Medieval Academy of America, Fred C. Robinson, after noting th e wildly divergent periods considered medieval, is forced to conclude th a t “in light of all this disagreem ent. . . perhaps we should content ourselves with saying th at our period extends from the close of the classical period to the beginning of the Renaissance.”80 This point is crucial to understand: The medieval period is not an absolute construction, but entirely relational or structural. Its core semantic function is to facilitate drawing a contrast between this period, an earlier classical, and a later m odern era.81 Thus, although some have characterized th e notion of a medieval period as an “ultim ately m eaningless generalization,”82 the consistent, contrastive use of these term s indicates that, to the contrary, they are far from “meaningless.” They are in fact redolent w ith well-established m eanings—m eanings th a t are encoded in the structural relationship of these three term s. In particular, the asymmetry of authority in the series classical, medieval, m odern should be transparent—as should be the fact th at this periodization is precisely about authority. Much like th e dis­ courses th a t construe Tantrism as a narrative beginning, there is a clear binary (or ternary) th at structures discourse on the medieval. I have attem pted to schematize this in table 2.2. What is clear is th at the positive poles of the ternary are the classical and the m odern. The medieval represents a regrettable lapse in cultural quality. We have seen the association of Tantra, magic, and the medieval earlier. One prim ary contrast of this discourse is thus th at between religion (characteristic of the classics and moderns) and the magic or superstition of the medieval. This latter is associated with unreason, another key duality insofar as the classical and m odern periods are associated w ith reason. Similarly, the

t a b l e 2.2

Conceptual binaries of rhetoric of the medieval

MEDIEVAL

Religion

Magic Superstition

MODERN

CLASSICAL

Reason

Reason (reborn) Rational

Irrational Violent

Non-violent

?!

Catholic

Evangelical

Protestant

Ritual

Doctrine

Body

Mind

Spawned

Born

Bloated

(Normal)

(medieval-sized implements) Inhuman (medieval torture) Obsolete

Human Relevant

medieval (and, in recent works, Tantric Buddhism) is associated w ith another aspect of unreason: violence (allegedly the province of the irra­ tional). Violence is fu rth er associated w ith th e body and the latter (in term s of religion) with ritual; ritual, in turn, has well-established nega­ tive religious associations insofar as it is correlated with unreason, and (further) with th e allegedly rational, doctrinal nature of evangelical (clas­ sical) and Protestant (modern) Christianity. Fred Robinson, president of the Medieval Academy, him self notes th at “medieval is m ost often used in M odern English simply as a vague pejo­ rative term m eaning ‘outmoded,’ ‘hopelessly antiquated,’ or even simply ‘bad.’ Renaissance and classical. . . are never used in this pejorative way.”83 Robinson fu rth er enum erates the variety of alternative names for w hat came to be called the Middle Ages—w hat he calls a “host of also-rans like Barbarous Age(s), Dark Age(s), Obscure Age(s), Leaden Age(s), Monkish Ages, Muddy Ages, and of course the eighteenth-century favorite, Gothic Period!'84 The negative connotations of the first six should be evident. Lest the last should appear less pejorative than the others, one may recall th a t—as an

early-nineteenth-century com m entator noted—Gothic was “used at first contemptuously, and in derision . . . deprecating the old Medieval style, which they term ed Gothic, as synonymous with everything th at was barbarous and rüde.”85 Nor is this usage limited to English. Rather, the same connotation is to be found throughout European languages: It is integral to the concept itself.86 Some scholars—European medievalists, in particular87—may here object th at term s th a t have negative connotations in conventional discourse may be used nonetheless in scientific writing. For instance, the use of the term “cult” in studies on religions is an example of a word with negative connotations in the colloquial language, used (generally) responsibly in Professional literature. In m atter of fact, Indologist Andre Wink and Buddhologist Ronald Davidson both m aintain that the term “medieval” is not so com prom ised and stress th at legitim ate scholarly use can be made of this periodization in th e context of Indian history. Both scholars moreover are critical of some uses of the term . Davidson, for instance, echoing similar rem arks made by Wink, notes th at the word “comes w ith much baggage and can become a tool for dubious strategies.”88 Wink similarly stressed the m ethodological problems th at arise, insofar as the use of this historical analogy has (in his view) frequently driven the interpretation of the evidence, rather th an the reverse.89 On the other hand, while criticizing its ideological use by “Indianist historians,” Wink him self claims to use th e term in an “innocent” way as a “purely neutral denom inator” of “particular clusters of centuries”90 Davidson likewise claims to be able to use th e category of th e medieval in an ideologically-neutral fashion, asserting th at “we employ the nom enclature of periodization as a convenient rubric—and nothing more.”91 However, at least in th e case of Davidson, this disavowal is belied by clear patterns of usage discernible in the rhetoric he employs. Through­ out his work, one finds clusters of term s connotatively related to the medieval—such as chivalry and feudalism —being invoked to describe the culture of esoteric Buddhism. His discourse is further peppered w ith tongue-in-cheek com m ents about violence as “th e Standard medieval mode of sectling disputes”92 and the “medieval-sized” (i.e., presum ably bloated or otherw ise grotesque) implements wielded by esoteric monks.93 W hether intentional or not, this “convenient rubric” sets in m otion networks of interrelated rhetorics, such th a t—much as Wink said of the Indi­ anist historians—the interpretation of the evidence seems to proceed in Service of the m etaphor, rather than the reverse.94 The association of the

medieval with violence, for instance, reflects precisely a m etaphorical understanding, expressed (among other places) in the recent American usage “to get medieval”—defined in the Oxford English Dictionary as “to use violence or extrem e measures on, to become aggressive.”95 That the negative connotations of Davidson’s rhetoric are not lost on his scholarly readers—and are understood as being entirely of a piece w ith his in terp retatio n of Tantric Buddhism—may be observed in the m anner in which his conclusions have been read, understood, and recast by others.96 One very w ell-regarded scholar, in sum m arizing the schol­ arly contribution of Indian Esoteric Buddhism, States th at it establishes th at the Tantric scriptures were “spawned by medieval Indian säm anta feudalism.”97 However innocent or well-meaning this reader—and w hether or not it was consciously registered by him —it is clear th at th e trend of Davidson’s rhetoric has had a distinctive and predictable effect. For one, we again see connotatively interrelated term s (“m edieval” and “feudalism”) foregrounded98—in short, the predictable constellation of associations w ith the European Middle Ages. Perhaps more notably, however, in recasting Davidson’s message this author revealingly uses the word “spawn” to refer to its mode of production. Needless to say, this is not a discursively innocent term , and I believe it reflects a cogent reading of Davidson’s prose, which ff equently drips with contem pt for its subject. Twice in the Oxford English Dictionary*s definitions of spawn as a verb, the m eaning involves an attitude of condescension toward the object: that is, “in contem ptuous use, to give birth to” and “in contem ptuous use with reference to literary work, utterances, etc.” (pre­ cisely the usage found in this passage). Similarly, in its use as a noun spawn denotes “a person contemptuously regarded as the offspring of some parent or stock.” Furtherm ore, being spawned is typically correlated with teem ing or swarming (neither a flattering term ). In case there is any fur­ th er doubt about its connotations, consider the objects typically spawned in the OED’s many quotations illustrating the historical usage of the word. These include fish and fungi (the primary m etaphors for the broader usage) and, metaphorically (as in our passage), tornadoes, hurricanes, flash floods, “a race obscene,” “ill ones,” falsehoods, “seeds of sinne,” “cursed spawne of serpents,” “mischiefs and outrages,” apostates (a “spawn of Beelzebub”), those who “practice the frauds of courts,” “w retched Heretics,” heresies, libels, atheism, “monstrous and misbegotten fantasies,” and evil." Paraphrasing Richard King’s observations about the mystical, the following m ight consequently be said:

Defining the medieval then is never a “purely academic” activity (in the sense in which one means “of no real consequence”), nor can it ever be completely divorced from the historical remains of past definitions of the term .100 Rather, th e scholarly usage of medieval in reference to Tantric Buddhism trades precisely on these negative connotations. When James H. Sanford describes th e antinom ian, left-handed practices as “the m ore baroque forms of tantrism ,”101 it is abundantly clear th at we are dealing not with an in nocent tem poral periodization—a convenient rubric—but rath er with a them atics of style intim ately interwoven with an elaborate network of associations w ithin the W estern historical imagination, whose applicability to Buddhist Tantrism is based on little more th an the most tenuous of historical analogies. A central elem ent o f th e semiology of this term (as noted by Robin­ son) is th a t th e m edieval is obsolete.102 While th e ancient may be recovered (in a renaissance) for th e benefit of th e m odern, th e m edieval cannot (or should not) be. It is m erely to be transcended (or, at best, regarded as quaint). One may fu rth e r observe th a t—although some do speak of ancient India and even of classical India—one never encounters the expression Renaissance India. Given th at th e idea of the Middle Ages in th e W estern historical im agination is inextricable from th a t of th e Renaissance—w ithout th a t contrast, its use in the European context is m eaningless103—one m ay have legitim ate doubts about the coherency of speaking of m edieval India.104 That this asym m etry is integral to its Indian application is m ade abundantly clear in the rem arkable (and by no m eans uncom m on) historiography found in James Fergusson’s influential Tree and Serpent Worship (1873): Three hundred years after Buddha Aśoka did for Buddhism exactly what Constantine did for Christianity. . . . Six hundred years after Buddha, Nâgârjuna and Kanishka did for the eastern faith what St. Benedict and Gregory the Great did for the western, . . . We must go on further still for four centuries more . . . before we find our Mediæval churches quite complete . . . In the sixteenth Century after Christ came the reformation, and with it the restoration of Evangelical Christianity. In the sixteenth Century after Buddha came a reformation, but it was one of extermination of the faith, so far as India was concerned.. . . Whatever may have been the abuses and corruptions that had crept into Buddhism in the eighth

and tenth centuries of our era, they were replaced by a faith much less pure, and far fuller of idolatrous absurdities than that which it superseded. What the western reformers aimed at, was to restore the Christian Hinayâna. In the east this was not thought of, hence the different fate of the two faiths. In Europe Christianity was invigorated by the struggle, in India Buddhism perished altogether.105 In a significant respect, then, the narrative of the medieval as it is used to describe Indian history—lacking as it does any redeem ing renaissance or reform ation—ultim ately appears as merely a more subtle (and more plausibly objective) version of the model of decay. Like the models of decline and prim ordial undercurrents, it is one of a lim ited fund of stock n ar­ ratives th a t recur in m odern historiography; and the choice of this over either of the others is essentially arbitrary.

HISTORICAL NARRATIVE A N D IDEOLOGICAL IMPLICATION

What, then, remains of the project of making sense of Indian Tantric Bud­ dhism through narrative historiography? Paraphrasing Richard Evans, Davidson claims th at historians who take structuralist or postm odern discussions seriously, “som etim es. . . dem onstrate a confusion between th eir theories, on th e one hand, and m ethod and evidence, on the other.”106 It should be abundantly clear, however, th a t as long as several narratives can account for established events, there is no m ethod or evidence th at can serve as a valid criterion for preferring one among several plausible alternative narratives over others. As noted earlier, there is an irreducibly ideological elem ent in the construction of historical narrative; all the posturing about rigorous historical m ethod in the world will not im pact th at fact in the least. The move from historical evidence to historical narrative is one th a t is essentially and inescapably characterized by a precritical (or, at least, arbitrary) choice am ong available narrative forms. Such a choice is not ultim ately generated from th e data them selves, but is ra th e r a product of th e theoretical fram ework w ithin which authors have already chosen to in terp re t th a t data. As W hite puts it, “th e best grounds for choosing one perspective on history ra th e r th an an o th er are ultim ately æ sthetic or m oral ra th e r th an epistem ological.”107 That is, w hether one casts the history of Tantric Buddhism as a sad decline into a m oral cesspool (end,

bad), or th e story of th e tragic loss of a glorious religious civilization at the peak o f its creativity (end, good), as a sinister and prim al religious u n d ercu rre n t (beginning, bad), as th e survival of the pure religious intuitions o f th e early days of m ankind (beginning, good), or as a vio­ lent and superstitious interlude in the passage from the glories of clas­ sical tim es to th e achievem ents of m odernity (medieval, alm ost always bad108): ultim ately, this choice has little or nothing to do w ith the raw historical record, and everything to do with the norm ative posture the historian has chosen to adopt in interpreting th at data as data. The em ploym ent of fam iliar narratives from European historical im agination—and of analogies between historical events and figures (as seen in th e passage from Ferguson)—allows these discourses to appear as natural and unproblem atical. To readers socialized into communities in which these narratives are canonical, th at is, the m odern West and those educated in its regimes, such stories do not seem contrived. These narratives are part of th e architecture of our understanding; their use in structuring histories of unknown cultures like th at of Tantric Buddhism intuitively make sense. Or, rather, they do so until one sees them juxtaposed one w ith others such th a t the deep structure of the m odern his­ torical im agination and its role in shaping historical narrative is exposed. This is precisely w hat I have attem pted to do in this analysis. It may be hoped that, by developing a greater awareness of the literary and rhetorical elem ents th at structure their work, historians of religions may contribute to liberating th eir imaginations from the ruts of the easy, seemingly natural narratives th a t have been deployed to interpret the history of Tantric Buddhism. It may not be possible to escape narrative altogether, but at least we can endeavor to avoid the coarse, analogical models th at only allow us to see the same thing over and over again. His­ tory, it m ight be said, does not repeat itself; historians repeat themselves. In advancing this project, however, there is another perspective th at is w orth considering: M odern scholars are neither the only nor the first to craft narratives by which to identify and interpret the esoteric traditions. Long before the first scholars in the nineteenth Century articulated narratives to make sense of Tantric Buddhism for themselves and their contem poraries, this task was undertaken by the authors of the esoteric Buddhist traditions themselves. A comparative look at the structure and function of these narratives may thus allow the critical historian greater purchase on the project of narrative historiography across cultures.

[3] GOING NATIVE T R A D IT IO N A L H IS T O R IO G R A P H Y OF T A N T R IC B U D D H IS M

o d ern sch o la rs

M

h av e o n ly r e c e n tly b e g u n to d e v o te a tte n tio n

to th e q u e s tio n o f h o w in d ig e n o u s so u rc e s u n d e r s ta n d a n d d e p ic t th e o rig in s a n d h is to ry o f th e ir tra d itio n s , a lth o u g h th is

q u e s tio n fig u re s p r o m in e n tly in e s o te ric B u d d h ist lite r a tu r e . V irtu a lly n o n e o f th e S ta n d a rd sc h o la rly w o rk s o n I n d ia n B u d d h ism d e v o te a n y a tte n tio n w h a ts o e v e r to th e q u e s tio n o f h o w th e c o m m u n ity th a t h o ld s th e m sa c re d u n d e r s to o d th e h is to ry o f th e s e te x ts , b u t m e re ly pass o v er th e q u e s tio n in s ile n c e .1W h a t little c o m m e n t o n e d o es e n c o u n te r o n o ccasion is la rg e ly c o n fin e d to d e p ic tin g th e in d ig e n o u s v iew as re d u c ib le to th e claim t h a t th e T a n tra s w e re ta u g h t b y th e (h isto ric al) B u d d h a h im self, c e n tu rie s b e fo re th e C h ristia n e r a —a fa n ta sy c le a rly b e n e a th th e d ig n ity o f m o d e rn h isto ric a l Science e v e n to e n te r ta in .2 We h av e p re v io u sly se e n how B e n o y to sh B h a tta c h a ry y a a t tim e s m a in ta in e d (a lth o u g h n o t w ith o u t s u b s ta n tia l s e lf-c o n tra d ic tio n ) th a t su ch T a n tra s as th e M a ñ ju śrīm ã la ka lp a 3 d eriv e “p ro b a b ly fro m th e tim e o f th e B u d d h a him self.”4 In m o re c a u tio u s m o m e n ts, th o u g h , B h a tta c h a ry y a b ra c k e ts th is as th e n a tiv e view ; th a t is, h e claim s th a t th e h isto rio g ra p h ic a l c o n c e it o f th e e s o te ric B u d d h ists w as to “in tro d u c e th e ir d o c trin e s in to B u d d h ism b y th e c o m p o sitio n o f a n ew S ańgīti o r co lle c tio n o f v erses, all o f w h ic h w e re to b e ta k e n to h av e b e e n d e liv e re d by th e B u d d h a in a n a ssem b ly o f th e faith fu l.”5 As w e sh a ll see, th e r e a re in d e e d p a s sa g e s in th e T a n tra s a n d in th e in d ig e n o u s n a r r a tiv e s o f th e h is to ry o f e s o te r ic B u d d h ism t h a t a s s e r t th a t th e s e s c rip tu r e s w e re ta u g h t b y Ś ā k y a m u n i h im s e lf.6 H ow ever, it is im p o r ta n t to u n d e r s ta n d t h a t th e r e is c o n s id e ra b ly m o re to in d ig e ­ n o u s h is to rio g r a p h y o f e s o te ric B u d d h ism th a n m e re ly a b a ld ly -s ta te d a n a c h ro n is m . We w ill, a c c o rd in g ly , b e g in b y c o n s id e r in g th e r ä n g e o f

narratives crafted by Tantric Buddhists through which they conceived of the innovative ritual and cultural forms th at united their com m unities as entirely of a piece w ith the older Buddhist traditions. To do so is not only essential to u nderstanding the attitude adopted by esoteric Bud­ dhist groups toward cultural innovation, but provides an invaluable per­ spective on th eir m ost fundam ental self-understanding. The narratives of th e revelation and spread of esoteric Buddhist teachings constitute an essential resource in reconstructing the “distinctively hum an world, [the] ‘second en v ironm ent’”7 of Indian esoteric Buddhists—the shared world of m eaning in which they im agined them selves and into which they socialized new m em bers. Much as Georges Dreyfus has argued regarding th e role of scholastic treatises such as the Abhisamayãlamkāra in Contem porary Tibetan m onastic com m unities,8 the fram e stories of Buddhist scriptures serve to locate those revelations (and th eir target audience) in m eaningful contexts—contexts th a t reinforce communal identity and w ithin which the various semiotical transactions th at con­ stitute th e religious and social life of the group are understood as reasonable.9That is, ju st as we have noted in chapter 2 th at the use of historical narrative in m odernity reveals more of the contours of the m odern cul­ tural im agination th an Inform ation about events in the past, so too does the use of narratives by Tantric Buddhists gives us insight—beyond their own ostensive claims—into the subtending historical imagination of reli­ gious com m unities in late first-m illennium India. Our prim ary focus of critical attention begins to shift here from the scholarly subject to its object. The introduction and the preceding two chapters were largely devoted to the task of “w hat Victor Shklovsky term ed ‘defam iliarization '—making the familiar seem stränge in order to enhance our perception o f the familiär!'10 That is, we took as the object of our rhetorical analysis the very discourses used to represent Tantric Bud­ dhism in order to dem onstrate th at the models taken for granted in mod­ ern academic research are themselves not only contingent and historical, but reflect ra th e r m ore of the constitutive im agination11 of the m odern interp reter th an of th e object they purport to explain. In so doing, the aim was to identify and warn of ruts in the scholarly road, so we can be free to follow new approaches. The task for the rest of the book is to bring an equally acute analysis of rhetoric to the study of Tantric Buddhism itself as an object. Accordingly, we m ust take up the other challenge in the analytical dialectic of th e history of religions: to make the stränge seem familiar, in order to enhance our understanding of the stränge.

Until ra th e r recently, th e study of th e indigenous historiography of Indian Buddhism has not been marked either by its sophistication or its herm eneutical generosity.12 It m ight be said (albeit uncharitably) th at it has often been undertaken w ith th e attitude of an adolescent toward parents: Get w hat resources you can from them , but contradict (or deride) them w henever possible. Yet, there is significantly more to be garnered from these docum ents than m erely lineage lists or other nuggets of his­ torical data th at can be freed from the dross of their narrative contexts. These narratives are them selves a crucial subject of inquiry in their own right. Just as the m odern narratives analyzed here contribute to constructing and m aintaining a meaningful world for th eir readers w ithin which to situate and interp ret Tantric Buddhism, so too do the traditional n arra­ tives perform m uch the same function for th eir own intended audience. True, they invoke divine and semidivine agents and heavenly locations th at th e m odern critic cannot countenance as real. The m ere presence of the miraculous, the divine, or the transcendent in these accounts does not, however, ipso facto set them beyond the appropriate ränge of serious scholarly attention.13As Bruce Lincoln has observed: Processes of authorization that invoke the divine or transcendent at some crucial point of their Operation seem typical of societies in which the foundational assumptions... made normative by the European Enlightenment have not acquired hegemonic Status. It would be foolish—not to say presumptuous and ethnocentric—for those of us who stand on one side of this divide to underestimate the complexity, seriousness, efficacy, and importance of the differing ideological styles more commonly employed by our counterparts located on the other.14 Although the constitutive im agination of m odern historical research does not countenance the divine or the miraculous, th at of the esoteric Bud­ dhist com m unities of the late first m illennium clearly did. The ideologi­ cal style thus constituted is, as we shall observe, by no m eans as simple as a claim th a t “Buddha really did say it”; and the “complexity, serious­ ness, efficacy, and im portance” of these strategies in the docum ents of Indian esoteric Buddhist com m unities have a significant role to play in the scholarly reconstruction of these traditions and their relationship to other contem poraneous Indian knowledge Systems. In this chapter, we will survey a ränge of prim ary sources th a t bear on the indigenous historiography of the Buddhist Tantras. We will begin

by reviewing the conceptions of history and scripture shared commonly by Indian Buddhist com m unities before attending to the particular innovations o f th e early Mahāyāna. Against this background, we will examine th e historical indications to be found w ithin the esoteric scriptures themselves, noting th eir general ideological consonance with the preexisting Buddhist symbolic order. Subsequently, we will tu rn to the more developed and individualized narratives to be found in the parallel Cor­ pus of esoteric Buddhist literature not considered to have been revealed by buddhas.15 As the national literatures of the countries to which eso­ teric Buddhism was exported in this period—China, Tibet, Japan—also contain sim ilar narratives, these too will be explored for th eir insights into th e stru ctu re of esoteric Buddhist historiography. These observations will form th e basis for reflection on trends in Indic historiography more broadly, and the socio-epistemic role of such discursive practices in the form ation and m aintenance of the traditional Sciences. Continuities between the intellectual practices of the ancient Indians and Greeks, and those o f m odernity, will be discussed in light of the overarching problematics of history and cultural Interpretation.

