JIABS

Journal of the International Association of Buddhist Studies

Volume 32 Number 1-2 2009 (2010)

Obituaries
Hu Hayan VON HmuBER In memoriam, ~~** Ii Xianlin (6. August 1911 -11. Iuli 2009) ............................

Articles
James A. BENN The silent sa1!Lgha: Some observations on mute sheep monks. .. Vincent ELTSCHINGER Ignorance, epistemology and soteriology - Part I Frances GARRETT Eating letters in the Tibetan treasure tradition .
11

39

85

Sarah H. JACOBY "This inferior female body:" Reflections on life as a Tibetan visionary through the autobiographical eyes of Se ra mkha' 'gro (bde ba'i rdo rje, 1892-1940). . . . . . . . . . . . . . . . . 115 Andrew MCGARRITY Aryadeva's gradual stages: Their transmission from India to Tibet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Jan-Ulrich SOBISCH Interpreting the tantras: A Tibetan debate on the numbers of adepts admissible to tantric consecration. . . . . . . . . . . . . 213

Contents

Tibetan scholasticism in the 11th and 12th centuries

Contributions to a panel at the XVth Congress of the International Association of Buddhist Studies, Atlanta, 23-28 June 2008

Guest editors: Pascale Rugon and Kevin Vose

Pascale RUGON and Kevin VOSE Introduction - Unearthing the foundations of Tibetan Buddhist philosophy . . . . . . . . . . . . . . . . . . . . . . . . . . 237 Kazuo KANO rJiTog BIo ldan ses rab's position on the Buddha-nature doctrine and its influence on the early gSmi phu tradition. . . 249 Kevin VOSE Making and remaking the ultimate in early Tibetan readings of Santideva . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Pascale RUGON The origin of the theory of definition and its place in Phya pa Chos kyi sen ge's philosophical system . . . . . . . . . . . . . . 319 Jonathan STOLTZ Phywa pa's argumentative analogy between factive assessment (yid dpyod) and conceptual thought (rtog pa). . . . . . . . 369
~

Georges DREYFUS and Drongbu TSERING Pa tshab and the origin of Prasangika . . . . . . . . . . . . . 387 Thomas DOCTOR In pursuit of transparent means of knowledge - The Madhyamaka project ofrMa bya Byan chub brtson 'grus . . . . . . . . 419 Chizuko YOSHIMIZU Zan Than sag pa on theses (dam bca', pratijfia) in Madhyamaka thought. . . . . . . . . . . . . . . . . . . . . . . . . . . . 443

Contents

Heidi KOPPL Ron zom Chos bzan on Mahiiyoga and Madhyamaka . . . . . . 469

Notes on the contributors

483

The Journal of the International Association of Buddhist Studies (ISSN 0193-600XX) is the organ of the International Association of Buddhist Studies, Inc. As a peer-reviewed journal, it welcomes scholarly contributions pertaining to all facets of Buddhist Studies. nABS is published twice yearly. Manuscripts should preferably be submitted as e-mail attachments to: editors@iabsinfo.net as one single file, complete with footnotes and references, in two different formats: in PDF-format, and in Rich-Text-Format (RTF) or OpenDocument-Format (created e.g. by Open Office). Address books for review to: nABS Editors, Institut flir Kultur- und Geistesgeschichte Asiens, Apostelgasse 23, A-1030 Wien, AUSTRIA Address subscription orders and dues, changes of address, and business correspondence (including advertising orders) to: Dr Jerome Ducor, lABS Treasurer Dept of Oriental Languages and Cultures Anthropole University of Lausanne CH-1015 Lausanne, Switzerland email: iabs.treasurer@unil.ch Web: http://www.iabsinfo.net Fax: ;:t-41 21 692 29 35 Subscriptions to nABS are USD 55 per year for individuals and USD 90 per year for libraries and other institutions. For informations on membership in lABS, see back cover. Cover: Cristina Scherrer-Schaub Font: "Gandhari Unicode"designed by Andrew Glass (http://andrewglass.org/ fonts.php)
Copyright 2010 by the International Association of Buddhist Studies, Inc.

EDITORIAL BOARD KELLNER Birgit KRASSER Helmut Joint Editors BUSWELL Robert CHEN Jinhua COLLINS Steven COX Collet G6MEZ Luis O. HARRISON Paul VON HINUBER Oskar JACKSON Roger JAINI Padmanabh S. KATSURA Sh6ryii KUO Li-ying LOPEZ, Jr. Donald S. MACDONALD Alexander SCHERRER-SCHAUB Cristina SEYFORT RUEGG David SHARF Robert STEINKELLNER Ernst TILLEMANS Tom

Print: Ferdinand Berger & Sohne GesmbH, A-3580 Horn

In memoriam

*iJt# Ji Xianlin
(6. August 1911 - 11. Juli 2009)

Haiyan Hu-von Hiniiber

Ji Xianlin ist am 11. Juli 2009 kurz vor seinem 98. Geburtstag in Peking gestorben. Ais einer der ftihrenden Indologen, Tocharologen und Buddhismusforscher hinterHisst er der wissenschaftlichen Welt ein reiches Erbe. Ji Xianlin wurde im Dezember 2002 anlaBlich der 13. Konferenz der lABS in Bangkok zum Ehrenmitglied gewiihlt. Ais Ji Xianlin am 06. August 1911 in der nordchinesischen Provinz Shandong (Llf~:) geboren wurde, fand gerade eine Wende in der chinesischen Geschichte statt, im Zuge derer das Kaisertum durch die Xinhai-Revolution (*~1.iI:r) gesttirzt wurde. In den darauf folgenden Jahrzehnten befand sich China in einem Umbruch wie selten zuvor. Ais Zeitzeuge hat Ji Xianlin die Entwicklungen des Landes beinahe ein Jahrhundert miterlebt und mitgestaltet. In seinem Lebenslauf spiegelte sich diese bedeutende Epoche der geisteswissenschaftlichen und gesellschaftlichen Wandlung Chinas vom 19. bis ins 21. Jahrhundert wider. Der aus einer armen Bauernfamilie stammende Ji Xianlin wuchs bei seinem Onkel in der Provinzhauptstadt Jinan (~Jj) auf, der ihm eine solide Schulbildung ermoglichte. Sein Weg zur Indologie und Buddhismusforschung ftihrte tiber das Germanistik-Studium. 1930 begann er an der Eliteuniversitat Qinghua em) in einem englischsprachigen Programm das Fach "Europaische Literatur" mit Schwerpunkt "Deutsche Philologie" zu studieren, das er 1934 mit der Bakkalaureus-Arbeit "The Early Poems of HOlderlin" abschloss. Durch die Lehrveranstaltung "Ubersetzungsliteratur buddhistischer Texte" (f~~tl~)(), die Prof. Chen Yinke C~~'11
Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 5-9

Haiyan Hu-von Hintiber

1890-1969), der erste im Westen ausgebildete Buddhologe Chinas, dort abhielt, kam der junge Ji Xianlin zum ersten Mal mit Sanskrit und Buddhismus in Beruhrung. 1m Spatherbst 1935 traf Ji Xianlin, dank eines Austauschprogramms mit Deutschland, in G6ttingen ein, wo er im Frtihjahr 1936, als Ernst Waldschmidt (1897-1985) gerade an die GeorgAugust-Universitat berufen worden war, Indologie als Hauptfach wahlte. Da sein Doktorvater E. Waldschmidt zu Beginn des Zweiten Weltkrieges eingezogen wurde, ubernahm sein bereits emeritierter Vorganger, Emil Sieg (1866-1951), die Betreuung des einzigen Doktoranden der Indologie und las mit ihm ~gveda, paJ:.linis A~tadhyayI, Patafijalis Mahabha~ya, Dandins Dasakumaracarita usw. Daruber hinaus hielt es E. Sieg fur wichtig, dass der werdende Indologe aus China Tocharisch erlernte, und brachte Ji Xianlin daher sowohl das Ost- als auch das Westtocharische bei. 1941 wurde Ji Xianlin mit der Dissertation "Die Konjugation des finiten Verbums in den Gathas des Mahavastu" promoviert. Kriegsbedingt musste er die Zeit bis zur Heimkehr mit einem Lektorat im sinologischen Seminar uberbrucken, das er bereits 1937 angetreten hatte. Der zehnjahrige Aufenthalt in Deutschland bildete die Grundlage fUr Ji Xianlins Leben als Orientalist. 1m Spatherbst 1945 verlieB er Deutschland und kehrte 1946 nach elf Jahren uber die Schweiz, Frankreich, Vietnam und Hongkong nach China zuruck, wo er von der Peking-Universitat als ordentlicher Professor mit dem Auftrag berufen wurde, eine Fakultat fUr Orientalistik aufzubauen, deren Grundungsdekan er war; er fuhrte das Dekanat bis 1983 selbst. 1956 wurde Ji Xianlin zum ordentlichen Mitglied der Chinesischen Akademie gewahlt. Seine wissenschaftliche Tatigkeit, die in G6ttingen begonnen hatte, dauerte fast 70 Jahre. Seine Forschungsarbeit in China entwickelte sich in zwei Abschnitten: (a) von 1946 bis zur "Kulturrevolution" im Jahr 1966, und (b) vom Jahr 1977, dem Jahr, in dem Reform und Eroffnung Chinas begannen, bis zu seinem Tod im Jahr 2009. Die durch die zehnjahrige Kulturrevolution verursachte Lucke lasst sich aus seiner Bibliographie deutlich erkennen. In den 80er und 90er Jahren hingegen, bis ins hohe Alter hinein, hat Ji Xianlin auBergew6hnlich viele wissenschaftliche Arbeiten ver6ffentlicht.

~~* Ii Xianlin (6. August 1911 - 11. Juli 2009)

Nach der Berufung an die Peking-Universitat stellte Ji Xianlin bald fest, dass dort die Grundvoraussetzungen nicht vorhanden waren,. urn philologische Forschungen, wie er diese aus Gottingen kannte, fortzufiihren. Deswegen musste er sich auch anderen Forschungsgebieten zuwenden. Dieser Wechsel der Forschungsrichtung hat im Laufe der Zeit zu einer stetigen Ausweitung seiner Arbeitsgebiete geftihrt. Bereits 1998/99 erschien sein damaliges Gesamtwerk in 24 Banden. 1 An dieser Stelle seien nur einige Schwerpunkte aus Ji Xianlins indologischer und buddhologischer Forschung in Erinnerung gerufen: (1) Altindische Philologie bzw. Sprache des frtihen Buddhismus; (2). Tocharisch, bes. Maitreyasamitinataka; (3) Chinesische Ubersetzungen aus der Sanskrit-Literatur; (4) Bearbeitung von Sanskrit-Handschriften aus China; 2 (5) Geschichte des Buddhismus in Indien, China und Zentralasien; (6) Geschichte Indiens; (7) Kulturhistorische Beziehung zwischen China und Indien. 3
1 Beginnend mit 2009 veroffentlicht der Verlag Waiyu Jiaoxue yu Yanjiu (:7r;g.~*~1iJfntJ!t&H) in Peking das Gesamtwerk von Ji Xianlin in 30 Banden *ll**~~U 30 ~. Diesen Hinweis habe ich Prof. Seishi Karashima (ili,(wf~ii!;;), ebenfalls einem personlichen Schiller von Ii Xianlin, zu verdanken. 2 Zu Ii Xianlins Publikationen in dies en vier Bereichen vgL Hu-v. Hiniiber, *ll** IiXianlin (1906-2009), in: Tochanan and Indo-European Studies, vol. 12, Kopenhagen 2010 (in Druck). Seine Publikationen von 1929 bis 1991 wurde von Li Zheng (*~) zusammengetragen in Papers in Honour of Prof Dr. Ii Xianlin on the Occasion of His 80th Birthday, ed. by Zh. Li & Zh. Jiang et aI., 2 vols., Nanchang 1991, S. 1-35; vgl. ferner *ll***ifl~1tjg:~U Ii Xianlin Xueshu Lunzhu Zixuan Ii [Von Ji Xianlin selbst ausgesuchte wissenschaftliche Abhandlungen], Beijing 1991, S. 670-675. 3 In die folgenden Sammelbande wurden die wichtigsten Publikationen von Ji Xianlin, die er seIber fiir nennenswert hielt, aufgenommen: *~ ** (Ii Xianlin), J:j:l~p)c1t~m1*~M!i)c~ Zhongyin Wenhua Guanxishi Lunwenji [Abhandlungen zur Geschichte der kulturellen Beziehung zwischen China und Indien], Beijing 1982; *~** (Ii Xianlin), ~P~ t11~;g.M!i)C~ Yindu Gudai Yuyan Lunwenji, Ausgewiihlte kleine Schriften zur altindischen Philologie / Selected Papers on the Languages of ancient India, Beijing 1982; *~** (Ji Xianlin), 1*~l15t~iJlffir,,~

Haiyan Hu-von Hiniiber

Trotz seiner vielfaltigen Verpflichtungen betrachtete sich Ji Xianlin in erster Linie als Hochschullehrer. Daher hat er bis zum 85. Lebensjahr seine Lehrtatigkeit ausgetibt. In der von ihm beinahe 40 Jahre lang geleiteten Fakultat fUr Orientalistik, die als Chinas groBte Institution fUr das Fachgebiet gilt, werden heute wie damals Sanskrit, Hindi, Urdu, Bengali, Tamil, Persisch, Arabisch, Japanisch, Koreanisch, Birmanisch u.a. unterrichtet. Mehr als 6.000 Fachkriifte wurden hier ausgebildet, darunter etwa 30 chinesische Botschafter fUr den diplomatischen Dienst. Ji Xianlin ktimmerte sich besonders intensiv urn die Heranbildung talentierter Nachwuchskrafte und scheute keine Mtihe, den von ihm betreuten Studierenden Fortbildungsmoglichkeiten im Ausland zu eroffnen. Er setzte sich ebenso fUr seine Mitarbeiter in der Fakultat wie fUr seine Studenten ein. Seine Zielstrebigkeit in der Forschung und seine personliche Bescheidenheit machten ihn beliebt und bertihmt. 1953 wurde Ji Xianlin Mitglied des ersten Volkskongresses der Stadt Peking (:ltJ?:-mlSll:t5b). 1956 wurde er in die kommunistische Partei Chinas aufgenommen. Er war Mitglied des Nationalkomitees der zweiten (1954), der dritten (1959), der vierten (1964) und der ftinften (1978) Politischen Beratenden Konferenz Chinas (:i:~ lSlI:t5b), eines der hOchsten politis chen Gremien in China. 1983 stieg er zum Mitglied des Stiindigen Ausschusses des Nationalen Volkskongresses Chinas CA:A:1t~fn auf. Ji Xianlins Zivi1courage und sein wissenschaftspolitisches Engagement zeigten sich auch darin, dass er seine Meinungen sehr offen auBerte, ohne sich jemals von offiziellen Vorgaben der Zentralregierung beeinflussen zu lassen. Ais Botschafter von Kultur und Wissenschaft bereiste Ji Xianlin zahlreiche Lander Asiens und Europas wie Indien, Nepal, Japan, Korea, Thailand, Syrien, .A.gypten, Irak, Burma und Deutschland,
@ Yuan~hi Fojiao de Yuyan Wenti [Zur Frage der Sprache des friihen Buddhismus], Beijing 1985; ~~***#JJJHj::~~ Ii Xianlin Xueshu Lunzhu Zixuan Ii [Von Ji Xina1in se1bst ausgesuchte wissenschaftliche Abhandlungen], Beijing 1991; ~~**1Jt~*#J~JlijJt~ Ii Xianlin Fojiao Xueshu Lunwenji [Kleine Schriften zum Buddhismus von Ji Xianlin], Taibei (Series of the Chung-Hwa Institute of Buddhist Studies 4) 1995.

~~** Ii Xianlin (6. August 1911-11. Iuli 2009)

auch die ehemalige DDR, die ehemalige Sowjetunion und Taiwan. JiXianlin legte besonders groBen Wert auf die freundschaftliche Beziehung zwischen Indien und China als zwei Lander mit graBen Kulturen, die sich immer gegenseitig bereichert hatten. Er besuchte Indien mehrmals (1951, 1955, 1978 und 1985), manchmal fur langere Zeit, und traf mit mehreren Prasidenten der indischen Republik zusammen. Er war auch ein gesuchter Gesprachspartner fur indische Medien und Chinaforscher. Uber zwei Jahrzehnte war Ii Xianlin Schirmherr der "Stiftung fUr Indisch-Chinesische Freundschaft", die von der britischen Schriftstellerin Han Suyin ins Leben gerufen wurde. Wissenschaftliche Kontakte pfiegte Ii Xianlin hauptsachlich mit Europa und Japan. Eine enge Freundschaft verb and ihn ferner mit der thailandischen Prinzessin Maha Chakri Sirindhorn, die an der Peking-Universitat studiert hatte. In China selber wurde Ji Xianlin mit Forschungspreisen fUr seine oder von ihm geleitete Projekte uberhauft. Hinzu traten zahlreichen Ehrungen aus dem Ausland. 1m November 1992 verlieh ihm die Sanskrit-University Varanasi den hOchsten Lobpreis. 1998 erhielt er von der Universitat Teheran die Ehrendoktorwurde fur seine Beitrage zur Zentralasienforschung. Am 5. Juli 1999 uberbrachten ihm der Prasident der Sahitya Academi und der indischen Botschafter in China die Ehrenurkunde eines "Honorary Fellow of India's National Academy of Letters". 2000 uberreichte ihm die deutsche Botschaft in Peking die goldene Medaille seiner alma mater, der Universitat G6ttingen. 1m Dezember 2002 wurde er anlasslich der 13. Konferenz der lABS zum Ehrenmitglied gewahlt. Besonders hervorzuheben ist, dass die indische Regierung am indischen Nationalfeiertag 2008 Ji Xianlin bisher als dem einzigen Chines en den Orden Padma Bhushan verlieh. Am 12. Mai desselben Jahres ernannte die Japanische Akademie Ji Xianlin als ersten Chinesen aus der Volksrepublik China und als dritten Nichtjapaner nach Sylvain Levi und Louis Renou zum Ehrenmitglied. Als eine Leitfigur fur die wissenschaftliche Welt Chinas reicht Ji Xianlins Einfluss weit uber seine Leistung als Fachgelehrter hinaus. Die Erinnerung an ihn wird weiterleben als an einen neuzeitlichen Bodhisattva, der die Weisheit des Ostens mit der Wissenschaft des Westens zu verbinden wusste.

The silent saTflgha: Some observations on mute sheep monks*

James A. Benn

Introduction

The question of how to define a Buddhist monk or Buddhist community is one that is not only of interest to scholars of monasticism; it also occupied the subjects of our studies. This article takes as its starting point the ways in which the monastic community (sarrtgha, sengqie {~{tJa) was classified in canonical and non-canonical Buddhist texts that circulated in medieval China (from roughly the fourth to the tenth century CE). When we look at some authoritative scholastic texts that Chinese Buddhists translated we may be surprised by their characterisation of the types of monks that were included in the sarrtgha. These materials may challenge some of our commonly-held notions about Buddhist monks and the Buddhist monastic ideal. In particular, so-called "mute sheep monks" who are ignorant of scripture, unable to distinguish right from wrong, or even ugly from beautiful, are said to form part of the normative sarrtgha. When some Medieval Chinese practitioners
This article is based on a paper I gave at the conference on Buddhist monasticism held at the University of British Columbia in 2003. It was organised by Professor Chen Jinhua, and sponsored by the Tzu Chi Foundation, Canada. A number of other papers given at the conference have now been gathered together published in revised form as Benn, Meeks and Robson, eds. 2010. I thank Chen Jinhua for inviting me to participate and for his many useful comments on an earlier draft of this article. Shayne Clarke provided valuable advice as I first started to wrestle with the problem of mute sheep monks. I should also like to thank Robert Sharf for his helpful suggestions as I revised the paper for publication.
Journal of the International Association of Buddhist Studies Volume 32' Number 1-2 2009 (2010) pp. 11-38

12

James A. Benn

combined these visions of what the salJ1gha was, or what it could be, with other, sometimes unrelated, doctrines, the result was an idealistic conception of a separate assembly of mute sheep monks that was in many ways quite the opposite of that described in scripture. These mute sheep monks were presented as being diligent in keeping the precepts and in performing austerities, self-effacing but not afraid to discipline their peers. Unlike some mute sheep monks described in scripture, who were unable to resolve disputes in the larger Buddhist community, these mute sheep were supposed to act as the leaders of the salJ1gha. How, and why, did some medieval practitioners arrive at this somewhat unexpected ideal of the mute sheep monk?

The mute sheep salJ1gha in the Great Perfection of Wisdom Treatise

The Dazhidu lun *~Il~ (Great Perfection of Wisdom Treatise) attributed to Nagarjuna, is a text that defies easy categorization, but no one would question its importance for scholars who seek to reconstruct the background to the worldview of many early medieval Chinese Buddhists.! The Treatise contains the following pas! Lamotte 1944-1981. While commonly referred to by that title, the Dazhidu lun was known by a variety of names in medieval China. Among the Dunhuang manuscript versions and in citations by medieval authors we find such titles as Dazhi lun *~~ (Great Wisdom Treatise), Dazhi shilun *~~~ (Great Wisdom Explanatory Treatise), Dazhidu jing lun *~ m'i\!~ (Treatise on the Great Perfection of Wisdom Scripture), Moheyan shilun ~~iiJ1>J~~~ (Explanatory Treatise on the Mahayana), Mohe boruo boluomiduo jing shilun ~~iiJB~:;fi5~~!fl~i\!~~~ (Explanatory Treatise on the Mahaprajiiaparamita scripture). The last form is also attested in abbreviated form in the catalogue Kaiyuan shijiao lu mr:TC~l~~)R (Record of Siikyamuni's teachings compiled during the Kaiyuan reign-period) 4, T no. 2154, 55:513a, and appears at the head of Sengrui's preface to the Taish6 canon version (no. 1509). Dazhi shilun *~~~~, which is found in an earlier catalogue Chu sanzang jiji tf:\=~~2.~ (Collected records on the translation of the Tripitaka) 10, T no. 2145, 55:74c, appears to be another abbreviated variant. Mohe boruo boluomiduo jing shilun is probably the oldest title, and corresponds to the Sanskrit Mahiiprajriapiiramitopadeasiistra. The term upadea also appears in the biographies of Nagarjuna, see T 2047, 50.l84c.l8

The silent sal1lgha

13

sage, which, when I first read it, seemed to me to be rather remarkable. In response to the question, "what is the meaning of the term 'sa1'[lgha'?" (yun he ming sengjia ~fiif~{~{jJo) the Dazhidu lun supplies us with the following answer: sa1'[lgha in Chinese means "assembly" (zhong m.); several bhik~us (biqiu ~ci:) in one place are called a sa1'[lgha, in the same way that the term "forest" indicates a collection of large trees in one place. Considered separately, the trees are not a forest, but if one takes them away one by one the forest ceases to exist; just so, an individual bhik~u is not the sa1'[lgha, but the sa1'[lgha no longer exists if the bhik~us are all removed. Having defined the term, the Dazhidu lun goes on to expound on four types of sa1'[lgha, here using the abbreviated term seng {~. Given that this single glyph has also been used in China since early medieval times to designate an individual monk (and this remains the conventional meaning of the character today) there is inevitably a certain amount of slippage in this passage and throughout our discussion between the corporate sa1'[lgha and the monks who make up that sa1'[lgha, and hence between the concepts of "mute sheep monk" and "mute sheep sa1'[lgha."
~1'1WiZ:911 ~~1'1W ~~1'1W
0 0 0 0 0 0 0

ujl:::1'1W jlf{'1W
0 0 0

-z::fBJ;g~~1'1W 1-if~~1i1Z
0 0 0 0 0

!rD5~5* ijEJ3Ufrfl,W *1~lli ~;g~~1'1W -z::1PJ;g~~1'1W 1i1Z~ o!r D~5* ~~~11= ~;g~~1'1W -z::1PJ;gJIjl:::1'1W 9!lE~1i1Z~ ~t!it~~ ~ ~J3Ufrfl,W ~%I'~~ ~%I~$~$ ;:fi~1'1W~ o=A~~ ~ij5
0 0 0 0 0

~)~ I!!J:~~ ":'frDB:::7J~Ajf&~ij511=iit ~;gujl:::1'1W 1'1W ;:fi~A;:fi~~Ao 1.1iZ:9:lf1:cr 1TiZ:9ioJlli ~;g'lf1!W

0 0 0 0 0 0 0

-z::1PJ;g'lf

There are four kinds of sarrzgha: the sarrzgha that knows shame (youxiu seng ~~1'1W, literally "sarrzgha that has shame"), the shameless sarrzgha (wuxiu seng ~~1'1W), the mute sheep sarrzgha (yayang seng ujl :::1'1W), and the true sarrzgha (shi seng 'lf1'1W). Why is the sarrzgha that knows shame so called? [The monks of this sarrzgha] keep the precepts and do not break them. They are pure in body and speech. They can distinguish between beautiful and ugly,

and T no. 2058, 50.318bI6. Since the Dazhidu [un is a discussion of doctrine in question and answer format, it would seem to belong to the genre upadesa. See Demieville 1973,470 n. 1; on the attribution of the text to Nagarjuna see pp. 475-476.

14

James A. Benn but they have not yet attained awakening. Thus it is called the sa1'(tgha that knows shame. Why is the shameless sa1'(tgha so called? [The monks of this sa1'(tgha] break the precepts. They are impure in body and speech. They commit every kind of evil. Thus it is called the shameless sa1'(tgha. Why is the mute sheep sa1'(tgha so called? Although they do not break the precepts, [the monks of this sa1'(tgha] have dull faculties and lack wisdom. They cannot distinguish between beautiful and ugly. They do not know what is trivial and what is serious. They do not know what is considered a transgression and what is not. If there is some issue within the sa1'(tgha and two people are arguing, they cannot resolve it. They are silent and do not speak, like white sheep who cannot make a sound even when they are slaughtered. Thus it is called the mute sheep sa1'(tgha. Why is the true sa1'(tgha so called? [The monks of this sa1'(tgha] are whether they are those who abide in the four fruits (phala, guo saik/ia (xueren ~}J or asaik/ia (wuxue ren ~~)J - and those who cultivate the four accesses. 2 Thus it is called the true sa1'(tgha. 3

*) -

If we restate this definition in somewhat crude terms, the Dazhidu lun appears to claim that the only "true saJ?1-gha" consists of those

monks who are at the end of the path of practice (those who "abide in the four fruits"), or actively working towards that goal (those who "cultivate the four accesses"). But other (inferior?) saJ?1-ghas also exist: there is a saJ?1-gha composed of good monks who have a sense of shame, a saJ?1-gha of bad monks who are shameless, and finally a saJ?1-gha which consists of stupid monks who stand mute like sheep when confronted with disputes in the community. We may sum up the qualities of the mute sheep monk (or member of the mute sheep saJ?1-gha) in the Dazhidu lun as follows: they do not break the precepts but they have a low spiritual capacity. They are
2 The four fruits are the attainment of the states of stream-enterer, oncereturner, non-returner, arhat. The four accesses refer to the accesses to those fruits. The term saik/ia applies to one who is still studying and has not attained the stage of arhat, while asaik~a refers to one whose path is complete: an arhat. 3 Dazhidu {un *~Jj'Rfi Tno. 1509, 25:80a. Cf. the translation of Lamotte 1944-1981: 202.

The silent sarp.gha

15

incapable of distinguishing between right and wrong, and cannot resolve the disputes of others. Their silence on those occasions is like that of sheep about to be slaughtered, hence their name. The shameless monks, who will also be important to us at a later stage of this discussion, are simply those who break the precepts and are evil in every way. This passage was intriguing to me for a couple of reasons. First, this appears to be the only definition of the sGlpgha contained in the Dazhidu lun, and frankly it is a somewhat baffling one: why is such detailed attention paid to such seemingly redundant members of the Buddhist community as the mute sheep monks? What is their role or purpose? How should they be treated by other monks or by the laity? The text offers us no answers to these questions. Second, until encountering this passage I had only known of mute sheep monks in the later context of the Sanjie jiao -=.~~~ (The Teaching of the Three Levels), a somewhat unusual group of Chinese monks and nuns that flourished during the sixth to eighth centuries, and I had not expected to find them mentioned in an earlier scholastic treatise. 4 In search of clarification, I followed the reference in Etienne Lamotte's footnote to his translation of the text and consulted the entry on "ayi5 o::F" in the first fascicle of Hi5bi5girin .~# (Dictionnaire encyclopedique du Bouddhisme d'apres les sources chinoises et japonaises).5 There, the origin of the Chinese term yayang O::F is explained thus: "une interpretation fantaisiste du sk. e<;lamuka "sourd-muet", ou Ie mot e<;la "muet" a ete confondu avec e<;la "mouton"" (a far-fetched interpretation of the Sanskrit et;lamiika, where the word et;la "deaf" has been confused with the word et;la "sheep"). This explanation seemed plausible to me. But if Hi5b6girin's explanation was correct, then, since the term is explained in the Dazhidu lun by means of an extended analogy comparing the monks to white sheep who remain mute up to the point of death, it

5*

4 In addition to the discussion of mute sheep monks in Jamie Hubbard's work, mentioned below, see also the somewhat different approach taken by Lewis 1990, especially pp. 223-25. 5 Lamotte 1944-1981: 202 n. 4; Demieville and May 1929-1999: 45.

16

James A. Benn

would appear that this passage may not originally have been part of a Sanskrit text. It would not make much sense to gloss the term "deaf-mute" with a reference to "sheep." My preliminary hypothesis was, then, that an apocryphal term had not only made its way into the Dazhidu lun, but was actually glossed within the text. Does this explanation of the mute sheep monk or mute sheep saf!lgha accord with what we know about the Dazhidu lun? Po-kan Chou's recent dissertation argues that the Dazhidu lun is an edited account of KumarajIva's (Jiumoluoshi M~*~{t, 344-413 or 350-409) translation and explanation of a commentary on the *Mahiiprajfiiipiiramitiisutra made between 402 and 405. 6 The editor was Sengrui {~m (352-436), who also acted as KumarajIva's amanuensis.? Because of the collaboration of KumarajIva and Sengrui, along with active participation in the making of the Dazhidu lun by the project's sponsor, the non-Chinese ruler of the Later Qin, Yao Xing 1it~frJ~~ (365-416), Chou favours a "partly Chinese" authorship for the work. B This theory seems plausible, and there are a number of features of the text that would tend to support that attribution. To give one readily apparent example, the Dazhidu lun's explanation or commentary on the *Mahiiprajfiiipiiramitiisutra follows the Chinese word order rather than the Indian. 9 Also, the style of the Dazhidu lun as a whole is that of a dialogue. As Chou points . out, this is a characteristic not only of the Sarvastivadin commentarial literature with which KumarajIva was familiar but also of what he calls contemporary Chinese "Neo-Daoism."lo But it also seems to reflect the actuality of the translation process: that is to say, KumarajIva answered the questions of Yao Xing and Sengrui while translating the text, and these questions and answers were written into the text itself as the team went along. A good example of the back and forth nature of the text is the explanation of term

Chou 2000; see also Chou 2004. On Sengrui's dates see Chou 2000: 19, n. 1. B Chou 2000: 62. 9 See Chou 2000: 68 10 Chou 2000: 74-77. Generally xuanxue ~~ or "Dark Learning" is to be preferred to the misleading term "Neo-Daoism."
6
7

The silent saI]1gha

17

"bhagavat" in three alternative ways.ll Sengrui wrote down all that was said, but due to his involvement with other projects he did not consistently replace old Chinese translations of Buddhist technical terms with new ones, hence the presence in the Dazhidu lun of both old and new formsP As Chou concludes, the Dazhidu lun "is the witness of the translation, not the translation itself."13 It seems possible that the section of the Dazhidu lun concerning mute sheep monks might have been the result of just such a dialogue. KumarajIva himself could have supplied the gloss on the term "mute sheep" in response to a question from Sengrui or Yao Xing. But when we look further into the literature we find two factors which appear to rule out the possibility of the term "mute sheep monk" being only an artefact of the composition process of the Dazhidu lun. First, the division of the sa7'[tgha into four or five types, each consisting of different kinds of monk, is not unique to the Dazhidu lun. Second, the term "mute sheep monks" is attested in a number of other scriptures, some of them translated by Xuanzang ~~ (600-664). Is it likely that a mistaken translation or apocryphal term could have been perpetuated by other translators?

The fourfold division of the saT{lgha.

The fourfold division of the sarrtgha according to the spiritual capacity or type of practitioner is striking in the above quotation from the Dazhidu lun, but it is far from unique in Buddhist canonical sources. Although our mute sheep monks are not always present among the four sarrtghas, we can find the term in quite a few texts known in medieval China. Xuanzang's translation of the great repository of Sarvastivada Abhidharma lore, the Abhidharmakosabha~ya, explains that the term "bhik~u" covers four types of monk: the bhik~u in name only, the self-proclaimed bhik~u, the bhik~u who is a mendicant, and
11 Chou 2000: 78; Dazhidu [un T no. 1509, 25:70b; Lamotte 1944-1981: 115-116. 12 Chou 2000: 81-84. 13 Chou 2000: 80.

18

James A. Berm

the bhik/ju who has destroyed the passions. 14 The commentary on this work by Xuanzang's disciple Fabao 5!. (fl. ca. 703) amplifies these definitions.is According to him, a bhik/ju in name only is a layman who calls himself a bhik/ju. A self-proclaimed bhik/ju is a fully-ordained bhik/ju who repeatedly transgresses the precepts. In actuality he is not a bhik/ju, but only proclaims himself one. A mendicant bhik/ju is one who supports himself by begging. Only those sages who have destroyed the outflows (aniisrava, wulou ~ 5~) can be called bhik/jus in the highest sense. While this list does not include the mute sheep monks, it is not entirely dissimilar in form or meaning to the kinds of practitioner distinguished by the Dazhidu lun. The self-proclaimed bhik/ju is somewhat reminiscent of those monks in the shameless sa1Jlgha of the Dazhidu lun, and the bhik/ju in the highest sense reminds us of the monks of the true sa1Jlgha found in that same text. It seems that the Abhidharmakosabhii/jya and the commentary both draw here on the four-fold definition of the term "bhik/ju" that appears in the SarViistiviida Vinaya (Shisong iii +m~), a text also translated by KumarajIva,16 To these four types of bhik/ju, the Abhidharmakosabhii/jya and the Sarviistiviida Vinaya add a fifth: the monk who has received regular bhik/ju ordination by a fourfold act of public declaration. The Abhidharmakosabhii/jya contrasts this figure with the true bhik/ju who is infallible and cannot lose his status, unlike the ordinary monk who can be expelled from the sa1JlghaY The same series of five types of bhik/ju appears again in canonical texts such as Xuanzang's translation of the Yogiiciirabhiimi (Yujia shidi lun fM11mjfititiltj).18 None of these typologies of monks features the mute sheep monk, although they do betray a noticeable fear or obsession with the possibility t~at

14 Apidamo jushe lun JlaJ~jIH~*~ 15, T no. 1558, 29:79b20-21. de la Vallee Poussin 1988: vol. II, 614-615. is lushe lun shu fJ:l:*~iFfrt 15, Tno. 1822, 41:654c. 16 Shisong lii 1, Tno. 1435, 23:2a-b. 17 Apidamo jushe lun 15, T no. 1558, 29:79b22-23. i8 Yujia shidi lun l'Mftmgffl:tiB~ (*Yogilcilrabhumisilstra) 29, T no. 1579, 30:447a.

The silent sa.Q1gha

19

laymen could pose as monks and that monks who persistently broke the precepts could continue to call themselves bhik~us. We can detect in these scholastic treatises some echoes of what must have been a real and continuing concern with policing the boundaries of the saf(lgha. Probably, the ways and means by whiCh men and women could take precepts or enter the saf(lgha were often more dependent on immediate circumstances and cultural norms or precedents than they were on canonical regulations. The definitions of the term "saf(lgha" or "bhik~u" that were available to medieval Chinese seem to hint that it took more than ordination, or even observance of the Vinaya, to make one a true monk. In two other Sarvastivadin works, the Apidamo shunzhengfi lun [loJ1~E5ImJl~IEf!'[ffij (*Abhidharmanyiiyanusarasastra) and Apidamozang xianzong fun [loJf'E.~m;)i~*[ffij (*Abhidharmasamayapradipika or *Abhidharmakosasastrakarikavibhii.yya, Revealing the Tenets ofthe Abhidharma Treasury), both attributed to Sarpghabhadra

(Zhongxian ~jf) and translated by Xuanzang between 651 and 654, we find a five-fold division of the saf(lgha as follows:
{~{f]a~ , 53UIII!~EfI 0 -M,~M~ ,=u${~ , :='M~{~ , [1;]ti!:{%{~

, Elm
0

~{~ M,~~{~:tf ' 1i;~~ftl(; mHtt$~:fm~{f]a~

0 0 0

U~{~:tf, 1i;~:='~' M,pJT7~:fmW~{f]a~ '=:5<Du~, M,~mffl W<

* *

u:tf' ,liM,)t5~: ffl5~)t$' ,liiAA~$ffl ~P,liJl:t~:fIM~{f]a~'~:=. ~r:piAA~mffl M~{~:tf ' ~~}1lfmz: ' ~J%f'j~ h~;g15 ' ~5.M~ :fm~{f]a~ Jl:t =~5J-~~F5~* ti!:{%{~:tf ' 1i;;g~~ lfU5W{'F5~~F$ lm~{~:tf ' ~~iAA~$ , '&~PJT{:&~:fm~{f]a~ Jl:tJEM,?G~~F$
0 0 0 0 0 0 0 0
0

Er:pl&B~~pJT!W{:&

As for distinctions in the saY[lgha, they are broadly of five types: 1. "shameless monks;" 2. "mute sheep monks;" 3. "partisan monks;" 4. "vulgar monks;" 5. "monks in the supreme sense." "Shameless monks" refers to those who break the precepts, but wear dharma robes. "Mute sheep monks" refers to those who understand nothing in the tripitaka. They are like mute sheep, who have no way to talk. Some call them "mute," showing they have no way to preach the dharma; some again call them "sheep" indicating that they have no ability to listen to the dharma. Thus it is clear that these kinds of people have no means of hearing or speaking about what is in the tripi!aka.

20

James A. Benn "Partisan monks" refers to those who form factions and put all their efforts into arguing; they join together for expedient means. These three categories (above) mostly create karma that goes against the dharma. "Vulgar monks" refers to good worldlings. They are able to perform actions that either conform to the dharma or go against it. "Monks in the supreme sense" refers to those who [engage in the study and practice of] the dharma of saik:ja and asaik:ja, along with the conditions on which the saik:ja and asaik:ja rely. Thus this type of person definitely cannot form karma that is contrary to the dharma. Of these five, the supreme is the one in whom to take refuge. 19

Along with some clarification of how laypeople might be expected to treat these different kinds of monks (they should take refuge only in the monks in the supreme sense, for example), we can see taking shape here a certain pessimism about the nature of the sa1'(lgha. The shameless monks and mute sheep monks of the Dazhidu lun are now joined by monks who enjoy factional disputes and, taken together, these categories now outnumber those of the vulgar monks and the supreme monks. In Xuanzang's translation, although the word "sheep" implies not being able to hear (i.e., deaf), he understands etjamuka as "mute sheep" and not as "deaf mute." Specifically, the mute sheep is someone who has no ability to hear or speak about what is in the Buddhist canon (tripitaka). Xuanzang's translation gives us a further clue as to how to understand the apparent confusion between "deaf" and "sheep." Although Franklin Edgerton in his Buddhist Hybrid Sanskrit Dictionary (s.v. etjamuka) is of the opinion that "stupid (lit. dumb) as a sheep" is the regular meaning of the compound in Pali and Sanskrit Buddhist literature, I am inclined to infer from Xuanzang's translation that the compound is quite likely to be a pun.20 It means "deaf-mute" but suggests also the stupidity of the sheep. In this text the monks in question are quite clearly ignorantas to the contents of the Buddhist canon as much as they are sheep-like in their behaviour. It is noteworthy that another skilled
19

Apidamo shunzhengli lun (*Abhidharmanyiiyiinusiirasastra), T no. 1562,

29:557c. Cf. the parallel passage at T no. 1563, 29:870a. 20 Edgerton 1970.

The silent saIp.gha

21

translator, Yijing ~5- (635-713) also understood er;lamuka as mute sheep in his translation of a passage on the term "sarrtgha" from the *Mulasarviistiviidavinayasarrtgriiha that echoes those definitions in Xuanzang's translations. 21 We have seen then that the category of mute sheep mollk appears in a number of Indian scholastic texts in the Buddhist canon, mainly connected with the Sarvastivadin or Miilasarvastivadin tradition and translated by such luminaries as Kumarajlva, Xuanzang and Yijing. But it was not primarily through these texts that mute sheep monks became part of the religious scene in medieval China. The mute sheep monks described in the texts above are unable to discriminate between right and wrong, and unable to understand the tripitaka. They seem most unlikely models for monastic practice, and perhaps they were not intended to be emulated. Nevertheless, "mute sheep monk," somewhat perversely, was taken up not as a pejorative term, but as a badge of honour by some medieval Chinese monastics. Mute sheep monks in the Sanjie movement Sanjie jiao =~~~ (The Teaching of the Three Levels) is the term we use today for a Buddhist movement that was centred around the charismatic monk Xinxing 1i'T (540-594).22 This organization (or cult, perhaps) was very popular and influential under the dynasties of both the Sui ~ (581-618) and the Tang g (618-907) although it was officially proscribed a number of times. While the Sanjie movement was not as antinomian as has sometimes been made out, it certainly had a distinctive, and often separatist, attitude towards monastic practice. Xinxing's favourite Buddhist sutra was Da fangguang shi lun jing *:1J.+'~~ (Great Square and Broad Scripture of the Ten Wheels, *Dasacakra[k~itigarbhalsutra; hereafter *Dasacakra),

21 Genben sapoduodu lii she f~*i\i~~Jm1~:jJl7, Tno. 1428, 24:566b. 22 The definitive works on the movement are - in Japanese - Yabuki 1927 and Nishimoto 1998 and, in English, Hubbard 2001.

22

James A. Berm

and he quoted frequently from this work in his writings. 23 There are two translations of the scripture: the first attributed to an unknown translator of the Northern Liang jt5ffi (397-439), and the second to Xuanzang. However, the earlier translation does not appear in catalogues prior to Fajing's 5*~~ (d. u.) Zhongjing mulu z~t* (Catalogue of Scriptures) of 594, and Yancong's (557-610) work of the same title produced in 602.24 It is entirely possible that it postdates the translation of the Dazhidu lun and thus could have borrowed terms and ideas from it. It does not appear to be apocryphal, although as a translation it is somewhat rough and ready. 25 Xinxing himself seems to have been the first influential teacher to cite from the *Dasacakra, although he was later followed by such learned doctors of the Chinese Buddhist tradition as Daoxuan (597-667), Daoshi (fl. 656-668) and Fazang )*~ (643-712).26 In particular, these authors cited the sutra in discussions of government interference in the affairs of the saf!lghaY The *Dasacakra took a firm line on this issue, insisting that even bad monks had the power to allow the laity to make merit. Let us look at the following passage from the sutra, which forms part of a lengthy discussion on Vinaya:

;:'r*

m:-tit

m:'

~-jI:,
{~o

iffifriE-=r,

~[9fj{~ {EJ~1!rr[9? ~-~{~, )-{~, U5l~{~, ~'fiWfJrl

0

~{EJi:l~-~{~? ~{~tlt#, *ifi\fj*~OJi\fj, g,t{~, ~-tJJ)t:, ~1~ :tt 0 \liiJ*I~, \liiJ13~-a-, M'Wf:-a-, ~J[~B5g, ~-t:TlA, i:l1!rr~-~{~ 0 ~

~:tt*~5t:B~~, ::ffJ5~~5J1l*lm*JZ..ftlG,
1~:*e, ?f]\i:l~-~{~
~i:l5-{~
0

::fA;ffi"i\fj,

~ffif1~:*

~{EJi:l1!rr5-{~? ~~fJ5ff5J1l*lm*JZ..~JEftlG~, f(Dl$~~11:' ~{~::f38,

0

23 Yabuki Keiki found 130 quotations from the *Dasacakra in the Sanjie literature (Yabuki 1927: 594 and 644). More recently, Fran<;oise WangToutain (1998) has found more than 200 (p. 114, n. 474). 24 Wang-Toutain 1998: 19. Tno. 2146, 55:120b, 121c; Tno. 2147, 55:152c5. 25 Cf. Wang-Toutain 1998: 70-71. 26 Wang-Toutain 1998: 52-53. 27 Wang-Toutain 1998: 55-63.

The silent sarp.gha

-2{PJ;g02~{~? ::f9illf~:2fs:~,
0

23
::f9ill1T&~ffi~ffiJPJ

::f9ill3'B::f3'B, ::f9illt~,

'~'~ ~~;fm9ill, ::f5E,~~PJ~, F::f*Jl:~:);D~:st~ ::f~~JHLI~9ill

~:st~t&, ::f~Et-aFo,~cp)~~--fiiJ=tf~~~F~, {~1tfoJ~i,~, fiiJ~i,~ :t<O~~if;go2~{~
0 0

fiiJ=tf3'B, {PJ=tf3'B~~,

-2foJ;g~'~t\l{~? :E~Ai,El)E1fP, *A{~)t, ffiJ'St5~~m*JZ.1lX;, ~ ?'1~3'B, llZ;fOE{~, ~~'I1i'tl ::f~13ti!!:, pq'~~1i' ;!:t::t<Dft ~ ~~, m'~~ffi, ~~'rl'~' ~Mi-=*, 1E1.i,fUlr n)(~7\'~, ~if1i~
0 0

-s~Il}j(, fIj~F~IE5t 0 3<O~~A, '~'tl{~o

{:&Jl:iJt5t, JG,;fm'I1i'tl

~t&;gi,~

Next, sons of good family, there are four kinds of monks. What are these four? They are the monks in the supreme sense, the pure monks, the mute sheep monks, and the shameless monks. Why are the monks in the supreme sense so called? These are all the buddhas and world-honoured ones, the great bodhisattva-mahiisattvas, and pratyekabuddhas, who have comprehensively attained mastery of all dharmas, arhats, aniigiimins (non-returners), sakrdiigiimins (once-returners), srotiipannas (stream-enterers). These seven types of people are called monks in the supreme sense. [Also there are] those who are householders and do not have the dharma robe, who cannot receive in full the precepts of the priitimok,ya, who do not enter the po~adha nor praviiralJii, and yet by relying on themselves still gain the fruits of sagehood. 28 All those who have attained the fruits of sagehood are called monks in the supreme sense. Why are the pure monks so called? All those who are able to uphold the full precepts of the priitimok~a, who practice in accordance with the Vinaya and do not transgress the decorum, these are called pure monks. Why are mute sheep monks so called? They do not know the root transgressions. They do not know the difference between transgression and non-transgression. They do not know if a transgression is minor or major. They do not know that it is appropriate to repent even the slightest transgression. They are deluded and ignorant, and do not see that if there is a transgression they should be ashamed. Also, they do not depend on good friends (kalyiilJamitras). Because they are not often close to good friends, they cannot debate the profound mean-

28 Po~adha is the meeting of the saYJ1gha held every fifteen days. The praviirar:u'i is a rite of confession performed at the end of the rainy season.

24

James A. Benn ings found in the sfitras, [such as] what is wholesome, and what is not wholesome, what constitutes breaking a major precept, and what constitutes breaking a minor precept; whether cultivating this is good, or cultivating that is evil. Because of these characteristics they are called mute sheep monks. Why are shameless monks so called? Suppose there is someone, who for his own livelihood, enters the buddhadharma and receives the precepts of the priltimok(fa and breaks all of them, and destroys the united sal'{lgha without any shame. He is not afraid [of retribution] in a later lifetime, in his very bosom he is defiled. His voice is like [the sound of] a conch shell, his speech is rough and hard. He always nurtures envy and he is ignorant and dull. He dispenses with the three trainings and just values profit. He gives himself over to the six emotions, and is subject to the five desires of form, sound, smell, taste and touch. He slanders the true dharma. People like this depend on egotistical methods, and in their minds have no shame. Thus they are called shameless monks. 29

We can see that although the sa1'[tgha is here divided into four, as in the Dazhidu lun, this is a different list and quite a different kind of list. We may note in passing that in this text laypeople who have attained awakening are classified as monks in the supreme sense, a claim that hints at interesting possibilities for how the category "monk" could be understood, especially in East Asia. But to judge from the length of the descriptions and the type of language used, the passage is intended to be read primarily with reference to the mute sheep monks and the shameless monks. This choice of emphasis is perhaps not surprising given the *Dasacakra's major themes: the idea of the universal dharma, the belief that the degen(the decline of the dharma) had already beerate era of mofa gun, and the subsequent importance of the Vinaya and the respect due to monks in those desperate times. 30 In the early Sanjie jiao literature the citations from the text come from the older translation, but in the later Sanjie fofa miji - ~~{5/15

*5*

29 Dafangguang shi lunjing :k::0Jt+,illl$J[ (Great square and broad scripture of the ten wheels, *Daiacakra[k(fitigarbhalsutra) 5, T no. 410, 13:703ab. 30 See Wang-Toutain 1998: 15-51.

The silent saqIgha

25

5*z.f~G (Secret Record of the Buddhadharma of the Three Levels) references are made to Xuanzang's translation of 652, Dasheng daji dizang shilun jing ::k*::k~ttt~+tnfi~~ (Mahayana Great Assembly Satra of the Ten Wheels of K~itigarbha). It is quite possible that Xuanzang or members of his translation team had some contact with members of the Sanjie movement, and that Sanjie ideas may have influenced the translation. 31 Let us now compare Xuanzang's translation of the passage in question. &::X~~T, 1lZ]f.I{~ {6J~lO IZ]? -:tfJm~{~, =:tf-tt!:{i:l-{~, -'f::{~, rz:g:tf~'[$kt*{~
0 0

= :tf0 2

~fiiJ;gJm~{~? ~{~-tt!:~, ;:fimiif[ij~~DJ[ij4'J., ~t~~rJ, ~-i;)]5~,

1~ErtE:tf;:fi~Jm:Jt, ;:firoJm~, ;:fi/Fim, ;:fi-3J(, ;fim5frL 3<D~tf.i :fmt~{1JOm, MJ~{~~ ;:fimi1'~, wtE*i', /F~U~~, /FBbi~~ ~ /F1~~-i;)]W*53Umg>t1t<;, -Wmm:ffiil,i, El;:l\~~~~, ffii1~$ 1~~*i&:, MJ~{~~ ~;gMJ~{~
0 0 0 0 0

~{6J;g-tt!:{:fr{~? ~Jlj~U~~, :fbZBbi~~,

{~
0

gl(gtW*53Umg>t1t<;

~;g-tt!:{i:l

~fiiJ;g02-'f::{~? ~/F7~i'&*~*, 38~/F38/f"9(D'~m, 3l~38f.if.l/J\

~Jj\*, /f"~~B'[jHapM8 ;f1;~w~, ~1J'&/J\*,
0 0

/f"Je,/F{'i! o/F{:&If!
0

13F3~ffiH /f"8~8~Fs'tt~~~If!I3F3:tfPJT, ~JHli7J($ W\/f"')['l)[~~ ~!Fp' - ~{6JlO~? ~{OJ/f"~? ~{6J1*? ~{OJ~*? {~{6JJ&frj>? 1''FfOJlO

@? 3<D~-i;)]:fmm{;lmm,

02-'f::{~~ 0 ~;go2-'f::{~

~fOJ;g~'[iWftl{~? ~Jlj;:fi1'~, lO5j!i$i&: 0 iW{:&flG5~, ffii>JCW* 0 1~W* e, ~pJT~~53Umg>t:ttX;, -i;)]~38, ~'[iWf~'tl /FJe,/F{'i!1~-tt!:N* i*J
0 0

'tf, IfJfE=MJ*o~~;fU.,
0

'~~:, 3<D~!HIiIJ~!i

~~;g~1tt)llii7\~, frr~fr~)~, ~W:-ES, w~1Ii~ 3<D~-i;)]:fmm{f]Dm, ~'[iWf{~~ 3l~WJIE$, ~;g~'[iWf'tl{~

0 0

&'l'3B]iT, 1it~r~15, ~~-. '~~~"9P, ~.'~

0

Next, good sons, there are four kinds of monk. What are these four? 1) monks in the supreme sense; 2) vulgar monks; 3) mute sheep monks; 4) shameless monks. Why are monks in the supreme sense so called? Buddhas and worldhonoured ones, all the bodhisattva-mahasattvas, whose virtue is so honourable and lofty as to have mastery of all dharmas, pratyekabuddhas, arhats, non-returners, once-returners, stream-enterers: these seven types of people are monks in the supreme sense. There are also
31

Wang-Toutain 1998: 115-119.