HISTORIOGRAPHY A N D COSMOLOGY IN EXOTERIC BU D D H ISM

In order properly to com prehend the indigenous historiography of eso­ teric Buddhism, it is necessary first to grasp th e cosmological, “buddhalogical,” and historiographical realignm ents effected by the early M ahāyāna m ovem ents.16The conceptualization of the Situation of suffering hum anity in th e cosmos underw ent at their hands a major spatial and tem poral realignm ent th at deeply conditioned the subsequent course of the tradition. However, one m ust also be careful not to exaggerate the degree of innovation in the Mahāyāna cosmos. The remarkable innovations in Buddhist historiography witnessed in the years ca. 2 0 0 b . c . - a . d . 500 took shape in th e context of an already well-established program adopted by several currents within the early communities. According to some trends in early Buddhism, prior buddhas—including Gautama after his passing—were in general unavailable for further spiri­ tual assistance. The incredible salvific power attributed to the Buddha— such th a t th e liberation of beings from the endless round of suffering was frequently thought to have been occasioned by his mere presence—was considered (with few exceptions) confined to th e short window of his

limited sojourn of eighty years in his final incarnation. After his passing, this power was thought to persist in an attenuated fashion in his bodily relics, speech relics (in the form of the teachings of dharm a), and (for a few generations at least) those considered liberated saints in his Com­ m unity (sańgha). All of these, it was thought, would fade in th e course of some centuries, leaving the world a spiritual wasteland until the coming of th e next buddha, M aitreya.17 Such enlightened beings, however, were believed to come only w hen the fluctuating lifespan of hum anity was one hundred years: a circum stance th at was thought to transpire at most every thousand years.18 Thus, th e Buddhist com m unities who adopted this perspective considered themselves as living in a world between buddhas, preserving the precious few, rem aining fragm ents of Gautama Buddha’s legacy in text and m onum ent, and seeking to em ulate his ideal of moral discipline and m ental developm ent for the good of seif and other, so as to prepare themselves for the advent of Maitreya in some future life. This particular form of Buddhist cosmological historiography contains im plicit assum ptions regarding the relationship of the faithful to the divine and the nature of the canon of revealed word (buddha-vacana) considered authoritative by the community. This latter was in principle conceived of as a closed corpus of those teachings given by the Buddha during his lifetime and rem em bered by reliable sources, as signaled in the expression found at th e beginning of all scriptures “thus have I heard” (evam mayã śrutam). Likewise, in the absence of a living buddha between the passing of Gautama and the arrival of Maitreya, Buddhist com m uni­ ties were led to focus th eir efforts on presencing th e departed Buddha through media such as corporeal images, production and recitation of texts, and construction of reliquary shrines (stüpat caitya). Through interaction w ith these m ediated forms of the enlightened presence, it was claimed, one could obtain “welfare and happiness for a long tim e” or “birth in a happy state or a heavenly world.”19 However, these perspectives were by no m eans universally held by early Buddhist com m unities.20 In term s of authoritative scripture, for instance, a num ber of exceptions were adm itted. W hether or not anything in any of the extant collections was actually taught by the person alleged to have initiated th e Buddhist traditions (“th e Buddha”) is an open and ultim ately insoluble question for m odern scholars.21 Significant disputes existed even am ong th e early com m unities. However, th ere were also clear cases in which these com m unities agreed to include in the developing corpus of Buddhist literature teachings th a t were not even alleged to

have been spoken by th e Buddha. In fact, a considerable degree of flexibility was allowed in this regard. For instance, in the Samyutta Nikäya of the Theraväda canon (päh'),22 num erous teachings are cast narratively as having been given by worldly divinities—even on such elevated subjects as the proper practice of the monk (bhikkhu), how the properly religious (here called th e brahm in) should conduct himself, and the nature of the enlightened saint (arahant). Similarly, in the Digha Nikäya, a powerful spell of protection to be used by monks m editating in the forest is given to the Buddha by four kings of th e gods.23 Consequently, as Paul Harrison has noted “even the M ainstream canons contained teachings believed to have been preached by deities, but nevertheless accepted as buddhavacana [Buddha Word].”24 Etienne Lamotte has likewise noted a universal acceptance in literature of the Buddhist monastic discipline (vinaya) of authentic dharm a being taught by a wide ränge of special beings other than th e Buddha.25 The Buddhist communities of the waning centuries of the first millennium b .c . also witnessed a major revolution in scriptural thought, in which a large corpus of m uch later m aterials was incorporated into the Buddhist canonical collections. I am referring, of course, to the Abhidharma, the writings th a t system atized the sometimes diffuse teachings of the sütras into a uniform vocabulary and intellectual framework. Although clearly a later synthetic product, most Buddhist communities came to accept these works as (in some sense, at least) authentically the word of the Buddha, and they were incorporated conceptually as a third “basket” of scriptures placed on a position of equal (or even greater) authority to the sütras and vinaya.26 Thus was born the concept of the Tripitaka (“Three Baskets” of scripture) we know today, which superseded the older twofold Buddhist corpus of Dharma-Vinaya (“Teaching-Discipline”). Accordingly, as Peter Skilling has observed: If we examine the history of the transmission of the Abhidhamma closely, we see a process in which a set of texts was gradually naturalized and canonized . . . This is only one example of the inadequacy of the “very idea” of canonicity as a primary tool of analysis in the historical study of Buddhist literature. “Canons” were continually redefined, refreshed, and reinvented. Yesterday’s apocryphon becomes today’s canonical text.27 Notably, th e Theraväda tradition took a unique course in rationalizing their acceptance of this corpus. Alone am ong the early traditions, the

Mahāvihāravāsins (whose texts com prise virtually the whole of the Con­ tem porary Theravāda canon) innovated a remarkable narrative of revelation for th eir Abhidhamma in which the Buddha was said to have taught these scriptures in the heavenly realm of Trayastrim śa while visiting his late m other th ere.28 In so doing, they joined o th er com m unities in creating a distinctive narrative by which to conceive of the intervention of the Buddha in the world. That is, his salvific power was not confined merely to th e forty-five years of his teaching career in the quotidian hum an world, but could also reach Buddhist com m unities through the m ediation of divine realms. The narrative is careful, however, to link this interven­ tion with (and situate it in the accepted teaching career of) the Buddha. In addition to accepting a more expansive conception of the pedagogical reach of the Buddha than typically assumed, early Buddhist com m uni­ ties were likewise not uniformly as buddhalogically parsim onious as the common ideal-typical characterization would suggest. All th e Buddhist communities of which we know allowed for the existence of a num ber of buddhas other than Gautama. In fact, in the view of many early Buddhist schools (with the notable exception of the M ahāvihāravāsin branch th at came to dom inate later Theraväda), buddhas were considered “infinite in both space and tim e”29—a view th a t becam e norm ative for the later Mahāyāna movements. However, eyen am ong Contemporary Theraväda com m unities—who only adm it of one buddha of the present—the following verse appears in widely recited liturgies: The buddhas of the past, and those yet to come, Those [pl.] of the present, too—[to these] I pay homage always!30 All of which suggests th at throughout the course of history, by far the m ajority of Buddhist com m unities considered them selves to inhabit a world in which th ere were m ultiple buddhas not only in the past and future, but also in the present. With the rise of the cultural currents th at coalesced into the Mahāyāna movement, both of these aspects of the early traditions became central com ponents of an evolving perspective. On th e one hand, the conceptual architecture of th eir cosmology considered the universe (made up of num erous worlds) to be inhabited by countless buddhas and aspiring buddhas (bodhisattvas)—all active in teaching, disciplining, and liberating beings. They were also notably prolific in the creation of new scriptural materials: So m uch so, th a t at least one Contemporary scholar has

characterized th e M ahāyāna as consisting at its core prim arily in the production of new scriptures.31 In so doing, the authors of the Mahāyāna accounted for this efflorescence of scriptural composition as the product of processes of ex post facto (yet authoritative) revelation similar to those we have seen here. Some, for instance, em ulated the Mahāvihāravāsins in attrib u tin g new scriptures to revelation in a heaven. The scripture called th e “C hapter on Trayastrim śa” (Trayastrimśatparivarta) is a classic instance of this. This work locates its own revelation during the very same journey o f th e Buddha to teach his late m other in heaven. In term s of content, however, it presents a distinctively Mahāyāna perspective on the teachings th e Buddha gave at th at time, attributing lectures on voidness and the perfection of wisdom to this occasion.32 There were, very likely, M ahāyāna com m unities th a t claim ed th a t their scriptures had come to them through conventional channels, having been tau g h t by th e Buddha during his lifetime, then learned, compiled, and transm itted. However, it would seem as if this m odel—which we may call th e “bare historical” m odel—was not the predom inant nar­ rative created to account for th e M ahāyāna, its com m unities, and the revelation o f its scriptures. It is im portant to recognize in this regard th at not all narratives th a t feature teaching by th e Buddha Śākyamuni reflect an identical historical perspective. That is, th e m ere assertion th at groups considered th e ir teachings to be “tau g h t by th e Buddha” does n o t evince an adequate level of critical analysis. As we have seen here, b o th Śrāvaka and M ahāyāna apocrypha had had recourse to the notion o f m iraculous teachings by gods or in heavens to account for scriptural innovation. However, even in cases in which th e teaching is said to have taken place on Earth (Jambudvīpa)—and thus conform spatially to th e conventional n arrativ e—th e tem poral param eters of the M ahāyāna narratives are shifted in such a way as to fundam entally reconfigure th e ir m eaning. In line w ith th e ir greatly expanded vision of th e cosmos and th eir equally ram ified buddhalogy, th e burgeoning corpus of M ahāyāna literature added a num ber o f new features to th e world presupposed by their narratives. Essentially, these authors incorporated into these new Works a substantially refigured m etanarrative of revelation, in which Śãkyam uni’s Ś rāvakayāna teac h in g —although still accorded g reat respect—is decen tered in such a way as to provincialize th e conven­ tional life o f th e Buddha as a site of scriptural revelation.33 That is, the Primary focal point of revelation is no longer m erely the forty-five-year

teaching career o f th e Buddha Śākyamuni. Rather, th e locus of articulation o f th e d h arm a is reconfigured to encom pass all th e m any buddhas o f past, present, and future. For instance, in the Śūramgamasamādhi Scripture, a divine being is said to arrive from another world, th e Abhirati Universe of th e Buddha Aksobhya. In characterizing th a t world for his disciple, th e Buddha observes th a t “there, [Aksobhya] always expounds th e Śūramgamasamādhi. 0 D rdham ati, all buddhas w ithout exception expound the Śūramgamasamādhi.”34 That is, the tru e dharm a (saddharma) is being constantly taught by bud­ dhas in an array of universes. Śākyamuni, in this perspective, is merely a local—and som ewhat peripheral—tran sm itter of this teaching. Further, this tru e dharm a consists of th e teachings of the way of the bodhisattvas (bodhisattvayäna), not the m ainstream teachings of the way of the śrāvakas (śrāvakayāna). These latter, in contrast, are depicted as the unique teach­ ings of Śākyamuni, th e expression of his distinctive pedagogical genius, devised for the limited and unim aginative beings of his especially recalcitrant buddhafield.35 The Scripture ofthe Lotus Blossom ofthe True Dharma (Saddharmapundarika Sütra; th e “Lotus Sütra”) is exemplary in this regard. It contains a variety of M ahāyāna narratives th at speak of new scriptures as having always been taught, or (at least) as having been taught long ago by th e great buddhas of th e past.36 The widely (and rightly) praised w ork of Paul H arrison has draw n a tte n tio n to th e rem ark ab le c o n trib u tio n of th e Scripture o f the Samädhi o f Face-to-Face Confrontation with the Buddhas o f the Present (Pratyutpannabuddha-sammukhãvasthita-samādhi-sūtra) to th e M ahāyāna articulation of the source of new scriptures.37 This work makes sim ilar claims about being th e eternal teaching of all th e buddhas.38 Its novel co n trib u tio n , however, consists in th e fact th a t this relatively early M ahāyāna scrip tu re39 provides a system atic presentation of two modes of revelation th a t may be used to rationalize novel scriptures. On th e one hand, as its title indicates, it contains a m editational m ethod by w hich a p ra c titio n e r may invoke th e presence of a buddha, receive teachings, and re tu rn with them to th e ordinary world.40 If practitioners who are well-disciplined “concentrate th e ir thoughts w ith undistracted m inds on th e Tathāgata Amitäyus for seven days and nights, then, w hen a full seven days and nights have elapsed, they will see th e Lord and T athägata Amitäyus.”41 This is presented as a real and direct encounter w ith th a t buddha, for it specifies th a t if th a t practitioner does not m eet the buddha they are invoking th a t day, they will receive a dream vision

at night. This co n centration notably claims the specific benefit of providing new teachings, for they see that Lord and Tathāgata Amitäyus . . . and also hear the Dharma. And they retain, master and preserve those dharmas after hearing them expounded [0]n emerging from that samädhi the bodhisattvas expound at length to others those dharmas, just as they have heard, retained, and mastered them.42 The Samädhi o f Face-to-Face Confrontation also speaks of scriptures being sequestered in caves or other secret places to be extracted and revealed to the world in later times by bodhisattvas who have pledged to reincarnate th en for th at purpose. In particular, it is w orth noting th at the sütra describes its text as being placed inside a stüpa after the death of the Bud­ dha, to be guarded by gods and snake-spirits (näga) until the time arrives for them to be re-revealed by the predestined bodhisattvas.43 In th e course of th e developm ent of this new cosmo-historical per­ spective, th e degree of flexibility allowed in the visionary revelation of new scripture at tim es reached extremes. For instance, the Dharmasamgīti Scripture says: For the one whose spiritual inclination is perfect, Lord, if there are no buddhas, the sound of the Dharma issues from the vault of the sky and from walls [and] trees. For the bodhisattva whose spiritual inclination is pure all instructions and precepts issue from his own internal dialogue.44 That is, ultim ately speaking, because from a fully-developed Mahāyāna perspective—such as is adopted throughout the esoteric Buddhist tradi­ tions—the entire fabric of reality is made of buddhas (buddhamaya), real­ ity is only m ind (cittamätra), and the minds of all beings are ultim ately enlightened (possessed of tathägatagarbha), the power of the enlightened ones need not be m ediated through so-called “historical buddhas. It radiates from the very substance of a world th at is mind and buddha. This conception of the om nipresence and omni-availability of divine revelation may be found expressed throughout various esoteric ritual practices. The vast preponderance of these rituals is precisely oriented toward m anipulating the porous boundary between quotidian reality and the divine presence. For instance, the early Buddhist esoteric work, the Mañjuśrīmülakalpa (MMK) devotes considerable attention to its own ritual

for th e p ractitioner to translate herseif into the presence of buddhas. The suprem e practice (uttamasädhana) set forth in this scripture accomplishes much the same goal as we saw in the Samädhi ofFace-to-Face Confrontation.45 In this case, the visualization (and translation) is aided by the use of a depiction of the deity (in this case, Śākyamuni himself) painted on a cloth scroll. During the rite, the practitioner sets this scroll up in front of him self and makes offerings of various substances before it. The scroll is th en supposed to blaze with light, w hereupon the practitioner circum am bulates it thrice. Taking hold of th e scroll, he is th en able to soar up to th e heavens, w here he will m eet and receive teachings from the Buddha Sańkusum itarājendra and thousands of bodhisattvas, including Mañjuśrī.46 It is w orth noting th at this procedure too may be used to account for the existence of teachings attributed to heavenly realms. Likewise, the ritual of consecration (pratisthä) of divine images also (in principle) allows devotees to interact with the divine face-to-face and is predicated on th e notion th at buddhas are available for intervention in the ordinary world.47 However, rath er than bring the practitioner to the divine (as in the suprem e practice of the MMK), consecration brings the divine to th e practitioner by inviting a buddha or bodhisattva to inhabit h er/h is likeness in painted or sculptural medium.48 It is most fascinating to observe that, among the scriptures th at trea t of consecration, there are some th at explicitly address the mechanisms through which it is thought to be efficacious. In this connection, th e question arises concerning the very need for consecration in a reality th a t is always already sacred.49 That is, in line with the notion th at buddhas are om nipresent in reality, why would one need to consecrate an image th at is already a locus of divine power ipso facto its m ere existence (as, indeed, are prosaic things such as rocks or old rags)? The Concise Consecration Ritual Tantra States, “all th e buddhas are present (*pratisthita) w ithout any ‘presencing’ ritual (pratisthä: a.k.a. consecration); like space they are everyw here. . . prim ordially unborn—how could [they] be ‘presenced?’ [The rite] is only done so th at beginners will understand.”50 In sum, th e co nstitutive im agination of the M ahāyāna com m uni­ ties was quite different from th a t reflected in th e ideal-typical notion of a Buddhism dependent upon the ephem eral teachings of Śākyamuni. Rather, these groups conceived of the divine power of the buddhas as being om nipresent in space and time, and of the teachings of the dharm a as being universally available to those spiritually so attuned. There were evidently m any among the contem poraneous Buddhist communities who

were not entirely comfortable with the direction things were moving. The rise of widely-accepted visionary revelations was recognized as a new innovation and troubled the more conservative elem ents.51 However, the Mahāyāna com m unities in which esoteric Buddhism developed took this vision of the world and revelation for granted.