26

James A. Benn those beings who keep the appearance of being householders, who do not shave off their hair nor wear the monastic robe (ka~aya), although they have not received all the separately preached precepts for renunciants, nor [do they participate in] the po~adha or pravaralJ-a, yet relying on themselves they succeed in completely ridding themselves [of all outflows], gaining the sagely dharma and attaining the fruits of sagehood. Because they belong to the category of supreme monks then they too are called monks in the supreme sense. Why are vulgar monks so called? They are so called because they shave off their hair, and wear the monastic robe. They then perfect. all the precepts separately preached and explained for renunciants. Thus they are called vulgar monks. Why are mute sheep monks so called? They are so called because they do not know the root transgressions, or whether they violate them or not. They do not know minor or major transgressions, nor that for every kind of minor opportunity or transgression, one should arouse [the mind to] repent for one's transgressions. They are stupid and dull. As for the smallest transgression they do not see it, they do not fear it. They do not depend on the wise good friends (kalyiilJ-amitras). They do not from time to time visit the places of the learned and clever to approach and serve them as teachers. Also, they do not have the occasion to respectfully ask, "What is good? What is not good? What [action] involves a transgression? What does not involve a transgression? What should one cultivate for good? What cultivation is bad?" Thus, all these people are included among the mute sheep monks. And therefore they are called mute sheep monks. Why are shameless monks so called? They are so called because some beings for their own livelihood take refuge in an egotistic dharma, and seek to leave home. Having left home they receive and uphold all the separately preached precepts, but break them all without shame and without fear. They do not see or fear the bitter fruit of a later life. Inwardly they nurture rottenness and filth like slimy snails. They talk like conch-shell trumpets and they walk like dogs. They love empty speech with nothing of substance. They are stingy, jealous, stupid, arrogant, and they avoid the three supreme forms of karma. They crave and are attached to profit, reverence and fame. They are addicted to the six realms of rebirth and they enjoy sex and licentiousness. They crave form, sound, smell, taste, tactile and mental objects. All persons

The silent sarp.gha

27

like this are included among shameless monks. They slander the true dharma and thus are called "shameless monks."32

Let us now sum up the characteristics of the mute sheep monks as presented in the two versions of the *Dasacakra. They do not know the basic transgressions, nor what constitutes transgression, nor whether a transgression is minor or major. They do not realize that they ought to be ashamed, nor know to repent if they do transgress. Because they have little contact with good companions in the dharma they cannot talk about the meaning of scripture. They are not described as silent per se, but they have few opportunities to discuss the siitras with the wise. What is remarkable about the Sanjie movement is not their interest in the apocalyptic literature of the latter days, which painted such a bleak picture of the capacity of the saf!1-gha to understand and practice the dharma, but rather their enthusiastic identification with the mute sheep monks of the *Dasacakra. The presentation of these practitioners in the two translations is anything but positive: the mute sheep monks are apparently only marginally superior to the shameless monks by virtue of their utter ignorance of everything. Nowhere in the scripture itself do we find any sustained discussion of the role and significance of the mute sheep monks beyond this passage. The scripture describes the mute sheep monks, but does not appear to advocate any particular response to their particular situation. What evidence do we have for the Sanjie movement's sense of identity with the scriptural mute sheep monks? The Zhi zhongshi xufa t~.4iljl~5~ (Assorted rules for Community Regulation, hereafter ZhiJa) is a text that appears to be a Sanjie disciplinary manual that was probably written by Xinxing himself. It has been preserved in a Dunhuang manuscript (Pelliot 2849) along with two other texts. Nishimoto Teruma lZ"*~~ has made the most extensive study of this text and I follow here the edition made by him in his recent monumental study of the Sanjie jiao. 33

32
33

Tno. 411, 13:749c-50a.

Nishimoto 1998: 578-601.

28

James A. Benn

For Xinxing, it seems, it was in fact the very stupidity of the mute sheep monks that prevented them from falling into error. In Xinxing's vision the most profound problem afflicting beings of his day was that they held views. The mute sheep monks, however, could not be guilty of having views on anything, since their minds were not capable of discrimination. In contrast to the ideal nondiscriminating wisdom of the bodhisattva, we might perhaps call this quality "non-discriminating stupidity." It is possible that this quality possessed by the mute sheep monks is the key to understanding why Xinxing was so keen to promote them as an ideal. Eric Greene, in a recent article, has begun to explore the historical connections between Sanjie jiao and the early Chan movements of the eighth century.34 Some of the eighth-century documents found at Dunhuang and now associated with Early Chan (for example, Wuxin lun ~iL.,~and Jueguan lun ~5e~) champion practices of no-thought - glossed variously as "mindlessness," (wuxin ~ iL.I) "cessation of thought," (jueguan ~5e) "no thinking" (wunian ~~) etc. 35 These Chan exhortations to cease discursive thinking may represent a trend that ran in parallel with Xinxing's vision of mute sheep monks who held no views and thus could not fall into discursive thought. Clearly, more research is required to determine the lines of connection that seem to run between Chan and Sanjie jiao ideas. Setting aside the immediate historical context for a moment, it is clear that Xinxing's concern to promote mute sheep monks reflected attitudes to monastic practice that were by no means unique to his time and space. As Luis Gomez has shown, the Pali Atthakavagga also urges monks to shun all views and thus avoid disputes with others.36 We might to do well, then, to think of the mute sheep monk as a particular way of conceptualising a perennial issue for Buddhists concerned with the correct mental attitude for practitioners.

34 Greene 2008. 35 These texts and others of their genre are briefly introduced in Sharf
2002, 47-5l. 36 Gomez 1976.

The silent sarp.gha

29

Elsewhere in his writing, Xinxing had explicitly excepted the mute sheep monks from the class of beings in the third stage of the Buddhadharma who were all tainted by views of eternalism or nihilism, and this exception is probably what made the mute sheep monks most significant for his vision of the possibilities for Buddhist practice in the situation he and his contemporaries knew, one that was far removed from the Buddha in time and place. 37 But in Sanjie monastic life the separation of mute sheep monks from beings who held views was not merely doctrinal or theoretical. The Sanjie monastic community lived, and for the most part practiced, apart from the rest of the medieval Chinese sa1J1-gha. The first two items in the ZhiJa are explicitly concerned with mute sheep monks. For the sake of space, let me concentrate on these two items in the following discussion, and leave the rest of the document for a later date. Let us see if we can make sense of what they say, and compare this vision of the mute sheep monk with that found in the canon.
5JU=~5*' ~~1RX:*U' 5tgifi5tgifi~[B] , {*gifi{*gifi~[B],~1i ~1i~[B] j4e* ' m1'T;f~1tf '~5JU~~ 1T~~~{~1tf{iX~~{~ ~'1'T~~N[~{~{iX!&~{~~ j5t-~, 1it{iX~~{~~j: ::f1~m~g: i'113z*!E Di~~=~tf~ , ::f1:E~~~ x.~f ~f ~>tti!<:B~1:E1ii:E9l1=~ 5fO[B]~5t~1tf ' ::f1:E~~~
0 -, 0 0 0 0
0 0

One: the rule of dividing the two assemblies.

Item: there are distinctions among the sages. 38 Dharma masters belong with other dharma masters; Vinaya masters belong with other Vinaya masters, [those who practice] seated meditation belong with [others who practice] seated meditation. From now on, one should associate oneself with those who are of one's level in understanding and practice. One whose practice is equal to that of a mute sheep monk Nishimoto 1998: 410, quoting from the Japanese copy of the Sanjie fofa (Buddhadharma of the Three Levels) found in Yabuki 1927: 257. Hubbard 2001: 89. On the third stage (or level) of the Buddhadharma see pp.
37

=~I'{~$

84-89.
38 Compare this with Zhiyi's ~iJ'i (538-597) manual of community regulations, which opens by stating that people's natures are different. Guoqing bailu ~5jlfB~-R (One hundred records of Guoqing monastery) 1, T no. 1943,

46:793c.

30

James A. Benn should associate himself with the assembly of mute sheep monks; one whose practice is equal to that of a wise monk should associate himself with the assembly of wise monks. Although the assembly is divided in two one should rely on the mute sheep monks as the leaders. The two assemblies should not be mixed together with the exception of the two assemblies meeting in general congregation, which does not fall under this stricture. Also excepted is the case where the two assemblies at the time of the recitation of precepts at the middle and end of the month congregate in the place of solitude (ara7Jya) to perform together some matters of dharma. This case does not fall under the stricture. 39

According to the first item in the Sanjie disciplinary code, it was fundamental to Sanjie practice that the two assemblies of the mute sheep monks and the wise monks not be mixed, and yet of the two groups it was the ignorant and clueless mute sheep monks who were to be in charge of the sa'!lgha as a whole. We can see immediately that this is in some contrast to the portrayal of mute sheep monks in the *Dasacakra. In that satra they are depicted as having scarcely the ability to direct their own affairs, let alone those of the larger community. The *Dasacakra mute sheep monks did not even have the sense to consult good friends in the dharma, but here we find them functioning quite happily as an independent unit. We know from other sources that the Sanjie members were indeed physically separated from other monks. In 725 the emperor Xuanzong ~* (r. 712-756) issued orders to end this practice:
9;01El&:~t~~iB$UITL Ifflft+-=-~Z:B:~7\A-=B *lJ~~-=~\i'~3t :illL9~~*~~Jlf ~.wjd3t~mLW.1~~]j5::f1~53U1' PJTff~~-Rj~J$tIT~~
0 0 0 0

3[~

Knowing that they are contrary to the truth and incite falsehood, an edict was issued prohibiting them: on Kaiyuan Ifflft13.6.3 (February 18, 725) an imperial edict was issued to all of the cloisters of the Sanjie ordering the removal of all barriers. The followers will live in the main monastery with the community of monks. They may not maintain separate residences. The collected works of Xinxing are all prohibited and should be destroyed. 40 Following Nishimoto's edition. Nishimoto 1998: 578, lines 8-15. Kaiyuan shijiao lu 18, T no. 2154, 55:679a; translated in Hubbard 2001:

39
40

The silent saqIgha

31

The edict strongly suggests that prior to 725 there were physical walls or fences between the Sanjie cloisters and other parts of monasteries in which they were located. Reading this edict together with the Sanjie regulations might give us pause to think about what a Tang monastery actually looked like. As we know, many monasteries inthe capital were not purpose-built edifices, but converted from the donated mansions of the aristocracy. These sprawling compounds must have offered rather easy opportunities for smaller groups such as the Sanjie who wished to live and practice in common. But whether or not the architectural layout of the monasteries was a significant factor or not, it is clear that the mute sheep monks of the Sanjie jiao did not mingle with other monks and maintained separate residences, presumably only coming together for matters of Vinaya, as the Zhija specified. The second item in the Sanjie disciplinary code tells us a little more about the community of mute sheep monks. Unfortunately the text is somewhat poorly transcribed and garbled in places. I have made what sense of it I can, and my translation is tentative in places.
&~f~~Jl,i1j~A5t ,~=

- , 1T1Tf~~ ,~~tM~f~ '~~~~f~ 0l~~1;~~1;I:I}*e*/f"

0

38~)]1i=1iii!<;~+-lm~'8 X:~~~;fEl-=$j::~t~, ~~;;Jj\~~~~~

0

ffi~~t~ 0 0l~[liiJI!lF~1;E51~H~:'&Wf~~J1;)5E1'irlli~ii!<;~t;[e:9~ , 0l1~ El

~El~El~-W~'/f"~El~El~El~-W~'Ol~~~~~-W~'
/f"1~~fili~fili-t)]~ 0

{iiI ~ift Ij)-iJJS=~~RlG~M1li~~ , OlH~-t}]IE~A$m1T{'F-iJJS=~

0

~A$m1T ' OlH~-t)]S=~~A5tm1T{'F-t)]IE~A5tm1Tift

0

/f"llJiJ~I:I}*Ao /f"n~I:I}*Ao /f"Elfl;~m~fjUxtt~:t~ ;;Jj\/f"~~f~

~tt~:t~Wf~Ao {~~[5j/f"m, {iiI5mf~~5t~ 0 /f"{'F+-fl~ 0
0

3<DIIt

f~~fi5~f~~txtE~1l"1~~fI~ X.~~~ifW/f"38ii!<;~~ , 3[~EfiM~ A~1TWI.l0J~ ,;;Jj\1~*~~~ X.1;E~~Fr:r~1T~~jli ' ~J!~fi5WLIOJ

0

1l" , 1~~tE~

214. See also Nishimoto 1998: 412.

32

James A. Benn
Two: Method of selecting people for the mute sheep monk assembly.

Item. The practitioner monks of the single practice are sometimes called meritorious monks, and sometimes called mute sheep monks. It is just that some from birth, and some from when they left home, have not transgressed the precepts of the first and second chapters, and have studied the twelve dhutagul'}as. 41 For the text, they should study the seated dhyana of the formless samadhi. For the meaning, they should each study the empty seated dhyana of the marks of emptiness and non-emptiness. With the exception of times when they chastise disciples, or even discipline those within the saf!lgha who break the precepts, they should always see and speak only about their own evils; they should not see or speak about any kind of good on their own part. They should only see and speak about good on the part of others, and should not see or speak about any kind of evil in them. Why is this so? It is because all of the perverted sentient beings who have realized heterodox views take all of the truly virtuous people, teachings, understandings, and practices as false people, teachings, understandings and practices and take all of the false people, teachings, understandings and practices to be truly virtuous people, teachings, understandings and practices. Do not chastise or curse those who have left home. Do not hit or bind those who have left home. Do not personally consume or use the food and drink or possessions of the saf!lgha. Also, do not give the saf!lgha's food and drink or possessions to lay people. If the saf!lgha's property is not for use, then surely one cannot use the property of the Buddha or dharma. Do not act as one of the eleven types ofleader.42 This kind of monk may, however, act along with the community as a kind of leader. Also, there are some with no focused essence, who do not transgress the precepts, or who having transgressed are able to repent. They may be placed in the category above. They can also be leaders of the assembly.43

41 The twelve ascetic practices. Sanjie literature seems to have followed the standard canonical list. See Hubbard 2001: 26-27. 42 The reference is presumably to the types of leader mentioned at *Dasacakra 6, Tno. 410, 13:710a. 43 Reading .. for 1:.

The silent sarp.gha

33

Also, as for disciples of the community, do not ask whether their practice is longstanding or recent. If they display [evidence of] study that matches that described above, they may become disciples. 44

We may now arrive at the following characteristics of mute sheep monks as prescribed in the Sanjie jiao. 45 They were to live separately from other monks, regularly coming together only for the recitation of the precepts. Either from birth onwards or at least after leaving home, they should not have broken the precepts. Those who had broken the precepts and had sincerely repented were still allowed to join the community, whereas, according to scriptural accounts, mute sheep monks did not even know that they ought to repent. They were ascetics who studied the twelve dhutaguIJas, and they were also meditators who studied the markless samadhi. Asceticism and meditation explicitly took the place of scriptural study and exegesis, which according to scriptural accounts mute sheep monks were unable to master anyway. They were to behave in a self-effacing and humble manner; to see all good as residing in others and all evil in themselves. Yet they were also to playa vital disciplinary role and were not to refrain from chastising disciples who had broken the precepts. They were to respect other members of the saf(lgha, and not slander, strike or tie them up. They were not to use the property of the saf(lgha for themselves or give it to laypeople. They were not to serve as religious or secular leaders, but they could serve as leaders within the monastic community. We are thus presented with a very different picture of the mute sheep monk than that described in the *Dasacakra, or any other canonical source we have considered here. In fact, in this conception, the mute sheep monk is an austere, sincere, practice-oriented monk who diligently keeps the precepts, and this is far from the scriptural presentation of the clueless monk who cannot even understand the most basic Buddhist texts. Actually, as Nishimoto has suggested, the Sanjie jiao mute sheep monks are almost the exact opposite of the violent, precept-breaking, licentious, shameless

44 45

Nishimoto 1998: 579, lines 16-33. See Nishimoto 1998: 413-414.

34

James A. Benn

monks as presented in the *Dasacakra. 46 But we may imagine that Sanjie models were taken from closer to home as well. The ideals of the mute sheep monks probably offered a monastic path that was the opposite of what the Sanjie community perceived around them: monks associated too closely with the state, scholar-monks tied up in book learning, a sarrtgha that was licentious and poorly disciplined. In the *Dasacakra, the mute sheep monks are described as those who had little contact with good friends in the dharma. However, in Sanjie practice it was the mute sheep monks who were themselves the good friends. In another major Sanjie work, Practice that Arises in Accord with Capacity (Duigen qixing fMR!Eef'J), we find the following passage:
~*-t)];g9ill~~~' :!K[1;]/,Ef= -~x~{J'VF1W@~{~
0 0 0 0 0

o=1lf1Wx

::f~~~@~{~ Jit=flt@~{~;g9ill~ , ~{OJPJ~ !l6\JJ.)~,~L' PJ9ill:Jik ~F :!K[1;]/,Ef1\~1' -1lf=*tt) ' t:21:.e*, :!K{~-t)]$R1:. ' ::fmz::ft:{~

fImif:JJ' 7J~::fmz:~{~ o=1lf'ftEl~lm' 1:E~::f~5{'F~-&' n~::f~5f'F g~ =1lf'ftEl{\Z~' ::f{'F+~' %::fJllitt<> 12]1lft:2tH~e*' 'rtEl:j:\f
0

ffZ~~ 1i1lft:2tH~e*' m~lm' ::f~{'FA)~~ 1\1lf~~+ lmWf:; , m~~"T ' m~~~ 1J1:.n~tH~e*, J'l,Jit1\flt ' m:ft:fEJ2 ' {;&{'F;g9ill~ ~~/,EfJ8ffZ ' ~P~H$r'fl ' 'f&::fm~ , 9l!H8~ffZ ' ~::f~lH8 ' 5R:f'J'!Pt.tfEJ1lf '%{;&{'F;g9ill~ =1lffU;jt~A[1;] '/,Ef1J&U~'Mf~t:)~1\
0 0 0 0
0

)~, n[EJ.L@~{~{\z~~1J&U1lf' %1~*{'FtHti!:;g9ill~

Explaining the exhaustive search for good friends.

Within [this category] are three kinds [of good friends]. The first is the mute sheep monk who cannot understand the text or the meaning. The second is the mute sheep monk who can understand the text but not the meaning. How can one recognise these two kinds of good friends who are mute sheep monks? One must test them by means of the Dharma, then it is possible to know if they are [good friends] or not. Within [their actions] there are six levels [of attainment]. First, from birth onwards, their nature is pliant in performing the three forms of karma [acts of body, speech and mind] towards all other sentient beings. They do not dare to fight with others, or to hit them, they do not even dare to hate others.

46

Nishimoto 1998: 414.

The silent sa!p.gha

35

Second, their nature is self-effacing; as householders they do not be. come the heads of households, or even seek to take an official position. Three, by nature they fear their own misdeeds, they do not commit the ten evils, or even break the precepts. Four, from the time they leave home by nature they uphold the precepts and avoid wrongdoing. Fifth, from the time they leave home they are constantly self-effacing, and do not dare either rule men or the dharma. Sixth, they enjoy studying the twelve dhiitagu1}as they always enjoy receiving inferior things, and they always enjoy receiving bad things.
If from birth up to after they have left home they have all these qualities, one can rely on them as good friends together with the community. If having broken a precept they are full of remorse and do not try to conceal it; and even if they break the prohibitions and precepts but do not re-offend and their other practices are as above, then one may rely on them as good friends.

The third is the person of sharp faculties, who has comprehensively studied the six teachings or seven teachings in accordance with his capacity, if he also fears wrongdoing etc. in the same way as the above mute sheep monks, then one can rely on him as a good friend on the renunciant path. 47

If we compare this list of qualities with the ideal mute sheep monk as presented in the Zhifa, we see exactly the same kinds of concerns, expressed slightly differently, possibly because Zhifa represents a slightly later stage in the development of Sanjie monastic practice. According to Duigen qixing, the mute sheep monks are passive and pacific from birth. As both householders and later as monks, they should avoid taking positions of responsibility. They should not break the precepts out of a fear of wrongdoing. They should be self-effacing ascetics. They may not know the text or its meanjng, but they were to practice the paths of asceticism and meditation. Although the scriptures describe them as dull-witted and spiritually backward, the mute sheep monks of the Sanjie movement are

47 Nishimoto 1998: 489. Translation in Hubbard 2001: 90-91, with some modifications.

36

James A. Benn

presented as a spiritual elite, conforming to the very highest standards of morality and discipline, and providing a model and source of support for the beings of the third stage or level. But to attempt to comprehend Sanjie jiao in terms of its doctrinal justifications may be to misapprehend the whole phenomenon. It would be better perhaps to consider the movement as a cult within Chinese Buddhist monasticism that may have arisen initially out of devotion to the charismatic figure Xinxing, but that perpetuated itself as a saJ?lgha within the saJ?lgha under a somewhat esoteric and particular slogan drawn from the founder's favourite text - the mute sheep monk assembly. Conclusions The term "mute sheep monk" may have appeared first in Kumarajlva's translation but it also appears in a number of later texts. Evidence from these translations indicates that the term may be a clever pun (mute sheep for deaf mutes). Puns are notoriously hard to translate, even if one understands both meanings of a term, and one wonders how many other examples there are of this phenomenon in Chinese translations of Indic texts. The four-fold saJ?lgha of which the mute sheep monks are a part is not unique to the Dazhidu lun but is echoed in Sarvastivada and Miilasarvastivada texts. Why exactly the saJ?lgha was described in this way remains somewhat unclear, but certainly in these scholastic texts the mute sheep monks were not ascribed with any significant role or purpose. In the more apocalyptic world of the *Dasacakra we find an amplified description of the mute sheep monks. Although this text clearly supplied the inspiration for Xinxing's ideal monastic community, the image of the mute sheep monk that we can reconstruct from Sanjie literature is overwhelmingly positive, and quite different from the negative portrayal in the *Dasacakra. Does a rhetoric of the decline of the dharma, such as we find in the *Dasacakra, necessarily imply more lenient standards for the saJ?lgha? It seems not. Although the Sanjie mute sheep monks were told to replace text and exegesis with meditation and ascetic practice, this does not necessarily imply any more than a similar rhetorical attitude

The silent sarp.gha

37

towards the written word in Chan. In other words, the mute sheep monks, like the lineages of Chan, look like a particularly Chinese solution to the problem of being located far from the Buddha in space and time. The mute sheep monk assembly may also represent a sincere attempt to work even harder at the monastic path, even when all seemed lost. All we know of Sanjie practice comes from texts, and the texts used by the Sanjie movement were prescriptive rather than descriptive. Can we really say then that the Sanjie monks were as diligent in practice as they were supposed to be? The insistence on moral conduct and humble demeanour represented by the figure of the mute sheep monk ought to make us at least a little suspicious that Sanjie rhetoric may not have translated into particularly high standards of conduct in the community. What these documents do show, I think, is that it is quite possible for a self-consciously elite group dedicated to upholding certain ideals of practice and decorum to arise within the larger monastic body. Standards in Buddhist monasteries have always varied widely, and no doubt similar events have occurred at other times and places across the range of Buddhist monasticism in Asia.

Abbreviations
T

= TaishO

shinshii daizokyo *lE~M~*~~ (100 vols., eds. Takakusu Junjiro ~OOJI&[.:x~G and Watanabe Kaigyoku )!tll~:hEL et al.). Tokyo: Taisho issaikyo kankokai, 1924-1932.

Bibliography
Benn, James A., Lori Meeks, and James Robson, eds. 2010. Buddhist Monasticism in East Asia: Places of Practice. London: Routledge. Chou, Po-kan. 2000. The Translation of the Dazhidulun: Buddhist Evolution in China in the Early Fifth Century. PhD, Committee on History of Culture, University of Chicago, Chicago. ___ . 2004. "The Problem of the Authorship of the Mahiiprajfiiipiiramitopadeia: A Re-examination." Taida lishi xuebao (34): 281-327. de La Vallee Poussin, Louis. 1988. Abhidharmakoabhii~yam. Translated by Leo M Pruden. Berkeley, Calif.: Asian Humanities Press.

38

James A. Benn

Demieville, PauL 1973. "Sur Etienne Lamotte, Le traite de la grande vertu de sagesse de Nagarjuna, tome II," In Choix d'etudes bouddhiques (19291970),470-490, Leiden: E. J. Brill. Demieville, Paul and Jacques May, et al., (eds) (1929-) HobOgirin: Dictionnaire encyelopedique du bouddhisme d'apres les sources chinoises et japonaises. Fascicules I-VIII (to date). Tokyo: Maison FrancoJaponaise. Edgerton, Franklin. 1970. Buddhist Hybrid Sanskrit Grammar and Dictionary. 1st Indian ed. Delhi: Motilal Banarsidass. G6mez, Luis O. 1976. "Proto-Madhyamika in the Pali Canon," Philosophy East and West 26, no. 2: 137-65. Greene, Eric. 2008. "Another Look at Early Chan: Daoxuan, Bodhidharma, and the Three Levels Movement." T'oung Pao 94, no. 1-3: 49-114. Hubbard, Jamie. 2001. Absolute Delusion, Perfect Buddhahood: The Rise and Fall of a Chinese Heresy. Honolulu: University of Hawai'i Press. Lamotte, Etienne. 1944-1981. Le traite de la grande vertu de sagesse de Nagarjuna (Mahaprajiiaparamitaastra), 5 vols., Louvain-La-Neuve: Universite de Louvain, Institut Orientaliste. Lewis, Mark Edward. 1990. "The Suppression of the Three Stages Sect: Apocrypha as a Political Issue." In Chinese Buddhist Apocrypha, edited by Jr. Robert E. Buswell, 207-238. Honolulu: University of Hawai'i Press. Nishimoto Teruma 'E9*~~. 1998. Sangaikyo no kenkyu =~~~(/')1iJf~ A Study of the San-chieh-chiao, Tokyo: Shunju sha. Wang-Toutain, Fran<;:oise. 1998. Le bodhisattva Ksitigarbha en Chine du Ve au X1IIe sieele, Paris: Ecole fran<;:aise d'Extre~e-Orient. Yabuki Keiki 9<:::~m~-'. 1927. Sangaikyo no kenkyu =~~~Z1iJf~ [A Study of the Sanjie jiao] , Tokyo: Iwanami Shoten.

Ignorance, epistemology and soteriologyl Part I

Vincent Eltschinger

For Jacques May

1.1. Introduction
Increasingly fewer scholars would side with Edward Conze, who regarded the Buddhist epistemological school (6 th_13!h c. India) as, at best, paying merely lip service to the Buddha and the Buddhist ideal of salvation. But in spite of seminal pieces of scholarship by Ernst Steinkellner, Eli Franco and Helmut Krasser that offer discussions to the contrary, many if not most Dharmakirtian scholars seem to be satisfied with conceding Dharmakirti's (6 th century?) Buddhist persuasion in a purely marginal or superficial way. Many of them discuss the various aspects of Dharmakirti's logic and theory of knowledge as if these had sprung out of the head of a philosopher trained in Vienna, Cambridge or Harvard (or in earlier days, Konigsberg). Quite often, scholars label as "Buddhist" whatever theoretical posit can be located in a literary document known to have been authored by a Buddhist, and fail to wonder in
1 This study has been made possible by the generous financial support of the Austrian Science Fund (FWF-Projekt P19862 "Philosophische und religiose Literatur des Buddhismus"). Most sincere thanks are due to Helmut Krasser and Ernst Steinkellner. Lambert Schmithausen also deserves my wholehearted gratitude for having gone through this essay with incomparablygreat care and erudition. My most sincere thanks are due to Cynthia Peck, who kindly corrected my English.

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 39-83

40

Vincent Eltschinger

which way these tenets can be said to be Buddhist at all. I readily agree that Dharmakirti's at times frightening philosophical genius amply deserves to be brought into a comparative and/or systematic philosophical light. It would indeed be regrettable to deprive such thought of its inspirational force by studying and preserving it as a museum piece or only in the spirit of cultural and historical relativism. Those who are willing to take Dharmakirti's Buddhist persuasion seriously are repeatedly confronted with fascinating areas to examine, as in Dharmakirti's constant merging of epistemology with soteriology, or in the manner he fits his theory of knowledge to the demands of his religious thought. Buddhism may be defined minimally as the tracing of the unsatisfactory condition of living beings back to ignorance (avidyii, ajfiiina, moha) and the prescription of a therapy against suffering (duJ:tkha) as its eradication. 2 It would indeed be surprising for an allegedly Buddhist philosopher, all the more one particularly concerned with episteme, to address gnoseological questions without trying to locate ignorance in the psychological machinery he describes and discuss the part played by ignorance in human error and overall dissatisfaction. And indeed, there is hardly a single page in the first two chapters of Dharmakirti's Pramiil}aviirttika (PV) that does not mention ignorance and the evils it is responsible for. In the present essay, I would like, first, to draw a systematic picture of ignorance in Dharmakirti's philosophy, and second, to examine how far Dharmakirti's reception of Dignaga's (480-540?) two-headed epistemology might have been dictated by his conviction that perception (pratyak:ja) and inference (anumiina) provide all the epistemic requirements for one's achieving final liberation. In doing so, I hope I shall succeed in showing not only that the very heart of Dharmakirti's system is permeated by -let's say - religious schemes, concepts and preoccupations, but also that it evidences,
2 Previous studies on avidyii include La Vallee Poussin 1913: 6-9, Wayman 1957 and 1980, Matila11980, Mejor 2001 and 2002. On DharmakIrti's notion of avidyii, see Vetter 1990: 22-26, Franco 2001: 289-300, Eltschinger 2007: 503-510. The present essay does not take into consideration Dharmaklrti's Yogacara account of ignorance as duality; on this point, see Eltschinger 2005: 162-175.

Ignorance, epistemology and soteriology (1)

41

especially as far as perception and inference are concerned, an organic relationship to soteriology. 1.2. Ignorance as erroneous cognition (mithyopalabdhi) 1.2.1. Let me start with one of the most important stanzas Dharmaklrti ever dedicated to ignorance: "Ignorance (ajiiiina) is [nothing but] erroneous cognition (mithyopaZabdhi)[, and this is for three reasons:] Because it counteracts vidyii; because, since it is a mental factor [associated with the mind, it can only be] perception; [and] because [the Buddha himself] has declared [this in diverse SfItras. Any] other [account of ignorance] is ill-founded."3 1.2.2. In their commentaries onPV 2.213a (vidyiiyii!:t pratipak$atviit), Devendrabuddhi, Prajiiakaragupta and Manorathanandin criticise a (presumably) Vaibhii~ika opponent in whose opinion ignorance should be viewed as a mere lack of clarity (asamprakhyiina).4 In order to do so, they take up an interesting elaboration ofVasubandhu's (PrSVyS and) AKBh on the meaning of the negation or privative prefix "a" in the Sanskrit word avidyii. Dharmaklrti's commentators first define "knowledge" (vidyii) as the perception (darsana) or grasping (grahalJa) of something real (sadartha, bhatiirtha),6
3 PV 2.213: vidyiiyii!:t pratipak:jatviic caittatvenopalabdhita!:t I mithyopalabdhir ajfiiinam uktei ciinyad ayuktimat II. 4 Dharmaklrti's opponentes) is/are, according to Devendrabuddhi, (a) certain co-re1igionist(s) (svayathya) (PVP D91a5/P105bl: ran gi sde pa 'ga' zig), vaibhii:jikiidi according to PVT D13SbS/P167a7-8 (ran gi sde pa ni bye brag rab tu smra ba la sogs pa'o II). Note AKBh S6,6 on AK 2.26ac (standard definition of the six kleiamahiibhamika): fatra moho niimiividyii 'jfiiinam asamprakhyiinam. asamprakhyiina (Tib. mi gsal ba) in PVP D91aS/PlOSa8 (mi ses pa mi gsal ba'i no bo), PVT D13Sb4-5/P167a7 (mi gsal ba'i no bo zes bya ba'i rtogs pa med pa'i no bo'o II), PVV 85,S-6 (asamprakhyiinamiitra). 5 On the PrSVy, see Frauwallner 19S8: 43-48, Tucci 1971, de Jong 1974: 145, Muroji 1991 and 1993, Mejor 1997 and 2002. See also below 1.3.S and nn. 100 and lOS. 6 PVP D91a7/plOSb3: yan dag pa'i don mthon ba ni rig pa yin (*sadarthadarsanaT[! vidyti); PVT Je D113b7-114al/P134b2-3 = PVSVT 209,20-21: bhatiirthagrahm;aT[! vidyii. Note also PYA 14S,22: vidyii nairiitmyadr:jti:ju

42

Vincent Eltschinger

and then ask: Does ignorance consist (1) in something other (anya) than knowledge, (2) in the mere lack (abhiiva) of knowledge, or (3) in an independent cognitive event counteracting (vipak~a) or contradicting (viruddha, virodha) knowledge? First, ignorance cannot consist of something other than knowledge, for then it would ensue that dharmas such as visible things (rupiidi) or sense faculties (cak~uriidi) would be ignorance too. Moreover, dharmas such as visible things or sense faculties continue to exist beyond (sopadhi.e~a-)nirviir:ta, and his would be incompatible with their being ignorance since (at least some form of) ignorance is eliminated at the time of nirviir:ta. Second, ignorance cannot consist in the mere absence of knowledge, as the Vaibha~ika would have us believe, for, being a blank nothing like a rabbit's horns, it could not function as a cause or a condition. But since ignorance forms the first link in the twelve-membered chain of dependent origination (pratztyasamutpiida) and serves here as the causal factor for karmic forces (sa1!lskiira), it cannot consist of a causally non-efficacious absence. Moreover, the mere lack of knowledge (in the form of undefiled ignorance, akli~tiijfiiina) still characterizes Arhats and Pratyekabuddhas, i.e., persons who have already reached nirviir:ta.l We are left, then, with no other possibility than to hold ignorance to be an independent dharma that conflicts with, counteracts or opposes knowledge (vipak~a[bhiita], pratipak~a, pratidvandva[bhuta] ,
viruddha). 8

(sic); PVV 85,5: vidyaya nairatmyadmeJ:t; PVV 85,7: vidyayaJ:t sadarthatvtit. 7 PYA 145,23: na tavad vidyabhavo 'vidya tadabhavasya* nirvalJe 'pi bhavat /. * tadabhavasya according to PVAtib D Te 136a3 (de med pa), against tadbhavasya Sankrtyayana. AKBh 1,13-15: pratyekabuddha api kama1J1 sarvatra hatandhakaraJ:t / kli~tasa1J1mohatyantavigamtit / na tu sarvatha / tatha hy e~aYf! buddhadharme~v ativiprakr~tadeSakale~v arthe~u canantaprabhede~u bhavaty evakli~tam ajnanam /. On the ignorance of the Arhat, see AKVy 4,12-5,18, Bareau 1957 and laini 200l. 8 This manner of defining ignorance can already be found in Vasubandhu's PrSVy 8a3-b5 (with the conclusion, 8b4: rig pa'i mi mthun pa'i phyogs kyi chos gian ni ma rig pa yin no ies bya ba rig [P rigs] par bya ste I); but here, Vasubandhu seems to discard six hypotheses before reaching this conclusion (1. rig pa'i dfws po med pa tsam 8a3-7; 2. rig pa las gian pa gan yin pa 8a7-8; 3. rig pa dan 'dra ba'i chos gian 8a8; 4. smad pa'i rig pa 8a8-b3;

Ignorance, epistemology and soteriology (1)

43

The examples adduced by Vasubandhu, Prajfiakaragupta and Manorathanandin are worth considering. 9 According to Vasuban5. rig pa chwi ba 8b3; 6. rig pa dan bral ba gan yin pa 8b3-4). This explanation is based on grammatical explanations on Pa 2.2.6; in this connection, Vasubandhu quotes the following unidentified stanza: prati:jedhe ca sattiiyii[lz], anyatve, sadrse 'pi ca I kutsiilpavirahiirthe:ju vipak:je ciipi nafi bhavet II. "This negative particle is [used] in [the sense of] opposition to being [i.e. absence], difference, similarity, contempt, smallness, separation, and contrariety" (Sanskrit text as quoted in Mejor 2002: 90; translation Mejor 2002: 91 and 94). On this point, see Matilal 1980: 156-157, Mejor 1997: 155-156 and especially Mejor 2002 (which contains an English translation of the entire PrSVy passage). For scholastic literature, see AKBh 140,26-141,5: athiividyeti ko 'rthalz I yii na vidyii I cak:juriidi:jv api prasangalz I vidyiiyii abhiivas tarhi I evaYfl sati na kiYflcit syiit I na caitad yuktam I ... kuta etat I pratyayabhiivenopade.1iit I; PrSVy 8a7-8: rig pa las gian pa gan yin pa de yan ma rig pa ma yin te I mig la sogs pa la'an thal bar 'gyur ba'i phyir ro II, and PrSVy 8a3: re iig rig pa'i dnos po med pa tsam ni ma rig pa ma yin te I 'di ltar dnos po med pa ni 'du byed skye ba'i bdag po'i rkyen fiid du mi rigs pa'i phyir ro II; Yasomitra's comments at AKVy 301,2-13, among which AKVy 301,3-5: na kiYflcit syiid iti I yadi vidyiiyii abhiivo 'vidyii 'dravyaYfl syiid ity arthalz I na caitad yuktam iti I pratyayabhiivenopade.1iit I, as well as AKVy 301,11-13: kuta etat I pratyayabhiivenopade.1iit I avidyiipratyayiilz saYflskiirii iti I na ciividyiiyii anye:jiiYfl cak:juriidfniiYfl pratyayabhiivenopadeso yuktalz I arhatiiYfl saYflskiiriisambhaviit I na ciibhiivasya I sasavi:jiilJiidfniim apratyayatviit I). For epistemological literature, see PVP D91a7-b21 P105b3-6: de la re iig yan dag pa'i don mthon ba ni rig pa yin la I ma rig pa ies bya ba'i rtogs pa de las gian pa'am I de dan 'gal ba'am I de med pa'i chos yin gran na I re iig gian ni ma yin te I gzugs la sogs pa yan de lta bur thal ba'i phyir ro II med pa (P pa: D pa'i) yan ma yin te I dnos po'i mtshan fiid kyi gnas skabs yin pa'i phyir ro II mi mthun pa'i phyogs su 'gyur gran na I mi mthun pa'i phyogs fiid kyan 'gal ba'i ran biin fiid yin no II (see also Siikyabuddhi's comments in PVT D135b6-136al/PI67b2-4). PYA 145,23-25: na tiivad vidyiibhiivo 'vidyii tadbhiivasya nirviilJe 'pi bhiiviit I na ciibhiivo hetulz I niipi tadanyo rupiidfniim avidyiiprasangiit I tasmiid vidyiiviruddho dharmo 'vidyii 'dharmiinrtavat I tac ca satkiiyadarsanam eva /. PVV 85,5-6: vidyiiyii nairiitmyadr:jter vipak:jo 'vidyii I sa ciiprakhyiinamiitraYfl vii rupiidi vii na bhavati I nirviilJe 'pi tayor bhiiviit /. See also PVP D91b2-4/P105b6-106al. 9 AKBh 141,1-5: tasmiit - vidyiivipak:jo dharmo 'nyo 'vidyii 'mitriinrtiidivat II (= AK 3.28cd) yathii mitraviparyayelJa tadvipak:jabhutalz kascid amitro bhavati na tu yalz kascid anyo mitriin niipi mitriibhiivalz I rtaYfl cocyate satyam I tadvipak:jabhutaYfl viikyam anrtaYfl bhavati I adharmiinarthiikiiryiidayas ca dharmiidipratidvandvabhutiilz I evam avidyiipi vidyiiyiilz pratidvandvabhutadharmiintaram iti dra:jtavyam /. PYA 145,22

44

Vincent Eltschinger

dhu, ignorance stands in an analogous relationship to knowledge, just as amitra (enemy) to mitra (friend), anrta (untruth) to rta (truth), adharma (immorality) to dharma (morality), anartha (non-profit) to artha (profit), and akarya (crime) to karya (duty). Words like "enemy," "untruth," "immorality," "non-profit" or "crime" do not (just) denote something other than, or the lack of, a friend, truth, morality, profit or duty. To put it in other words, all these cases testify to the use of the negation (nail.) in the sense of a contradiction or incompatibility (virodha), and not in the sense of mere exclusion (paryudasamatra).l0 1.2.3. In PV 2.213b, Dharmaklrti puts forth a second argument to prove that ignorance indeed consists in an erroneous cognition. His argument relies on the Abhidharmic classification of ignorance as a mental event (caitta, caitasika) associated with the mind (cittasamprayuktadharma)Y He says: "[And] because [ignorance] is a perception since it is a mental factor [associated with the mind, it has the nature of] an erroneous perception[, and not of a mere lack of knowledge]." In their explanation of Dharmaklrti's argument, the commentators again refer back to Vasubandhu, for whom "[t]he mind and its mental factors are associated by reason of five equalities or identities, identity of support (asraya), of object-support (alambana), of aspect (akara), of time (kala), and equality in the number of dravyas."12 According to Yasomitra, "a [sensory] cogni(adharmanrtavat) and 145,25-27: tatha hi paranugrahalak:WlJo dharmas tadabhavamatraf!! nadharmo 'pi tu tadanyamatram I api tv* anugrahaviruddha upaghilto 'dharmal;t I tatha na bhutarthapratipadanabhavo 'nrtam api tv abhutapratipadanam asatyavacanam I. * PVAtib D Te 136a5 (chos ma yin na ni de med pa tsam ma yin la de las gian tsam yari ma yin gyi I 'on kymi. ..) suggests .. .tadabhilvamatraf!! nadharmo napi tadanyamatram I api tu .. . ,PVV 85,7 (adharmanrtavat). 10 AKVy 301,5-6: tadvipak~a iti I virodhe nan iti darsayati I na tu yal;t kascid anyo mitrad iti / na paryudasamatra iti darsayati I. 11 See AKBh 56,2-6. 12 AKBh 62,9 on AK 2.34d2 : pancabhil;t samataprakarair asrayalamban akarakaladravyasamatabhil;t I. Translation in Pruden 1988-1990: 206, with "factors" instead of "states." PVP D92al-2/P106a6-7: de ltar (D ltar: P ltar ni) rten dan dmigs pa dan rnam pa dan dus dan rdzas mtshuns pa dag gis

Ignorance, epistemology and soteriology (1)

45

tion (vijfiiina) cognizes, i.e., perceives something blue or [something] yellow; affective sensation (vedanii) experiences [on an affective level] this very thing that [has] thus [been made] the object-support [as pleasant, unpleasant, or neutral]; ideation (saf!1.jfiii) determines [its identity, and] volition (cetanii) conditions [the mind with regard to it], and so forth."13 To put this briefly, every phase (k~a1Ja) of the mind (or: cognition) is associated with a complex machinery of concomitant mental events sharing various features with this cognition, e.g., the object that has given rise to it, but also its temporal and numerical determinations. What does this argument amount to? According to Devendrabuddhi, the point at stake is that as a mental factor associated with the mind, ignorance "grasps an aspect and an object-support[, and hence] does not have the nature of a [mere] lack of cognition."14 As Sakyabuddhi argues, "if ignorance had the nature of a [mere] lack of cognition, then it would not be equal [to the mind] as regards [its] object-support and aspect (...). For this reason, since it is a mental factor, ignorance does not have the nature of a [mere] lack of cognition, but has the nature of a cognition."15

mtshwis par ldan pa'i phyir* mtshuns par ldan pa yin par 'dod do 1/. PYA 146,3-4: iisrayiilambaniikiirakiiladravyasamatiibhil; samarrt prayuktii iti samprayuktiiJ:t I. PVV 85,8-9: iisrayiilambaniikiirakaladravyasamatiidibhil; samarrt prayuktiiJ:t samprayuktii iti samprayuktalak~a7Jam I. * PVT D136a5-61 P168a2-3: rten dan dmigs pa dan [rnam pa dan] dus dan rdzas mtshuns pa dag gis mtshuns par ldan pa'i phyir ies bya ba ni chos mnon pa las sems las byun ba mtshuns par ldan pa yin par bied do II. 13 AKVy 142,1-3: vijfiiinarrt hi nflarrt pftarrt vii vastu vijiiniity upalabhata ity arthaJ:t I tad eva tathiilambanarrt (?) vastu vedanii 'nubhavati I sarrtjfiii paricchinatti I cetanii 'bhisarrtskarotzty evamiidi I. 14 PVP D92al/P106a6: dmigs pa dan rnam pa 'dzin pa yin (D pa yin: P om. pa yin) te I rtogs (P rtogs: D rtog) pa med pa'i no bo ni ma yin no II. Note Sakyabuddhi's definition of iilambana and iikiira in PVT D117b2/P143b5: dmigs pa ni yul yin la de'i 'dzin pa (D pa: P pa'i) rnam pa yin no II. 15 PVT D136a6/P168a3-4 and D136bl/P168a6-7: gal te ma rig pa rtogs pa med pa'i no bo fiid yin (yin em.: DP om. yin) pa de'i tshe dmigs pa dan rnam par mtshuns par mi 'gyur TO II ... de bas na sems las byun ba fiid yin pa'i phyir ma rig pa ni rtogs pa med pa'i no bo fiid ma yin gyi 'on kyan rtogs pa'i no bo yin no II.

46

Vincent Eltschinger

1.2.4. Dharmaklrti adds a scriptural argument against his

Vaibha~ika co-religionist: "[And thirdly] because [the Buddha himself] has declared [in diverse Siitras that ignorance has the nature of a cognition grasping aspects that are. contrary to those grasped by discernment]'''16 Though I would like to postpone my inquiry into Dharmaklrti's sources concerning ignorance,17 let me just mention that in this connection his commentators quote from two important Mahayanasiitras, viz. the Candrapradfpasamiidhisutra (CPSS) and the Dasabhumikasutra (DBhS). The CPSS says: "What does the elimination of ignorance consist in? [The elimination of ignorance consists in the fact] that [every kind of] superimposition on dharmas as they really are is [definitely] eliminated."18 As for the excerpt from the DBhS, it runs as follows: "The one who looks upon this world's coming into existence and annihilation thinks in this manner (evaf!l bhavati, 'di sfiam du sems te): 'As many as the worldly transactions/purposes [can be], all of them originate from the adherence to a self; from the elimination of [this] adherence to a self, worldly transac-

16 According to PVP D92a5-6/P106b3-4: gSUli phyir beam ldan 'das kyis kyan mda de dan de dag tu (= tatra tatra sutre PYA 145,5) ses rab kyis bzun ba'i rnam pa las phyin ei log tu 'dzin pa'i no ba iiid ni ma rig pa yin no ies gsuns pa yin te I, to be compared with PYA 146,4-5: uktal'[! ea prajiiakaraviparftapratipattir[u]paivavidyeti I tatra tatra sutra ukte~* /. * ukteJ:t according to PVAtib D Te 136b3 (gsuns pa'i phyir ra), against ukta~ Saillq-tyayana. On the "aspectual" opposition between avidya and prajiial vipasyanii, see below 1.2.5. 17 See below 1.3.5. 18 CPSS 298,28: tatra katama liianavigama~ / yad idal'[! yathiibhutiiniil'[! dharmalJiim adhyiirapavigamaJ:t II. PVP D92a4-5/P106b4-5: 'di Ita ste 'Phags pa zla ba sgron ma'i tin ne 'dzin las de la mi ses pa med pa gan ie na I gan yan dag pa ji lta ba biin gyi ehos rnams la sgro 'dogs pa med pa yin no ies gsuns pa lta bu'a II. PYA 146,5-6: tatra katama liiiinavigam[aJ:t /] yo yathiibhutiinal'[! dharmiilJiim adhyaropavigamaJ:t* /. * For adhyiiropiidhigama~ in Sailkrtyayana's edition.

Ignorance, epistemology and soteriology (1)

47

tions/purposes do not arise [anymore].' And he [also] thinks in this manner: 'Here indeed [all] these [living beings] who with infantile minds adhere to a self are screened by the darkness of ignorance, crave for existence and non-existence, devote themselves to improper reflection, have wrong conduct and are sinful; [all these human beings] accumulate performances(!accumulations of karman) that are meritorious, demeritorious, or leading to immovable states. 19 For these [ignorant persons], the seed of mind that has been planted by these conditionings, which is under the sway of cankers and clinging, will one day become capable of producing birth, old age, death, and re-existence."'20 As Devendrabuddhi rightly
19

On pU1JyiipUlJyiinefijyiin abhisaf!!skaran upacinvanti, see BHSD 1953:

57b.

20 DBhS VI C 48,1-10: tasya lokasya sambhavaf!! ca vibhavaf!! ca vyavalokayata evaf!! bhavati I yavatyo lokasamudaciiropapattayaf:t sarviis tii iitmabhiniveato bhavanti I atmiibhiniveavigamato na bhavanti lokasamudiicaropapattaya iti I tasyaivaf!! bhavati I tena khalu punar ime biilabuddhaya iitmiibhinivi~ta ajfianatimiriivrtii bhiiviibhiivabhilii~i1Jo 'yonisomanasikaraprasrta vipathaprayiitii mithyanuciiri1Jaf:t pUlJyapu1Jyiinefijyiin abhisal'[tskariin upacinvanti I te~af!! taif:t saf!!skarair avaropital'[t cittabfjaf!! siisraval'[t sopiidanam ayatyal'[t jiitijaramara1Japunarbhaviibhinirvrttisambhavopagataf!! bhavati I. PVP D92a7-b3/P106b5-107a2: de skad du Sa bcu pa'i mdo las de 'jig rten gyi 'byU1i ba (P ba: D om. ba) dan I 'jig pa la (D La: P las) rnam par Ita ba'i tshe 'di sfiam du sems te I 'jig rten du 'dogs pa 'byun ba ji sfied pa de dag thams cad ni bdag tu mnon par zen pa las 'byun ste I bdag tu mnon par zen pa dan bral na 'byun ba yan med par 'gyur ro sfiam mo II de 'di sfiam du sems te I byis pa'i blo can bdag tu mnon par zen pa mi ses pa'i rab rib kyis bsgribs (D bsgribs: P sgribs) pa yod pa dan med pa la mnon par 'dod pa can tshul bzin pa ma yin pa yid la byed pa las byun (D byun: P 'byun) ba lam log par zugs pa log (D log: P ldog) pa'i rjes su 'bran ba de dag ni bsod nams dan bsod nams ma yin pa dan mi g.yo ba'i 'du byed rnams kyan sogs te I 'du byed de dag gis yons su bsgos pa'i sems kyi sa bon zag (D zag: P zad) pa dan bcas pa fie bar len pa dan bcas pa ni phyi ma la skye ba dan rga si'i ran bzin yan srid pa mnon par grub pa 'byun ba fie bar 'gro ba yin te zes gsuns so II. Note also Sakyabuddhi's valuable explanations in PVr D136b2-4/P168a7-168b3: 'dod pa na spyod pa'i dge ba'i las ni bsod nams so II bzlog ni bsod nams ma yin pa'o II gzugs dan gzugs med pa na spyod pa'i las ni mi g.yo ba'o II 'du byed de dag gis zes bya ba ni bsod

48

Vincent Eltschinger

points out, "in the [present excerpt] too, teaching the link of ignorance (avidyiiliga) amounts to teaching the false view of a self (iitmadr$!i)."21
1.2.5. Reasoning and scriptures lead to the same conclusion: ignorance IS to be viewed as a mental factor associated with the mind that counteracts the perception of reality as it really is. Far from being a mere lack of knowledge, ignorance consists of an "antiknowledge," or, better, a "counter-knowledge." The commentators' explanation of this contradictoriness is consistent throughout their glosses on PV 2. A given cognition's erroneousness consists in its grasping an unreal (abhuta) or contrary (viparUa) aspect. 22 Can we be more precise regarding "contrariness" in this context? For two cognitions, being mutually contradictory means that they display contrary object-supports and aspects (viparuaZambanakara[tva]).23 In other words, ignorance and knowledge are cognitions grasping contrary aspects.24 We already know from DharmakIrti's commennams dag gis so II yons su bsgos pa'i sems kyi sa bon la brten nas ses pa'i nus pa byed pa'i phyir ro II sa bon de yan zag pa dmi bcas pa yin te I de la dmigs pa'i fion mons pa rnams der rgyas par 'gyur ba'i phyir ro II sred pa la sogs pa'i grogs can ni 'bras bu 'byin pa'i phyir fie bar len pa dan bcas pa'o II skye ba dan rga si dan ldan pa'i 'du byed kyi rgyun ni skye ba dan rga si'i ran biin srid pa'o II de mnon par 'grub cin de 'byun ba ni de fie bar len pa'i nus pa las byun ba yin te I de las byun ba yin pa'i phyir ro II der fie bar 'gro ba ni ldan pa'o II. PYA 146,6-7: tatM ya!:t kiiScana lokopacaropapattaya!:t sarvas ta atmabhiniveSato bhavanti I atmabhiniveSavigamato na bhavanti I. PVV 85,10-13: bhagavatapy uktalJ1 ya!:t kiiScana lokavyavaMropapattaya!:t sarvas ta iitmabhiniveSato bhavanti I iitmabhinivesavigamato na bhavantfty anena sattvadr~tir eva janmahetur ukta I atmabhiniveSalak~a1}atvat tasya!:t I. 21 PVP D92b33/P107a2: 'di fiid la yan ma rig pa'i yan lag bstan pas bdag tu lta ba bstan pa yin no II. 22 J;>VT D133b4-5/PI64b6-7: .. .yod pa ma yin pa'i rnam pa 'dzin pa'i phyir ro II. PVT D133b5/P164b8: ...phyin ci log gi rnam pa la dmigs pa'i phyir ro II. 23 PVP D56bl-2/P64a7: .. .'gal ba yin I dmigs pa dan rnam pa phyin ci log pa fiid yin pa'i phyir ro II. PVP D56b4/P64bl-2: . . .'gal ba yin te I dmigs pa dan rnam pa phyin ci log pa fiid du bstan pa'i phyir ro II. PVV 58,5-6: viparftalambanakiiratvat I. 24 PVP D91bl-2IP105b5-6: de ltar na rnam pa gan rig pas gzun ba de las

Ignorance, epistemology and soteriology (1)

49

tators that knowledge is to be understood as the perception of a real object, which is tantamount to sayin~ that it grasps the real or true aspects of an object. According to Sakyabuddhi, the aspects that are to be grasped (grahya) by knowledge consist chiefly in selfless(ness) and painful(ness), to which we may safely add empti(ness) and momentari(ness).25 Knowledge is, then, nothing other than discernment (prajfia, vipasyana).26 One of Devendrabuddhi's proofstatements puts the whole argument in a nutshell: "If it is characterized as contradictory to [a mental factor] y, a mental factor x grasps [aspects that are] contrary to the aspects grasped by y, like the [mental cultivation of the] loathsome (asubha) grasps [aspects that are] contrary to the aspects grasped by desire (raga). Now, ignorance is characterized as contradictory to discernment, which grasps the true aspects [of things]."27 Ignorance can be described, then, as that mental factor which, when it exists, obstructs the perception of true reality (tattvadarsana).28

phyin ci log pa'i rnam pa 'dzin pa can ni ma rig pa zes bya bar 'gyur ro II. PVP D91b5-6/P106a2-3: log par dmigs pa'i mi ses pa ni ma rig pa yin no II rig pas gzun bar bya ba'i (D ba'i: P ba) rnam pa las phyin ci log tu rtogs pa'i no bo ilid ni ma rig pa'o II. Note also the example adduced by Devendrabuddhi, PVP D91b6-7/P106a3-4: dper na bde ba'i 'du ses kyis gzun (D gzun: P bzun) ba'i rnam pa las 'dus byas rnams la sdug bsnal gyi 'du ses ni phyin ci log tu rtogs pa'i no bo yin pa Ita bu'o II. 25 PVT D136a3-4/P167b8: rig pas gzun bar bya ba'i rnam pa ni bdag med pa dan sdug bsnalla sags pa'o II. 26 On prajfiiilvipasyanii, see Part II, 2.2.2. Note PVP D92a2-3/P106a7-8: rtogs pa'i no bo yin du zin kyan bden pa'i rnam pa 'dzin pa can nam brdzun pa'i rnam pa 'dzin pa can du 'gyur gran na I re zig de ni bden pa'i rnam pa 'dzin pa 'dod pa ma yin te I de ltar 'dzin pa ni ses rab kyi mtshan fiid yin pa'i phyir ro II de'i phyir log (D log: P ldog) par dmigs pa ni mi ses pa yin no II. 27 PVP D92a4-5/PI06bl-3: sems las byun ba gan zig gan dan 'gal ba'i mtshan fiid can yin pa de ni des bzun ba'i rnam pa las phyin ci log tu 'dzin pa yin te I dper na mi gtsan ba ni 'dod chags kyis bzun ba'i rnam pa las phyin ci log tu 'dzin pa yin pa Ita bu'o II mi ses pa yan bden pa'i rnam pa 'dzin pa'i ses rab dan 'gal ba'i mtshan fiid can yin no ... Devendrabuddhi goes on stating that the logical reason is a svabhavahetu. 28 PVP D91b4-5/P106al-2: gan yod na don gyi de (D de: P om. de) kho na iiid mthon ba'i 'gegs byed par 'gyur ba de ni ma rig pa yin no ...