HISTORIOGRAPHY A ND COSMOLOGY IN ESOTERIC BU DD H ISM

Based upon th e foregoing, one can abstract five models available to eso­ teric Buddhist groups (of which the Tantric movements were a part) by which to construct an account of the revelation of their innovative new scriptures. Following com m on precedent, esoteric Buddhists could (l) merely assert th a t they were taught by Śākyamuni, period; or, (2) claim that they were taught by Śākyamuni in some other world. Alternatively, following tren d s populär am ong the various prolific M ahāyāna “proclaimers o f D harm a” (dharmabhänaka),52 they could (3) m aintain th at the esoteric M ahāyāna m ethod of m antra (mantranaya) represents the ancient and honorable teachings of all the buddhas, more venerable and profound th an the transitory teachings of either the Śrāvaka Way or the exoteric Mahāyāna m ethod of the transcendent virtues (pāramitãnaya). To this general model they further could add the details th at (4) the scrip­ tures were retrieved from sacred caskets hidden in stüpas or caves for future, predestined revelation, or (5) th at they represent teachings presented anew in direct encounters with buddhas other than Śākyamuni (or even from Śākyamuni himself).53 The various historical conceptions put forth by th e esoteric Buddhist com m unities took shape in relationship to these precedents. In m any ways, as we shall see, the esoteric tradi­ tions did not innovate beyond the basic models of scriptural revelation accepted already by the exoteric Mahāyāna. Except in local detail (particular names, places, etc.) the esoteric narratives follow these five wellestablished patterns. The Ganapatihrdaya, for instance, takes a rather prosaic view of things, presenting th e bare historical viewpoint (no. l). This teaching presents a special recitation th a t purports to accomplish ju st about anything. It is explicitly stated to have been delivered by the Buddha to his faithful butler Änanda in the conventional m ainstream scriptural setting of Räjagrha.54 The Amoghapãśa Mahākalparāja, which represents a kind of transitional exo/esoteric scripture, also conforms in its basic narrative

stru ctu re to th e models accepted in M ahāyāna cultural circles. In this work, Śākyamuni clearly appears to be th e teacher; however, th e action is set on m ythical M ount Potalaka, in th e palace of th e bodhisattva Avalokiteśvara, thus conform ing to the m otif of revelation in a heavenly realm (no. 2).55 At first glance, like the Ganapatihrdaya, the Mahãśītavatī Dhāranī seems to depict itself in a conventional narrative framework, also being set in Räjagrha, and addressed to th e Buddha’s son Rähula. However, a closer look reveals th a t this work clearly adopts strategy no. 3, positioning itself as a transcendental teaching given by all the buddhas. This is stated explicitly at conclusion: Furthermore, this very Mahāśītavatī knowledge-spell was spoken [in the past], will be spoken [in the future], and is spoken [in the present] by Lord Buddhas equal to the number of grains of sand in ninety-one Gan­ ges rivers.56 Likewise, th e very in flu en tial Chanting the Names o f Mañjuśrī (Mañjuśrīnãmasamgīti) seems to advance a bare historical model insofar as its authors claim th a t it was “spoken by the Lord, the transcendent lord Śākyamuni.”57 The location of th e teaching is not specified, though it is requested by and given to the esoteric hero Vajrapäni, which m ight imply a transcendental location of some sort. Although an aggressively Tantric work in many ways, it articulates the general Mahāyāna model th at these traditions are the overarching, core teaching of the buddhas of all times. This is clear at the outset of the work w here th e provenance of the teach­ ing is clarified: The supreme Chanting of Names, good in the beginning, middle and end, Profound in meaning, of exalted meaning, great in meaning, unequalled, auspicious— That which was spoken by the buddhas of the past, and will be spoken by future buddhas Is also taught by the buddhas of the present again and again.58 This is not the exclusive teaching of Śākyamuni, but is an eternal revela­ tion th at is given not only by num erous buddhas in past and future, but by m any in the present (grammatically, at least three, but presum ably countless) all the tim e.59

Vajrapāni also appears in the Vajravidāranū this tim e as the teacher of the scripture. However, in doing so, Vajrapäni acts as a Surrogate for all buddhas and bodhisattvas in a m anner accepted even in early communi­ ties.60 In the opening sequence, the Lord (bhagavän) constitutes the world as made of adam ant (vajra) and by means of his spiritual power (anubhâva) causes Vajrapäni to become absorbed in the adam antine samädhi. Then, by th e power of th e Buddha, Vajrapäni declares special m antras th at derive from th e blessing of all the buddhas and bodhisattvas.61 As we shall see, this Vajrapäni becomes a favored character in esoteric Buddhist accounts of the revelation of their scriptures. In addition to the shared model of the eternal teaching of the Mahāyāna, however, th e esoteric traditions also advanced narratives of revelation cast in m ore prophetic or visionary term s (modes 4 and 5). The extremely influential Mahävairocana Tantra (MVT), for instance, takes its narrative completely out of the ordinary world, presenting itself as the teaching of the cosmic buddha Mahävairocana, not Śākyamuni. The setting is clearly divine and reflective of a notion th a t this revelation was not one th at occurred in the conventional hum an world, but in a miraculous realm: Thus have I heard at one time the Lord was residing in the residence of all transcendent lords, the great and extensive adamant palace of the realm of reality, the assembly-hall of all vajradharas, the playful emanation of the zeal of the transcendent lords, without center or periphery, [with] a soaring dorne, a storied palace made of and extremely beautifully adorned with [veritable] kings of gems, seated on a lion-throne composed of the bodies of bodhisattvas. This is not to say, however, th at this scripture does not place its teach­ ings in the historical framework of the life of Śākyamuni. Rather, the MVT claims—as several o th er subsequent scriptures do—th at it was precisely with this teaching th a t the Buddha became enlightened: The m antra methods are incomparable. By their means, the savior Śākyasimha overcame the irresistible host of Mära and [his] vast army.62 Nonetheless, this claim needs to be understood in context. Rather than stressing the ownership of Śākyamuni (here called Śākyasimha, the Lion °f the Śākyas”), this claim serves two functions: Most coarsely, it validates the new esoteric m ethods as central to the enlightenm ent of the prime

(current, or local) exemplar of buddhahood; more subtly, it serves further to decenter and provincialize Śākyamuni, who is here cast as a disciple of o ther buddhas.63 This them e is given detailed elaboration in the Sarvatathägatatattvasamgraha (STTS: “Compendium of the Reality of All Transcendent Lords”). This scripture foregrounds a narrative of the enlightenm ent of Śākyamuni presented at th e outset of th e work. In this tale, th e m an destined to become the historical Buddha is encountered before his enlightenm ent experience as the bodhisattva Sarvärthasiddha. He is seated at the foot of the bodhi tree, absorbed in the unshakable samādhi (ãsphānakasamãdhi), w hereupon he is visited by a great com m unity of transcendent lords appearing in th eir m iraculous forms (sambhogikakäya), invisible to ordinary beings. They ask him, “How can you become a buddha w ithout the knowledge of all the transcendent lords?” Sarvärthasiddha emerges from his samädhi and respectfully requests, “Make me understand, Lords! How shall I find out w hat reality is?” All the transcendent lords urge him, “Find out, 0 Noble One, by means of the m editative focus th a t attends to your own mind.” They th en teach him five m antras, initiating him as the eso­ teric bodhisattva Vajradhätu, in which form he th en becomes enlightened. Upon his enlightenm ent, they escort him to the pinnacle of Mt. Sumeru to a storied palace w ith an adam ant-, gern-, and jew el-encrusted dorne, where they seat him on the lion throne of all transcendent lords.64 That is, he is taken to a place very much like the setting for the revelation of the Mahävairocana Tantra. In fact, the homology is straightforward: The four transcendent lords other than Vairocana seat themselves in the Cardinal directions, forming a maņdala around Śākyamuni, who is subsequently referred to as the transcendent lord Vairocana. In this acccount, the Buddha is not merely presented as a conduit for the ancient and eternal teaching of all the buddhas; rather, he becomes a disciple of all th e buddhas who occasion his enlightenm ent through teaching and initiation. As a consequence of this refram ing, Śākyamuni becomes a subordinate, albeit significant, character from a spiritual point of view: Luke Skywalker to th e Jedi Council of the cosmic buddhas, as it were. The transcendent lords are the real teachers, who reveal them selves to th e spiritually prepared in miraculous visions in th eir beatific forms (sambhogikakäya). There is an im plicit homology: Sarvärthasiddha was initiated into the practice ju st as the Contemporary disciple/practitioner is. Thus, ju st as the Buddha obtained a visionary teaching and initiation, the same is available to anyone who is prepared for the practice.

After th e Mahãvairocana Tantra, m ost of the esoteric scriptures cast their narratives using elem ents from modes two, four, and five: Revela­ tion in heaven, ancient teachings hidden away, or revealed in visionary encounters w ith buddhas in th eir beatific forms. Thus, for instance, the Sarvadurgatipariśodhana Tantra (“The Purification of All Bad Rebirths Tan­ tra”) is taught by “the Lord” (bhagavân) in the “grove of the highest delight of all th e gods” (sarvadevottamanandavana). This venue is presumably in the heaven of Trayastrim śa, because the entire audience is composed of deities and Indra inquires about the fate of a fellow god who passed (died) “from here, the abode of the gods of Trayastrimśa.”65The Lord in this case turns out to be Śākyamuni, but this is clearly another example of teach­ ings delivered outside the conventional, m ainstream Buddhist historicocosmic framework. The historical narrative of the Catuspttha (“Four C hapters”) is most interesting as it reveals a three-fold typology of sites of revelation. According to this scripture: The great king of Tantras, the glorious Four Chapters [Catuspītha] in 180,000 [verses], was uttered by the Lord, Glorious Vajradhara, in the congregation of deities of the Śuddhāvāsa Heaven.66 Then Vajrapāni, having Condensed [this Ur-Tantra] into the King of Tantras in 12,000 [verses] in the glori­ ous land of Odiyäna67 [where it was comprehended?] by eight hundred million yogins and yoginīs. Then Glorious Nāgārjuna having gone there [to Odiyäna] and very secretly heard the Root Tantra of 1,200 [verses], he disseminated [it] in the world.68 Thus, th e scripture is said to have been revealed in three stages from the divine to th e hum an. It is initially revealed in a divine realm by the somewhat im personal Lord Vajradhara (the leader of “all the transcendent lords”). It is th en redacted by Vajrapäni (here considered distinct from Vajradhara) and taught in this form to yogins and yoginis in the special land of Oddiyäna. As we shall see, this latter m otif of revelation of Tantras in Oddiyäna becomes a central elem ent in m uch subsequent indigenous histori°graphy. As one can infer from its location in the Catuspītha narrative, °d diyäna occupies a place in term ediate betw een divine realm s and the hum an realm. It is n eith er fully divine (as it is located on Earth), nor fully hum an (it is difficult of access and inhabited by semidivine beings). In Oddiyäna, the Tantra is again redacted into a shorter version

(suitable for lesser beings), in which form it is learned by a great saint, Nāgārjuna, who returned to teach it in the ordinary world. It may also be observed th a t this narrative appears to be organized as well around the concept of th e th ree buddha bodies, such th a t Vajradharasdharm akäya, Vajrapānissambhogikakāya, and Nāgārjuna=nirmānakāya. This structure is also utilized in Chinese and Tibetan sources. A sim ilar account may be found in an esoteric Buddhist com m entary from around th e sam e tim e.69 Jñ ān a m itra’s com m entary on th e Prajñãpāramitãnaya-pañcaśatikā provides a rem arkable tale of the revela­ tion of Tantras such as the Guhyasamäja after the death of the Buddha. It is a remarkable docum ent and rewards close attention. Earlier, during the eighty years when the Buddha lived in the human realm, there were none in the human realm of Jambudvlpa who were [suitable] to be disciplined by or to be vessels of the [spiritual] ways of the Sarvabuddhasamäyogaythe Guhyasamäjayor the like. Thus, the gods and fortunate bodhisattvas in the heavens of the Four Kings,70the Thirty-three (Trayastrimśa), Tusita, and so on, being [suitable] vessels, at that time those scriptural collections resided there.71 The narrative thus makes quite clear th at the esoteric scriptures were not revealed during th e lifetime of Buddha Śākyamuni in th e context of his ordinary teaching career, as the world was not ready for them . The text m aintains, however, th at they did exist at th a t time, being the provenance of th e divinities of Trayastrim śa and o th er heavens. It is not explicit w hether they were taught there by Śākyamuni or predated him. Subsequently, after the Buddha passed into nirväna [i.e. died], there was a king of Zahor72 who, along with his retinue, had amazing faith in the Dharma. Having the good fortune to be disciplined by that [spiritual] way and being a [suitable vessel], the Eighteen Great [Scriptural] Collections, such as the Sarvabuddhasamäyoga and the like, came to the country of Zahor, by the blessing of Vajrapäni.73 Thus, th e esoteric Buddhist scriptures m iraculously appeared in the hum an world some tim e after the passing of the Buddha. They were brought th ere—evidently in the form of texts—by the power of Vajrapäni, who perceived th at the tim e was ripe for th eir revelation, because there

was a suitable disciple. However, the scriptures were incomprehensible to the king, so he needed to find someone who could explain them. The King of Zahor, Indrabhüti, inspecting those scriptures, [could] not decipher the writing. Then, having obtained clairvoyance by the force of prior karma, looking, [he saw] the Ācārya Kukkura in the Central Land (Magadha) in the region of Malapa.74 During the day, [the Ācārya] would teach the Dharma to about a thousand dogs; at night, he enjoyed the sacraments with those dogs and acted as [Tantric] mentor to the dogs. [Seeing that the Ācārya] was a [suitable] vessel for that [spiritual] way and seemed also to have the fortune to discipline him, the King sent an ambassador to invite the Ācārya to come.75 Thus, th e king made overtures to Kukkuräja, the “king of the dogs,” evidently an antinom ian yogin of the kind we will explore in chapters 5 and 6. There is a slight anachronism here, insofar as Kukkura is depicted as a kind of p ractitioner of nondual Tantric yoga, enjoying sacram ents (samaya: presumably the consum ption of revolting substances we will discuss in chapter 4, but also perhaps implying sexual yogas) with a pack of dogs (one of the m ost spiritually polluting of animals according to regnant brahminical mores). Yet, these practices—being the provenance of exactly the Tantras whose history is being narrated—had not yet appeared in the world. However th at may be, the teacher (ãcãrya) sensed th at he could be of Service, but he w anted to check out the goods first. As the Äcärya, due to the force of prior karma, had the five clairvoyances, he investigated [therewith] whether or not the king had the fortune to be disciplined by him, and whether or not he himself was a [suitable] vessel for those scriptures, and concluded that the king did have the fortune to be disciplined by him and that he himself was a [suitable] vessel for the scriptures. [The Ācārya thought,] “I may [be able to] eliminate the king s doubts; however, since I’ve never seen those scriptures before, if at some point761 [am not able to] eliminate his doubts it would be extremely bad. So, he sent a missive back with the ambassador, [saying,] “As I would like to see those scriptures beforehand, please send them here.” When the scrip­ tures arrived and he inspected [them,] as he did not know which end [of the text] was the beginning and which the end,77he sank down right there, crying “I have no savior! No refuge!”

Glorious Vajrasattva came in person and asked, “What do you desire?” [The Ācãrya] begged, “I wish to understand these profound scriptures by merely looking at them.” [Vajrasattva] having said, “I grant that it be so,” thenceforth without even opening the texts of the Sarvabuddhasamãyoga and so on, the meaning of those became clear to the mental direct perception [of the Ācārya]. Then, the Ācārya went to Zahor and taught those Dharmas to the king and his retinue.78 Thus, in th e end, Kukkuräja is blessed to understand the esoteric revelations by means of a m iraculous vision of th e esoteric buddha Vajrasattva. Based on this divine intervention and using the texts th a t miraculously appeared, Kukkuräja becom es—like Nāgārjuna in the Catuspītha n arra­ tive—th e first tran sm itter of the Tantras in the ordinary hum an world. In sum, this narrative represents a hybrid of models 4 and 5: The texts appear physically (no. 5), yet require divine revelation for th eir proper Interpretation. The final centuries of Indian Buddhism witnessed the rise of a new eso­ teric revelation th a t had its own historical vision. The Kälacakra (“Wheel of Tim e”) Tantra returns to a m ore realistic vision, insofar as it explicitly claims to have been taught by Śākyamuni Buddha. The Vimalaprabhä (“Stainless Radiance”) Commentary, which forms an integral p art of the basic text, gives th e following tale of its origins, which places the chronology of its revelation quite precisely in a m ainstream Buddhist framework: In this Land of the Nobles (India), the Lord Śākyamuni was enlightened at dawn on the full moon day of the month Vaiśākha. At the end of the fifteenth lunar day of the waxing moon, upon the waning moon commencing, [He] set the Wheel of Dharma in motion and created the Three Ways. In the twelfth month, on the full moon day of [the month] Caitra, at Śrīdhānyakataka [He] displayed . . . the six-chapter Ädibuddha [Tantra].79 However, the realism of this tradition should not be overestim ated. Even though this narrative thus strives to locate the revelation of the Kälacakra in th e teaching career of Śākyamuni—one m onth after his enlightenm ent, in South India at th e site of th e Amarāvatī Stūpa—it is not clear th a t this revelation was th e proxim ate source of the tradition in India. For th e teaching is said to have been redacted and taken to Sambhala (or Śambhala)—a special realm much like Oddiyäna, also im agined to reside

in the northw est—where it was transm itted for centuries during which time the Stainless Light Commentary was composed by one of its kings. Only then was it, like th e Catuspltha, redacted and taught in the hum an realm.80 However, in the Kälacakra traditions too the historicity of the revelation is fu rth e r challenged (or, at least, augm ented) by more visionary models. This same Stainless Light Commentary itself refers to the verse of the Nāmasamgīti we considered earlier, and explicitly invokes the same Mahāyāna notion we have seen throughout these histories—that the later revelations are the eternal, ancient, and constant teachings of all the bud­ dhas, not ju st Śākyamuni: Just as the Nāmasamgīti was spoken, will be spoken, and is spoken by the transcendent lords of the past, future, and present, just so is the Ädibuddha [the XJr-Kälacakra Tantra], The ädi- [in the word ädibuddha] means without beginning or end, that is “in beginningless time, it was taught, will be taught, and is taught [by] the primordial buddha.” It is not taught only by Śākyamuni or the transcendent lord Dīpańkara. [Here the text cites Nãmasamgīti, vv. 12-13.] Hence . . . all the transcendent lords taught the Way of Mantras.81 A later com m entator, Sādhuputrapandita Śrīdharānanda, suggests th at there m ight in fact be sim ultaneous teachings taking place elsewhere. Citing a verse (presumably from the larger [divine, Ur-] Kälacakra Tantra):82 At Vulture’s Peak, Maitreya [will teach] the Prajñāpāramitā; [And] Buddha will teach the pure way of mantra at Glorious Dhänya. From this passage [we know] that the Lord’s teaching of the way of mantras [took place] at Glorious Dhänya. However, elsewhere, as [or “when”] needed by those oriented toward that [way],83a great bodhisattva of the tenth stage [and/]or another revealer of scripture give the Tantrateachings in detail.84 That is, Sädhuputra claims th a t although, yes, Śākyamuni did teach the Kälacakra at the Dhānyakataka Stüpa in South India, its revelation is not limited thereto. For those who do not have access to th at tradition, the Tantra is taught in various locations by various revealers whenever there Is someone who can benefit from these teachings. In addition to the Tantras themselves and the scholastic authors mentioned here, there survives one (and, apparently, only one) Sanskrit work

on th e history of esoteric Buddhism in a specifically narrative genre. The so-called “Sham sher M anuscript”85 begins its history of the esoteric scripture(s) with Buddha delivering m ainstream teachings to śrāvakas, but th e scene th en quickly moves to the South (daksinäpathe) where he is said to have em anated a divine world (maņdala) including a retinue of bodhisattvas.86 Much of the text is devoted to lists of lineage m em bers (paramparä). These strongly suggest the idea of visionary revelation insofar as the lists are very short (there only ever occur th ree or four names before nam es datable to the late first millennium ) and the lineages do not begin with Śākyamuni, but ra th e r w ith a divine buddha em anation (buddhanirmäna). This docum ent serves as a bridge linking the Indian literature we have explored so far and the m any such narratives also found in the East and Central Asian sources. One encounters esoteric saints (siddha) traveling to Oddiyäna to fetch Tantras to bring back to India. One sees miraculous visions of deities, instructions and prophecies in dreams, saints retiring to sacred m ountains to do antinom ian practice, and students following them th ere for instruction. One sees them perform ing rituals as well as pursuing education in Śrāvaka Buddhism, Mahāyāna, and th e Tantras, and composing com m entarial works thereupon. The stories related in this m anuscript are thus very much of a piece both with its Indian precursors and w ith those we find in Tibetan works composed around the same tim e or shortly thereafter. As Sylvain Lévi noted, “the work of which we have a fragm ent here gives the history, naturally legendary, of this tradition, its transm ission from m asters to disciples, and its ritual. It is a curious specimen of th e docum ents that m ust have served as the basis f o r . . . compilations in Tibetan.”87 For th eir part, the Tibetans seem to have been well aware of all of the historical models by which Indian Buddhists conceptualized th eir tradition. For example, Tibetan authors could (and would) eite prior revelatory antecedents such as the Abhidharma in defense of the continuing openness to scriptural innovation found in the Indian and Tibetan prac­ tice of scriptural revelation.88 A notable analysis of this issue is found in th e General Presentation of the Tantras (Rgyud sde spyi mam) of the Sa skya scholar Bsod nams rtse mo (1142-1182). This work abstracts four modes of esoteric revelation: 1. Those taught by Śākyamuni in his regulär form 2. Those taught by Śākyamuni and revealed at a later time (after his death)