50

Vincent Eltschinger

1.2.6. Dharmaklrti's understanding of ignorance as erroneous cognition, i.e., as a cognition displaying unreal aspects, has farreaching implications. On the one hand, to claim that ignorance is responsible for our grasping of unreal aspects amounts to saying that it has us superimpose erroneous constructs on the real. The entire realm of superimposition (iiropa, samiiropa, adhyiiropa) is traditionally held to be co-extensive with the scope of Sa7pvrti, i.e., concealment, covering or convention. On the other hand, the expression "erroneous cognition" comes very close to a central concept in the epistemological theory, namely pseudo-perception (pratyak:jiibhiisa). Considering that Dharmaklrti consistently defines "concept(uality)" (vikalpa, kalpanii, etc.) as pseudo-perception, we can surmise that concept(uality), as such and as a whole, falls into the realm of ignorance. This is indeed what DharmakIrti explicitly says. According to him, terms like "concealment," "su,. perimposition," "concept(uality)" and "pseudo-perception" are conceptually equivalent. Needless to say, Dharmaklrti's frequent use of these expressions allows one to discern how deeply his epistemological doctrine is structured along such key Buddhist metaphysical assumptions. 1.2.7. Dharmaklrti was by no means the first Buddhist epistemologist to resort to the notion of saf!lvrti(sat) or "convention(ally existent)" in this context. Dignaga had already made use of it in his treatment of pseudo-perception both in his NM and in PS 1.7cd8ab. Let me first quote NM 8c19-24 in Junjie Chu's translation: "Thus it is said that memory, inference, desire, doubtful cognition, erroneous cognition, etc., [namely, the cognition of water] in respect to a mirage, etc., are not perception, because they occur as the conceptual construction of what has been experienced formerly (*purviinubhutakalpanayii). In the same way all the cognitions of a jar (*ghata), etc., a number (*saflkhyii), etc., [movement such as] lifting (*utk:jepalJiidi), etc., existence (*sattii), etc., and jarness (*ghatatva) that belong to the conventionally existent are pseudoperceptions, because they occur as conceptual constructions in assuming a different form (*anyarupa) [and] superimposing another object [that is not substantially existent] (*arthiintariidhyiiropiit)

Ignorance, epistemology and soteriology (1)

51

with regard to something substantially existent (*dravyasat)."29 Let me turn now to PS 1.7cd-8ab. A lot of ink has been spilled over these four padas, where Dignaga, after defining perception as "free from conceptual construction" (kalpanaporJha, PS 1.3c) and presenting the different types of perception (PS 1.6-7ab), spells out his views on pseudo-perception. 30 In Chu's translation, these two halfverses run as follows: "Erroneous cognition, [cognition of] conventionally existing [things, and cognitions such as] inference [and] the result [of inference] as well as recollection [and] desire are pseudo-perceptions, which have the special character of ignorance."31 All interpreters, both traditional and modern, seem to agree that the first three compounds (bhrantisaY[tvrtisajjfiana, anumananumanika, smartabhilaifika) refer to three types of conceptual/mental errors and that they make no mention whatsoever of the sensory illusions which might result from an impairment of the sense faculties. As for the expression sataimira, Chu's recent in-depth study has suggested that it should be interpreted here along the lines of Jinendrabuddhi,32 i.e., as pertaining to ignorance, and not as being related in any way to the eye-disease known as timira. 33 Thus if Jinendrabuddhi's (and Chu's) interpretation is correct unlike that of Dharmaldrti, Dignaga ascribes the whole range of conceptual error (among which is included the cognition of conventionally existing things) to ignorance. Dignaga's own explanation of saY[tvrtisajjfiana is unfortunately too short to derive any clear idea of what he had in mind: "[As for the cognition bearing] on conventionally existing [things, it is a pseudo-perception], because it superimposes extraneous things,
29 Chu 2004: 135. For the Chinese text of this passage, see Chu 2004: 135n. 85; see also Tucci 1930: 51-52, and Katsura 1993: 68. 30 See Hattori 1968: 95-98, Franco 1986, Funayama 1999: 76-79 and Chu 2004. 31 PS 1.7cd-8ab: bhriintisaY(lvrtisajjiiiinam anumiiniinumiinikam II smiirtiibhilii~ikaY(l ceti pratyak~iibhaY(l sataimiram /. Translation Chu 2004: 143144; see also Hattori 1968: 28. 32 PST 64,8-9: timirasabdo 'yam ihiijiiiinavacanaJ:! ... 33 See Chu 2004: 143-144. On timira, see Jean Filliozat's remarks in May 1959: 226-227n. 779.

52

Vincent Eltschinger

[and] thus functions through the conceptual construction of their [conventionally existing] natures [as something real]."34 Let us, however, consider one part of Jinendrabuddhi's commentary on this passage. According to him, a cognition bearing on conventionally existing things is a pseudo-perception "because this cognition superimposes extraneous things like pots on the [real] bases 35 of the designations (prajiiaptivastu), such as corporeality and the like, but does not cognize the bare basis of [these] designations."36 In bringing together these expressions, Jinendrabuddhi is most likely relying on an earlier work of Dignaga, the HVP. 37 In this short treatise, Dignaga attempted to undermine externalist/realistic accounts of reality by proving that the three spheres of existence in the traditional Buddhist cosmology are mere illusion/error Ckhrul pa tsam).38 Stanza 2 runs as follows: "When one examines/inquires into the nature [of things] with reference to all the [different] bases of [their] designations, [one comes to the conclusion that,] as many objects of conventional cognition (sarrr-vrtijiiiinagocara) there can be, [all of them] are Uust] designated/designations on [the basis of]/from other(s)[, provisionally more real things]."39 Let us now consider Dignaga's own explanation of this verse: "When one exa34 PSV 3,19: saJ?lvrtisatsv arthantaradhyaropat tadrupakalpanapravrttatvat.

35 "Basis" is one of the two meanings of vastu in the Tattvarthapatala of the BoBh; in this meaning, it is regularly (but by no means systematically) translated by Tib. gii. See also Vasubandhu's distinction between five meanings of vastu in AKBh 94,11-16 on AK 2.55d2 36 PST 58,9-10: tad dhi prajfiaptivastuo!u rupadio!v arthantarabhutan ghatadfn adhyaropayati, na prajfiaptivastumatraJ?l pratipadyate I. 37 On the chronology and teachings of the HVP, see Frauwallner 1959: 127/803-130/806. 38 HVP 155/831,8. 39 HVP 154/830,1-4 (= k. 2): I btags pa'i dnos po thams cad la II ran gi no bo brtags pa na II kun rdzob ses pa'i spyod yul ni II ji sfied yod pa gian las btags I. The reading gian las is not soundly established: (1) the wording of Dignaga's own commentary reads (154/830,11-12) gian du ies bya ba ni I don dam pa'i fiid las so II; as for the translation by DanaSIla/Dpal 'byor sfiiIi po, its padas 3-4 read (Frauwallner 1959: 154/830): gian du kun rdzob ses pa'i II spyod yul ji sfied btags pa yin I.

Ignorance, epistemology and soteriology (1)

53

mines/inquires into [things] such as a certain rope by analysing [itinto its] parts, etc., one does not find the nature [of these things any longer]; like the [mistaken] cognition of a snake [in place of a rope], the cognition of [something] like a rope, too, is nothing but a mere error. In the same way, as many as can be the objects of conventional/practical knowledge (vyavahiirajfiiina) such as pots and goblets, which [only] exist on the basis of [their] parts, all of them exist as [mere] designations, too.''40 As the evidence provided by the HVP suggests, Dignaga's notion of sa1?Zvrtisaj(jfiiina) can be traced back to the standard Abhidharmic interpretation of the two realities as can be found, e.g., in AK 6.4 and AKBh 334,3-14 thereon. 41 According to the Vaibha~ika Abhidharma at least, conventionally existing things are those whose idea (buddhi) ceases as soon as they are broken into pieces (as in the case of a pot) or their different properties (dharma) are mentally (buddhyii) subtracted (apoha) from them. 1.2.8. There is, however, strong evidence to suggest that this does not reflect, at least not primarily, Dharmaklrti's line of interpreting sa1?Zvrti. In PVin 1.33acl' Dharmaklrti claims pseUdo-perception to be twofold, namely conceptual construct (vikalpa), "because something unreal appears in it" (avastunirbhiisiit),42 and (sensory, ak~aja/indriyaja in PVin 1.29,11-12) illusion (upaplava), "because it is deceptive/unreliable" (visa1?Zviidiit). It is generally acknowledged that this dual account of pseudo-perception was already responsible for the framework of Dharmaklrti's interpretation of PS 1.7cd-8ab: "There are four kinds of pseudo-perception: three
40 HVP 154/830,5-10: ji ltar cha sas la sogs pa'i dbye bas tha dad pa'i thag pa la sogs pa la brtags pa na rari gi rio bo ma dmigs te 1thag pa la sogs pa'i blo yari sbrullo siiam pa'i blo biin du 'khrul pa tsam du zad pa de biin du 1rios cha la sogs pa la bltos nas yod pa rdza ma dari phor bu la sogs pa tha siiad pa'i ses pa'i spyod yut ji siied pa thams cad kyari btags par yod pa yin te I. 41 On AK 6.4, see Katsura 1976 and 1993: 69; on the Vadavidhi's notion of sal'J'!vrtijiiana, see Katsura 1993: 68-69. 42 Or, according to Dharmottara's gloss, "because it does not display the [real] entity [itself]" (cf. PVinT D124a4/P143b2: de yari drios po mi snari ba'i phyir drios po snari ba dari brat ba'i phyir).

54

Vincent Eltschinger

[of them] consist in conceptual cognitions, and one [of them, being] born of an impairment of the [sensory] basis [itself], is nonconceptual.''43 Commenting in general terms on the three conceptual kinds of pseudo-perception, Devendrabuddhi explains first that they function through conceptualizations based on conventions and that they superimpose extraneous things on reality.44 Focusing then on saY(lvrtisajjfiiina, he adduces five examples directly borrowed from Dignaga's NM, where they were meant to match Vaise~ika ontological categories: pot, number, lifting/raising, existence, potness. 45 His explanation of the conceptual pseudo-perception of a pot is as follows: "[Things such as] pots, since they perform one [particular] effect (ekakiiryasiidhana) such as carrying water, etc., commonly differ from [all the things] other than them which do not have the [said] effect (tadanyebhyo 'tatkiiryebhyalJ')"'46 As is obvious, Devendrabuddhi's explanation of conventionally real entities (at least partly) substitutes the foundations of the so-called apoha theory for the standard Abhidharma-like account favoured by Dignaga (at least in his HVP and NM). A short glance at Dharmaklrti's theory of apoha should allow us to understand his notion of saY(lVrti(sat).47
1.2.9. Be it because of their spatio-temporallocation or the intrinsic

nature they owe to ever changing causal complexes, bare particulars are irreducible to one another: An individual entity can only 43 PV 3.288: trividhal'Jl kalpaniijfiiinam iiSrayopaplavodbhavam I avikalpam ekal'Jl ca pratyak~iibhal'Jl caturvidham II. 44 PVP D212a2/P248b2: rtog pa'i ses pa gsum yan brda'i rten can dan sgro 'dogs pa'i rnam par rtog pas 'jug ste I. Cf. PV 3.290ab: sanketiiSrayiinyiirtha samiiropavikalpane I. 45 PVP D212a2/P248b2-3: dper na bum pa dan grans dan 'degs pa dan yod pa dan bum pa fiid la sogs pa lta bu'o II. On the NM passage, see above 1.2.7 and n. 29; note also Vibh. 205n. 2: sattval'Jl dravyal'Jl gharasankhyiik~epar;asattiigharatviidi~u I.
46 PVP D212a2-3/P248b3-4: bum pa la (P la: D las) ... chu la sogs pa 'dzin pa la sogs pa 'bras bu gcig sgrub par byed pa'i sgo nas de las gian pa'i de'i 'bras bu can ma yin pa dag las tha dad pa fiid tha dad pa med pa yin no II. 47 On Dharmaklrti's theory of apoha, see Frauwallner 1932, 1933a, 1933b, Vetter 1964: 41-63, Katsura 1991, Dunne 2004: 113-143.

Ignorance, epistemology and soteriology (1)

55

be identical with itself, and this during one and the same phase of its transient existence. Dharmaklrti spells this out in the following stanza: "Since they are fixed/established in their own [strictly individual] intrinsic nature, all entities are naturally excluded [i.e., distinct] from those whose being is [similar to] their own (svabhiiva) and from those whose being is other [than their own] (parabhiiva)."48 That bare particulars can only be identical to themselves does not mean that they cannot be regarded as roughly similar to certain others in a specific respect. The causal complexes and processes that bring them into existence do indeed endow them with similar causal efficiencies or functionalities (arthakriyii). In other words, similarities and dissimilarities can only be accounted for in terms of causal efficiency, and not by resorting to real similarities (as the Sankhyas would have it) or to universals inhering in individuals (as the Vaise~ikas would assume). For instance, plants such as the gUi;liicf (Cocculus cordifolius), the dhiitrf (= iimalakf, Emblica officinalis) or the abhaya (Terminalia chebula) can differ extremely from one another (parasparam atyantavibhinnamiirtaya~, TSP 296,21), but they are possessed with a similar capacity to cure diseases (rogasiinti) or to calm fever (jvarasamana).49 "In the same way - Dharmaklrti says - individuals like a sif(lsapii[, a khadira or a nyagrodha tree], although they are not related to one another [i.e., are devoid of any real universal such as treeness], naturally produce a uniform recognition [of them as trees], or [any] other function that according to conditions can be performed by wood, such as burning or [building] a house. But [things] like water, though they no less differ [from these individuals than these individuals among themselves], are not [capable of burning or being used in the making of a cart], in the same way as [something] like the auditory [sense faculty is not capable] of

48 PV 1.40 (leaving yasmat untranslated): sarve bhava/:t svabhavena svasvabhavavyavasthite/:t / svabhavaparabhiivabhya1'[t yasmad vyavrttibhagina/:t II. 49 See PV 1.74, TS 5722-724/K723-725 and TSP thereon, TS 510041 K1005 (TSP 5390,21/K317,11: harftakf). See also TS 51034/K1035 and TSP ad loco

56

Vincent Eltschinger

[producing] a cognition of something visible, etc."so In other words, our classificatory schemes, both conceptual and linguistic, do not refer back to commonalities in re(bus), to real universals or similarities, but to our own way of accounting for the fact that things which have nothing in common perform roughly similar functions. No matter how different they are as to their colours, shapes, materials or sizes, we call "pots" or "receptacles" all those things that are capable of carrying liquids, and ipso facto single them out from all things that do not perform this function (atatkiirya). Most important in the present context is Dharmaklrti's claim that mutually exclusive entities can give rise (in a mediate/indirect way, as we shall see) to a conceptual cognition that he variously designates as a unitary judgement (ekapratyavamarsa) or recognition (pratyabhijiiiina): "Such is the nature of [real] entities[, namely,] that some [of them], although they differ [totally from one another], are naturally bound to perform one [and the same] effect such as a unitary judgement."51 This judgement is of a determinat(iv)e character and has for one of its properties the display of a unitary image of irreducibly diverse particulars. At this stage, however, ignorance has already stepped into the process: As Dharmaklrti insists, this conceptual construct covers or conceals (saf(Li'VJJ.) the bare particulars' diversity with its own unitary aspect. Whereas this cognition is to be considered saf(Lvrti itself, its pseudo-objects are saf(Lvrtisat or conventionally existing things with no counterpart in reality. Dharmaklrti spells this out in the following passage: "On the basis of entities that are [entirely] distinct [from one another], a cognition arises which, having the appearance of a unitary object, conceals with its own [unitary] form the form of the others [i.e.,

50 PVSV 41,3-6: evarrz sirrzsapadayo 'pi bhedaJ:t parasparananvaye 'pi prakr,tyaivaikam ekakararrz pratyabhijfianarrzjanayanty anyarrz va yathapratyayarrz dahanagrhadikarrz ka~thasadhyam arthakriyam / na tu bhedavise~e 'pi jaladayaJ:t I srotradivad rupadijiiane /. 51 PVSV 40,22-25 (together with PV 1.73ad1): ekapratyavamarsarthaj iianadyekarthasadhane I bhede 'pi niyatah kecit svabhaven[a] ... II. Note also TSP S297,14-16/K240,8-1O: ... [a]nye 'pi ghatadayo bhavaJ:t svahetupratyayebhyas tathotpatteJ:t prakrtyaivaikakarapratyavamarsadihetavo bhavi~yant[i] ...

Ignorance, epistemology and soteriology (1)

57

the radically distinct entities; the latters' ultimate] diversity being [thus] concealed by this [cognition, i.e., by] concealment, [these] entities, [though] in themselves distinct [from one another], appear as indistinct under a certain form [i.e., as universals constructed by their exclusion from dissimilar entities with reference to the effects they produce]. The universal is [only] reported to be real with reference to this [superimposing cognition, but] in the way in which this cognition imagines it, it is ultimately unreal."52 It is to be emphasized that real entities are only indirectly or mediately the cause of this distorting cognition. What they bring about directly is a true perceptual awareness of themselves, which in turn is the cause for the awakening or actualization (prabodha) of a latent tendency (viisana) that has been left/imprinted (iihita) in the mind by the previous experience of functionally similar particulars. "Latent tendency" or, better, "imprint" is consistently defined as a capacity or efficiency (jakti, siimarthya);53 in the present case, the latent tendency consists in the capacity to generate recognition. New perceptual experiences of functionally similar entities provoke the actualization of this latent tendency, i.e., its readiness to bring about 54 52 PV 1.68-70: pararupa'!1 svarupelJa yayii sa'!1vriyate dhiyii I ekiirthapratibhiisinyii bhiiviin i1sritya bhedinaJ:! II tayii sa'!1vrtaniiniirthiiJ:! sa'!1vrtyii bhedinaJ:! svayam I abhedina iviibhiinti bhiivii rupe1}a kenacit II tasyii abhipriiyavasiit siimiinya'!1 sat prakfrtitam I tad asatparamiirthena yathii sa'!1kalpita'!1 tayii II. On this passage, see van Bijlert 1989: 139-140 and Katsura
1993: 67-68. 53 See PVSV'f 156,8 (ad PVSV 35,6-7), 206,19-20 (ad PVSV 49,22), 212,29-30 (ad PVSV 51,12-13), 298,16-17 (ad PVSV 76,9), 434,27 (ad PVSV 119,19-20), 539,19-20 (ad PV 1.286). On the metaphor of perfuming, see MS 1.15, AKVy 329,19 and Siddhi 1.120 (et passim). To be compared with the Sautrantika doctrine stated in AKBh 278,20-24 (distinction between kleiinusaya and klesaparyavasthiina). 54 The prabodha of a latent tendency is consistently defined as kiiryotpiidana'!1 praty iibhimukhyam (PVSV'f 206,19-20 ad PVSV 51,12-13 = PVSVT 540,24 ad PVSV 151,27), kiiryotpiidaniinugwJyam (PVSVT 156,8 ad PVSV 35,6-7), and iinugulJyam (PVSVT 212,29-30 ad PVSV 51,12-13). Note also PYA 124,4: prabodhaJ:! katha'!1cit kadiicit kenacid dhetuneti I and PVP D68a7/P78a4-5: de dag ni rnam par ses pa sna tshogs pa gian daJi gian
bag chags rab tu sad par byed pa dag gis rab tu sad par 'gyur ro II gaJi gi tshe rkyen thag fie ba de'i tshe 'gyur ba ... (short comments in PVT D123a5-61

58

Vincent Eltschinger

the unitary judgement. 55 As Dharmaklrti himself points out, this latent tendency towards conceptual construct is the real cause of error (viplava), or has error as its result. 56 We are now in a position to understand the following statement concerning how conceptual construction conceals ultimate reality: "On the basis of entities that are distinct from [those] other than them[selves with respect to their functionality], a conceptual cognition arises; true to the nature of its own latent tendency[, which is its real cause and consists in the capacity to produce a unitary cognition of mutually distinct things, this cognition] obliterates the latters' [mutually] distinct nature, ascribes [to them] its own unitary appearance, and presents them by unifying them [i.e., as if they were of a single nature]. And such is the nature [both] of the entities [themselves], which differ from others in that they share the same aim and instrument,57 and of the latent tendency [that is the cause] of this conceptual construct: that the [cognition] produced [by them in an indirect and a direct manner, respectively,] appears like that [i.e., by superimposing its own aspect on something else]. And it is this [conceptual cognition that we call] 'concealment' in that it conceals the form of [mutually distinct] others with its own form [i.e., with its own unitary aspect]. As for the [entities] whose differences have been concealed by this

P151a3-4). Here, de dag refers to defilements. 55 See PVSV 42,13-16: yad etaj jiiiinal'{t vastusvabhavagriihi1:Ziinubhaven iihitiil'{t viisaniim ii.ritya vikalpakam utpadyate 'tadvi~ayam api tadvi~ayam iva tadanubhavahitavasaniiprabhavaprakrte/:z . .. PVSV 49,20-23: tiim ekiil'{t jiianiidikiim arthakriyiil'{t te~u pasyato vastudharmatayaiviinyebhyo bhidyamiina bhiiviis tadvyiivrttivi~ayadhvanisal'{tsr~tal'{t tad eve dam iti sviinubhavavasaniiprabodhena sal'{tsr~tabhedal'{t mithyiipratyayal'{t janayanti I. 56 See PVSV 43,2-3: sarvas ciiyal'{t svalak~m:liiniim eva darsaniihitaviisaniikrto viplava iti ... PVSV 50,2-3: ta eva bhiiviis tadekarthakiirilJo 'nubhavadviirelJa prakrtyii vibhramaphaliiyii vikalpaviisaniiyii hetutviin nimittam/. 57 According to Sakyabuddhi (PVr D Je 87a3-4) and Kar:gakagomin (PVSVr 198,13-14), siidhyasiidhanao is a dvandva compound; siidhya is explained as udakiididharalJiidi, and siidhana as mrtpilJiidi. But as Lambert Schmithausen has rightly pointed out to me, the compound may also be interpreted as: "in that they are the means(jinstruments) for one and the same aim(!end)."

Ignorance, epistemology and soteriology (1)

59

[conceptual cognition], they appear [in the latter] under a certain [superimposed] form as non-different [from one another,] although they are [indeed] different in themselves."58
1.2.10. Let us now try to summarize Dharmak'irti's understanding of saY(lvrti and saY(lvrtisat. Basically, saY(lvrti is a cognition (buddhi, dhf, pratyaya) of a conceptual character (vikalpika). Dharmakirti frequently calls it a unitary (eka, ekiikiira, abhinniikiira) judgement (pratyavamarsa[pratyaya]), a recognition (pratyabhijiiiina), or a perceptual ascertainment/judgement (niscaya[jfiiina]).59 This cognition is variously described as erroneous (mithyiiO) or as error (bhriinti, vibhrama, upaplava, vip lava). As for conventionally existing things (or rather, pseudo-things), they consist in the objects (artha) or images (iikiira, riipa, pratibimbaka)60 of the object as they appear in or are displayed by the conceptual cognition Ufiiinao, buddhipratibhiisin, jfiiinanivi$ta).61 This object is of a purely "cognitional" nature (bauddha).62 The object as it appears in this cognition is basically unreal (nistattva, ni/:lsvabhiiva, asamfk$itatattva),63 but several properties are systematically ascribed

58 PVSV 38,17-24: buddhil:t khalu tadanyavyatirekirtalJ padiirthiin iisrityatpadyamiina vikalpika svavasanaprakrtim anuvidadhatf bhinnam e$aT(l rupaT(l tirodhaya pratibhiisam abhinnam iitmfyam adhyasya tan saT(lsrjantf saT(ldarfayati / sa caikasadhyasadhanatayanyavivekiniiT(l bhavanaT(l tadvikalpavasaniiyiis ca prakrtir yad evam e$ii pratibhtiti tadudbhava / sii ceyaT(l saT(lVrtil:t saT(lvriyate 'nayii svarupelJa pararupam iti / te ca taya saT(lvrtabhedalJ svayaT(l bhedina 'py abhedina iva kenacid rupelJa pratibhiinti /. 59 On perceptual ascertainment/judgement, see Katsura 1993 and Kellner 2004. 60 (artha)pratibimba(ka) is the word most frequently used by Santarak~itain this context (see TS SlO05b/KlO06b, S1016b/K1017b et passim); KamalaSIla (TSP S391,8/K317,16) glosses it with arthiibhiisa; see PV 3.l65b (together with PVP D182al-2/P211b8-212a1), PVSVT 182,9, 212,17 et passim. Note also TS S1182/K1l83: asmabhir ukta akaralJ pratibimbaT(l tadabhata / ullekhalJ pratibhiisas ca saT(ljiiiibhedas tv akiiralJam //. 61 jiiiinanivi$!a in PV l.78a; vikalpabuddhyiiruq.ha in TSP S391,21/K318,6. 62 PV 1.88a.
63 nistattva in PV l.77c, nilJsvabhiiva as a gloss of the latter in PVSVT 190,8, asamfk$itatattva in PV l.85c.

60

Vincent Eltschinger

(adhyavasita, a.ropita, adhyasta) to it: thought erroneously interprets it as being external (bahya) , real (vastu), unitary (eka) , or suitable for performing a function (arthakriya.ka.rin).64 Conceptual construction's (kalpana., vikalpana) basic and most congenital error is to conceal reality's diversity by ascribing uniformity to it, thus construing universals or commonalities (sa.ma.nya[lak~a~a]) that it takes to be real: Many distinct things with similar functionalities are erroneously thought to share real common features that exclude them from those which lack the said functionalities. Conversely, indivisible (nira1?1sa, eka) particulars are erroneously divided into numerous real properties (dharma) and a substantial property-bearer (dharmin), this being the basis for "co-reference" (sa.ma.na.dhikara~ya) and inference itself.65 To put it in other words, the intrinsic error of a conceptual construction consists of unifying what is multiple and dividing what is ultimately one. Conceptuality (i.e., sa1?1vrti) distorts reality, superimposes its own constructs (i.e., the sa1?1vrtisat) on it, and falsely ascribes erroneous aspects to it. According to an interesting remark by Sakyabuddhi, sa1?1vrti is to be defined as a cognition having the appearance of a contrary aspect (viparfta.ka.ra).66 Let us recall here that the hallmark of ignorance is precisely that it does not grasp reality itself (avastugraha~a), that it consists in a cognition displaying aspects which are contrary to the ones grasped by vidya.. In the same way, a conceptual construct does not grasp reality (na ... vikalpasya vastugraha~am);67 all of them are endowed with an erroneous object (mithya.rtha eva sarvo vikalpalJ),68 and this is, according to Dharmaldrti, their intrinsic or natural error prakrtivibhrama.d vikalpa.na.m).69
64

See, e.g., PVSV 42,18-20, PV 1.77.

65 See PV 1.78-79 and PV 1.82-83. Note also PVSVT 183,7-8: ekiikiirelJa pratibhiisaniit siimiinyavyavahiiraf:z I anekiikiirelJa caikasya pratibhiisaniit siimiiniidhikaralJyavyavahiiraf:z I. See also Part II, n. 125.
66 PVT D146b6/P181a7-8: kun rdzob la ies bya ba ni phyin ci log gi rnam pa can gyi blo la'o II. 67 PVSVT 176,26-27. 68 PVSVT 183,9. 69 PVSV 39,2-3: tam eva grhlJatf sii prakrtivibhramad vikalpiiniirrz vastugriihilJfva pratibhiiti I. "Though this [cognition] grasps this [conceptual rep-

Ignorance, epistemology and soteriology (1)

61

It comes as no surprise, then, that Dharmaklrti explicitly equates this conceptual construct with ignorance: "That conceptual constructs are incapable [of grasping reality as it is] comes from ignorance. Errors do indeed not depend just on external [things], but are also due to an internal confusion (viplava) [consisting in: one's grasping erroneous aspects], just like the error of [one's cognition displaying] a hair-mesh, etc?O [Objection: But] if confusion is produced by ignorance, it ensues that [ignorance will also] affect [perceptual awarenesses] such as visual cognitionsL which you do not accept. Answer:] No, because this [ignorance, i.e., a cognition that superimposes unreal aspects,] has the conceptual construct for its definition.71 Ignorance indeed only consists in the conceptual construct, [and] this [ignorance] is in error by its very nature."72 A
resentation] only, it appears as if it would grasp the thing [itself, and this] due to the intrinsic error of conceptual constructs[, i.e., due to the error consisting in grasping things without distinguishing them from their own aspect]." Note also PVSV 39,5-8 and 9-10: vyakhyiitiiraJ:t khalv evaY(! vivecayanti na vyavahartiiraJ:t I te tu sViilambanam eviirthakriyayogyaY(! manyamiinii drsyavikalpyiiv arthav ekfkrtya pravartante I tadabhipriiyavasiid evam ucyate I ... pratibhasabhediidibhyas tu tattvacintakii niibhedam anumanyante /. "[Only] the theoreticians distinguish in this way[, i.e., between the conceptual image and the thing], not those who are engaged in practice. As for the latter, they hold their own [cognitional] object to be able to perform a function and act by unifying both objects, the perceptible one and the conceptualized one. C ) On the contrary, [the theoreticians] who investigate the true nature [of things] by distinguishing between [their respective] modes of appearance [or causal efficiencies, these] do not assent to the identity [between perceptible and conceptual]." 70 Note also TSP S396,19-20/K322,6-7: na hi sarvii bhriintayaJ:t siidharmyadarsaniid eva bhavanti I kiY(! tarhy antarupaplaviid ap[i]. .. On internal and external causes of error, see Part II, nn. 116 and 139. 71 Note also TSP S632,22/K519,20-21: vikalpasya cavidyiisvabhiivatvi'it I. n PVSV 50,16-51,1: asaktir e~ii vikalpiiniim avidyiiprabhaviit I na vai biihyiipek~a eva bhriintayo bhavanti I kiY(!tu viplavad iintariid api keiidivibhramavat I avidyodbhaviid viplavatve cak~urvijfiiiniidi~v api prasmigaJ:t I na I tasyii vikalpalak~a1Jatvi'it I vikalpa eva hy avidyii I sa svabhavenaiva viparyasyati /. According to his final, Yogacara position, Dharmaklrti admits that perception, inasmuch as it presupposes a duality between an object (griihya) and a subject Cgriihaka), is affected by ignorance. See Eltschinger 2005: 162-175.

62

Vincent Eltschinger

stanza quoted by KamalasIla, not yet identified, nicely summarizes Dharmaklrti's understanding of conceptual construction as ignorance: "The conceptual construct falls into the nature of ignorance by its very nature, because it functions/exists by superimposing its own aspect [on reality] as [being] of an external nature."73 1.3. Ignorance as personalistic false view
(satkiiyadr~ti)

1.3.1. Dharmaklrti's understanding of ignorance as conceptual concealment reveals itself as a highly sophisticated account of the two truths, the dogmatic structure that underlies his whole philosophical enterprise.74 But as "religiously" relevant as the two truths

73 TSP S633,9-1O/K519,21-22: vikalpaJ:! svayam evayam avidyarupatarrz gataJ:! I svakararrz bahyaruperza yasmad aropya vartate II. 74 Most conspicuously in PV 3.1-3 (without 3.1c2 d): manarrz dvividharrz vi.Jayadvaividhyac chaktyasaktitaJ:! I arthakriyayarrz... sadrsasadrsatvac ca vi.Jayavi.JayatvataJ:! I sabdasyanyanimittanarrz bhave dhfsadasattvataJ:! II arthakriyasamartharrz yat tad atra paramarthasat I anyat sarrzvrtisat proktarrz te svasamanyalak.Jarze II. "There are two [and only two] means of valid cognition[, viz. perception and inference], since there are two [and only two kinds of] objects[, and this for four reasons]: because [one] is capable of performing a function, whereas [the other] is incapable [of it] (...)*; because [one] is common [to other things] whereas [the other is strictly] particular**; because [one] is the object of words, whereas [the other] is not the object [of words]; because the cognition [of the one] exists when causes other*** [than the object itself] are present, whereas [the cognition of the other] does not exist [when causes other**** than the object are present but the object itself is absent]. In this [treatise, we] term 'ultimately real' the [kind of object] that is capable of performing a function[, is particular, verbally inexpressible and cognized only when it is present, and we term] 'conventionally real' the other [kind of object]. These are[, respectively,] the svalak.Jarza and the samanyalak.Jarza." * Cf. PVSV 87,4: yad arthakriyakari tad eva vastv ity uktam I. ** sadrsa is explained gian dan thun man ba in PVP D123b51 P144al, and sarvavyaktisadhararza in PVV 112,3-4; asadrsa is explained phan tshun ldag pa thun man ma yin pa in PVP D123b6/P144al-2, and sarvata vyavrtta in PVV 112,4. *** According to PVP D124a2/p144a6 = PVV 112,9: manaskara[t]atsa(d)gurzyasanketagrahalJanam (yid la byed pa de dan rjes su mthun pa dan brda 'dzin pa); Sakyabuddhi (PVf D152b6-71 P188b6) explains sa(d)gurzya as 'bras bu skyed pa la nus pa yad pa, and sanketagrahalJa as snan gyi brda dran pa. **** According to PVP D124a31

Ignorance, epistemology and soteriology (1)

63

may be, the theory of apoha does not seem to account for the most prominent and. dramatic function of ignorance within Buddhism, i.e., its being the first link in dependent origination (pratftyasamutpiida) and the factor ultimately responsible for suffering. To be sure, the conceptual cognition of a pot, or even the mistaken notion of a snake, can hardly be held responsible for one's being a slave to rebirth and sarrtsiira. But how is it with the mistaken notion of a self, which owes its existence to psychological conditions very similar to those responsible for universals? As we shall now see, Dharmaklrti calls "personalistic false view" that part of ignorance which, superimposing a fictitious self and that which is believed to belong to this self, gives rise to all defilements and commits living beings to action and rebirth,l5 To be more precise, Dharmaklrti consistently equates ignorance with the personalistic false view. Considering the fact that "personalistic false view" is by no means synonymous with "conceptuality," how can Dharmaklrti equate ignorance with the personalistic false view and with conceptuality? This identification raises a second problem: Though he assents to it in his PrSVy, Vasubandhu strongly rejects the identification of ignorance with a false view (dr$ti) in his AKBh. To sum up, .Dharmaklrti's position faces both the problem of internal consistency and of Abhidharmic orthodoxy.
1.3.2. Dharmaklrti equates the personalistic false view with ignorance/ignorance/delusion in several sections of his pv. Let us review a few of them. The identification of satkiiyadarsana with ignorance occurs in the following stanza: "The birth of all kinds of [moral] faults is due to the [false] view of self [i.e., to the clinging to self and one's own, and] this [false view of self] is ignorance [itself]."76 Its identity with ignorance is stressed in Dharmaklrti's commentary on the same stanza: "All [moral] faults are born from the [false] view of self, and it is this [false view of self] that is

P144a7-8, these causes consist of cak~us, aloka, manaskara, etc. 75 On the satkayadr~ti, see Rahder 1932 and Kosa 5.15-17.

76 PV 1.222ac sarVasa1!L do~ajatfna1!L jatil:t satkayadarsanat / savidya . ..

j :

64

Vincent Eltschinger

called 'ignorance' [in our doctrinal system]."77 As for equating personalistic belief and delusion, this occurs at least twice in PV 2: "Delusion is the root[-cause] of [moral] faults, and this [delusion] consists in the belief in a [personal] being."78 Dharmakjrti was well aware of the fact that his interpretation of ignorance as the false view of self might incur criticism from many of his co-religionists, even though, as we shall see, he was by no means the first Buddhist scholar to interpret it in this manner. Whereas his general understanding of ignorance in terms of erroneous cognition was closely in line with Vasubandhu's position, his identification of a false view with ignorance could be rejected on the grounds that Vasubandhu held the false view to be associated (samprayukta) with ignorance and not identical with ignorance;79 consequently, this identification could easily be found guilty of contradicting the scriptures (agamavirodha). Dharmakjrti was thus compelled both to explain how something can be said to be associated with itself, and to account for such a loss of the meaning of "associated" (samprayuktartha). Dharmakjrti seems to have entrusted (hypothetical) followers and/ or commentators with the task of answering these questions of a more dogmatic character. His exegetical justification for this identification occurs in the following stanza: "The contradiction [with scriptures] which [seems to ensue] on this point has to be explained. However, since [the perception of emptiness] is contradictory to this [personalistic false view], it is established that the perception of emptiness contradicts all the [moral] faults [too], which are born of this [personalistic false view]."80 The argument underlying this justification is the following. Most Buddhist teachings hold ignorance/ignorance/delusion to be the cause of defilements and suffer77 PVSV 111,19-20: satkliyadarsanajlil:t sarvado~lil:t / tad eva clijfilinam ity ucyate /. 78 PV 2.l96abj : mohas ca muZarrt do~li1Jlirrt sa ca sattvagrah[al:t]. .. PV 2.212cd: tanmuZlis ca mallil:t sarve sa ca satkliyadarsanam //. ''All the defilements have this [delusion] as [their] root[-cause], and this [delusion] is the [false] view of self." 79 AK 3.29c: dr~tes tatsamprayuktatviit. 80 PV 2.214: vylikhyeyo 'tra virodho yas tadvirodhlic ca tanmayail:t / virodhal:t sunyatiidr~tel:t sarvado~ail:t prasidhyati //.

Ignorance, epistemology and soteriology (1)

65

ing. Since the perception of emptiness (unyatadr~ti), or the perception of selflessess (nairatmyadr~ti), is the only means capable of uprooting defilements, it must be capable of eliminating their cause too, because the elimination of something presupposes the destruction of its cause. 81 That the perception of selflessness elimInates the cause of defilements can only be due to the fact that its objectsupport and aspect are contrary to those displayed by the cause of defilements. The only cognition whose aspects are contra(dicto)ry to those displayed by the perception of selflessness is the false view of self. The ignorance held to be the cause of defilements can, therefore, only consist in the personalistic false view. The following stanza follows the same line of argument: "Delusion is presented as the [principal] cause of moral faults [in one Sfitra, and] in another one, it is the personalistic false view, because [moral faults] are eliminated when the [personalistic false view] is eliminated."82 In his commentary,83 Dharmaklrti argues that if the Buddha has taught delusion to be the cause of defilements in one Slitra, and in another one the personalistic false view, he could only have the same primary cause (pradhiina[hetu]) in mind, because in both cases he is pointing out the factor that, when eliminated, leads to the elimination of the defilements, i.e., the "material" cause (upadana). In other words, "delusion" and "personalistic false view" are synonymous. How do commentators with this type of doctrinal background explain away the contradiction to the scriptures that seems to ensue from this identification? What does "associated" mean in the Abhidharmic statement "ignorance is associated with the false

81 Or, to put it in the reverse: the factor that is capable of eliminating the cause of another factor indirectly (arthat) eliminates the latter factor too, just like fire eliminates a chill by eliminating the cold. Cf. PVP D93a6/P107b6-7: dper na gran ba dan 'gal ba'i me ni gran ba'i khyad par ba mo'i reg pa la sogs pa dan gran ba'i 'bras bu spu Ion ies byed pa'i khyad par la sogs pa de'i ran biin dan Ihan cig 'gal ba Ita bu'o II. To be compared with PVV 86,4-5: sftaviruddhasyilgner iva tatkilry[ai] romahar~ildibhi/:1 ... 82 PV 1.223 (leaving ata eva untranslated): moho nidilnaf!! do~iir;.iim ata evilbhidhfyate I satkiiyadr~tir anyatra tatprahar;.e prahar;.ataJ:t II. 83 See PVSV 111,23-112,5, Dunne 2004: 372-373, Eltschinger 236-239.

66

Vincent Eltschinger

view(s)" if ignorance and the personalistic false view are one and the same thing?84 According to Devendrabuddhi, "associated" points here to a relationship between the parts and the whole (ekade.faikade.fibhava);Sakyabuddhi explains it as being like the relationship between the body and its limbs (migaftgibhava).85 According to Prajfiakaragupta, "associated" refers to a relationship between the universal and specific instance (siimiinyavise~abhiiva).86 Ignorance and the personalistic false view stand in the same kind of relationship as a forest (vana) and paliisa-trees (Butea frondosa) in the expression: "The forest possesses paliisa-trees" (paliisayuktalfL vanam iti), or the body (sarfra) and limbs such as hands (piilJi) in the expression: "The body possesses [limbs] such as hands" (piilJyiidiyuktalfL sarfram iti).87 The apparent contradiction with scriptures can then be explained away easily: To say that ignorance is associated with the personalistic false view means that ignorance as a whole, a body or a universal possesses the personalistic false view as a part, a member/limb, or a specific instance. 88 As 84 PVP D92b7-93a1/P107a7-8: gal te 'jig tshogs su Ita ba fiid ma rig pa yin pa de'i tshe Ita ba dan mtshuns par ldan pa ma rig pa yin no ies bya ba'i mtshuns par ldan pa'i don 'thad pa ma yin te I bdag fiid bdag fiid dan mtshuns par ldan pa ma yin pa ... To be compared with PYA 146,14-16 and PVV 85,15-17. 85 PVP D93al/P107a8'-b1: phyogs dan phyogs can gyi dnos po, but PVP D93a2-3/p107b2: phyogs gcig dan phyogs gcig pa can gyi dnos po. PVT D137a5/P169a6: phyogs yod pas na phyogs can ni tshogs pa ste I yan lag dan tshogs pa'i dnos po ies bya ba'i don to II. Prajiiakaragupta and Manorathanandin reflect Devendrabuddhi's interpretation in PYA 146,14 (moh[aik]ade1abhiitii [satkiiyadr~t*]) and PVV 85,16-17 ([satkiiyadr~tis tu] tadekade1aJ:t). 86 PYA 146,20; see also PVV 85,20. 87 paliiSayuktaY(! vanam iti in PVP D93a1-2/P107b1, PYA 146,20, PVV 85,20, Vibh. 85n. 10. piir:zyiidiyuktaY(! sarfram iti in PVP D93a2/P107b1, Vibh. 85n. 10. The second example obviously does not fit together with the explanation of "associated" as siimiinyavisqabhiiva, and hence has been dropped in PYA and PVV. 88 PVP D93a3/P107b2-3: de biin du 'dir yan ma rig pa ni phyogs gcig pa can yin la I de'i phyogs gcig pa'i no bo ni 'jig tshogs su Ita ba yin no II. PVT D137a5-7/P169a6-b1: de la spyir log pa 'jug pa tsam fiid yin pa'i phyir fion mons pa thams cad kyi bdag fiid can gyi ma rig pa ni tshogs pa'o II 'di'i

Ignorance, epistemology and soteriology (1)

67

Prajiiakaragupta concludes, "by stating [that the personalistic false view is] a specific instance (vi.e~abhidhanena), ignorance consisting in this [i.e.; a false view,] is pointed out as being primarily the cause [of defilements]."89 Or, according to Manorathanandin, "[Dharmakirti's] aim [in defining ignorance as the personalistic false view] is to show that ignorance consisting in a false view is primarily the cause of defilements."90 1.3.3. According to Dharmakirti as well as Vasubandhu, ignorance is an anti- or counter-knowledge displaying and superimposing erroneous aspects on reality. Though this understanding of ignorance is perfectly consonant with the Buddhist traditions' insistence on conceptual construction and language providing a biased account of reality, it does not explain why true aspects such as selflessness, though perceived, never become the objects of determinate cognitions. Moreover, this account of ignorance as superimposition does not satisfactorily explain in which sense traditional accounts hold ignorance to be the cause of defilements, clinging, action and bondage. Dharmakirti's specification of ignorance as the personalistic false view aims at answering these two questions. This false view is made responsible for an ordinary person's superimposition of erroneous aspects such as self and one's own (atmfya, i.e., what belongs to the self), which are the causes of the defilements and actions leading to painful existence. In this perspective, "personalistic false view" might well be just an arbitrary designation referring to that part of ignorance which, insofar as it superimposes such aspects, is primarily the cause of subsequent defilements.

phyogs gcig ni 'jig tshogs su lta ba la sogs pa'o II de bas na nag gi don ni 'jig tshogs su lta ba dan mtshuns par ldan pa'i ma rig pa 'jig tshogs su lta ba ies bya ba phyogs gcig pa ni phyogs gcig pa dan ldan par gyur pa yin pa de bas na I phyogs gcig pa can gyi ma rig pa de'i 'jig tshogs su lta ba ni phyogs gcig pa yin no ies bya ba 'di yin no II. Note PYA 146,21: paramiirthata!:t paliisasvabhiivataiva kathitii ... 89 PYA 146,21-22: evaY(! vise~iibhidhiinena tatsvabhiiv[ii] 'vidyii nidiinabhatii priidhiinyena nirdi~tii /. 90 PVV 85,20-21: dr~tisvabhiivii 'vidyii priidhiinyena kleiahetur ity upadarsanaY(! ... prayojanam I.

68

Vincent Eltschinger

1.3.4. Various terms refer to this kind of ignorance in Dharmaklrti's writings: "personalistic false view" (satkiiyadr~ti, darsana), "false view of a self" (iitmadarsana), "belief in/adhesion to a self" (iitmagraha, iitmiibhiniveSa), and "false view of/belief in a living being" (sattvadr~ti, sattvadarsana, sattvagraha).91 According to Yoga-

cara and Sautrantika definitions, the personalistic belief consists in one's considering the five constituents to which one clings (upiidiinaskandha) either as a self (iitmataJ.z) or as one's own (iitmzyataJ.z).92

91 For the various designations of ignorance in Dharmaklrti's works (especially in PV 2), see Vetter 1990: 23. satkayadr~ti in PV 1.223c, PVSV 111,24, PV 2.199; satkayadarsana in PV 1.222b, PVSV 111,19, PVSV 112,3-4, PV 2.212; atmadarsana in PVSV 111,18, PV 2.140; atmagraha in PV 2.135, PV 2.211, PV 2.235, PV 2.256; atm(iitmfy)abhiniveia in PVSV 8,20, PV 2.218; sattvadr~ti in PV 2.256; sattvadarsana in PV 2.200; sattvagraha in PV 2.196. 92 That satkaya should be interpreted as the five upadanaskandhas is obvious from the following definitions of this false view. (1) Maulf Bhamil:z of the YBh (as edited in Ahn 2003: 62; for a German translation, see Ahn 2003: 169-172): tatra satkiiyadr~til:z katama / asatpuru~asal'{!seviim iigamya asaddharmasravalJam ayonisomanaskiiral'{! nail:zsargikal'{! vii punal:z smrtisampramo~al'{! paficopiidiinaskandhiin iitmato viitmfyato samanupasyato yii nirdhiiritii viinirdhiiritii va kli~tii prajfiii /. (2) AS 7,8 (as quoted in Ahn 2003: 170n. 27; for a German translation, see Ahn ibid.): satkiiyadr~til:z katamii / paficopiidiinaskandhiin iitmata iitmfyato va samanupasyato yii k~iintf rucir matil:z prek~ii dr~til:z /. (3) Vasubandhu's treatment of the satkiiyadr~!i in the PSk echoes these Yogliclira definitions (PSk [4.1.26.1] 9,12-13; for a French translation, see Dantinne 1980: 15): satkiiyadr~!il:z katama / paficopiidiinaskandhiin iitmata iitmfyato vii samanupasyato yii kli~ta prajfia /. (4) Note also Sthiramati's comments on Tr 6cd (TrBh L23,12-14/B*14,14-16; on these four kleias of the [kli~tal'{!] manas, see Siddhi 1.255-256): upadiinaskandhe~v iitmeti darsanam iitmadr~til:z satkiiyadr~tir ity arthal:z / moho 'jiiiinam / iitmany ajfiiinam iitmamohal:z / iitmavi~aye mana iitmamiino 'smimiina ity arthal:z / iitmani sneha iitmasneha iitmapremety arthal:z /. (5) Slikyabuddhi defines satkiiya (in the expression 'jig tshogs la chags pa = *satkiiyasneha PVP D90b5/P104b7) as follows: zig pas na 'jig pa ste / rnam par 'jig pa ies bya ba'i don to // tshogs pa ni fie bar len pa'i phwi po lna 'dus pa ste / mi rtag pa'i ph un po lna la chags pa ies bya ba'i don to //. This matches the Sautrlintika definition provided by Vasubandhu in AKBh 281,20-21 on AK 5.7: iitmadr~tir iitmfyadr~tir va satkiiyadr~til:z / sfdatfti sat / cayal:z kiiyal:z sal'{!ghiital:z skandha ity arthal:z / sac ciiyal'{! kiiyas ceti satkiiyal:z paficopadiinaskandhiil:z / nityasal'{!jfial'{! pilJt/asal'{!jfiiil'{! ca tyiijayitum eval'{! dyotitii[l:z] / ... satkiiye dr~ril:z

Ignorance, epistemology and soteriology (1)

69

People who are deluded by this false view hold a basically transient (sat < sfdati) collection or cluster to be both permanent nityasar(Ljiiii) and unitary pilJrjasar(Ljiiii). Note should be made that this view does not have this perishing multitude (satkiiya) as its content, but bears on it as an object in an erroneous way. Texts like the YBh and the AKBh view the personalistic false view as twofold. In its speculative (vikaZpita, parikaZpita) form, it characterizes the non-Buddhists' (Sarikhya, Vaise~ika) doctrines of the self and is to be eliminated by the path of vision (darsanaheya). As for the innate (sahaja) personalistic belief, it afflicts all living beings including animals, and can only be eliminated by the path of cultivation (bhiivaniiheya).93 In other words, only Buddhas and SraI ... atmatmfyadr~tir eva tu satkayadr~tir ukta /. Sahaja satkayadr~ti/:t in PV 2.199 (see below) and PV 2.200 (sukhf bhaveyal'(! du/:tkhzva ma bhuvam iti tr~yataJ:t I yaivaham iti dhf/:t saiva sahajal'(! sattvadarsanam II). Viniscayasaligrahal}f of the Savitarkadibhumi (P Zi 112b6-113a1, as quoted [together with the Chinese of T. 1579: 621b6-1O] in Kritzer 2005: 293): de la 'jig tshogs la lta ba gan ze na I fie bar len pa'i phun po dag la bdag gam bdag gir ba Ita ba dan I mnon par zen pa dan sems la 'jog pa gan yin pa de ni 'jig tshogs la Ita ba zes bya'o II de 'an rnam pa gfiis su rig par bya ste I lhan cig skyes pa dan kun brtags pa'o II de la lhan cig skyes pa ni byis pa so so'i skye bo thams cad dan tha na ri dags dan bya rnams kyi yan yin no II kun brtags pa ni gzan mu stegs can rnams kyi yin par blta bar bya'o If. The text continues with the distinction between avyakrta and akusala. PrSVy IIIv5 (Muroji 1991: 83): sahajatmadr~ti; PrSVy IIIv6-7 (Muroji 1990: 81): parikalpitatmadr~ti. AKBh 290,19-20: sahaja satkayadr~tir avyakrta I ya mrgapak~i1Jam api vartate I vikalpita tv akusaleti purvacarya/:t f. According to Kritzer (2005: 292), Sailghabhadra identifies this as the opinion of the STItra-master. AKVy 463,17-18: ya atmavadibhiJ:t kapilolukadibhir vikalpita I. See also LAS 117,17-118,13 and ASBh 62,3-4. PV 2.199adJ: satkayadr~ter vigamad adya evabhavo bhavet I marge cet sahajahaner na ... "[Objection:] Non-[re]existence (abhava) [i.e., liberation from saI'(!Sara,] should occur [already] on the initial path [i.e., during the path of vision], for the [false] view of self[, which is the cause of the connection to a new birth,] ceases [at that time. Answer: ] No, because [at that time] one does not rid oneself of the innate [false view of self]." PVP D85bl-2/P98a5-7: bdag tu Ita ba'i rnam pa gfiis te I kun brtags pa dan lhan cig skyes pa'o II kun tu brtags pa de ni de dag gis sin tu kun tu spyod pa'i chos ma yin pas na spans pas de ni skye ba'i kun nas 'chin ba'i rgyur mi 'gyur ro II 'jig tshogs su lta ba lhan cig skyes pa yan srid pa'i rgyu gan yin pa de ni de dag fiid kyis spans pa ma yin no II de ma spans pa'i rgyu'i phyir bdag tu chags pa ma log pa fiid yin
satkayadr~ti/:t
93

70

Vincent Eltschinger

vaka-Arhats/Pratyekabuddhas have rid themselves of it. 94 No matter how important the speculative belief in a self can be in a philosophical and "heresiological" perspective, it is only of marginal importance in the context of Dharmak"irti's religious thought: 95 All he has to say about the personalistic false view pertains to its innate or natural form. According to him, the innate false view of personality is to be equated with ignorance, is the root cause of all defilements and depravities, and is the ultimate cause of suffering. As we can expect, the personalistic false view supplies the first link in Dharmaklrti's account of dependent origination. To rid oneself of it is tantamount to freeing oneself from bondage, rebirth and
saY[Lsiira.