3. Those revealed at a later time that were not taught by Śākyamuni 4. Those taught by him while emanated in another form89 The first category includes Tantras taught to ordinary disciples of the Buddha’s life in ordinary places such as Vaiśālī and so forth. We have seen examples of these here, such as the Mahāśītavatīytaught to Rähula at Räjagrha. The second includes Tantras such as the STTS th at specify th at they only descended to the hum an realm later; hence, these were specifically taught by him (it narrates his particular enlightenm ent story, after all) in a heavenly realm and only revealed in the hum an realm some time later. The third, those revealed at a later tim e th at were not taught by Śākyamuni, include th e Cakrasamvara Tantra. To explain this, Bsod nams rtse mo invokes an Indian authority90to the effect that such teachings are always available: As the Commentary on the Samvara Tantra explains, this teaching has existed since beginningless time; it was around before Śākyamuni. Since the Tran­ scendent Virtue ofWisdom and so forth disappear during the fiery and other apocalypses at the end of each age, [although it had been taught before,] the Lord Śākyamuni had to teach it again. This is not so for the glorious Samvara. It never declines since it resides with and is practiced by heroes and heroines and so forth throughout inexpressibly many buddha fields.91 Hence, this type of scripture does not require intervention by the his­ torical Buddha. It is always available in countless alternate universes. The final category—those taught by Śākyamuni through an em anation other than his ordinary one—is said to include m ost of the Tantras such as the Guhyasamäja and so forth. It is interesting to note th a t at th e end of his discussion Bsod nams rtse mo relates a story of the revelation of the Guhyasamäja th at becomes extrem ely influential in Tibetan historical thought. ln this account, Śakyamuni Buddha him self reveals the GST to a King Indrabhüti of Oddiyäna, who asked for a m ethod of liberation th at did not require him to leave worldly life and its pleasures, w hereupon the Buddha creates the Guhyasamäja to order.92In fuller retellings than the one given in Bsod nams rtse mo, th e Tantra (w ritten on gold sheets w ith m elted beryl ink) ends up concealed in a magical Heruka chapel (he ru ka’i gtsug lag khang) under a lake which dries up at a later time, exposing the chapel, and allowing the revelation of the Tantra. Hence, it does not actually fit neatly in any

of Bsod nams rtse m o’s categories, as this complex narrative presents the GST as tau g h t by Śākyamuni himself, em anated as Guhyasamäja, in the special land of Oddiyäna and not revealed in th e ordinary hum an world until quite some tim e afterward, w hen the miraculous tem ple appears. Bsod nams rtse m o’s com m ents on the provenance of this narrative are particularly interesting. Before discussing this narrative, he first rejects an o th er conception of the circum stances of th e revelation of the GST, saying th a t it is “not taught in any tantra, ritual manual, or śāstra—not even in oral trad ition—[and] hence should be known to be a fabrication by innovators.”93 Immediately thereafter, he gives this Indrabhüti story— which he ultim ately accepts as accurate—prefacing it with the com m ent th at others claim this narrative is part of the oral tradition.94 If this is true, it would account for the fact th a t—however populär it may be in Tibet— this narrative does not seem to appear in extant Indian sources. It is also in teresting to note th a t Bsod nam s rtse mo goes out of his way to reject a narrative in w hich th e Tantras fall on th e roof of the palace of a King Ja, who does not understand them , enlists the aid of a m onk nam ed Kukkuri (“Bitch”), who him self gains knowledge through V ajrasattva’s Intervention. This should look familiar, as it is basically th e n arrative given in Jñ ān a m itra’s com m entary. Bsod nam s rtse mo points out issues th a t he thinks make it implausible to his view and also dismisses it as “a fabrication by innovators.”95 Although this may seem ra th e r a serious blunder in light of the fact th a t a very sim ilar narrative occurs in Jñ ān am itra’s work, in his defense Bsod nam s rtse mo is not entirely off base in this assessm ent. The version of this narrative featuring King Ja ra th e r th an Indrabhüti seems not to derive directly from an Indian source, but ra th e r is a hybrid narrative m ediated through a late first-m illennium 96 Tibetan apocryphon called th e Scripture Collecting the Intention (Mdo dgongs ’dus).97 It is perhaps w orth noting th a t this latter esoteric scripture describes itself in such term s th at it would fall w ithin our and Bsod nam s rtse m o’s th ird categories insofar as it claims to be tau g h t eternally in th e “m ind stream of th e V ictors” (rgyal ba’i dgongs brgyud) in dharmakäya and sambhogikakäya forms.98 Tibetan literatu re is filled w ith m arvelous narratives of many sorts concerning th e revelation of various Tantras: Narrative historiography of this sort seems to have been particularly appealing to Tibetans and— in contrast to the Indian Situation—m any of these have been preserved. Although this is a fascinating area of study, because we are concerned herein with Indian Buddhism, we will restrict ourselves to com m enting

on those narratives th a t have clear continuities with Indian traditions. However, there is one further narrative type that bears m entioning in this regard, insofar as it occurs in two, seemingly independent extra-Indian sources. In Tibet, this narrative is found in the work of the celebrated polymath Bu ston Rin chen grub. In several of his works on the esoteric traditions, this scholar addresses the history of these revelations. He summarizes his knowledge of th e Indian sources, referring to three paradigms, all of which conform to the “taught in heaven” model (no. 2). Änandagarbha, for instance, m aintains th a t the Tantras were taught in Paranirmitavaśavartin heaven for the sake of the gods—much like the Trayastrimśas model and consonant w ith th e Statem ents of Jñānam itra’s narrative th at the Tan­ tras existed only am ong the gods during Buddha’s lifetime. Alamkakalāśa taught th a t the revelation of the Tantras was a miraculous display, conducted while th e Buddha was still residing prenatally in Tusita, by an em anated form (nirmäna) produced from the elem ents of his own body. Others, he notes, assert th at all the Tantras were taught in Tusita.99 Like Bsod nams rtse mo, Bu ston too Privileges the Indrabhüti taught by the Buddha him self in Oddiyäna narrative. However, he goes on to elaborate a second narrative of a King Visukalpa, born in the era when the Heruka chapel is said to have first appeared.100 This monarch, from various scriptural indications, infers th at there m ust be a special superMahāyāna th a t uses th e passions as means. In a dream, he is told by eso­ teric angels (dakini-s) to go n o rth to Oddiyäna. There, he is told, he will find the Heruka chapel containing the scriptures of the Guhyasamäja that emerged after the desiccation of the lake that had formed over it. Having found the sacred chapel, the king proceeds to prostrate at all four doors,101 returns to the first (eastern) door and prays. A girl emerges and asks, “Are you Visukalpa?” Replying in th e affirmative, he is brought inside and granted en try into th e actual102 Guhyasamäja maņdala, given initiation, Instruction, and so forth. The text of the Tantra too is entrusted to him. This narrative bears striking sim ilarities to the more populär of the two Sino-Japanese traditions concerning the revelation of the esoteric scriptures: th a t they were recovered from a miraculous library stüpa.103 The Chinese esoteric patriarch Amoghavajra “goes into considerable rit­ ual and textual detail concerning the opening of the ‘iron stūpa’ in India from which the teachings of the STTS were purportedly recovered, but identifies th e adept in question only as a ‘great w orthy’ (dade, Sanskrit bhadanta).”104 Amoghavajra begins by declaring th at the “great worthy

received the esoteric Mahãvairocana Sütra in a direct vision (mode no. 5) of Vairocana, “who m anifested his body and a m ultitude of bodily forms. In mid-air, [Vairocana] expounded this teaching together w ith its textual passages and lines. He had [the great worthy] write them down.”105 Later, seeking a great esoteric scripture “enclosed in an iron stüpa in south India” (mode no. 4), he circum am bulated th e stüpa for seven days. Gaining entry, “he obtained the instruction of all the buddhas and bodhisattvas and these [he] rem em bered and held and did not forget.”106The similarity of th e stüpa and th e Heruka chapel should be plain, especially because this particular stüpa had th e special characteristic of being secured by “iron gates and locks” and space behind th e door containing “incense and candles” as well as “exquisite flowers and jeweled canopies hung in sum ptuous array.”107This is certainly an unusual stüpa and sounds rath er like a chapel, albeit m any stüpas, like maņdalas, consist of four gateways (toraņa) leading to a central, dom ed structure.108 Another, similar tale is related of the fifth-century apocryphal Chinese Consecration Sütra (Kuan-ting ching, Taishö 1331).109 This work “is quite explicit about its claimed origins as a treasure,110 describing at length its initial preaching by the historical Buddha, its subsequent concealm ent, and its eventual discovery from a grotto, w here it had been hidden in a jew eled casket, w ritten in letters of purple and gold upon sandalwood tablets.”111 It does not take a great leap of th e im agination to grasp the homologies between this tale and th at of the Guhyasamäja, taught by the Buddha to Indrabhüti, w ritten in m elted beryl on gold plates, concealed in a m iraculous chapel, and rediscovered. There is a fu rth e r sim ilarity in th e two narratives insofar as both claim th a t th eir scriptures existed for a period during and after th e Buddha’s lifetim e until, due to th eir extrem e efficacy, they had liberated everyone in their proximity and went “underground.”112 The Chinese traditions (like th e Indian and Tibetan) also reflect the short lineage th at is the unavoidable consequenct of scriptures revealed recently. Zhao Qian, a disciple of Amoghavajra, “recount[s] th e transm ission of th e STTS from an ‘iron stü p a’ in a lineage encom passing Mahāvairocana, Vajrasattva, Nāgārjuna, Nāgabodhi, and Vajrabodhi.”113 The last in this list is the teacher of the au th o r’s teacher. He accordingly com m ents th a t “from the origin flows a single tradition perhaps comprised of only some ten persons and th at is it!”114 Likewise, a late eighthcentury (781) stele inscription proclaims th a t “from Vairocana to the m onk [Amoghavajra] are a total of six ‘petals.’”115 It may also be noted

th at Zhao Qian “makes plain . . . t h a t . . . this doctrine was not preached by Śākyamuni.”116 The Japanese esoteric Shingon tradition similarly advances the notion th at th e MVT w as/is taught by Dharmakäya Buddha, not the nirmäna Śākyamuni. Kōbō Daishi distinguishes the exoteric teachings as the reve­ lation of nirmãnakāya and sambhogakäya buddhas, while the esoteric is the dharmakäya teaching.117 This is essentially the position advanced by the younger bro th er and heir of Bsod nams rtse mo, Grags pa rgyal m tshan (1147-1216). The latte r w rote a brief work on the “Origins of Heruka,” in which he foregrounds the Mahāyāna theory of the three bodies with a herm eneutical spin, asserting th at in definitive meaning (nītãrtha) the Buddha attained enlightenm ent in Tusita, while in interpretable meaning (neyärtha) he “tu rn ed m any wheels of dharm a in m any places having em anated m any teachers.”118 On this model, then, the Tantras are the teaching of one or another nirmänakäya, an d /o r of the sambhogakäya. There are num erous such Tibetan accounts that are marked for their tendency to stress th e quasi-Docetic119 M ahāyāna cosmology of the three bodies, allowing for a consequently more flexible notion of Buddhic authorship, insofar as buddhas such as Śākyamuni are regarded as merely one of myriads of buddha-em anations that appear from the same source (dharmakäya) throughout m ultiple universes.120 Although all this historiography may seem rather bizarre to m odern sensibilities—the mystical rationalizations of innovating Buddhist communities or charlatans—the Buddhist esoteric traditions were not alone in India in advancing such claims. Much the same was taking place in other, contem poraneous Indian religions th at were participating in the efflorescence of Tantric movem ents in late first-m illennium India. For instance, the m eans to reveal in the world the Śiva Sütra—an im portant scripture for the Śaiva Tantric Trika traditions—is said to have been given to the sage Vasugupta in a dream vision.121 The work, he was told, is w ritten on stone, th e text concealed by being placed face downwards. The sage was instructed to travel there to extract the text from its place of concealment (mode no. 4). In an alternate tradition, Vasugupta is said to have received the scripture directly from Śiva in the dream (mode no. 5).122 Here we see utilized for one and th e same scripture, both of the means of revelation described in the Buddhist Face-to-Face Confrontation Scripture-dream rev­ elation and physical concealm ent. I n d e e d , j u s t a s t h e B u d d h is ts w e r e e n c o u r a g e d t o d o in t h a t s c r ip tu r e , th e T a n tr ic Ś a iv a s a ls o a d o p t e d t h e p r a c tic e o f a c t iv e ly s e e k in g v is io n a r y

revelations. As Alexis Sanderson has noted, th e Krama literature attests to “th e practice of seeking revelation by propitiating th e goddess during a period of ascetic re tre a t at a rem ote PItha [sacred site].”123 The Kãlikãkramapañcãśīkãy for instance, claims to be the result of such a vision quest, “tran sm itted by Niskriyänanda through a [disembodied] voice to the crem ation-ground dwelling Siddha Vidyänanda w hen he was propi­ tiating th e goddess in a cave in th e m ountains of the Pītha Śrīśaila.”124 Numerous sim ilar examples of visionary revelation may be found in the Śaiva Tantras. The third chapter of the Rauravasütrasamgraha relates the origins of th e Rauravägama, th e “decent of the Tantra [into the hum an world]” (tantrãvatãra), in which the “supreme guru of the world” Ananteśa reveals the scripture in the form of a smokeless blaze of light to Śrīkantha [Śiva], who transm its it to the goddess Devī, and so to Nandīśa, Brahmā, and ultim ately to various sages and ordinary hum an disciples.125 The conception th a t esoteric scriptures were to be found in—and could be retrieved from —th e wonderful country of Oddiyäna is also widespread in Śaiva religious culture. Jñ ān an etran āth a, for instance, to whom all Krama authors trace th eir lineage, is said to have “received the Krama revelation in Uddiyäna directly from its Yoginīs, known as Pītheśvarīs, or from th e ir leader (cakranāyikā) Mańgalā.”126 These Pītheśvarīs of Oddiyäna play im portant roles in m any of th e origin narratives of the Śaivas. Such narratives are used to account for revelation of such esoteric scriptures as the Vätülanäthasütra127 and others. Thus, the historiography of the Śaiva Tantric traditions m irrors much of w hat was being articulated by the Buddhists at the same time. Although, because th eir chief object of devotion was unequivocally im m ortal, the historiography of Śaiva Tantra was correspondingly simplified. One need not bother oneself with bare history in the case of an eternal deity. Thus, in some sense, w hat required significant historiographical restructuring by the M ahāyāna—conceiving of divinity as active in all pasts, presents, and futures—was a fait accompli for the Śaivas. However, they still needed to account for th eir history, to place the tradition in a narrative trajectory. In th e account of the Rauravasütrasamgraha, one can clearly see the mode of eternal divinity (no. 3) articulated in concert with th a t of revelation in heaven (no. 2): The eternal (ananta) lord (īśa) Ananteśa transm its the teaching to the deity Śrīkantha [Śiva], he to his divine wife, she to her son, and he to Brahmä, before the latter revealed it in the hum an world. Likewise, the Śiva Sütra concealed on a hidden stone tablet (no. 4) or in a dream vision (no. 5) give am ple evidence th a t there was considerable

historiographical consonance among the various esoteric traditions of the Tantric Age.

OBSERV A TIO NS ON STRUCTURE, FUNCTION, A N D HISTORIOGRAPHY

Thus, m uch as we saw of m odern historiography in chapters 1 and 2, in the case of the indigenous historiography of esoteric Buddhism one can likewise discern clear patterns in the articulation of historical narratives. Just as m odern historiographers drew on the resources of their own cul­ tures for conceptual models and stock narrative forms with which to make sense of esoteric Buddhism for th eir own readers, so too did the authors of the Tantric traditions tu rn to earlier Śrāvaka and Mahāyāna precedents in seeking to craft through historiography a corporate understanding that would make sense to their contem poraries. Both groups of historiog­ raphers employ well-established narrative prototypes; they merely have different prototypes, deriving from differing discursive traditions. They both sought to make th e objects of their narratives comprehensible, by situating them in culturally familiar frameworks th at provide a readymade interpretative context. In the preceding, we have abstracted five basic narrative modes used throughout Indian esoteric Buddhist historiography and its reflections in the esoteric literatures of Tibet, China, and Japan. As I have indicated elsewhere,128 th ere is a rem arkable structural continuity in the form of these narratives insofar as they all in some fashion account for scrip­ tural revelation subsequent to the lifetime of Śākyamuni Buddha. Bare historical assertions may suffice for less significant works (such as the Ganapatihrdaya) th a t merely articulate spells of protection or prosperity. These could easily be conflated w ith the quite large and well-accepted corpus of m ainstream Buddhist protection texts (raksät paritta). As Skilling has observed, “the chanting of certain auspicious verses or texts for protection against disease and m alignant spirits and for the prom otion of welfare was no doubt a 'pan-nikāya practice.”129 Such a strategy was, how­ ever, impracticable for more significant expressions of the Tantric genius, such as the MVT, STTS, and GST-capacious works that present complex ritual program s quite distinct from prior Buddhist practices. Thus, spe­ cial narratives were found necessary in these cases; and, faced with this challenge, th e esoteric traditions were able to resort to well-established precedents th a t had seen service w hen the task had been to gloss over

the canonization of the huge and novel literature of the Abhidharma and Mahāyāna. With th e exception of th e bare historical model (no. l ) —which in the case of the M ahāyāna traditions is not “b are” at all insofar as it had been radically tran sfig u red in co n cert w ith a new buddhalogy in to m ode no. 3—th e oth er four all involve revelations th at did not come through the usual channels. They were either taught to the gods, from which they later descended to th e hum an world (no. 2), an d /o r they were the eternal teachings of cosmic buddhas (or Śākyamuni in cosmic form) th at could be accessed at any tim e (no. 3), a n d /o r they had been revealed before and hidden for later re-revelation (no. 4), a n d /o r they were taught anew in m iraculous visions (no. 5). These conceptions allowed the Buddhist tradi­ tions to account for scriptural anachronism by providing a conceptual structure w ithin which even newly revealed teachings may be considered as possibly authentic. M otivating this accom m odation was a central intellectual value at the core of Indian culture of the time. As Granoff has noted, Tantric religion “required . . . a definition of scripture th a t would adm it the validity of new religious visions and of individual beliefs and practices” insofar as Tantric practices were “explicitly acknowledged to be new rituals.”130Yet, these new visions and individual innovations sought acceptance in a cul­ tu ral context in which the tru e is generally understood to be found in the old. In the thought of the tim e—much like in the classical cultures of Europe—tru th s had to have always been true. New tru th s were not to be made. When expressed in a Buddhist idiom, this axiom requires th a t if a religious technique or philosophical concept were true, it had to have been taught by the Buddha or, at least, a buddha. That this culture of tru th was pan-Indian is clear w hen one notes th e consonance of Buddhist thought w ith Śaiva. Although the cultural im aginations of th e various Śaiva traditions differ in certain significant respects from those of the Buddhists, we have seen th at they structured the histories of th eir revelations in rem arkably similar fashion. Like the esoteric Buddhists, the new literature and novel practices of the Śaivas em erged in dialog w ith already existing ideas and the people and institutions th at perpetuated them . These narratives were thus crafted in a context of cultural contestation—and contested they were. As in the case of th e Buddhists,131 the claims of the esoteric Śaivas were not m et with unanimous acceptance. Those who were not of a m ind to accept the new approaches “expressly declared th eir hum an origin and consequent

unauthoritativeness.”132 For those more positively inclined to the new form ulations, however, th e narratives we have explored provided an imaginative fram ework in which these religious practices could appear plausible to Buddhist and Hindu communities alike. To com prehend properly the function of the historical discourses of M ahā/Vajrayāna Buddhists in context, it is perhaps helpful to think about the Buddha less as a concrete hum an being, and instead to understand the m anner in which “Buddha” served an ongoing epistem ic function in Buddhist cultures. For the esoteric Buddhist traditions, clearly, Bud­ dha was not a historical Buddha (and certainly not “th e ” historical Bud­ dha of some m odern scholars and Buddhists), but an epistemic Buddha. Throughout th e Buddhist world, the word of the Buddha is considered equivalent to tru th and vice versa.133 Hence, the famous Statement that “w hatever is well said [i.e., true] was said by the Buddha.”134 On the one hand, this allowed th e co-option of ideas from other groups in the cul­ ture. Phyllis Granoff has, for instance, drawn attention to passages in the Mañjuśrīmülakalpa th a t forthrightly claim th at esoteric teachings of the Vaisnava and Śaiva traditions “were actually proclaimed by Mañjuśrī disguised as one of the Hindu deities.”135 On the other hand, however, it also required th at w hatever claimed to be the word of the Buddha have been sufficiently vetted for consensual acceptance by the community. Such historiography is thus best understood as one feature of an ongoing process of epistemological negotiation in Indian cultured circles. In this process o f consensus building, new ideas could be tested in actual practice, and perhaps articulated in part through comm entaries or works attributed to nam ed persons, providing for some discussion and debate. The process of getting som ething accepted as scripture and thus (at least theoretically) binding on larger communities, however, required more substantial vetting through consensual critique. In this sense, it is worth bearing in m ind th at not all new scriptures were accepted by all or even a majority of com m unities, as can be inferred from the prophylactic passages in M ahāyāna scriptures (exo- and esoteric) referring to those who would reject them as “not the word of the Buddha.”136 In such cases, hard-edged realism ” was the order of the day. As Matthew Kapstein has observed concerning similar negotiations in Tibet, there were in all Bud­ dhist communities tw o o p p o sin g te n d e n c ie s : th e te n d e n c y to a n a th e m a tiz e and th e te n d e n c y to c a n o n iz e . T h e fo r m e r w as su p p o rte d b y e le m e n ts o f a realistic h isto rica l

orientation.. . . The latter, by contrast, drew its strength from the belief that the canon could be held closed only on pain of self-contradiction, and from a fundamentally idealist vision of the Buddhist world.137 That is, Buddhists were well aware of th e possibility of works being misrepresented as scripture and were concerned not to be led astray from the correct teaching. Thus, no doubt, m any contenders for com m unity approval were anathem atized as not th e word of the Buddha. Others, how­ ever, were found suitable and—based on the precedents we have reviewed here, including the idealist notion th at enlightened speech was available anyw here, anytim e—certain w ell-elaborated literary Statem ents were accepted as valid articulations of the perspective of th e community. By this process was gradually created a respected body of knowledge. It is essential to understand th a t this process is by no m eans lim ited to m ystical religious traditions such as Tantric Buddhism and Śaivism, but was a fundam ental feature of Indian cultural discourse in virtually all fields. During the same first-m illennium period th at witnessed the revela­ tion of th e M ahāyāna and M antrayäna scriptures, similarly consensual understandings deriving from years of social experience across th e cul­ tural spectrum were crystallized into m ajor Statements of theoretical and practical knowledge. Yet, due to the aforem entioned axiom th at all th a t is good and tru e m ust have been known of old by divine or sem idivine beings, these knowledge Systems too were attributed to miraculous revelations. As Sheldon Pollock has noted, in India śāstra [knowledge] must exist primordially. Extant śãstrasy consequently, come to view themselves as either the end-point of a slow process of abridgement from earlier, more complete, and divinely inspired prototypes; or as exact reproductions of the divine prototypes obtained through uncontaminated, unexpurgated descent from the original, whether through faithful intermediaries or by sudden revelation.138 Both of these forms of understanding will be quite fam iliar from the aforem entioned esoteric Buddhist narratives. The form er (the slow pro­ cess of abridgem ent) is precisely the narrative we saw in the case of the Catuspītha Tantraf which decreased by stages from a 180,000-verse divine edition, through a 12,000-verse Oddiyäna redaction, to 1,200-verse text th at is known in the hum an world. The sam e is related of the work on social duties and etiq u ette (dharmaśãstra) th e Näradasmrtiy which is

said to have been composed in 100,000 verses by the divine sage Manu, abridged by Närada to 12,000, eventually to emerge in the hum an world as the extant 2,700-verse redaction.139 We have also observed th at the latter process (direct revelation) is also very common in Tantric histo­ riography. In exactly the same cultural register, key works on dramatics, astrology, cuisine, and archery—the Bharatanãtyaśāstra, Süryasiddhãnta, Pãkadarpana, and Brhatśārńgadharapaddhati—are attributed to direct rev­ elation to hum ans by the Brahmanical gods Brahmā, Sürya, Yama, and Śiva, respectively.140 Consequently, w hat is perhaps most remarkable about the indigenous historiography of esoteric Buddhism is th at it is not remarkable at all. During the first m illennium in India, every branch of hum an knowledge and practice (śāstra) was given identical treatm ent. W hether the subject was religious history (puräna), erotics (kāma-śāstra), poetics (alamkära-), architecture (śilpa-), astronom y (jyotisa-), dram atics (nätya-), cookery (pãka-śãstra), or m edicine (äyurveda), all teil similar stories of the derivation of th eir key literary Statements from divine revelation. Although we have traced the Buddhist antecedents to this practice to illustrate the particular resources on which our authors drew, ultimately what we have described is a pan-Indian phenom enon common to all the knowledge Sys­ tems of the time, w hether religious or secular. In all of these areas, these corpora were negotiated, w ith some being anathem atized and others being canonized. The com m entarial process allowed subgroups to negotiate sometimes conflicting canons and to create their own synthetic tradi­ tions drawing on th e m aterials provided in the authorized scriptures.141 This conception of system atic presentations of com m unity knowledge as deriving from divine revelation was culturally sanctioned across the board in literate Indian cultures, providing a space in which people could understand, authenticate, and share innovative new approaches without challenging the privilege of the ancient. Of course, Indians (and Tibetans) often were skeptical of such claims. The use of this device was subject to critical reception; limits did exist outside of which these exceptions were not granted w ithin an otherwise quotidian perspective. The attitude toward the miraculous in Indian cul­ ture is not entirely dissimilar from th at taken in the classical Greek tradi­ tions where, as Veyne has noted: A m o n g th e lea r n e d , critica l cred u lity, as it w ere, a lter n a te d w ith a glob al sk e p tic ism an d r u b b ed sh o u ld e r s w ith th e u n r e fle c tin g c re d u lity o f th e

le ss e d u c a te d . T h e se th r e e a ttitu d e s to le r a te d o n e a n o th er, and p o p u lä r c re d u lity w as n o t c u ltu r a lly d e v a lu e d .142