1.3.5. Let me turn briefly to the vexing question of Dharmaklrti's sources for this equation of ignorance with the personalistic false view. As we have seen above,96 Devendrabuddhi as well as several Indian commentators regard the DBhS as providing scriptural evidence for Dharmaklrti's position. But many other Siitra and siistra texts may have been influential. In his pioneering study of Dharmaklrti's religious thought, Tilmann Vetter has drawn the attention of scholars to an oft-quoted passage from the Siilistambasutra, where ignorance is equated with notions such as those of a self (iitmasaY[Ljfiii) , of a (substantial) living being (sattvasaY[Ljfiii) , of

pa de ltar na ymi srid pa yod pa yin no //. PVT D131b6-7/P162a7-bl (ad loe.): kun tu brtags pa ni bstan beos sems pa la sogs pa'i sgo nas byU/i ba'o // lhan cig skyes pa ni thog ma med pa'i bag ehags las byun ba'o // de dag Hid kyis spans pa ma yin no ies bya ba ni thog ma med pa'i bag ehags las byun ba de ni bsgom pa'i lam fiid kyis span bar bya ba Hid yin pa'i phyir ro //. 94 As Prajfiiikaragupta rightly points out, the (innate) satkiiyadr~[i is nothing but defiled ignorance/delusion (PVA 145,12-13, together with stanza 886: kli~to hi mohaf:z satkiiyadarsanam eva / arhatiiY(! tu yad ajfiiinaY(! na tat kli~tam ato na te / mohe 'py ayuktasantiinii hfnasatkiiyadarsaniif:z II). 95 Except, of course, for Dharmaklrti's long polemics against the Nyaya conception of iitman and (the path to) release in PV 2.220sq. 96 DEhS VI C 48,1-10, quoted in PVP D92a7-b3/P106b5-107a2. See above, n. 20 and 1.2.4. For a brief statement about the DEhS's doctrine of ignorance, see Kritzer 1999: 97.

Ignorance, epistemology and soteriology (1)

71

an "I" (aharrzkiirasarrzIftii) and of a "mine" (mamakiirasarrzjftii).97 However, considering Dharmaklrti's doctrinal background, materials scattered throughout Yogacara works (including Vasubandhu's) are likely to have had a more direct impact on his ideas, though, as Lambert Schmithausen remarks, "the Yogacaras, like the Sarvastivadins (. ..), usually distinguish clearly between avidyii/ moha/ajfiiina and satkiiyadr:fti/iitmadr:fti-"98 According to Schmithausen, the Paramiirthagiithiis "lay considerable stress on the false view of Self (...) or the notion of Ego (...) and would thus seem to have, to a certain extent, affinities to a tradition like that of the Siilistambasutra (. ..) to which the (innate) false view of Self is the fundamental Defilement and not distinct from avidyii (...)"99

97 See Vetter 1990: 24 and 112-113n. 2. As quotedinPrP 562,14-563,1, this passage runs as follows: tatriividyii katamii I yai~iim eva ~a"(l"(liirrz dhiitiiniim

aikyasarrzjfiii pi"(lasarrzjfiii nityasarrzjfiii dhruvasarrzjfiii siisvatasarrzjfiii sukhasarrzjfiii iitmasarrzjfiii sattvasarrzjfiii jfvapudgalamanujamiinavasarrzjfiii* 'harrzkiiramamakiirasarrzjfiii I evamiidi vividham ajfiiinam ity ayam ucyate 'vidyeti I. * Cf. Schoening 1995: II.714n. 2b: sattvajfvajantupo~apuru~apud galamanujamiinavasarrzjfiii. "In that [connection], what is ignorance? The

beliefs that these same six elements are a unit, a uniform mass, permanent, eternal, unchanging, pleasurable, the self, a being, a life-force, a creature, thriving, a human, an individual, Manu's progeny, Manu's descendant, 'I,' 'mine,' and various types of ignorance such as that are called ignorance." Translation Schoening 1995: 1.296; for Sanskrit quotations of this passage, see Schoening 1995: II.713-717; see also May 1959: 269 and 466; for KamalasI1a's Siilistambatfkii on it, see Schoening 1995: II.499,7-500,7 (Tibetan text) and 1995: 1.296-299 (English translation). 98 Schmithausen 1987: II.517. See also Kritzer 1999: 34. 99 Schmithausen 1987: 11.518, and more generally 1987: I1.517-519n. 1421. On the Paramiirthagiithiis, see also Wayman 1961: 163-185. In this connection, one may quote (parts of) Paramiirthagiithiis 29-33: sarvabfje vipiike hi
jiiyate iitmadarsanam ... kalpayanty antariitmiinarrz tarrz ca billii ajiinakii/:t I iitmadadanam iisritya tathii bhavyas ca dr~taya/:t II pi"(lagrahiit svabfjiic ca piirviibhyiisiit sahiiyata/:t I srava"(liid anukiiliic ca jiiyate iitmadarsanam II snehas tatpratyayas caiva adhyiitmam upajiiyate I anugrahiibhilii~iic ca bahi/:t sneho mamiiyitam II (Sanskrit text as edited in Schmithausen 1987: 1.228).

"The [Result-of-]Maturation containing all Seeds comes to be looked upon as a Self (...). Not only do ignorant fools imagine this [Result-of-Maturation] to be [their] inner Self, but many [other false] views too arise on the basis of [this] view of Self. The view of Self arises from apprehending solid

72

Vincent Eltschinger

While dealing with dependent origination, however, the Yogacara(! Sautrantika) intellectuals paid a great deal of attention to another Siltra text, the Pratftyasamutpiidasiitra of the Sarrtyuktiigama. lOO In my opinion, this Siltra and/or its scholastic posterity is likely to have provided Dharmaklrti an even stronger scriptural foundation than the Siilistambasiitra or the Paramiirthagiithiis. The doctrinal line originating in the Pratftyasamutpiidasiitra permeates works like the Savitarkiidibhiimi of the YBh,101 the AS,102 the AKBh 103 as well as the AVS(N),104 and was commented upon at length by Vasu-

things, from its own Seed, from previous habit, from [bad] companion(s), and from listening [to doctrines] which favour [the arising of the view of Self], Conditioned by this [view of Self], there arises, furthermore, attachment to oneself, and, because one is eager to benefit [oneself], attachment [to what is] outside [consisting in] claiming (or coveting) [it] as one's own." 100 T. 99 (Sutra 298). A Sanskrit manuscript containing 25 Sutras from the SaJ!lyuktanikliya (Sutras 283-303 and 343-346) has been discovered in Turfan (see Waldschmidt 1957), and edited by C. Tripa!hI (see Tripa!hI 1962; on Tripa!hI's edition, see de Jong 1974). The original title of Sutra 298 seems to be Pratftyasamutplidlidivibhaliganirdeia (Adisutra in Tripa!hI's edition), "mais, dans l'usage courant, on Ie citait sous Ie titre de Pratftyasamutplidasutra" (de Jong 1974: 145). In addition to the interest of major Buddhist intellectuals (most of them Yogacara andlor Sautrantika), this Sutra seems to have enjoyed great popularity at least from the 5th century: a BrahmI inscription (2nd half of the 5th century) on a pillar near Dunhuang contains a fragment of it (see Gokhale 1944), the Sutra's entire text is found on both of two bricks (6 th century) found in Nalanda (see Chakravarti 1931-1932) as well as on two golden plates (between 650 and 800) located in Jakarta (see de Casparis 1956: 52). The Sanskrit text of the Nalanda brick inscriptions has been edited by Chakravarti (1931-1932: 197-199) and translated into German by Frauwallner (1958: 39-43); the Tibetan version of the Sutra has been edited by de Jong (1974: 146-149); the Chinese text of Xuanzang's version (T. 124) has been translated into English by Bagchi (1931-1932: 201-204). 101 .See YBh 204,1-212,4. 102 See Kritzer 1999: 33 (together with nn. 63-64) and ASBh 32,11-12. 103 See AKBh 140,6 (vibhalige mahlinidlinaparyliye caivaf[! nirdeilit I) together with Kosa II.85, and (at least according to AKVy 289,1: iheti pratftyasamutplidasutre), AKBh 135,8 together with Kosa II.70. 104 See AVS 5,15-7,1 (purporting to quote the whole section of the Pratftyasamutplidasutra dedicated to ignorance), as well as AVSN 203,1-211,3 (with many Sanskrit fragments of Vasubandhu's PrSVy).

Ignorance, epistemology and soteriology (1)

73

bandhu in his PrSVy.l0S In its vibhmiga section, the SfLtra provides a list of twentyl6 kinds of ignorance (ajfiiina). Although the original wording and sequence of the items are uncertain,107 those of the six initial ones are reasonably clear: (1) ignorance regarding the past [life] (piirviinte 'jfiiinam); (2) ignorance regarding the future [life] (apariinte 'jfiiinam); (3) ignorance regarding [both] the past and future [lives] (piirviintiipariinte 'jfiiinam); (4) ignorance regarding the internal [conditioned factors] (adhyiitmam ajfiiinam); (5) ignorance 105 See above, n. 5. The section on avidyli covers PrSVy P6a4-13b5. In his commentary, Vasubandhu holds the YBh, from which he quotes long excerpts, to be authoritative (see, e.g., PrSVy 11b3sq, and 13a1). 106 Twenty according to PrSVy lOa3-4 (yan lag bcu gfiis). Lambert Schmithausen informs me that while YBh 205,10-11 reads virrt.fatylikliram ajfilinam, the manuscript consistently reads thrice eklinnavilflsatiO, a reading confirmed by Xuanzang's Chinese translation (T. 30.1579, 322c20, etc.). In the second (= B) Nalanda brick inscription, the avidylivibhanga reads as follows (Chakravarti 1931-1932: 198; the numbering of the items as well as the references to the YBh are mine): yat (1 - YBh 204,2-5) purvlinte lfilinam (2 - YBh 204,6-8) aparlinte lfilinalfl (3 - YBh 204,8-10) purvlintliparlinte lfi.linam (4 - YBh 204,10-11) adhylitmam ajfilinalfl (5 - YBh 204,11-13) bahirdhlilfilinam (6 - YBh 204,13-15) adhylitmabahirdhlilfilinalfl (7 YBh 204,15) karmmJY ajfilinalfl (8 - YBh 204,16-17) viplike lfilinalfl (9 YBh 204,17-18) karmaviplike lfilinalfl (10 - YBh 204,18-20) buddhe lfilinalfl (11 - YBh 204,20-21) dharme lfilinalfl (12 - YBh 204,22-23) sanghe lfilinalfl (13 - YBh 204,23-205,1) dul;khe lfilinalfl (14 - YBh 205,1) samudaye lfi.linalfl (15 - YBh 205,2) nirodhe lfilinalfl (16 - YBh 205,2) mlirge lfilinalfl (17 - YBh 205,3-4) hetliv ajfilinalfl (18 - YBh 205,4-5) hetusamutpanne~u dharme~v* ajfilinalfl (19 - YBh 205,6-8) kusallikusale~u slivadylinavadye~u sevitavylisevitavye~u hf1Japra1Jftakr~1Jasuklasapratibhligap ratftyasamutpanne~u dharme~v ajfilinalfl ([20?] - YBh 205,9-10) ~arsu vli

punal; sparsliyatane~u yathlibhutasamprativedhe lfilinam iti / yat tatra tatra yathlibhutasylijfilinam adar.anam anabhisamayas tamal; salflmoho 'vidylindhakliram iyam ucyate 'vidyli /. * YBh salflsklire~[v] for dharme~v. If read-

ing the Siltra as involving 19 rather than 20 items, the list proper would stop . after hetusamutpanne~u dharme~v ajfilinam, the remaining part (kusallikusale~u ... ) being then a general explanation. 107 Whereas the YBh's explanations closely follow the order and wording of the Siltra (at least as far as the Sanskrit text of the Nalanda brick inscription and the Tibetan text edited by de Jong are concerned), AVS 5,15-7,1 as well as AVSN 103,2-110,11 diverge as to the order, the wording and the number of the items.

74

Vincent Eltschinger

regarding the external [conditioned factors] (bahirdhii liiiinam); (6) ignorance regarding [both] the internal and the external [conditioned factors] (adhyiitmabahirdhii liiiinam). These six items have been unanimously interpreted as referring to various features of the false view of self and one's own. Leaving items (3) and (6) out of consideration, let us have a closer look, first, at the two initial ones. According to the YBh, ignorance regarding the past (life) consists in "the ignorance of the [one] who improperly conceives the past conditioned factors [by wondering]: 'Did I exist in the past or did I not exist? Who was I, and (vii) how did I exist?"'10B As for the ignorance regarding the future (life), it consists in "the ignorance of the [one] who improperly conceives the future conditioned factors [by wondering]: 'Shall I exist in the future or shall I not exist? What shall I be? How shall I exist?"'109 Still worthier of notice is this tradition's explanation of items (4) and (5), with its emphasis on self and one's own, improper reflection and personalistic belief: "What does ignorance about internal [factors] consist in? It is the [kind of] ignorance that [is particular] to the [person] who improperly considers his conditioned factors to be a self. What does ignorance about external [factors] consist in? It is the [kind of] ignorance that [is particular] to the [person] who improperly considers to be his
lOB YBh 204,2-5: purvante 'jfianarrt katamat I atuan sarrtskaran ayonisaJ:t kalpayataJ:t kirrt nv aham abhuvam atfte 'dhvany ahosvin naham abhavam atfte 'dhvani I ko nv aham abhavam I katharrt vabhuvam iti yad ajfianam I. To be compared with Vasubandhu's quotation in PrSVy llb3-4. See also PrSVy 9b2-3, which provides Vasubandhu's own commentary on parvante 'jfianam: gan 'di snon gyi mtha' mi ses pa ni ci bdag 'das pa'i dus na byun ba iig gam ies (P om. ies) bya ba la sags par the tshom za ba'i phyir ro II. AVSN 103,5-6: purvantam atftarrt janma I tatrajfianarrt kirrt nv aham abhavam atfte 'dhvanfty evamadi vicikitsataJ:t I. See Samtani 1971: 103n. 3. 109 YBh 204,6-8: aparante 'jfianarrt katamat I anagatiin sarrtskiiriin ayoniso kalpayataJ:t kirrt nv aharrt bhavi:tyamy anagate 'dhvany ahosvin niiharrt bhavi:tyamy anagate 'dhvani / kirrt bhavi:tyami katharrt bhavi:tyamfty ajfianam /. To be compared with Vasubandhu's quotation in PrSVy llb4-6. See also PrSVy 9b2-3, which provides Vasubandhu's own commentary on aparante 'jfianam: phyi ma'i mtha' mi ses pa ni ci bdag ma 'Olis pa'i dus na 'byun bar 'gyur ram ies bya ba la sags par the tshom za ba'i phyir TO II. AVSN 103,6-7: aparantam anagatarrt janma I tatriijfiiinarrt kirrt nv aharrt bhavi:tyamy aniigate 'dhvanfty evamadi vicikitsataJ:t I.

Ignorance, epistemology and soteriology (1)

75

own the external conditioned factors which are not included among [entities constituting] living beings."l1o Vasubandhu explains these forms of ignorance as consisting in one's giving rise to the false view of a (substantial) living being (sattvadr~ti) regarding one's own as well as others' streams (svaO and parasantana).111

110 YBh204,1O-13 (Schmithausen 1987: II.518): adhyatmamajiiana1?1katamat I pratyatmikan sa1?1skaran ayonisa atmato manasikurvato yad ajiianam I bahirdhajiiana1?1 katamat I bahyan asattvasa1?1khytitan sa1?1skaran atmfyato 'yoniso manasikurvato yad ajiianam I. To be compared with Vasubandhu's quotations in PrSVy 11b6-7 (with sems can gyi grans su ma gtogs pa providing evidence for Schmithausen's asattvasankhyatan). See also Kritzer 1999: 96-97. Schmithausen (ibid.) quotes passages from the YBh testifying to a similar interpretation, as, e.g., BoBh DI67,5/W243,lOsq (.. .atmabhave 'ham iti va mameti va sa1?1mucjhaf:t .. .), and YBh 212,18sq (purvam titmabhave atmagrahadibhif:t sa1?1mucjhaf:t). DharmakYrti's claim that the first result of the notion of a self is the notion of the other, i.e., the external, provides an interesting echo to this passage. According to Kritzer (1999: 96), this interpretation provides the basis for a passage of the AKBh "in which Vasubandhu gives his own, as opposed to the orthodox Vaibha~ika, explanation of pratftyasamutpada" (AKBh 139,25-26: balo hi pratftyasamutpanna1?1 sa1?1skaramatram idam ity aprajanan ... "[A]n ignorant person, not understanding that what originates conditionally is merely the conditioning factors (...)" Translation Kritzer 1999: 96). According to Yasomitra, the word "merely" has here the meaning of a negation of self, whereas "not understanding" refers to the aver,tiky avidya* (AKVy 299,8-9: balaf:t prthagjanaf:t sa1?1skaramatram iti I matragrahar,tam atmapratiifedhartham I aprajanann ity aver,tikfm avidya1?1 dadayati I. See also Kosa 3.84). In Vasubandhu's own account of ignorance as the first link of dependent origination, the said ignorant persons thus give rise to the (false) view of a self (atmadrifti) and then engage in action in order to benefit this self (AKBh 139,27: atmanaf:t sukhartham ... karma arabhate), which matches perfectly DharmakYrti's view of dependent origination. * On the avenikf ignorance, see Siddhi 1.276-280 (note Siddhi 1.276: "L[e] PratYtyasamutpadasutra enseigne: 'L'Avidya dite aver,tikf, subtile, toujours en activite, offusque la realite, empeche qU'elle soit connue.'''), and below, n. 111. 111 PrSVy 9b5: nan mi ses pa ni ran gi rgyud la sems can du lfa bar bsams pa'o II. (Note AVSN 104,4-5: adhytitme 'jiianam iti I svasantane sattvadriftimato yad ajiianam I satkayadriftisamprayukta veyam avidyokta I.) PrSVy 9b5: phyi mi ses pa ni gi.an gyi rgyud la sems can du lta bar bsams pa'o II.

76

Vincent Eltschinger

As we can see, Dharmaklrti was by no means the first Buddhist intellectual to have connected avidyii with iitmadr~ti and the like,112 nor to have defined the former by means of the latter, but he seems to have been the first to have developed exegetical strategies to justify an equation that could easily be taken, at least among the Abhidharmikas, as unorthodox.

(To be continued)

Abbreviations and literature

Ahn 2003 - Sung-Doo Ahn: Die Lehre von den kleias in der Yogacarabhumi. Stuttgart 2003: Franz Steiner Verlag (Alt- und Neu-Indische Studien, 55). AK(Bh) - Prahlad Pradhan: Abhidharmakosabhil~yam of Vasubandhu. Patna 1975: K.P. Jayaswal Research Institute (Tibetan Sanskrit Works Series, 8). AKVy - Unrai Wogihara: Sphutarthil Abhidharmakosavyakhya, the Work of Yasomitra. Tokyo 1989 (19361): Sankibo Buddhist Book Store (The Publishing Association of Abhidharmakosavyakhya). AS - Prahlad Pradhan: Abhidharmasamuccaya of Asariga. Santiniketan 1950: Visva-Bharati (Visva-Bharati Studies, 12). ASBh - Nathmal Tatia: Abhidharmasamuccayabhil~yam. Patna 1976: K.P. Jayaswal Research Institute (Tibetan Sanskrit Works Series, 17). AVS(N) - N.H. Samtani: The Arthaviniscayasutra and Its Commentary (Nibandhana) (Written by Bhik~u Vfryasrfdatta of Srf-Nalandavihilra).

112 It may not be out of place to briefly mention the Sautrantika (AKVy 302,33-34: mayata punaJ:z sautrantikair avidya prakarabhinna vanJyate !) view held by Bhadanta Dharmatrata, according to whom ignorance is to be defined as an inclination for/engagement in (mayata? see Kosa 3.93n. 1) [the belief in] a living being in the form of "I am" (AKBh 142,9: asmfti sattvamayata 'vidyeti bhadantadharmatrataJ:z I). To be sure, Vasubandhu criticises Dharmatriita's view on the grounds that this mayata (mayana according to de La Vallee Poussin) cannot be different from the asmimana, and hence cannot define ignorance. That Vasubandhu was nevertheless in trouble with this definition is testified by his premature interruption of the discussion (AKBh 142,13-14: atra punar vicaryamalJe bahu vaktavyaYfljayate / tasmat ti~thatv etat I).

Ignorance, epistemology and soteriology (1)

77

Patna 1971: K.P. Jayaswal Research Institute (Tibetan Sanskrit Works Series, 13). Bagchi 1931-1932 - P.C. Bagchi: "A Note on the Pratltya Samutpada SUtra." Epigraphia Indica 21 (1931-1932), pp. 199-204. Bareau 1957 - Andre Bareau: "Les controverses relatives it la nature de l'Arhant dans Ie bouddhisme ancien." Indo-Iranian Journal 1 (1957), pp. 241-250. BHSD - Franklin Edgerton: Buddhist Hybrid Sanskrit Grammar and Dictionary. Vol. II: Dictionary. DeIhi 1970: Motilal Banarsidass (New Haven 19531: Yale University Press). van Bijlert 1989 - Vittorio A. van Bijlert: Epistemology and Spiritual Authority. The Development of Epistemology and Logic in the Old Nyiiya and the Buddhist School of Epistemology with an Annotated Translation of Dharmakfrti's Pramiiftaviirttika II (Pramii1}asiddhi) vv. 1-Z Vienna 1989: Arbeitskreis fiir tibetische und buddhistische Studien Universitat Wien (Wiener Studien zur Tibetologie und Buddhismuskunde, 20). BoBh - D = Nalinaksha Dutt: Bodhisattvabhumi/:t [Being the XVth Section of Asmigapiida's Yogiiciirabhumi/:t]. Patna 1978: K.P. Jayaswal Research Institute (Tibetan Sanskrit Works Series, 7). W = Unrai Wogihara: Bodhisattvabhrimi. A Statement of Whole Course of the Bodhisattva (Being Fifteenth Section of Yogiiciirabhumi). Tokyo 1971: Sankibo Buddhist Book Store. de Casparis 1956 - IG. de Casparis: Selected Inscriptions from the 7'h to the 9th Century A.D. Vol. II. Bandung 1956. Chakravarti 1931-1932 - N.P. Chakravarti: "Two Brick Inscriptions from Nalanda." Epigraphia Indica 21 (1931-1932), pp. 193-199. Chu 2004 - Junjie Chu: "A Study of Sataimira in Dignaga's Definition of Pseudo-Perception (PS 1.7cd-8ab)." Wiener Zeitschrift fiir die Kunde Siidasiens 48 (2004), pp. 113-149. CPSS - Candrapradfpasamiidhisatra = P. L. Vaidya: Samiidhiriijasutra. Darbhanga 1961: The Mithila Institute of Post-Graduate Studies and Research in Sanskrit Learning (Buddhist Sanskrit Texts, 2). D - Jikido TakasakifZuiho YamaguchifNoriaki Hakamaya: sDe dge Tibetan Tripitaka bsTan 'gyur preserved at the Faculty of Letters, University of Tokyo. Tokyo 1977-1981. Dantinne 1980 - Jean Dantinne: Le traite des cinq agregats (Paficaskandhaprakarana de Vasubandhu). Bruxelles 1980: Institut Beige des Hautes Etudes Bouddhiques (Serie "Etudes et textes," 7). DBhS - Johannes Rahder: Dasabhumikasutra et Bodhisattvabhumi, Chapitres Vihiira et Bhumi. ParislLouvain 1926: Paul Geuthnerfl-B. Istas, Imprimeur (Societe Beige d'Etudes Orientales).