Veyne provides the instructive example of Galen, the physician. Although skeptical of th e existence of centaurs in general (because no one had ever seen one), “when the same Galen no longer seeks to impose his ideas but to win new disciples, he seems to pass to the side of the believers.”143That is, w hen narrating the history of Greek medicine, he will readily describe its original teaching by Apollo to Chiron, the centaur. We m oderns, of course, cannot live in th a t world (we may not even w ant to); but the work of cultural criticism dem ands th a t we at least be able to understand the language used to articulate it. On th at basis, we can discern more subtle dynamics of their historiography. Failure to do so will inevitably result in m isconstrual of the object of our studies. On the one hand, to construe the indigenous historiography of esoteric Buddhism as engaging in the same cultural practice we do—to assume th at traditional authors m ean th e same thing we do w hen they attrib u te works to the Buddha—is to im port distorting assum ptions into our representations. This problem is pervasive in studies of esoteric Buddhism, and we have seen the results in these chapters. On the o th er hand, neither should we delude ourselves into thinking th at w hat we do and w hat they do are so terribly different. Frequently, m odern historians do not differ very significantly in th eir m ethods from th e ancient Greek m ythographers. Giuseppe Tucci, for instance, in discussing the narrative of the revelation of the Kälacakra, asserts th at “it is evidently a pious tale, w ithout the least historical fo u n d a tio n ,. . . but everything leads us to think th a t there is m uch tru th in the rest of the narrative.”144 In so doing, his m ethod is not at all distinct from th a t of Roman historians such as Livy who “lim ited them selves to rem oving details th at seemed false or, rather, unlikely or u n re a l. . . [and] presumed th at th eir predecessors were telling the tru th ” e r the Greek m ythographer Pausanias who sought to separate th e kernels of tru th from the “puerilities—nym phs and river fathers—th a t can easily be corrected.”145 Indeed, recent speculation th at the nonhum an spirits described as inhabiting rem ote m editation spots “reveals the social reality th at Buddhists began to encounter tribal and semi-nomadic peoples extensively in the early medieval period”146 reveals exactly the type of thinking that created euhem erism and o ther forms of Greek rationalization of myth. Much the same intellectual move can be seen in classical Indian w riters, as well.

“Vedottama,” for instance, “in his Pāñcarātra-prãmãnya has gone so far as to declare th at the original tan tra works of the Śaivas th at are believed to have been revealed by Maheśvara were compiled by an ordinary hum an being nam ed M aheśvara and some credulous people were mistaken to identify him w ith th e god Maheśvara only on the flimsy ground of the similarity of names.”147 Thus, th e C ontem porary in te rp re te r finds h erseif betw ixt and betw een—n eith er able to assume seamless continuity between herseif and th e culture th a t is her object, nor yet able to radically distinguish her own intellectual projects from those of the indigenous traditions. As Veyne answered his own question in Did the Greeks Believe in Their Myths?: Anyone with the slightest historical background would immediately have answered, “But of course they believed in their myths!” We have simply wanted also to make it clear that what is true of “them” is also true of ourselves and to bring out the implications of this primary truth.148 In some very im portant respects, the formal aspects of histories are always oriented more toward crafting the historical subject, than revealing the historical object. The discursively constructed position of the narrator and th e “ideal reader” suggest to the actual reader an orientation, an endorsed subjectivity, an interpretative perspective with regard to the phenom enon whose history is being narrated. The stock of narratives hegemonic in a com m unity encapsulate stylized orientations of the group and ten d —as we have seen in m odern and indigenous historiography—to be used again and again to interpret unfamiliar events and traditions, to make the unknown known. Scholars will do well, then, to eschew well-worn paths of origins and historical narrative until close, scholarly work allows a much better sense of the meaningful discourses and practices—the semiology—of these tra­ ditions. Much more spadework needs to be done before responsible his­ torical synthesis can be undertaken. If we do not fully understand who these com m unities were and w hat they thought they were doing—if we fail to understand th e constitutive im agination of esoteric Buddhists— how can we essay larger synthetic projects? We will just end up recapitulating well-worn rhetorics to fill in the gaps in our own ignorance.149 As Veyne has also suggested, after Foucault we can no longer naively teil stories of eternal figures like esoteric Buddhism, Śaivism, and so forth. Rather, we need to focus on relations (structures), for it is relations that

constitute w hat these things really are in history.150As Pollock has likewise suggested, such a mode of analysis “would include listening to the questions the texts them selves raise . . . rath er than, like inquisitors, placing th e texts in th e dock and dem anding th a t they answer th e questions we bring to them ; in o ther words, focusing on th eir critical processes rather th an on our critical positions.”151 The type of discursive analysis exemplified in the foregoing—and to be continued in subsequent chapters when we advance our interpretation of Tantric Buddhist discourses—can facilitate this process insofar as it can bring to the fore stylized rhetorical patterns in both the scholarly subject and its objects. If, as some have suggested, the real vocation of a historian is the analy­ sis of evidence,152 we may th en logically pass from the dom ain of histo­ riography to the interpretative challenges th a t confront the in terp reter of the literary rem ains of Indian Tantric Buddhism. As I suggested in the introduction, it is here th at a semiologically-inflected philology sensitive to the structures th a t p attern its rhetoric can provide some assistance. This is now where our path of investigation m ust lead—to a close and crit­ ical reading of the Buddhist Tantras. As I will endeavor to dem onstrate, th e critical lenses of semiological analysis allow us to approach the data with fresh eyes—to attend to patterns discernible w ithin the discourses and to interpret th eir significance w ithout recourse to an imagined social context or historical frame, operating on th e presum ption th a t the dis­ courses are coherent and, if a large enough sample is taken, will dem on­ strate reliable things about th e aggregate. The second p art of this book will focus on the analysis of prim ary sources for the study of Tantric Bud­ dhism. In particular, we will bring these tools to bear on th a t aspect of the esoteric Buddhist traditions th at has m ost frequently been associated w ith th e idea of T antra—th a t with which we began this book—the tran s­ gressive discourses and rites found in the Mahäyoga and Yoginī Tantras.

II INTERPRETATION

[4} THE SEMIOLOGY OF TRANSGRESSION

tattvam na paśyati hi so ’ksaramātradarśi candram didrksur iva cãńgulim īksamānah

|

The one who sees only the literal, does not see reality—like one who wants to see the moon, gazing a t the finger [pointing at it].

—Candrakīrti, Pradipoddyotana h e t im e

T

h a s n o w c o m e to r e tu r n to th e q u e s tio n p o se d a t th e

b e g in n in g o f th is w o rk : H ow to m ak e se n se o f th e fa c t th a t th e se e m in g ly an tiso c ia l, a n tin o m ia n b e h a v io rs a d v o c a te d in m u ch o f

th e la te r T a n tric lite r a tu r e —th a t se em a t first g la n ce to b e th e rav in g s o f m a d m e n —a re in fa c t “re c k o n e d to be th e sa c re d s c rip tu re o f m illio n s o f in te llig e n t h u m a n b ein g s.”1W h a t is o n e to m ake o f a tra d itio n w h o se m o st rev e re d s c rip tu re s se e m to c o u n s e l its d e v o te e s to v io la te n o t o n ly its ow n m o st b asic m o ra l p re c e p ts , b u t to v io la te all th e m o st e sse n tia l c o n tem p o ra n e o u s S tan d a rd s o f h u m a n d ec en c y ? W h a t m ig h t all th e s e o u tré S tate­ m e n ts (an d , p resu m a b ly , b e h a v io rs) m e an ? H ere to o a close a t te n t io n to r h e to r ic allow s a m o re n u a n c e d assessm e n t o f th e p la ce o f tra n s g re s s io n in th e social a n d sym b o lical eco n o m y ° f e s o te ric B u d d h ist c o m m u n itie s . We h av e s e e n in th e fo re g o in g how, in th e c a se o f h is to rio g r a p h y a t le a st, h u m a n c o m m u n ic a tio n is ra re ly if e v e r a s tr a ig h tf o rw a r d p ro ce ss. W h a t m ay se em a t firs t g la n c e to be firs t-o rd e r S ta te m e n ts a b o u t t h e o rig in s o r h isto ry o f e so te ric B u d d h ism — w h e th e r m o d e rn o r in d ig e n o u s —a p p e a r u p o n an aly sis as h ig h ly co m p lex an d ra m ifie d m o d e s o f e x p re ssio n w h o se u ltim a te re fe re n ts en c o m p a ss a ran g e th a t in c lu d e s e x p re s s in g in te r c u ltu r a l ju d g m e n ts , re in fo rc in g an d

celebrating shared Systems of valuation, and negotiating intracultural discussions over epistemic authority. The antinom ian discourses and practices of Tantric Buddhism are no less complex th an those used to comm unicate its m odern and traditional historiography. In th e face of such complexity, in order to understand how (and what) they signify in th eir proper cultural settings, one m ust again move beyond the particular expressions (paroles) and endeavor to grasp th e larger semiotical System (langue) of which they are but a part. Otherwise, as Pierre Bourdieu (among others) has warned, T h o se w h o tak e th e sh o rt c u t w h ic h lea d s d ir e c tly from e a c h sig n ifier to th e c o r r e sp o n d in g sig n ified , w h o d isp e n se w ith th e lo n g d e to u r th r o u g h th e c o m p le te S y stem o f sig n ifie r s w ith in w h ic h th e r e la tio n a l v a lu e o f e a c h ite m is d e fin e d (w h ic h h a s n o th in g to d o w ith an in tu itiv e ly g r a sp e d “m e a n in g ”), are in e v ita b ly lim ite d to an a p p ro x im a te d isc o u r se w h ic h , at b e st, o n ly stu m b les o n to th e m o st ap p a ren t sig n ific a tio n s .2

This chapter will begin the work of discerning this total System: to see how Tantric Buddhists can make sense w ith (and, we, of) th e odd and often repulsive Statem ents th a t appear in th e scriptures of th e Mahäyoga and Yogin! Tantras. The prim ary focus of analysis will be the so-called five m eats (mämsa) and five am brosias (amrta) as they appear in the m ost renowned and influential of the Buddhist Mahäyoga Tantras, the Guhyasamäja (Esoteric Community) Tantra.3 These two sets of five substances—beef, dog, elephant, horse, and hum an flesh, and feces, urine, blood, semen, and m arrow4—feature prom inently in the literature of the later Tantras, and are a conspicuous elem ent of their ritual performance. With this focus in mind, before turning to my own philological and semiological analysis, it may be helpful to review the positions m od­ ern scholars have taken concerning the interpretation of these disturbing elem ents of Tantric ritual and scripture: That is, on th e question of w hether or not this transgressive discourse is to be taken literally. I will subsequently suggest that, to truly appreciate the semiology proper to the antinom ian aspects of Tantric Buddhism, one m ust look beyond the level of plainly denotative (what I will also call natural) language. Those who argue th a t these injunctions should be taken literally and those who argue for a figurative interpretation both fail, I argue, precisely on account of a narrow focus solely on denotation. To the contrary, one may clearly discern in these traditions a coherent System of what Roland Barthes has

called m ythic speech or, more precisely, connotative semiotics. On the basis of the data of ritual prescription and analysis of Tantric scriptural narratives, I will endeavor to dem onstrate that, by recognizing that these antinom ian signs derive their significance from such a connotative semiotical system, scholars may b etter be enabled to address the fundamental question of th eir meaning(s) and, on this basis, reconstruct (in chapter 6) the social contexts th a t created and sustained them.

THE LITERAL A N D THE FIGURATIVE IN TANTRIC HERMENEUTICS

In general, in addressing the question of Tantric interpretation, m odern scholars have reduced the problem of interpretation to one of determ ining direct reference. This is, in seeking to resolve M itra’s quandary, the question is construed as: Did they do these outrageous things or not? Underpinning this approach is the assum ption th at words are found in the texts, th at they denote various meanings, and that consequently the question of in terp retation is simply one of deciding precisely w hat it is that they denote. Conforming to this general orientation, one may discern two m ajor tendencies. Many assert that the Tantras—being the secretive, esoteric scriptures they claim—express themselves via a kind of special code (twilight language or intentional language), which m ust be broken in order to understand w hat the real m eaning is behind w hat seem, taken literally, to be antinom ian Statem ents or references to exotic m eats or revolting bodily fluids. Others (currently among the m ost vocal) claim that th e Tantras say exactly w hat they m ean and this question of in ter­ pretation is ultim ately an artificial one born of naively giving credence to the later, “bowdlerizing,” “sanitizing,” an d /o r “semanticizing” tendencies found in th e com m entarial literature. This latter, it is averred, seeks to explain away the literal m eaning intended by the original (lay) authors in order to render them more palatable for a very different (monastic) audience.5That is to say, scholars have tended either toward literalism or figurativism. Scholars of the literalist tendency claim th at the Tantras were intended as straightforw ard, literal Statements, and th at this literal meaning m ust be taken as th e prim ary basis for cultural interpretation. They assert th at the authors of the Tantras m eant exactly and only what they said on the ostensive level of discourse. This is essentially the attitude one finds in the earliest m odern writings on the Tantric traditions. It was clearly with

a literal reading in m ind th at Eugène Burnouf (writing in 1844) made his now-famous Statem ent (later repeatedly attributed to T. W. Rhys Davids) th at “the pen refuses to transcribe doctrines as w retched in form, as they are odious and degraded in th eir foundations.”6 Rajendralal Mitra also— following w hat he considered to be the literal m eaning of the Guhyasamäja concerning th e m eats and am brosias—m aintains th at “the most appropriate food for devotees while engaged in [Tantric] worship is said to be the flesh of elephants, horses and dogs” and th a t “not satisfied w ith the order . . . to make offerings of excrem entitious m atter on th e hom a fire, the author [of the Tantra] goes to the length of recom m ending such substances as hum an food.”7 In general, it may be said th a t these scholars assume th a t the Tantric m ovement in Buddhism originated in a desire to loosen th e moral discipline enjoined by the tradition, to allow for w hat they consider a more natural enjoym ent of life’s pleasures. Many, m ost even, of these interpreters assume some model of the lustful monks etiology a n d /o r the degeneration narrative as a guide to th eir readings. As Monier Monier-Williams expressed it: “The eternal instincts of hum anity . . . insisted on making them selves feit notw ithstanding the u nnatural restrain t to which th e Buddha had subjected them .”8 M itra’s son, Benoytosh B hattacharyya— whom we have seen was a strong advocate of a “natural” life—goes so far as to praise th e Guhyasamäja for having “done Buddhism the Service” of elim inating all its disciplinary m easures. Although all sorts of luxuries were prohibited in the early days, he teils us in th e G uhyasam äja e v e r y th in g is p e r m itte d . N o t o n ly fle s h o f th e m o st h a r m less k in d b u t all kin ds o f fle s h -m e a t are p e r m itte d su ch as th e fle sh o f e le p h a n ts, h o r se s, d ogs, c o w s, nay, e v e n o f h u m a n b e in g s.9

Nor is this approach limited to these early Orientalists. Such notable recent scholars of th e Tantras as David Snellgrove and Ronald Davidson have made sim ilar claims and have argued quite stridently against those who take such term s to m ean anything other than w hat they literally denote. Snellgrove, for example, lam ents “a tendency nowadays, m uch prom oted by Tibetan lamas who teach in the Western world, to treat references to . . . worship carried out w ith ‘im pure substances’ (referred to usually as the ‘five nectars’) as symbolic.”10 Contrary to “w hatever later refined Inter­ pretation was placed upon such prescriptions,” he assures his readers, “there need be little doubt th at such ‘sacram ents’ were used in the circles

of tan tric yogins, w here these texts had th eir actual origin.”11 Davidson similarly dismisses those who disagree with the literalist approach as apologists, and devotes considerable attention to refuting the notion that the language of th e Tantras could bear significance beyond the literal. Like early Victorian scholarship on Buddhism—Davidson maintains that the Tantric scriptures were composed out of the desire of Buddhist monks to have scriptural w arrant for th eir lusts for “drinking wine and making love to nubile women.”12 Like Bhattacharyya, Davidson sees in the history of Tantric Buddhism the final act in “a narrative about the extended testing of preceptorial boundaries by the morally challenged.”13 Not all scholars of Tantric Buddhism, however, have been satisfied with this approach. Others have drawn attention to the im portant testim ony of the surviving Tantric com m entarial literature (many examples of which seem to suggest readings o th er th an the literal) as well as to noteworthy indications native to the Tantric “prim ary scriptures” (müla-tantra) themselves, which seem to indicate th at these works were not intended to be u n d ersto o d en tirely or exclusively literally. These scholars— evidently th e target of Snellgrove’s methodological ire—tend to describe the language of the Tantras as m etaphorical or symbolical. A. K. Warder, for example, in his 1970 work Indian Buddhism, noted th at “putting aside conjectures . . . th e com m entators are solidly in favor of the text[s] being m etaphorical.”14 This line of thinking may draw some support from the fact th at th e com m entaries on the Tantras offen do not accept the literal meaning as the intended sense. In many treatises of this sort, what might seem to be antinom ian term s or injunctions are interpreted as references to inner yogic processes. For instance, in several passages of Candrakīrti’s Pradīpoddyotana (an influential com m entary on the Guhyasamäja), expressions such as eating of feces and urine are glossed as “pacifying” the sense objects and th e sense organs.15 Responding directly to Snellgrove’s dismissal of the com m entators’ readings as representing a later trend than the root scriptures,16 Warder contends th a t “since the [Hevajra] Tantra itself stresses the m etaphori­ cal m eaning of its Statements we cannot accept his opinion.”17 This point has been made again more recently by Anthony Tribe, who noted th at this so-called symbolic interpretation cannot be attributed solely to later com m entators as th e Hevajra Tantra includes its own nonliteral exegesis w ithin itself.18 After a classical Tantric Statem ent th a t “you should kill living beings, speak lying words, take what is not given, consort with the wom en o f o th ers” (i.e., break four of the five basic Buddhist moral

rules), the Tantra itself interprets this passage to m ean th a t one “kills living beings” by “developing one-pointed cognition by destroying the life-breath of discursive thought;” th at one lies by vowing to save all sentien t beings; and so on.19 Such a move is by no means novel in the history of Buddhist exegesis for, as Tribe comments, “the whole device—of saying som ething th at appears to be shocking and th en explaining what is really m eant—is rem iniscent of passages from th e [exoteric Buddhist] Perfection of Wisdom sütras.”20 On the basis of such observations, Michael Broido articulated a general methodological critique of m odern Tantric studies, writing: One of the reasons for the weakness of current western work on the Tan­ tras is the almost complete neglect of the methods of interpretation which were used by the commentators and teachers who interpreted them. We may not have access to the methods used in oral instruction, but there is no good reason for this neglect of the methods used in the traditional commentaries.21 Thus, in recent decades, more attention has been paid to these traditional m ethods of in terp retatio n , w ith scholars such as Broido him self and Robert Thurm an exploring the complex, polysemous modes of Tantric interpretation found in the com m entarial and herm eneutical literature, such as the aforem entioned Pradīpoddyotana., which sets forth a system of interpretation th at allows for multiple, simultaneous readings of indi­ vidual passages—including, but not limited to, the literal meaning. As this herm eneutical system was considered authoritative in a wide ränge of later Indian and Tibetan Tantric circles, research into Candraklrti’s work has shed m uch valuable light on these historically-influential principles of Tantric exegesis.22 Although th e work of the figurativists has thus done m uch to advance discussion in th e area of Tantric herm eneutics, it is not in fact the only or even the best way to approach the issue. Each of these two approaches to interpretation has contributed to efforts to understand Buddhist Tant­ rism; yet, each also has rather serious limitations. For instance, although Broido and others are exactly correct to stress the necessity of docum enting and analyzing the historical actuality of particular instances of Tantric exegesis as found in th e surviving com m entarial literature, we cannot assume th at the surviving texts constitute a com prehensive catalogue of all such interpretations. Indeed, it m ight be argued, a literal

interpretation of the text does not need a com m entary to defend it; so the fact th at the only surviving com m entaries interpret the text in nonliteral ways appears neither surprising, nor significant.23 However this may be, som ewhat more to the point, I think, is the fact th at m any of th e surviving com m entaries do advocate literal readings of th e texts. Indeed, literal sense (yathäruta) is one of th e six exegetical alternatives outlined by Candrakīrti’s magnum opus on Tantric herm eneutics. The sex and death for which the Tantras are famous are by no means regularly and uniform ly excised by the commentaries. Although Davidson locates C andraklrti am ong those he considers “puritanical” com m entators, the Pradīpoddyotana includes numerous passages in which he details sexual rites in explicit and literal language—more so, even, than the prim ary scripture itself.24 Thus, there is certainly a place for literal interpretation, even according to the later com m entators. To assert th at the Tantras were w ritten com prehensively in code and were not to be understood literally at all is clearly untenable. Even a som ew hat attenuated form of literalism , however, is equally problem atical. Besides the difficulties m entioned here, the very notion of literal is not nearly so simple and straightforw ard as it m ight be made to sound. For instance, in chapter 8 of the Guhyasamäja Tantra, there is a half-verse which runs, roughly translated, “one should always sm ear feces, urine, water, and so on, in order to worship the Victors.”25 Here the literal m eaning seems clear as day: It is typical Tantric disgustingness, obviously, claiming th a t one should offer worship to the buddhas by the slathering of such foul substances as raw sewage. However, although th at might seem literal, it is in fact itself already interpretative. What is m eant in this passage by “feces and urine” is, in fact, feces and urine. However, unlike m any occurrences of these term s in the Tantra, in this context w hat is m eant (as confirm ed by the com m entaries in a gloss th at in no way seems forced)26 is cow dung and cow urine. Such a sm earing of feces, urine, water, and the like, is then (to an Indian eye) quite norm al and not foul or disgusting in any way. In orthodox Indian ritual contexts one routinely sm ears cow dung, urine, and w ater to purify a ritual site: There is nothing revolting, transgressive, or Tantric about it. As Freud is reputed to have quipped, “Sometimes a cigar is ju st a c ig a r. . . ” How, then, to resolve this scholarly quandary? Is it m erely the case, then, th a t determ inations of literal- or figurativeness m ust be resolved on a case-by-case basis? Is it simply not possible to elaborate a global theory of reference? Although ultim ately there is no com prehensive rule th at

can be applied across the board, I do believe th at much of this debate can be resolved by thinking m ore broadly about the nature of signification in this literature. What is striking about these approaches is th a t they take a remarkably narrow view of the possibilities (actualities, even) of hum an discourse. In approaching the question of interpretation as a choice betw een literal and figurative (or even as a polysemous m ixture of literal and figurative), earlier discussions all proceed from the assumption th a t these discourses are examples of directly denotative (natural) language. Starting from this premise, scholarly m ethod is reduced to the realist (one m ight even say positivist) project of attem pting to determ ine if th e Tantras really m eant w hat they said. That is, the fundam ental— even exclusive—question becomes “W hat signified or signifieds correspond to the signifiers found in Tantric discourses?” When it says beef, for instance, does th at mean (real) beef or som ething eise? The questions th at have guided research in this area have all been posed accordingly. It is by no m eans clear, however, th a t th e authors of the Tantras intended to use language in th e straightforw ard, prosaic way th a t this fram ing of th e question suggests. To take only one example, consider the following verse found in the Dhammapada (and its Sanskrit version, the Udänavarga), a work of impeccable pan-Buddhist authority: “Having killed m other and father as well as the king and two learned Brahmans, and having beaten th e kingdom along w ith [its] attendants, a m an is called pure.”27 As the tradition makes clear, this injunction was m eant to be in terp reted —a fact signaled by its hyperbolic (yet culturally precise) transgressiveness. “Mother, father” and so on, which one is to kill, refer to obstacles such as desire, which are to be overcome by the practitioner. Likewise, the Mahāyãnasamgraha makes Statem ents such as “the bodhisattva is th e suprem e slayer of living beings,” m eaning thereby (according to a herm eneutical etymology) th at he “cuts beings off from the round of rebirths.”28 These usages bear witness to a Buddhist penchant to use language in ways o ther than merely denoting literal m eanings in a direct, simple, and discursively naive way. In this exoteric context, it may be noted, no one questions the use of such literary devices, but there has been great reluctance to adm it th a t such may have been th e case w ith the Tantras. There are, I believe, very good reasons to believe th a t the discourses we find in the Mahäyoga Tantras are similarly complex in their semiotics; and th at one of the prim ary modes of signification employed is, in fact, not th at of natural language, but rath er the higher-order semio­ logical system of connotative semiotics.

C O N N O TATIV E SEMIOTICS AS EXEGETICAL METHOD

The notion of connotative semiotics was first advanced by the Danish linguist, Louis Hjelmslev, and later elaborated by the French semiologist, Roland Barthes. C onnotative sem iotics—w hat Barthes also called m ythology or m ythic speech—is a second-order system of significa­ tion. It presupposes th e conventions of n atu ral language, and uses them to indicate complex ideas, obliquely yet strongly. The basis of the model is th e stru ctu re of th e linguistic sign first set out in Ferdinand de Saussure’s Cours de Linguistique Générale. According to this model, a linguistic sign can be analyzed into an arbitrary signifier (usually one or more phonem es or graphem es) and a signified (a sense being indicated). The union o f th ese tw o is w hat is known as th e sign—a complex, dual phenom enon com prising the plane of expression (the signifier) and the plane of co n ten t (the signified). In ordinary, prosaic, directly referential language—the kind that, many assume, is used in Tantric scripture and ritual—this is all one has. Signi­ fication takes place directly and, generally, unambiguously: I speak of a “table,” and you know exactly what I mean. However, this is by no means the only level on which hum an beings express them selves—particularly when they take to expressing more complex meanings of th eir common culture or to signifying ideological (or otherwise highly-motivated) propositions th a t for one reason or other do not lend themselves to straightforward denotation. Transcending these first-order Systems of direct signification, or denotation, then, are two higher-order Systems. The first of these is th at used to speak about language and the structure of signs—w hat has been called m etalanguage. This type of second-order discourse is used by linguists such as de Saussure to describe the functioning of signs. (It is, in fact, the mode of discourse used in the passage .you are now reading.) In this case, a complete sign from natural language

figure

4.1 Structure of the linguistic sign according to F. de Saussure

2. S ig n ified

1. Sign ifier 3. Sign

1. S ig n ifier

2. S ig n ified

Language: 3. Sign I. SIGNIFIER

META­

11. SIGNIFIED

LANGUAGE: III. SIGN

becomes a signified in the m etalanguage, w ith term s such as sign, signifier, and signified serving as the signifiers. This mode of discourse may be schematized in figure 4.2. The oth er second-order system, connotative semiotics, is (arguably) a ra th e r more subtle and (certainly) a m ore pervasive m ode of hum an communication. Unlike m etalanguage, w hich—although accessible to the u n tu to red —is largely the province of Professional linguists, connotative semiotics (while no doubt also susceptible to professionalization) is wellattested as frequently used by ordinary Speakers. In this mode, a complete sign from th e natural language serves, not as a signified, but as a signifier in the higher-order system. Barthes famously gives two examples of this mode of signification in Mythologies—th at of a phrase serving as a gramm atical example in a textbook and th a t of a picture of a saluting French soldier on the cover of

f ig u r e

4.3 S tr u c tu r e o f c o n n o t a t iv e s e m io tic s

1. S ign ifier Language:

2. S ig n ified

— 3. Sign

CONN.

I. SIGNIFIER

SEM:

Source: Adapted from Barthes (1972), 115.

II. SIGNIFIED III. SIGN

Language:

1. quia ego nominor leo

2. “b ecau se m y n am e is lio n ”

3. Sign [th e m ean in gfu l phrase] CONN. SEM.:

I. SIGNIFIER

II. PREDICATE AGREEMENT

III. “I AM A GRAMMATICAL EXAMPLE.”

Paris Match. Each example highlights an im portant aspect of connotative semiotics, so we should examine each in turn. The first example Barthes gives is of the phrase quia ego nominor Ieof occurring in a Latin textbook as an example of the gramm atical rule of subject-predicate agreem ent.29 Here, it is im portant th at the signifier be a real sign produced out of natural language—a meaningful Statement rieh w ith its own significance—and not m erely an arbitrary signifier within th e natural language. This first level of signification has already been expressed in the denotative enunciation of the rule th at subject and predicate should agree. W hat is w anted in this case is a concrete example, which (to be effective) can only be such a sign. Here, what is being signified is not th e m eaning of th e phrase (“because my nam e is lion”), but the gram m atical rule it instantiates. Surely, this is a very different use of language th an the merely literal, yet one which we encounter in a variety of forms nearly every day, with nary an eyebrow raised. This usage may be schematized as shown in figure 4.4. B arth es1 second exam ple derives from his experience seeing a picture on th e cover of Paris Match of a French soldier of African heritage saluting th e tricolor. Here, th e basic signifier is a photograph, which depicts th e soldier ju s t described. However, this literal analysis does not cap tu re th e signification taking place on th e cover of th a t m agazine. It is n o t ju s t a p ictu re o f a soldier m eant to com m unicate his appearance innocently to those readers w ith a special in tere st in soldiers and th e ir appearance. As B arthes indicates, th e presence of th a t p articu lar kind of soldier displaying ju s t th a t kind of patriotism itself (as a signifier) expresses a h ig h er-o rd er c o n te n t—a c o n te n t th a t, in

1. P hotograph

2. Salu tin g Soldier

Language: 3. Saluting Soldier I. SALUTING SOLDIER

II. FRENCH IMPERIALITY

(QUA PATRIOTIC COLONIAL) SEM.: III. “THE FRENCH EMPIRE? IT’S JUST A FACT!”

fact, may be said to be th e (if not schem atically or tem porally, herm eneutically) prim ary signification of th e image. The signified expressed by th is im age—th e p atrio tic colonial—is, as B arthes insightfully indicates, “French im periality.” The sign th u s co n stitu ted serves ideologically to n atu ralize th e French colonial presence in West Africa: The view er is sem iotically seduced into a w orld o f m eaning w herein th e French em pire is “ju s t a fact.” Based on these two examples, Barthes says of connotative semiotics th at it “is a type of speech defined by its intention (I am a grammatical example) much more th an its literal sense (my name is Hon).” But he adds th at “in spite of this, its intention is somehow frozen, purified, eternalized, made absent by this literal sense (The French Empire? It’sju st a fact: look at this good Negro who salutes like one ofour own boys). This constituent am biguity of m ythical speech,” he says, “has two consequences for the signification, which henceforth appears both like a notification and like a Statem ent of fact.”30 This constituent ambiguity of connotative semiotics—th a t its significa­ tion is defined prim arily by its intention, yet this intention is obscured or m ystified in th e process of signification by the m anifest content of the natural language sign—is of central im portance to how I understand this mode of signification to be operative in Tantric Buddhist ritual and scripture. It is precisely this ambiguity which, I argue, makes connotative semiotics a powerful tool in ritual perform ances of the kind undertaken by its practitioners.

C O N NO TA TIVE SEMIOTICS IN TANTRIC RITUAL

The quotidian rituals of the esoteric Buddhist traditions that advocate consumption of the five meats and ambrosias consist essentially of variants on a basic rite called a sädhana, literally an accomplishing or effecting. As Yael Bentor very clearly dem onstrates in her work on Indian and Tibetan rites of consecration,31 the structure of the sädhana constitutes the ritual template for all Buddhist Mahäyoga and Yogin! Tantra rituals. W hether they be fire ceremonies (homa), offerings of ritual cakes (ba/i), rites of prosperity or curing (paustika- or śãntika-karma)t consecrations of statues or the like (pratisthä), all are not only based upon but actually nested within the overarching and prim ary ritual pattern of the sädhana. The sädhana rite is also called th e self-creation or, perhaps one might say, self-resurrection (ätmotpatti). We learn of its structure and nature from several sources, but in the various traditions of the Guhyasamäja, among the most authoritative are the self-creation rites attributed to Nāgārjuna and Candraklrti.32 It is these sources upon which I will base my presentation here. The central aim of this self-creation yoga is for the practitioner to do away w ith th e perception of herseif as ordinary—as well as the pride that is believed to be associated w ith th a t perception—and to replace it with a perception of herseif as a divine, enlightened being, with the sense of proud em pow erm ent and universal efficacy th a t characterizes such a being. Such a profound transform ation is not considered to be an undertaking th at can be accomplished ju st so; rather, it is a highly ramified process th a t involves m editatively dying from the previous, unenlightened em bodim ent and ritually taking rebirth with a new, perfected identity. The ritual texts ascribed to Nāgārjuna and Candrakīrti describe the following main stages of the rite: Determ ining the site where the ritual should be performed; focusing on great compassion as the motivation; m editating on a protective perim eter; focusing on voidness as a way of elim inating ordinary perception and its pride; creating the cosmic foundation for the maņdala world; constructing the divine palace and its m aņdala environs; entering into the m aņdala of ultim ate reality (i.e., death) and arising in the form of the Buddha Vajradhara; perform ing a series of yogic exercises to bless one’s new born enlightened body, speech, and mind; again en ter­ ing into ultim ate reality/death; arising in a new em bodim ent to benefit others; and perform ing various enlightened activities.33 The various rit­ ual activities besides self-creation (consecration, destruction, etc.) all are

thought to gain th eir efficacy on account of being enacted, not by an ordinary person, but by an enlightened being. It is through self-resurrection as a deity in the sädhana as ju st described, th a t the practitioner assumes this om nipotent ritual identity. The first th ree stages are quite straightforw ard, and fall w ithin the general p atterns of Indian and Buddhist ritual practice: Finding a suit­ able spot (lovely and somewhat off-the-beaten-track), setting th e correct m otivation of universal compassion as required of a Mahāyāna practitio­ ner, and delim iting th e site w ith a proper protective boundary (analogous to th e rite o f sīmabandha in the Buddhist ordination rite).34 The practitioner th en focuses on th e fact th a t all things (including herseif) are void of an intrinsically-real Status—which serves in this context as the epistemological precondition for such a rite of radical reenactm ent of th e cosmogony. She then im aginatively creates a divine environm ent for this recreated personality to inhabit: th e maņdala, w ith its glorious palace suitable as the residence of a fully-enlightened divinity. The yogini th en “enters th e maņdala of ultim ate reality” (i.e., dies, leaving her ordinary personality), and arises in a thoroughly-accom plished, perfected form whose mind is suffused with the great compassion and wisdom of voidness cultivated previously. A variety of yogas involving the arraying and recitation of m antras a n d /o r m anipulation of vital airs are th en prescribed to reinforce and consecrate this identity; w hereupon she enjoys the type of beatific body known as th e sambhogakäya, a special lucid em bodim ent in which she interacts w ith o th er enlightened beings. She th en dissolves this rarified form again into th e clarity of death, thereby reen terin g th e so-called dharmakāya, an enlightened form in which the practitioner-qua-deity pantheistically identifies with the entire universe. Having heard the pleas of enlightened angels to take birth once again in order to benefit others, she subsequently arises in a concrete bodily form visible to all (called a nirmānakāya) and perform s enlightened activities. According to the instructions found in th e rite ascribed to Nāgārjuna, it is in this final state of realization and com passionate direm ption th at “one perform s th e activities of eating [things] such as th e five am brosias and so on.”35 It is here, th e n -s itu a te d in the context of the culmination of the Tantric ritual of self-creation—th a t I suggest one look to try to understand th e m eaning of the five m eats and th e five ambrosias for the Mahäyoga traditions. In o rd er to grasp th e semiosis im plicit in th e consum ption of the m eats and ambrosias at the climactic m om ent of the Tantric sädhana, it is

essential to understand what these substances signify in the overarching discourse o f contem poraneous, m ainstream Indian culture. Bhattacharyya’s Suggestion th a t these meats were delicious luxuries much desired by a repressed Buddhist ecclesia could not be further from the mark. I do not believe we are justified in m aintaining th a t they appear in the Tan­ tras m erely because they are tasty and the monks were seeking scriptural legitim ation for an exotic barbeque.36 They appear, rather, because they signify (or, better, instantiate) the violation of ritual purity. All five of these meats are distinctive within first-millennium Indian culture as, for lack of a b etter word, taboo meats. In the compendia of the dharmaśãstra-s, which (among other things) discuss in detail the rules governing the preservation of ritual purity in Indian society, these foods are among those generally classified as svabhäva-dusta7“polluted (and, thus, polluting) by their very nature.”37 That these restrictions were not merely academic notions, limited to the textbooks on dharm a, but functioning social strictures,38 is confirmed by th e testim ony of th e Chinese Buddhist monk Xuanzang, who, in his seventh-century account of his visit to India, observes th at “the m eat of such anim als as oxen, donkeys, elephants, horses, pigs, dogs, foxes, wolves, lions, monkeys, and apes is not to be eaten as a rule. Those who eat the meat of such animals become despicable and detestable to the public and are expelled to th e outskirts of the city.”39 Although he does not specifically m ention hum an flesh here, we may safely take it for granted (given what eise we know of first-m illennium Indian society) th at this was also considered ritually im pure.40 Much th e same can be said of the five ambrosias. All five are bodily fluids, and thus polluting according to orthodox Brahmanical Standards operative at th e tim e. It is on account of th eir daily contact with such substances th a t physicians in ancient India were considered impure and excluded from the orthodox rites.41 Such bodily excretions may also be said to fall into th e polluting category of the sahrllekha7 th a t which is impure because it is “disgusting to the mind.”42 In short, contact with the five meats and the five ambrosias so absolutely violates the most central purity strictures in Indian society that reference to them in Tantric Buddhist ritual and scripture could only have constituted a deliberate semiosis. They signify that which is disgusting and polluting. In fact, an explicit awareness of this is revealed in the text of the Guhyasamäja Tantra itself: In chapter 14, the Tantra specifically refers to feces and urine as “foul-smelling and disgusting.”43Similar examples from Tantric Buddhist

literature in Sanskrit could be multiplied almost indefinitely. The them e of the revolting (jugupsa) is a consistent trope in these traditions. It is clear th at the authors considered these substances disgusting.44 Thus, perhaps som ewhat ironically, it seems to have been the muchm aligned early Orientalist scholars w ho—though reading the Tantras through the lenses of their own cultural presuppositions, and not those of first-m illennium India—were able to “read” (if not understand) the Tan­ tras correctly. To the early Indologists, the Tantras were full of the filthy and degrading, the foul and offensive. As I think should be very clear, a reader in first-m illennium India would have thought m uch the sam e.45 Nor, it is im portant to add, was this accidental—the disgusting nature of these substances was explicitly noted. The Tantric literature does not, therefore, as some would have it, reflect th e naive im portation of m ar­ ginal tribal magical techniques th a t ju st happened to be repulsive to the cultural m ainstream . Rather, th e authors of th e Tantras were speaking precisely the m ainstream cultural language of Indian society and pushing its buttons in such a systematic fashion th at it could only have been deliberate. In short, in disparagingly referring to th e Tantric Subhäsitasamgraha as “a caricature of both the teachings of earlier Buddhism and legitim ate Yoga,”46Cecil Bendall missed th e most salient point: That is precisely w hat it was meant to be. F urther evidence for this may be derived from observing the two verbs th at are consistently used in scriptural contexts related to the five meats and am brosias—eating and offering. On my reading, every instance in which th e five m eats and am brosias are m entioned in th e Guhyasamäja places them in one of these two contexts. Of the tw enty-nine references to m eats and ambrosias in this scripture, tw enty-one of them occur explic­ itly as th e objects of acts of eating or offering or some variant thereof. Fourteen of these instances involve actions of eating or consum ption;47 while two fu rth er examples can be inferred to be so.48 Seven occur in relation to acts of offering;49and one more is plausibly associated with this act in the comm entary.50 Of the five remaining, th ree do not refer to the five meats and nectars under discussion,51 leaving only two out of the twentynine th at are not explicitly associated in the prim ary text w ith eating or offering. A strong case, based on context and commentary, can be made to consider these two as also associated w ith consum ption.52 Hence, of the tw enty-nine instances, all twenty-six relevant references to the meats and ambrosias in the Guhyasamäja Tantra associate them w ith actions of oral consum ption an d /o r offering to divinities.

f ig u r e

4.6 S u g g e s t e d s e m io lo g y o f a n tin o m ia n d is c o u r s e in t h e M a h ä y o g a T a n tra s

M ain stream Indian culture:

MAHÄYOGA

1. B e e f

2. P ollution

3. P ollu tin g b e e f I. (EATING) POLLUTING BEEF

II. NONDUAL GNOSIS

III. “I HAVE ATTAINED COMMUNION (YUGANADDHA).” (“IT’S JUST A FACT!”)

What, then, does this entail with regard to our understanding of their semiology w ithin th e Tantric traditions they represent? I think it is safe to say th a t these two activities—eating and w orshipping/offering—are the quintessential m om ents of im portance to orthodox, Dharmic purity strictures: They are prim e occasions of danger, w herein one runs the risk of ritual pollution. In intercourse with the divine, much emphasis is placed on th e notion th a t proper protocols be observed, lest one’s Sta­ tus decrease—given th e gods’ transcendent purity, the postulant m ust be appropriately fastidious. Similarly, in the act of eating, wherein one accepts foreign bodies into one’s own—and, thus, one’s bodily Constitu­ tion is potentially com prom ised—a concern for purity strictures is param ount in th e Indian religious context. Thus, by interpolating these sets of polluting substances into the two archetypal liminal acts of the purity calculus, this literature seeks to ham m er home the fact th a t what is at issue in these contexts is ritual purity. Fundamentally, this is a discourse about purity and pollution (notably, an overtly pervasive them e in the later Buddhist Tantras)—not the special, intrinsic qualities of particular meats and bodily fluids. What does it mean, then, for a practitioner of the Mahäyoga Tantras, having gone through th e process of self-creation as an enlightened Bud­ dhist divinity, to eat from a skull a foul soup of polluting m eats and bodily fluids? In this semiosis (as can be seen schematized in table 4.6), the complete sign from the natural language of m ainstream Indian culture—the signifier beef, and so on in semiological union w ith its signified ‘ ritual Pollution”—acts as a signifier in the process of ritual consum ption con­ sidered as a discourse. The signified in this semiosis is the attainm ent

of the enlightened state of nondual gnosis (advayajñāna), called in some sources53 com m union (yuganaddha)—the ultim ate goal of th e practitioner in which the deluded perception of things as having an intrinsic nature (pure or polluting, good or evil) is transcended. This state of com m union is described thus in th e final ch ap ter of Nāgārjuna’s influential work on the practice of the Guhyasamäja, the Five Stages (Pañcakrama): Defilement and purification— Knowing them from the perspective of ultimate reality The one who knows [them as] one thing Knows [the] communion [stage].54 In particular, for our purposes, Nāgārjuna goes on to m ention th e fol­ lowing dualistic concepts which are likewise transcended by the accomplished practitioner. As oneself, so an enem y. .. As one’s mother, so a w hore,. .. As urine, so wine. As food, so shit. As sweet-smelling camphor, so the stench from the ritually-impure As words of praise, so revolting words ... As pleasure, so pain.55 Thus, by dram atically (and I use this term advisedly) dem onstrating th eir transcendence of conventional dualistic categories of purity and pollution in the concluding portion of the rite of self-creation, the practitioners of these traditions signify ritually th a t th eir attainm ent of the enlightened state—which, it is w orth rem em bering, is the starting point and the ending point of Buddhist Tantric practice—is, in fact, a fait accompli. In this way, the consum ption of the five m eats and ambrosias in these rituals constitutes an example of connotative semiotics. W hat im plications does this have for our und erstanding of tex t and ritual? If we re tu rn now m om entarily to th e examples given by Barthes, we will recall th a t th e re were two im p o rtan t points th a t he stressed w ith regard to th e effect of connotative semiotics. For one, he said, it is speech th a t is guided prim arily by its intention. That is, th e phrase serving as a gram m atical example m eans less its sense in natural language

than it signifies its in ten tio n to serve as an example of a gram m atical rule. This is th e first point. Second, this intention, which is the key elem ent of its signification, is occluded in the process of signification. “Its intention is som ehow frozen, purified, eternalized, made absent by this literal sense.56 One can see how this is a very effective technique in the kinds of manip­ ulative discourses of advertising and ideology th at Barthes took as his pri­ mary objects of study. Viewers of the 1950s issue of Paris Match on which our Afro-French soldier stood saluting, who may well have been experiencing a crisis of confidence regarding the French empire in Africa, were meant to come away reassured—it is this intention that is prim ary in the signification. Yet, th a t intention is in no way explicit; it is occluded. As Barthes reads the image: “The French Empire? It’s ju st a fact: look at this good Negro who salutes like one of our own boys.”57 The viewer is reas­ sured of the strength of French im periality via a profound, and seemingly ideologically innocent coup d’ceil, in a way impossible to achieve through the rhetorical persuasion of, say, an op-ed piece on the viability of the Situation in French West Africa. Yet this higher level of signification is shrouded by th e prim ary act of signification, ensuring deniability: It’s ju st a nice picture of a soldier, after a l l . . . This type o f signification is also present in advertising. Here the intention is obvious and clearly prim ary—to seil product. And it is this intention th a t it is also vital to keep occluded insofar as possible. If the rational m ind is alerted to the signification, it loses much of its power—it is demystified. Connotative semiotics are thus ram pant in the world of commercials: Products do not signify themselves, they signify ideas or pleasurable States. The SUV one sees climbing effortlessly into the garage of the Himalayan m onastery does not signify itself: It signifies freedom, peace, and power. The boy who begs his m other for the one Christmas present he really m ust have—a Cross Your Heart bra—has clearly been reading th e images he sees on television. Clever boy th at he is—skilled at reading commercial discourse through hours spent before the tube—he is unconcerned with the direct denotative signification of the brassiere with which he is confronted, but is completely taken up with its connota­ tive significance of total com fort and security. Similarly, in the ritual context of the sädhana—calling as it does for the practitioner to renounce her rational, discursive knowledge of her own ordinary and lim ited personality—connotative semiotics are used as a more direct, m ystifying mode of signification th an ordinary rhetorical

suasion. This latter had, in fact, been tried before in Buddhist pedagogical history. There is an extensive corpus of exoteric scriptures and philosophical literature devoted to advancing the notion th at all beings are intrinsically enlightened by nature, th at all are possessed of the tathägatagarbha. This is by no m eans th e m ost effective way of convincing som eone of th a t fact, however.58 There is simply too m uch evidence to the contrary available to the rational mind; ju st as, if one were to try to rationally convince a young boy of his need for a brassiere, one would be sorely pressed. However, in the ritual context of the self-creation rite, in which th e practitioner blissfully eats conventionally-defiling substances with impunity, having adopted th e attitu d e of th e overlord of the m aņdala (mandalãdhipati), there is no need for fu rth e r convincing. The Sugges­ tion is accom plished in a coup d ’æil—as (Vajra-) Barthes m ight say “The enlightened stage of communion? It’s ju st a fact: Look as I savor this soup of beef, dog, semen, and feces!” Elsewhere in his writings, in discussing an exhibition of shock photos, Barthes gives fu rther indications of the im portant signifying function of connotative semiotics. Speaking of photos th at deal with “the shocking,” he notes th a t “it is not enough for the photographer to signijy the horrible, for us to experience it.”59 What I believe he is getting at here, is the fact th at there is a distancing effect to the structure of natural language. Recall the adm onishm ent of a thousand w riting teachers, to show, not say: It is not enough for one to inform another th a t som ething is horrible for th at person to have a truly visceral, em pathetic experience of its horror. if th e intention is to share a taste of the horror and not merely to convince an o th er th a t A or B falls into a certain abstract category of experience called horror, one cannot use merely denotative discourse. Connotation is essential: It allows com m unication to be guided by an ulterior intention (to shock), and yet for th at intention to be occluded (so as not to make th e experience overly contrived). Otherwise, if the signifiers chosen are drawn solely from natural, denotative discourse, says Barthes, they “have no effect on us; the interest we take in them does not exceed the interval of an instantaneous reading: It does not resound, does not disturb, our reception closes too soon over a pure sign.”60 It is this understanding of th e contrast between the prosaical discourse of denotation and th e poetical discourse of connotative semiotics th at I believe has been leveraged in this aspect of Tantric ritual. Thus, although such direct signification is found elsewhere in the rite of self-creation— the practitioner recites the m antra om śünyatã-jñãna-vajra-svabhãvãtmako

’ham (“Om I am th e very adam antine nature of the gnosis of voidness”). Yet, in order to ensure the maximal experiential impact of the performance of self-resurrection, th e autosuggestion of inhabiting a divine identity, transcending purity and pollution, the authors of the rite have also chosen to employ the m ore visceral, more instantaneous mode of connotative semiosis. It is here, in this semiotic process, I believe, th at some of th e “m ystery” of the Tantras may be found.61 Seen in light of this dynamic, then, the original question of the meaning of th e five m eats and five ambrosias in Mahäyoga Tantra scripture and ritual would seem to call for some reconsideration. The question of w hether these words—cow m eat, dog meat, elephant meat, horse meat, human flesh, feces, urine, blood, semen, marrow—signify real beef, urine, and so on, I would suggest, is close to irrelevant. In the context of the selfcreation rite we have analyzed earlier, what is im portant is their semiotical function, th eir ability to instantiate ritual pollution as a lived fact. What is essential to th e signification of the rite are the five meats and five ambrosias as signs, insofar as they function as signifiers in the higher order System. In the natural language out of which th at sign is borrowed, the actual signifier is, as de Saussure insists, arbitrary.62 Thus, I would argue, th e question th a t has troubled m odern scholarship—is it “shit” or not?—is beside th e point.In fact, m uch the same seems to have been indicated by authors of the Guhyasamäja Tantra itself—even in its earliest stratum (chapters 1 to 12). In chapter 12, after enum erating a set of five yogic accom plishm ents th at correspond to eating each of the five meats, the text blithely notes th a t “if all these kinds of m eat cannot be obtained, while m editating, one should conceive [of them] as really existent.”63 The concrete reality of flesh as a denoted signified is extraneous; what matters is its significance w ithin the com m unity of Speakers of the Tantric yogin/L

CO N N O TA TIV E SEM IOTICS IN TANTRIC SCRIPTURE To frame th e question of the interpretation of antinom ian elem ents in Tantric ritual in term s of w hat its signifiers denote in natural language is—in th is case, at least—fundam entally to m isconstrue th e semiosis involved in th e ritual act of consuming defiling substances. The two clas­ sical positions on Tantric herm eneutics, the literalist and figurativist, ultimately fail to account for the signification observable in these tradi­ tions. Both in th e signs used in ritual perform ance and in the composition

of scripture, the antinom ian Tantras betray a clear semiotical structure. For reasons we will explore in chapter 6, th e leading lights of th e later esoteric movements sought to elevate a gnosis of nonduality as th e cen­ tral Buddhist goal, challenging (rhetorically, at least) concepts of purity and pollution (caste, astrological auspiciousness, etc.) widespread in the ritual and social mores of th e earlier esoteric dispensations (e.g., the Mañjuśrīmülakalpa64). The ritual consum ption of the five m eats and five ambrosias explored herein clearly reflect this central concern. Given th e sophistication of the Buddhist literary context out of which th e Mahäyoga Tantras evidently arose,651 think we can only conclude th a t th e notion th a t the literal m eaning m ust be presum ed to be original and prim ary can only be based on an unspoken assum ption th at Tantrism is prim itive—an assum ption w ith a long history in O rientalist scholar­ ship (some of which we have traced in preceding chapters), yet one th at would seem to be based on a failure to read the sources fully critically.66 On th e oth er hand, to suggest instead th a t th e discourses of taboo meats and foul fluids constitute m erely a code, hiding a secret transm ission of esoteric yogic techniques, is to miss th e historical resonance of these discourses in th e contexts of both ritual perform ance and contem poraneous culture. In approaching th e question of Tantric in terp retatio n in such a way, scholars of the figurativist tendency have paid little attention to aspects of Tantric discourse besides the denotative. Even if the signifiers “feces” and “u rin e” refer to the sense organs and th e ir objects (as Candrakirti claims), this signification is still well w ithin th e param eters of natural language. That is, th e code model m erely replaces the signified in th e sign relation w ith a variant elem ent, such th at one forms nothing m ore th an a simple sign composed, for example, of th e signifier “b ee f” and th e signified “th e form aggregate” (rüpa-skandha). This does not, I argue, capture th e essence of the m ode of com m unication used in the T antric discourses, although I adm it it is one suggested by some trends in th e com m entarial literature. I suspect th at much of the reason for the neglect of connotative modes of signification on th e p art of scholars of Tantrism has to do w ith the fact th a t these higher-order Systems are seen to operate m ost clearly in ritual—a notoriously neglected area for much of m odern religious (and, perhaps in particular, Buddhist) studies. Little attem pt has been made to situate the discourses of the Tantric scriptures within th eir proper ritual contexts, although there survives a wealth of Indian Buddhist literature on precisely this subject. This is all th e m ore paradoxical because—for

most of th e history of the m odern study of th e Tantras—it has been a scholarly m antra th a t the Tantras are primarily ritual (i.e., practical, not theoretical) texts. On th e o th er hand, I believe this semiosis can also be clearly and unm istakably discerned in the narratives of Tantric scripture. There are a num ber of im portant episodes internal to the Guhyasamäja Tantra itself, which I feel very strongly corroborate the view th at these discourses are not m eant to be taken as a direct, simple acts of denotative signification, but th a t—in scripture as well as in ritual—it is the experience of nondual gnosis th a t is the prim ary object signified. For instance, in a key passage that appears in GST chapter 5—a passage that has attracted a great deal of attention from m odern scholars in th at it is one of the most consistently and blatantly Tantric (i.e., transgressive) in the entire text—the Lord Bud­ dha Vajradhara teaches the assem bled buddhas and bodhisattvas th at “even those who com m it great sins such as the inexpiable sins (änantarya) will be successful in this buddha vehicle, the great ocean of the Universal Vehicle (mahãyãna).”67Further, he teaches th at those who violate the most basic Buddhist precepts—who take life, lie, steal, and are sex-maniacs— and even, notably, those who eat feces and drink urine, are considered by him to be “fit for th e sädhana” (bhavyäs te khalu sadhane). In a final flourish, he informs th e assembly th at those who commit incest with mother, sister, or daughter, will “attain vast success,” while the one who makes love to the Buddha’s own m other will attain buddhahood. At the conclusion of this pithy teaching, th e bodhisattvas in attendance are said to have been “amazed and astonished.” Why, they ask, is this bad speech (durbhäsita) being spoken in th e m idst of th e enlightened assembly? To this query, the buddhas in attendance reply th at they should not speak so: That this is the pure teaching of all the buddhas. Upon hearing this reply, the bod­ hisattvas are so overwhelmed th at they actually pass out, whereupon the Lord has to rouse them by the light rays of the meditative samädhi called (notably) the space-like nondual vajra (ākãśasamatãdvayavajra). This narrative is notew orthy in several ways; and a full unpacking of its im plications has m uch to contribute to our understanding of the literary techniques of th e Buddhist Tantras. First and foremost, it very clearly expresses a self-consciousness of the fact th at the teaching given by Vajradhara in this very passage in the Guhyasamäja Tantra is blatantly heretical. However, it is far too simple to consider this merely a device for giving scriptural sanction to deviant practices68 or as evidence of contem poraneous social Opposition.69 For this episode and similar passages

elsewhere in the literature do not merely suggest the sanction of one or the other unorthodox religious praxis. Rather, in this sermon, the Buddha Vajradhara systematically hits virtually every subversive note in the Bud­ dhist scale of religious values. Like the m eats and ambrosias, this is in no way a semiotically innocent list: The practices advocated by Vajradhara represent the precise inversion of m ainstream Buddhist ethical norms. The bodhisattvas, not surprisingly, are shocked and scandalized by this teaching, calling it bad speech (durbhäsita). This term , too, is significant, as it alludes to the Buddhist herm eneutical rule of thum b th at all th a t is well-spoken or good speech (subhäsita) is the revealed Word of the Bud­ dha.70 Equally resonant here, however, is the fact th at this term refers not merely to th at which is poorly spoken in some abstract sense, but rath er constitutes a distinct category of transgression of the Buddhist Monastic Discipline (vinaya).71 Thus, the bodhisattvas’ assessm ent of the teaching is th at it is not Buddha speech (subhäsita) but rath er heresy (durbhäsita); and w hen th eir enlightened classm ates insist th a t this is, in fact, th e “pure teaching of the buddhas,” th eir im aginations are beggared—they simply cannot process th e fact th a t th e pure teaching of the buddhas and the defiled teachings of the heretics are nondual seen from th e perspective of an enlightened being who has attained com m union—and they black out. In th e end, th e reader is told, the bodhisattvas are enabled to come around—to digest th e cognitive dissonance of this teaching, to tolerate th e signification enunciated by the Buddha Vajradhara, enough so as to regain consciousness—only w hen they are touched by the “light” of the gnosis of nonduality.72On my reading, once again, in scripture as in ritual, the transgressive elem ents of the Tantras reveal them selves to be motivated discourses, whose prim ary semiotical interest is to stress the Tan­ tric message of the nonduality of pure and im pure, sacred and profane, im m anent and transcendent. It is w orth stressing th at this is not an isolated instance. Exactly the same semiotical structure inform s a similar narrative th a t is the subject of chapter 9 of the Guhyasamäja. Both of these chapters, notably, are from th e earliest stratum of the scripture (chapters 1 to 12) and thus may be considered reflective of the core values of the antinom ian traditions. This chapter, too, consists of an initial serm on by Lord Vajradhara, astonishm ent of the bodhisattvas, and a rebuke and elucidation from th e tra n ­ scendent lords. It differs only insofar as the bodhisattvas do not in this case require spiritual light rays—a detailed explanation by the transcen­ dent lords suffices.

The teaching given by Vajradhara in GST chapter 9 consists of five sections. Each gives instruction in a visualization in which a maņdala is transform ed into one of the Five Transcendent Lords (Aksobhya, Vairo­ cana, Amitäbha, Amogha, and Ratnaketu), w hereupon the m editator is to envision perform ing a heinous act toward buddhas. In the first visualiza­ tion, th e buddhas of th e three times em anated by Aksobhya are pulverized by a blazing vajra and all beings, too, are similarly annihilated by the secret vajra. Ending on a more pleasant note, these all thereby become bodhisattvas (“sons of th e Victors”) in the vajra[-family] buddhafield [of Aksobhya]. In th e second visualization, Vairocana transform s into bud­ dhas appearing as jew eis th a t are then seized (i.e., stolen); these, too, become bodhisattvas, em inent among sages. The third involves visualizing Amitäbha transform ing into a space filled with buddhas who take on the form of women, w hereupon they are enjoyed sexually. In the fourth, Amogha becomes all buddhas, who are then lied to and betrayed. Finally, the fifth visualization consists of Ratnaketu who radiates all buddhas, who are th en verbally abused. Naturally, ju st as in GST chapter 5, the bodhisattvas in the audience are “amazed and astonished” and inquire why the Lord would say such an astonishing thing. The transcendent lords reply th at they should not consider it base (hxna) or revolting (jugupsita—note the recurrence of this term). They explain th a t this teaching is based upon the perception by buddhas of th e voidness (and, thus, nonduality and purity) of all things, using well-established Buddhist philosophical similes.73The chapter then concludes with the “amazed and astonished bodhisattvas, th eir eyes wide with wonder,” reciting: The conventional rings forth in the pure, the non-conceptual, In the most wondrous things that appear like space!74 Once again, we see a narrative in which the Tantric teaching involves violating basic Buddhist precepts (against killing, stealing, sexual misconduct, lying, and verbal abuse); the bodhisattvas are befuddled; and the issue is clarified through reference to the nondual voidness of reality. In case th ere should be any doubt about what is intended, the chapter ends with a verse of praise th a t specifically refers to nonconceptuality and purity: th e same complex of ideas th at informs the entire antinom ian Ori­ entation found in these Tantras.75 Lest the reader worry th at nonduality is not referenced, im mediately following this verse is the chapter title76

th at reads “the ninth recital, the chapter on the pledge of the m eaning of th e nondual reality, the ultim ate tru th ” (paramãrthãdvayatattvãrthasamayapatalo navamo ’dhyãyah).

❖ ❖ ❖

One can thus clearly see the utility of the model of connotative sem iot­ ics in the interpretation of the antinom ian aspects of the Tantric Buddhist traditions. Indeed, this approach shows great prom ise in the in terp re­ tatio n more generally of those cultural form ations we are in th e habit of calling “religious.” B arthes’ analytical m ethod allows a m ore critical approach to the uses of discourse in hum an society, which are rarely confined to literal denotation. This is true, evidently, not merely of the Contemporary advertising and pop culture w ith which Barthes was concerned, but for classical traditions as well. It is one very powerful tool in discerning the structuring principles th at inform discourses and prac­ tices—including m odern historiography—th a t seem on the surface to be about som ething eise, their m anifest content.77 It has been noted in another context th at a certain scholarly approach to th e study of myths mistook them for explanations of natural phenom ena, leading to the conclusion th a t they reflect a magical or pre-logical form of thought. Against this approach, it has been argued th at “m yths may think with natural objects or categories; they are alm ost never about natural objects or categories . . . the seasons may serve as a medium for thinking about periodicity, regularity, order, distinction, transform ation and place.”78 Likewise, I would suggest th at the currently prevalent, liter­ alist approach to the interpretation of the rhetoric of the Buddhist Tantric literature errs in assuming th at these discourses are about meats, fluids, and despised castes, thus failing to see beyond the ostensive content. Rather, I would argue, late-first-m illennium Tantric Buddhist scripture and ritual use these signifiers to think w ith—as a m edium for thinking about (and acting w ith reference to) ritual purity, freedom, and gnosis. Though it may seem a trivial observation to scholars of Hindu Tantrism th at ritual praxis should involve a calculus of purity and pollution, this is unfortunately not the case w ith m uch scholarship on Buddhist Tant­ rism. Much of this scholarship has been unfortunately limited to the cor­ pus of Mahäyoga and Yogini Tantras taken in isolation from their earlier esoteric forebears, and, am ong these, have focused alm ost exclusively

on the soteriological, ra th e r th an ritual, aspects. Accordingly, they have not been attentive to th e extent to which these traditions situate them selves against the purity concerns of the dualistic Buddhist Tantras and of broader currents in Indian society—an aspect of these traditions th at is counterintuitive to those who approach them as “Buddhist.” One is sometimes confronted, for instance, by stark claims such as “in the Buddhist Tantra of these periods [ca. A.D. 750-1200], caste purity and pollution are not fundam ental issues, in contrast to Hindu Tantra.”79 One wonders how one could come anyw here close to such a view, having read even a handful of Buddhist Tantras. Although we have heretofore restricted our analysis to examples of meats and ambrosias, in chapter 5 we will see that caste discourses figure very prom inently in antinom ian Buddhist Tantras as well. It is by grasping th e essential semiological connection of th e antino­ mian discourses to th e broader discursive context of Indian and Bud­ dhist concepts of purity and pollution—Bourdieu’s “com plete system of signifiers”—th a t one can avoid another m isleading notion—th a t Tant­ ric antinom ianism is concerned w ith “transgression as such.” As I think should be clear, th e simple notion of a “transgressive sacrality”—“the ritual inversion of social taboos, as a way of laying claim to psychological and physical powers repressed by social convention”80—is inadequate to in terp ret th e m aterials analyzed here. One does not see in these anti­ nomian traditions a m ethod to “overcome the fears th at kept uncharted parts of th e psyche repressed, with the aim of releasing those repressed forces and harnessing them for power and knowledge to be used in accomplishing specific ends.”81 This is merely to read the semiology of Contemporary “spirituality” into late-first-m illennium Indian religion, making sense of it through psychologization. The transgressions of this tradition are pointed and specific: They take their meanings from the cul­ tural context w ithin which they were deployed and are manifestly aimed at occasioning an experience (even if contrived) of nondual gnosis. Lest my m eaning and my Opposition to literalism be misconstrued, how­ ever, it is im portant to note that, in stressing the semiotical nature of the rite of consum ing the m eats and nectars, I do not necessarily m ean that these substances were not actually consumed. In fact, I would argue th a t— although I suspect th at actual consum ption was very rare in practice by any but virtuosi82—the possibility of such consum ption m ust be available (at least as a limit case) for the system of semiosis to function. Although the real world may be irrelevant in many cases of hum an signification, in

general it functions as the necessary horizon of possible experiences and signification. Consuming im pure substances ritually would seem to be of this latter kind: The notion th a t one could (and m ight) actually do it is im portant for the full impact of the semiosis of (non-)revulsion to occur. Consequently, w hatever the com m entators m ight say—and this is an area th at calls for considerable further research—a strongly figurative reading of these rites is as flawed as a strongly literal approach. These two interpretative camps both fail to grasp the fact that hum an praxis takes place in a reality th at is socially constructed and thus always already imbued with meaning. To put it in Lacanian term s, from the m om ent a child enters the symbolic order, her actions are inescapably significant; and yet the signifi­ cation of each person (generated from their personal “im aginaiy order”) is constrained by a concrete context (The Real). If one fails to account for either of these two aspects, one fails to account for a hum an phenomenon. What connotative semiotics offers is a way out of this dilemma. To take one final example, consider the traditions concerning th e Tan­ tric saint Virüpa, of whom stories are told of his excessive imbibing of alcohol. Taking a literalist tack, Ronald Davidson rejects th e “religious” interpretation advanced by the exegete M unidatta—which Davidson considers an example of later dom estication for m onastic consum ption— preferring to construe the song as “a hum orous acknowledgem ent th at [Virüpa] preferred to spend tim e in a bar ra th e r than in religious environm ents.”83 However, the approaches of Davidson and M unidatta both occlude the most essential aspect of the semiosis. On the one hand, Davidson’s strongly literalist reading fails to account for references in the song to “sixty-four jugs,” the “ten th door,” and other tendentious allusions that belie a semiotical regime alien to ordinary drinking songs. On the other hand, Davidson is exactly correct to note th at, in eliding the tippling altogether, M unidatta overshoots the interpretative mark. As he rightly observes “within India in particular, drinking is a low-status form of recreation.”84 In the context of Indian cultural norm s, then, the sign of a saint drinking alcohol speaks much the same language of purity and pollution as th at of a yogin consuming beef and sem en—it signifies a transcendent attainm ent of nondual gnosis. The drinking is as essential to the semio­ sis as is the religious interpretation: Were the drinking om itted, nothing would rem ain but a Standard exoteric saint; were there only the drinking, Virüpa would be nothing but a libertine. Different contexts might reveal differences of emphasis, but the integrated sign of a drinking saint is an essential and irreducible part of the discourse.85

T P -á j v

«1

(5] THE PRACTICE OF INDIAN TANTRIC BUDDHISM

A

s N o t e d in c h a p te r four, a lth o u g h th e a n tin o m ia n p ra c tic e s o f th e h ig h e r B u d d h ist T a n tra s fe a tu re a n ax ial se m io sis e s se n tia l to t h e ir p r o p e r u n d e r s ta n d in g ,1 th is d o es n o t e n ta il th a t th e s e

p ra c tic e s a r e n o t in fa c t to b e p e rfo rm e d . I su g g e ste d th a t a t le a st th e p o ssib ility o f su c h p e rfo rm a n c e is e sse n tia l to th e sem iosis: If c o n sid e re d ju s t a sy m b o l, th e full im p a c t w o u ld n o t be p o ssib le a n d th e sem io sis o f c o n n o ta tio n w o u ld b e re d u c e d to m e re flat d e n o ta tio n . In th is c h a p te r, we w ill a d v a n c e o u r d isc u ssio n o f th e tra n sg re ssiv e a sp e c ts o f th e B ud­ d h ist T a n tra s b y a p p ly in g m e th o d s o f s tru c tu ra l d isc o u rse an a ly sis to th e a n tin o m ia n (or, n o n d u a lis t) e s o te ric s c rip tu re s t h a t d escrib e th e p ra c tic e s fo llo w ers o f th e s e tr a d itio n s a re to e n g a g e in. T h a t is, w e w ill in v e n to ry re fe re n c e s to “p r a c tic e ” ac ro ss a la rg e c o rp u s o f p rim a ry so u rce s. Such an a ly sis w ill re v e a l t h a t th e tra n s g re s s iv e ac ts a d v o c a te d in th is lite r a ­ tu r e in v o lv e m u c h m o re th a n m e re c o n s u m p tio n o f re v o ltin g , p o llu tin g m e a ts a n d a m b ro sia s, c o m p re h e n d in g as w ell sexuality, in to x ic a tio n , an d a b ro a d r ä n g e o f o th e r b e h a v io rs c o n s id e re d im m o ra l. M oreover, it w o u ld seem as if th e e n tir e ty o f th e p r io r B u d d h ist tr a d itio n s is to b e c a st aside: T h ese s c rip tu re s p ro sc rib e b asic B u d d h ist d ev o tio n al acts, s c rip tu ra l recita tio n , a sc e tic d isc ip lin e s, a n d a w h o le rä n g e o f n o n a n tin o m ia n e s o te ric rite s, su c h as fire -o ffe rin g s, m a ņ d a la ritu a ls, a n d m a n tr a re c ita tio n . As w e sh a ll so o n d iscover, th e in te r p re ta tiv e ch a lle n g e s fac in g sc h o l­ ars o f T a n tric tra n s g re s s io n a re b y n o m e an s lim ite d to th o s e in v o lv ed in th e se m io tic s o f r itu a l a n d s c rip tu re . T h e re are f u r th e r issu es o f in te r p r e ­ ta tiv e m e th o d th a t a re c ru c ia l to g ra sp in o rd e r p ro p e rly to u n d e r s ta n d th e lite r a r y e v id e n c e available an d , o n th a t basis, to com e to co n c lu sio n s a b o u t h o w T a n tric p ra c tic e w as en v isio n e d a m o n g th e c o m m u n itie s th a t

produced these fundam ental sources. In this chapter, we will explore one such challenge in which once again we are confronted w ith signification th at involves more th an mere denotation. Throughout the scriptural Cor­ pus of th e transgressive Tantras, the Interpreter confronts an equally subtle challenge to in terpretation—th at posed by the pervasive em ploym ent of so-called term s of art. A term of art is “a word or phrase having a spe­ cial m eaning in a particular field, different from or more precise than its custom ary meaning.”2Although they are similar in some respects, term s of art are to be distinguished from what are called technical term s. These latter bear specialized meanings, but th eir use is generally restricted to the one field. Terms of art, on the other hand, are often common words in the general vocabulary. For instance, byte is a technical term in Computer Science referring typically to eight units of digital inform ation. Bit, how­ ever, is a term of art, referring to one such unit in the context of Comput­ ing, yet its prim ary m eaning is m ore general—simply, a small am ount or piece. In interp retin g Tantric Buddhist scriptures, it would seem th a t several m ajor term s of art have been almost entirely overlooked by m odern scholars. This is perhaps understandable, given the polysemic nature of these words. Unlike technical term s such as kotava (the nam e of a vital air in th e subtle body3), which are unique to th eir particular contexts, or the use of code words such as cam phor for semen, term s of art can seem like they are being used in their ordinary sense. Recognition of these term s as term s of art is, however, essential, insofar as failure in this regard creates and sustains broad and systemic m isinterpretation of Tantric literature and of the traditions th at produced (and were, in turn, produced by) these works.4 One term in particular, of crucial im portance to th e question of the nature of Tantric practice, is the term practice (carya) itself. In this chapter, I will argue th a t m odern scholarship has consistently and markedly m isconstrued the nature of practice in the antinom ian tra ­ ditions, insofar as references to caryä employed as a ferm of art have been understood instead as referring in the generic to Tantric practice. For instance, the tw enty-first chapter of the Samvarodaya Tantra (the chapter describing practice, or caryãnirdeśapatala) has been described as presenting th e religious practices of tan tric teachers and th eir disciples.5 More recently, the quite specific practice (caryä) I shall explain has been represented in quite general term s as “the post-initiatory practice which an initiate of tan tric Buddhism is perm itted to perform .”6 As will be clear from the evidence analyzed in w hat follows, the practice referred to in

these passages (and m any others) is by no means the practice of initiated Tantric Buddhists, but m erely one, very rarified, practice. That is to say, it is perhaps b etter construed as a proper noun—not practice, but “The Practice.” This crucial term of art appears across virtually the entire corpus of Bud­ dhist Mahäyoga and Yogini Tantras (and some śãstras) as well as in a number of Śaiva Tantras. It frequently occurs with the term vrata ([religious] observance) in the same contexts. Both term s are, of course, of extremely common usage throughout Indian and Buddhist religious parlance. Caryä, for instance, is by far th e m ost common generic term for the spiritual undertakings of buddhas and bodhisattvas. The Mahävastu, for instance, frames its trea tm e n t of the career of the Buddha Śākyamuni by referring to four types of bodhisattva practices (bodhisattvacarya) undertaken by the future buddha.7 The famous work of Śāntideva on engaging in the practices of enlightenm ent is called the Bodhicaryãvatãra, while chapter 16 of his Śiksãsamuccaya (Compendium of Learning) is devoted to the good conduct (bhadracaryä) of high resolve, dedication to the welfare of beings, and so on. The culm inating chapter of Asańga’s Mahãyãnasūtrãlamkãra, the stability in practice chapter (caryāpratisthãdhikāra), treats inter alia of four practices leading to enlightenm ent: The practice of the [six or ten] tran scen d en t virtues (pãramitācaryã), the practice of the [thirtyseven] accessories of enlightenm ent (bodhipaksacaryä), the practice of the superknowledges (abhijñãcaryā), and the practice of developing beings (sattvaparipākacaryã).8 These same four practices are treated in the prac­ tice chapter (caryäpatala) of the Bodhisattvabhümi9 Similarly, vrata appears regularly in Indian Buddhist literature, in even less marked a sense. For instance, in the Bodhisattvabhümi, the renunciant bodhisattva is said to be superior to th e householder bodhisattva on account of his m aintenance of vrata-niyama (i.e., celibacy and restraint).10 Thus, encountering the term caryä in Tantric literature, certainly the most obvious and natural understanding would be th at this term and related passages describe Tantric practice per se or in general, ju st as one would in terp ret the same word in works of exoteric Mahāyāna literature. Attentive reading em ploying structural analysis, however, reveals quite clearly th at this term of art recurs throughout antinom ian Tantric litera­ ture w ith a referent th at is quite specific and markedly consistent across a variety of sources, Buddhist and non-Buddhist.11 In this usage, caryä and vrata appear to be largely synonym ous and often occur in com pound one w ith the other, w ith eith er of th e two

taking the dom inant syntactical position. That is, one sees the term s caryävrata and vratacaryä, with identical m eanings.12 In addition to these forms (which are the most common), th e two also frequently occur in com pound w ith qualifiers related to ideas of secrecy or m adness, for example, guhyavrata (esoteric observance), guhyacaryä (esoteric prac­ tice), prachannavrata (concealed observance), unmattavrata (m ad/insane observance), and so on. There also exists a cluster of interrelated term s th at appear in the same contexts, and which seem to be largely synonymous, th at appear to be variant species of the same genus. These may be seen in table 5.1, together with the works w herein they occur.13 Of these, one in particular, vidyävrata (knowledge observance, spell observance, a n d /o r consort observance)—which is treated as essentially equivalent to caryävrata/vratacaryä in Buddhist and Śaiva sources—is w orth noting at this point as its Signal significance will become more evident as our analysis proceeds. All of these expressions refer to the same dass of ritual behaviors. This usage is consistent across a wide spectrum of texts, from which I conclude th a t this term of art is central to the ideology of the nondualist Tantras wherein they occur. The injunctions of the rite include certain very specific things th at are proscribed, things prescribed, sites w herein they are to be perform ed, specifications for the optimal tim e and duration of th eir perform ance, and specific accoutrem ents th a t are needed for or beneficial to th e rit­ ual acts. In w hat follows, we will examine each of these aspects of the caryävrata so as to discern the essential param eters of the concept in the Tantric traditions. I will dem onstrate th a t caryävrata/vratacaryä is (l) a highly specific term of art in the literature of the Buddhist Mahäyoga and Yogin! Tantras, signifying a very precise undertaking, (2) th at close atten ­ tion to th e semiology of th e rite reveals a very clear ritual intent th at is evident throughout the Buddhist literature, and (3) th a t the sources explicitly (if at times som ewhat obliquely) stress th at this rite is appropriate only in quite specific and elite ritual contexts with very specific prerequisites. I will also show (4) th at this term of art is also common to the contem poraneous nondual Śaiva Tantras of the Vidyāpītha, and th at the patterns of usage across the two traditions suggest an alternative way of understanding the interaction of these communities. Specifically, I argue th at close attention to the available literature suggests th at the semiology of the early Śaiva observance differs significantly from th at of the early Buddhists as outlined in (2), and th at the nature of the Buddhist and Śaiva variants fu rth er suggests th at (5) this distinctively Buddhist semiology

THE PRACTICE OF INDIAN TANTRIC BUDDHISM

[137]

T a b LE 5.1 C e n tr a l a n d r e la te d te r m s

CENTRAL TERMS AND SCRIPTURES WHEREIN ATTESTED caryä (“p ra ctic e ”): GST, CPAMA, BK, ST, HT, SU, YS, CMT, VĀ, GS, CMP, YRM, KMT vrata (“[ascetical] observance”): AVP40, (MVT), CPAMA, BK, KMT caryä-vrata (“practice observance”): ST, LS/HA, CMP vrata-caryä (“observance practice”): CPAMA, ST, HT, YS, YRM, KMT guhya-vrata (“esoteric observance”): ST, GS guhya-caryä (“esoteric practice”): GS tattva-caryä (“reality practice”): ST vīra-caryāvrata (“heroic practice observance”): LS/HA trividhā caiyā (“three-fold practice”): CMP, YRM prachanna-vrata (“concealed observance”): GS

RELATED TERMS (PROBABLY SYNONYMOUS, OR CLOSELY SO) vidyä-vrata (“consort observance”): (MVT), GST, KMT unmatta-vrata (“m ad/intoxicated observance”): ST, SU, GS, TD bhusuku-vrata (“observance o f eating, sleeping, and defecating”): CMP yoga-caryä (“yoga practice”): ST, SU samantabhadra-caryä (“universally good practice”): SU *avadhūti-caryā (kun ’dar gyi spyod pa) (“central channel practice”): ST *dig-vijaya-caryä (phyogs las rgyal ba’i spyod pa) (“practice victorious in all directions”): ST *ālińgana-caryā (’khyud pa’i spyod pa) (“embracing practice”): ST paricaryä (“entertainm ent”?): MK

c a m e u l t i m a t e l y t o e x e r t a p r o f o u n d in f l u e n c e o n t h e la t e r Ś a iv a u n d e r s t a n d in g o f t h e r ite (a n d , in d e e d , t h e i r u n d e r s t a n d in g o f T a n tr ic p r a c tic e in g e n e r a l) a f t e r t h e n i n t h C en tu ry. T h is c o n c lu s i o n fu r t h e r s u g g e s t s t h a t , c o n t r a t h e t h e o r ie s o f a “s u b s t r a t u m ” o r a t o t a l B u d d h is t d e p e n d e n c e o n Ś a iv is m , (6 ) t h e f e a t u r e s o f r e lig io u s o b s e r v a n c e (vrata) s h a r e d b y t h e s e t w o g r o u p s a r e t h e p r o d u c t o f a Z e itg e is t o f a n t in o m ia n p r a c t ic e w h e r e i n

(as is in evidence throughout Indian religious history), groups utilized a common vocabulary of term s and rites to which they gave th eir own dis­ tinctive inflections, and in which the borrowing was mutual. W hat, then, is the caryävrata? In short, in th e nondualist Tantric lit­ erature of the Buddhist Mahäyoga and Yoginī Tantras, this term and its equivalents come to encapsulate virtually all those features th a t have come most strongly to be associated with Tantrism (or so-called “Siddha T antrism ”) in the m odern mind: Sex, to be sure, but also eerie places (cem eteries, lonely fearsome forests, etc.), eccentric dress, and ecstatic behavior, including the W h o le sa le rejection of the m ainstream practices of exoteric Indian religion. This term is very prom inent in the later Tantric literature—so much so that frequently an entire chapter is seen to be dedicated to this observance. This is the case for the Guhyasamäja Tantra, as well as th e Mahãkãla, Buddhakapāla, Samputodbhava, Hevajra, Candamahärosana, Laghusamvara/Herukābhidhãnay Samvarodaya, Yoginīsamcāra, and Vajräralli Tantras. The ninth chapter of the Buddhakapãla Yoginī Tantra, for example, is devoted to the topic of caryä. It describes a rite th at a yogin undertakes w ith an “absolutely excellent wom an” (atyantavarãńganã) presumably for the purpose of engaging in sexual yogas. Taking a skull bowl (kapäla) in hand, the yogin wanders naked, with hair unbound, begging from house to house and eating w hatever is p ut in the bowl, regarding all things with equanim ous delight. The yogin is here called, as elsewhere, a vratin—a practitioner who has taken on a specific religious observance (vrata).14 Considering the data set in aggregate (especially tables 5.2 to 5.6), it can easily be seen th at the treatm ents as a whole in these Tantras foreground: (a) liminal, isolated spaces, and (b) funereal and horrific item s of dress. They fu rth er consistently (c) advocate certain behaviors (sex, wandering, commensality, song and dance, and consum ption of meats, alcohols, and bodily fluids) and (d) proscribe others (recitation, m editation, worship, b u rn t offerings, textuality, image devotion, and attention to astrological auspiciousness). A structural analysis of the ränge of these sites, accoutrem ents, prescriptions, and proscriptions is revealing. Consulting the chart on sites (table 5.2), one can see th at th e most com­ mon are the m ountain top, charnel ground, and either a generic uninhabited space (vijana) or varieties of liminal zones (the “suburban” pränta, crossroads, confluences of rivers, beaches, etc.). The Samvarodaya Tantra has quite an extensive list: charnel ground, a place with a lone lińga or tree (ekalmga, ekavrksa), forest, m ountaintop, riverbank, ocean shore, garden, broken well, em pty house, crossroads, city gate, palace gate, house of

X

04

2 W Q OC o