78

Vincent Eltschinger

Dunne 2004 - John D. Dunne: Foundations of Dharmakfrti's Philosophy. Boston 2004: Wisdom Publications (Studies in Indian and Tibetan Buddhism). Eltschinger 2005 - Vincent Eltschinger: "Etudes sur la philosophie religieuse de DharmakIrti: 2. L'tiSrayaparivrtti." Journal Asiatique 293-1 (2005), pp. 151-211. . Eltschinger 2007 - Vincent Eltschinger: Penser l'autoriti des Ecritures. La polimique de Dharmakfrti contre la notion brahmanique orthodoxe d'un Veda sans auteur. Autour de PramiiIJaviirttika 1.213-268 et Svavrtti. Vienna 2007: Verlag der Osterreichischen Akademie der Wissenschaften (Beitrage zur Kultur- und Geistesgeschichte Asiens, 56). Franco 1986 - Eli Franco: "Once again on DharmakIrti's Deviation from Dignaga on pratyak~iibhiisa." Journal of Indian Philosophy 16 (1986), pp.79-97. Franco 2001 - Eli Franco: "DharmakIrti's Reductionism in Religion and Logic." Pp. 285-308 in Raffaele Torella (ed.): Le parole e i marmi. Studi in onore di Raniero Gnolinel suo 70 compleanno. Vol. 1. Roma 2001: Istituto Italiano per l'Africa e l'Oriente (Serie Orientale Roma 92/1). Frauwallner 1932 - Erich Frauwallner: "Beitrage zur Apohalehre: 1. Dharmaklrti. Ubersetzung." Wiener Zeitschrift filr die Kunde des Morgenlandes 39 (1932), pp. 247-285. Frauwallner 1933a - Erich Frauwallner: "Beitrage zur Apohalehre: I. Dharmaklrti. Ubersetzung. (Fortsetzung)." Wiener Zeitschrift filr die Kunde des Morgenlandes 40 (1933), pp. 51-94. Frauwallner 1933b - Erich Frauwallner: "Beitrage zur Apohalehre: 1. Dharmaklrti. Ubersetzung. (Zusammenfassung)." Wiener ZeitschriJt filr die Kunde des Morgenlandes 42 (1933), pp. 93-102. Frauwallner 1958 - Erich Frauwallner: Die Philosophie des Buddhismus. Berlin 1958: Akademie-Verlag. Frauwallner 1959 - Erich Frauwallner: "Dignaga, sein Werk und seine Entwicklung." Wiener Zeitschrift filr die Kunde Sild- und Ostasiens 3 (1959), pp. 83-164. Funayama 1999 - Toru Funayama: "KamalaSIla's Interpretation of 'NonErroneous' in the Definition of Direct Perception and Related Problems." Pp. 73-99 in Shoryii Katsura (ed.): Dharmakfrti's Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of the Third International Dharmakfrti Conference (Hiroshima, November 4-6, 1997). Vienna 1999: Verlag der Osterreichischen Akademie der Wissenschaften (Philologisch-historische Klasse, Denkschriften, 281). Gokhale 1944 - v.v. Gokhale: "A BrahmI Stone Inscription from Tunhuang." Sino-Indian Studies 1-1 (1944), pp. 18-22. Hattori 1968 - Masaaki Hattori: Digniiga, On Perception, being the Praty0

Ignorance, epistemology and soteriology (1)

ak~apariccheda

79

of Digni'iga's Pramar;msamuccaya. Cambridge, Mass. 1968: Harvard University Press (Harvard Oriental Series, 47). HVP - Hastavi'ilaprakarar;a. The Tibetan version(s) have been edited by. E. Frauwallner. See Frauwallner 1959: 152-156. Jaini 2001- "On the Ignorance of the Arhat." Pp. 167-179 (= Chapter 9) in Padmanabh S. Jaini: Collected Papers on Buddhist Studies. Delhi 2001: Motilal Banarsidass (Buswell/Gimello [eds.]: Buddhist Soteriology. University of Hawaii Press: 1992\ pp. 135-146). de Jong 1974 - Jan Willem de Jong: "A propos du Nidi'inasaY(lyukta." Pp. 137-149 in: Melanges de Sinologie offerts a Monsieur Paul Demieville. VoL II. Paris 1974. (Gregory Schopen [ed.]: Buddhist Studies by JW: de Jong. Berkeley 1979R , Asian Humanities Press, pp. 237-249.) Katsura 1976 - Sh5ryii Katsura: "On Abhidharmakosa VIA." Indological Review 2 (1976), p. 28. Katsura 1991 - Sh5ryii Katsura: "Dignaga and Dharmakirti on apoha." Pp. 129-146 in: Ernst Steinkellner (ed.): Studies in the Buddhist Epistemological Tradition. Proceedings of the Second International Dharmakfrti Conference (Vienna, June 11-16, 1989). Vienna 1991: Verlag der Osterreichischen Akademie der Wissenschaften (Beitrage zur Kultur- und Geistesgeschichte Asiens, 8). Katsura 1993 - Sh5ryii Katsura: "On Perceptual Judgement." Pp. 66-75 in: N.K. Wagle/F. Watanabe (eds.): Studies on Buddhism in Honour of Professor A.K. Warder. Toronto 1993: University of Toronto, Center for South Asian Studies (South Asian Studies Papers, 5). Kellner 2004 - Birgit Kellner: "Why Infer and not just Look? Dharmaklrti on the Psychology ofInferential Processes." Pp. 1-51 in: Sh5ryii Katsura/ Ernst Steinkellner (eds.): The Role of the Example (dr~tanta) in Classical Indian Logic. Vienna 2004: Arbeitskreis fur tibetische und buddhistische Studien Universitat Wien (Wiener Studien zur Tibetologie und Buddhismuskunde, 58). Kosa - Louis de La Vallee Poussin: L'Abhidharmakosa de Vasubandhu. 6 vols. Bruxelles 1980: Institut BeIge des Hautes Etudes Chinoises (Melanges Chinois et Bouddhiques, 16). Kritzer 1999 - Robert Kritzer: Rebirth and Causation in the Yogi'ici'ira Abhidharma. Vienna 1999: Arbeitskreis fur tibetische und buddhistische Studien Universitat Wien (Wiener Studien zur Tibetologie und Buddhismuskunde, 44). Kritzer 2005 - Robert Kritzer: Vasubandhu and the Yogi'ici'irabhumi. Yogi'ici'ira Elements in the Abhidharmakosabhi'i~ya. Tokyo 2005: The International Institute for Buddhist Studies (Studia Philologica Buddhica, Monograph Series, 18). La Vallee Poussin 1913 - Louis de La Vallee Poussin: Bouddhisme, etudes

80

Vincent Eltschinger

et materiaux. Theorie des douze causes. Gand 1913: Librairie scientifique E. van Goethem (Universite de Gand, Recueil de travaux publies par la Faculte de philo sophie et des lettres, 40). Lamotte 1973 -Etienne Lamotte: La Somme du Grand vehicule d'Asariga (Mahiiyiinasarigraha). 2 vols. Louvain-Ia-Neuve 1973: Universite de Louvain, Institut Orientaliste (Publications de l'IOL, 8). LAS - Bunyiu Nanjio: The Larikiivatiira Satra. Kyoto 1923: The Otani University Press (Bibliotheca Otaniensis, 1). Matilal 1980 - Bimal Krishna Matilal: "Ignorance or Misconception? - A Note on Avidyii in Buddhism." Pp. 154-164 in: Somaratna Balasooriya et al. (eds): Buddhist Studies in honour ofWalpola Rahula. London, etc. 1980: Gordon Fraser. May 1959 - Jacques May: Candrakfrti: Prasannapadii Madhyamakavrtti. Douze chapitres traduits du sanscrit et du tiberain, accompagnes d'une introduction, de notes et d'une edition critique de la version tibetaine. Paris 1959: Adrien Maisonneuve (Collection Jean Przyluski, 2). Mejor 1997 - MarekMejor: "On Vasubandhu's Pratftyasamutpiidavyiikhyii." Pp. 151-161 in: Agata Bareja-Starzynska/Marek Mejor (eds.): Aspects of Buddhism. Proceedings of the International Seminar on Buddhist Studies (Liw, 25 June 1994). Warsaw 1997: Oriental Institute, Warsaw University (Studia Indologiczne, 4). Mejor 2001 - Marek Mejor: "Controversy on the mutual conditioning of avidyii and ayonisomanas(i)kiira in Vasubandhu's Abhidharmakosa." Journal of the International College for Advanced Buddhist Studies 4 (2001), pp. 49[/292]-78[/263]. Mejor 2002 - Marek Mejor: "On the sevenfold classification of the negative particle (nail,) (Grammatical explanation of a-vidyii in Vasubandhu's Pratftyasamutpiida-vyiikhyii)." Pp. 87-100 in: Early Buddhism andAbhidharma Thought: In Honor of Doctor Hajime Sakurabe on His Seventyseventh Birthday. Kyoto 2002: Heirakuji Shoten. MS - See Lamotte 1973: 1. Muroji 1991 - Yoshihito G. Muroji: "Vedana- und Tr~l).iivibhaiJ.ga in der Pratftyasamutpiidavyiikhyii." Mikkyo Bunka 173 (1991), pp. 74-98. Muroji 1993 - Yoshihito G. Muroji: Vasubandhus Interpretation des Pratff1asamutpiida. Eine kritische Bearbeitung der Pratltyasamutpiidavyiikhyii (Sarp.skiira- und VijfiiinavibhaiJ.ga). Stuttgart 1993: Franz Steiner Verlag (Alt- und Neu-Indische Studien, 43). NM - Nyiiyamukha (Dignaga). Referred to on the basis of Tucci 1930 and Chu 2004. P - Daisetz T. Suzuki: The Tibetan Tripitaka, Peking Edition, Kept in the Library of the Otani University, Kyoto. Tokyo/Kyoto 1957: Tibetan Tripitaka Research Institute.

Ignorance, epistemology and soteriology (1)

81

Pa - Pfu:1ini (A~tiidhyiiyf). Paramiirthagiithii - The 44 stanzas, together with their commentary, have been edited and translated by A. Wayman (see Wayman 1961: 163-185); stanzas 28-41 and their commentary have been edited and translated by L. Schmithausen (see Schmithausen 1987: I.223-241). PrP - Louis de La Vallee Poussin: Madhyamakavrtti/:t: Millamadhyamakakiirikiis (Miidhyamikasiltras) de Niigiirjuna avec la Prasannapadii Commentaire de Candrakfrti. Delhi 1992: Motilal Banarsidass (SaintPetersburg 1903-19131, Bibliotheca Buddhica, 4). PrSVy - Pratftyasamutpiidavyiikhyii (Vasubandhu). D no. 3995, Chi 1b6la, P no. 5496, Chi 1-71a. Pruden 1988-1990 - Leo M. Pruden: Abhidharmakosabhii~yam by Louis de La Vallee Poussin. English Translation by Leo M. Pruden. 4 vols. Berkeley 1988-1990: Asian Humanities Press. PS(V) - Ernst Steinkellner: Digniiga's PramiiIJasamuccaya, Chapter 1. A hypothetical reconstruction of the Sanskrit text with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments and the linguistic materials gained from linendrabuddhi's Tlkii. Vienna 2005: http://www.oeaw.ac.at/ias/Mat/dignaga_PS_l.pdf (last accessed 06/08/2010). PSk - Li Xuezhu/Ernst Steinkellner: Vasubandhu's Paiicaskandhaka. BeijingNienna 2008: China Tibetology Publishing House/Austrian Academy of Sciences Press (Sanskrit Texts from the Tibetan Autonomous Region, 4). PST - Ernst Steinkellner/Helmut Krasser/Horst Lasic: linendrabuddhi's Visiiliimalavatf PramiiIJasamuccayatfkii, Chapter 1. Part I: Critical Edition. BeijingNienna 2005: China Tibetology Publishing House/ Austrian Academy of Sciences Press (Sanskrit Texts from the Tibetan Autonomous Region, 1/1). . PV 1-4 - Yusho Miyasaka: "PramaI.1avarttika-karika (Sanskrit and Tibetan)." Acta 1ndologica 2 (1971-1972), pp. 1-206. See also PVV; for PV 2-3, see also PYA; for PV 1, see also PVSV; for PV 2.131cd-285, see also Vetter 1990. My numbering of the verses in PV 2 follows that of Vetter. PYA - Rahula Sankrtyayana: PramiiIJaviirttikabhii~yam or ViirtikiilaJikiiraJ:! of Prajiiiikaragupta (Being a Commentary on Dharmakfrti's PramiiIJaviirtikam). Patna 1953: K.P. Jayaswal Research Institute. PVinT - PramiiIJaviniscayatfkii (Dharrnottara). D no. 4229, Dze Ib1-Tshe 178a3/P no. 5727, Dze 1bI-We 209b8. PVP - PramiiIJaviirttikapaiijikii (Devendrabuddhi). Dna. 4217, Che 1-326b4/P no. 5717, Che 1-390a8. PVSV - Raniero Gnoli: The PramiiIJaviirttikam of Dharmakfrti. The First

82

Vincent Eltschinger

Chapter with the Auto-Commentary. Roma 1960: Istituto Italiano per il Medio ed Estremo Oriente (Serie Orientale Roma, 23). PVSVT - Rahula SaIilqtyayana: Kan:rakagomin's Commentary on the PramiiTJavilrttikavrtti of Dharmakfrti. Kyoto 1982: Rinsen Books Co. (Allahabad 19431: Kitab Mahal). PVT - PramiiTJaviirttikatikii (Sakyabuddhi). D no. 4220, Je 1b1-Ne 282a7/P no. 5718, Je 1b1-Ne 348a8. PVV - Rahula Sarikrtyayana: "Dharmaklrti's Prama1}avarttika with Commentary by Manorathanandin." Published as an appendix to the Journal of the Bihar and Orissa Research Society 24-26 (1938-1940). Rahder 1932 - Johannes Rahder: "La Satkayadr~!i d'apres\ Vibha~a, 8." Melanges Chinois et Bouddhiques 1 (1931-1932), pp. 227-239. Samtani 1971- See AVS(N). Schmithausen 1987 - Lambert Schmithausen: Alayavijfiiina. On the Origin and the Early Development of a Central Concept ofYogiiciira Philosophy. 2 vols. Tokyo 1987: The International Institute for Buddhist Studies (Studia Philologica Buddhica, Monograph Series, IVa/b). Schoening 1995 - Jeffrey D. Schoening: The Siilistamba Satra and its Indian Commentaries. 2 vols. Vienna 1995: Arbeitskreis flir tibetische und buddhistische Studien Universitat Wien (Wiener Studien zur Tibetologie und Buddhismuskunde, 35/1-2). Siddhi - Louis de La Vallee Poussin: Vijiiaptimiitratiisiddhi. La Siddhi de Hiuan-Tsang. 2 vols. Paris 1929: Librairie Orientaliste Paul Geuthner (Buddhica, Documents et travaux pour l'etude du bouddhisme publies sous la direction de Jean Przyluski, Premiere serie: Memoires, 5). T - Junjiro Takakusu/Kaikyoku Watanabe: TaishO shinsha daizokyo. Tokyo 1924-1932: Taisho Issaikyo Kankokai. TrBh - TriYfLsikilvijiiaptibhii~ya (Sthiramati). L = Sylvain Levi: Vijiiaptimiitratiisiddhi, deux traites de Vasubandhu: Vimsatikil et Trimsikii. Paris 1925: Librairie Ancienne Honore Champion (Bibliotheque d~ l'Ecole des Hautes Etudes, Sciences historiques et philologiques, fasc. 245). B = Hartmut Buescher: Sthiramati's TriYfLsikiivijiiaptibhii~ya. Critical Editions of the Sanskrit Text and its Tibetan Translation. Wien 2007: Verlag der Osterreichischen Akademie der Wissenschaften (Beitrage zur Kultur- und Geistesgeschichte Asiens, 57). Tripa!hI 1962 - Chandrabhal Tripa!hI: Funfundzwanzig Satras des NidiinasaYfLyukta =Sanskrittexte aus den Turfanfunden VIII. Berlin 1962: Akademie-Verlag. TS(P) - K = Embar Krishnamacharya: Tattvasaligraha of Siintarak~ita With the Commentary of Kamidasfla. 2 vols. Baroda 1984: Oriental Institute.

Ignorance, epistemology and soteriology (1)

83

= Swami Dwarikadas Shastri: Tattvasarigraha ofAciirya Shiintarak~ita with the Commentary 'Paiijikii' of Shri Kamalshfla. 2 vols. Varanasi 1981: Bauddha Bharati (Bauddha Bharati Series, 1). Tucci 1930 - Giuseppe Tucci: The Nyiiyamukha of Digniiga. The Oldest Buddhist Text on Logic, After Chinese and Tibetan Materials. Heidelberg 1930: Otto Harrassowitz (Materialien zur Kunde des Buddhismus, 15). Tucci 1971 - Giuseppe Tucci: "A Fragment from the Pratnya-SamutpadaVyakhya of Vasubandhu." Pp. 239-248 in: Giuseppe Tucci: Opera Minora. Parte I. Roma 1971 (JRAS, Calcutta 19301): Dott. Giovanni Bardi, editore (Universita di Roma, Studi Orientali Pubblicati a Cura della Scuola Orientale, 6). Vetter 1964 - Tilmann Vetter: Erkenntnisprobleme bei Dharmakfrti. Wien 1964: Hermann Bi:ihlaus Nachf., Kommissionsverlag der Osterreichischen Akademie der Wissenschaften (Osterreichische Akademie der Wissenschaften, philologisch-historische Klasse, Sitzungsberichte, 245). Vetter 1990 - Tilmann Vetter: Der Buddha und seine Lehre in Dharmakfrtis PramiilJ-aviirttika. Der Abschnitt iiber den Buddha und die vier edlen Wahrheiten im PramiilJ-asiddhi-Kapitel. Wien 1990: Arbeitskreis flir tibetische und buddhistische Studien Universitat Wien (Wiener Studien zur Tibetologie und Buddhismuskunde, 12). Vibh. - Vibhilticandra's notes to PVv. See PVv. VinSg - ViniscayasaJigrahalJ-f of the YBh. P no. 5539, Zi Ibl - 'i 142b8. Waldschmidt 1957 - Ernst Waldschmidt: "Identifizierung einer Handschrift des Nidanasa:q:tyukta aus den Turfanfunden." Zeitschrift der Deutschen Morgenliindischen Gesellschaft 107/2 (1957), pp. 372-40l. Wayman 1957 - Alex Wayman: "The Meaning of Unwisdom (Avidya)." Philosophy East and West 7/1-2 (1957), pp. 21-25. Wayman 1961- Alex Wayman: Analysis of the Sriivakabhilmi Manuscript. Berkeley/Los Angeles 1961: University of California Press. Wayman 1980 - Alex Wayman: "Nescience and Insight According to Asariga's Yogiiciirabhilmi." Pp. 251-266 in: Somaratna Balasooriya et al. (eds.): Buddhist Studies in honour ofWalpola Rahula. London, etc. 1980: Gordon Fraser. . YBh - = Yogiiciirabhilmi, or, followed by page/line numbers: V. Bhattacharya: The Yogiiciirabhilmi ofAciiryaAsariga. Calcutta 1957: University of Calcutta.

Eating letters in the Tibetan treasure tradition'

Frances Garrett

Across hundreds of years of Tibetan literature on occult technologies, we find a small but consistent canon of writing on a practice known as "edible letters" (za yig). These texts describe the consumption of small rolls of paper inscribed with Tibetan graphemes. Written with ink prepared from blood, musk or other ingredients, and often stuffed with such materials as aconite or fingernails, these edible amulets are meant to serve a wide range of practical needs, from increasing one's merit or wisdom or winning arguments, to protecting against thieves, contagious disease, spirit possession or dog bite. As we will see below, edible letter "recipes" may also prescribe Buddhist visualization practices. The practice of edible letters is strangely unexplored in secondary scholarship. It appears not to have been reported upon in ;.ethnographic literature on Tibet or the Himalayas, suggesting that the practice may nowadays be rare. In a journalistic autobiography by the nun Ani Pachen 2000, about her childhood as the daughter of a Khampa (khams pa) chief who later became a Tibetan resistance leader, Ani Pachen refers to her parents having eaten "Za yig, precious scriptures" in an effort to conceive a child. 1 In Buddhist Monasteries of Himachal, O. C. Ha1).<;ia refers to "edible charms (za yig)" as a "tantra-based psychological therapy" for healing. 2 Wad I would like to thank the Kluge Center at the U.S. Library of Congress for sponsoring this research, and Bryan Cuevas, Matthew King, Mona Schrempf, Gene Smith, Travis McCauley, David White and Kenneth Zysk for their advice on various aspects of this paper.
1 2

Pachen/Donnelley 2000: 42.

HiiI).<;ia 1987: 127.

Journal of the International Association of Buddhist Studies Volume 32' Number 1-2 2009 (2010) pp. 85-113

86

Frances Garrett

dell's 1895 work, The Buddhism of Tibet, or Lamaism, with its Mystic Cults, Symbolism and Mythology, and in its Relation to Indian Buddhism notes that the "eating of the paper on which a charm has been written is an ordinary way of curing disease," and he provides an image of such an "edible charm"3 - an observation that was repeated in the respected medical resource, Merck's Archives, in 1904.4 Douglas' catalog, Tibetan Tantric Charms & Amulets, contains a page of sixteen images of "edible charms," which he understands to be used for the curing of disease, and which he claims to be "probably derived wholly from Indian sources," although he offers no evidence for this (as I will discuss below, this appears not to be the case). He also observes the use of unusual scripts on some of the charms, which he calls "unpronounceable yet filled with potentiality."s Despite this paucity of reporting outside Tibet, within certain pockets of Tibetan literature the practice is not at all uncommon. In this article, I will present the corpus of edible letters literature as a uniquely Tibetan nexus of contemplative, ritual, astrological and medical traditions, with intriguing connections to similar Chinese practices. I will consider the dietary, pharmacological, contemplative and astrological features of the practice, as described in texts, and in particular; address its perspective on the power of embodying the alphabet.

4 5

Wadde111934: 40l. Merck's Archives (Merck & Co., 1904), vol. 6, ii. Douglas 1978: plates 11-26.

Eating letters in the Tibetan treasure tradition

87

A jeweled treasury of edible letters (Za yig nor bu'i bang mdzod)
I will begin by describing the late fourteenth-century "Za yig nor bu'i bang mdzod," by Rdo rje gling pa (1346-1405), the longest text devoted entirely to this topic that I have encountered. 6 After salutations to Padmasambhava and Ye shes mtsho rgyal, Rdo rje gling pa's edible letters text begins with an impassioned justification for the usefulness of edible letter therapy. The remainder of the text is a catalog of edible letter spells or, as I will refer to them, recipes. The recipes address a range of needs. Some focus on enhancing religious practice, by for instance aiming at prolonging life, generating wisdom, or guarding against defilements or vow infringement. A number of recipes focus on repelling the evil intentions of external attackers: there are methods for subjugating all kinds of harmful spirits, sorcerers or roadside thieves, as well as evil spells, curses, weapons, ritual cakes and daggers. A third type of recipe addresses a variety of difficult-to-treat medical problems, including heart pressure (snying rlung), madness (smyo 'bog), possession by planetary spirits (gza' nad), and failure to conceive, as well as several serious epidemic conditions, such as malignant intestinal ulcers (khong lhog), contagious fevers (rims), dysentery (rgyu gzer), and smallpox (shu thor). A miscellaneous assortment of useful prescriptions serve to treat bad dreams, increase wealth, bring others under your control, and ensure that you win arguments. The recipes in this text typically include several components. First, each comes with an interlinear drawing of the letters to be used on the talisman. The letters are stacked on top of each other vertically, and there may be a row of vertical columns (see Figure 1). Usually the letters are written in the block printing (dbu can) script, but in other texts, other scripts may also be used, as I will discuss further below. Next, the names of several plant, mineral or animal substances are listed, and in some cases it is explained that these should be mixed into a paste that is used as ink for writ6 Rdo rje gling pa, "Za yig nor bu'i bang mdzod ces bya ba bzhugs so" in Rin chen gter rndzad chen rna ed. 'Jam mgon kong sprul blo gros mtha' yas (Paro: Ngodrup and Sherab Drimay, 1976-1980), vol. 67, 625-40.

88

Frances Garrett

ing the letters. Many different ingredients are named in the different recipes, some of them common organic substances, such as camphor, saffron, musk, turmeric, cinnamon, vulture meat, calcite, frankinscense, aconite, crystal, wood, or pig fat, and some of them more esoteric materials, including human flesh, the bones of someone killed by lightening, the blood of someone killed with a knife, cemetery ash from a meritorious person's remains, a c()nsecrated hair-braid, soot from an old house, or the excrement of a powerful person, to name just a few. Then, recipes address the appropriate times for undertaking the different components of the practice. An astrologically recommended time may be given for writing the letters on a piece of thin, clean paper? Certain days of the month may be specified for eating the preparation. Next are instructions for use: after writing the prescribed letters using the special ink, the paper should be rolled up; sometimes that paste of special ingredients is also stuffed inside the paper. The paper roll should be swallowed by the person wishing its effects, at certain times of day, without touching the teeth. 8 Eating more than one roll per day may be advised; in that case, the paper rolls should be stored in a clean jar. Sometimes one is advised to avoid the company of certain inauspicious persons, such as lepers or widows, or to avoid certain types of places, while undertaking an edible letters practice or treatment. Finally, many remedies also prescribe an accompanying visualization exercise, which may also include the recitation of mantras, as I will describe further below. Edible letter remedies vary in length and detail, but the following are typical of Rdo rje gling pa's "Za yig nor bu'i bang mdzod." Each description is accompanied by a glyph showing the letters to be written.

7 Other texts' recipes specify thin birch bark, yellow paper, or white paper; see, for example, Mi 'gyur rdo rje, Gnam chos 13 vols. (Paro Kyichu, Bhutan: Dilgo Khyentsey Rinpoche, 1983), vol. 2, 341-42. 8 See Ibid., vol. 3, 198. Although it is most common to say that the charm should not touch the teeth, Mi 'gyur rdo rje elsewhere recommends first biting it, holding it for awhile in the mouth, or resting it on the palate behind the teeth before swallowing; see Gnam chos, v. 2, pp. 341-2.

Eating letters in the Tibetan treasure tradition

89

These are the edible letters for generating wisdom. Write them with a paste of camphor and saffron on the tenth day of the waning moon. Avoid defilement that worsens in the presence of male lepers, widows, and so forth. At dawn of the fourteenth and fifteenth day [of the month], eat ten [each day while reciting] the secret imperial mantra. Your wisdom will blaze like fire. 9 These are the edible letters for pacifying evil spells and curses. Mix saffron, musk and excrement which has been blessed, and write them when in harmony with the victory star. If you [put the paper with] a five-pronged vajra at your heart or eat it regularly at the two middle times [of midnight and noon, it will] cleanse you of possession by the otherworldly. Also, you will not be harmed in body, speech or mind by any spells or destructive charms.10 These are the edible letters for pacifying bad dreams. If you eat them when you wake up, all bad dreams will be pacifiedY These are the edible letters for protection against all kinds of weapons. Write them with a liquid of vermillion, saffron, musk, camphor, falcon blood and vulture meat. Roll it up and eat it like food. You will be free from the fear of weapons when you go to the battlefield. Eat seven each day, and it's best if you include mantras toO. 12 These are the edible letters for causing conception. At the time of the Victory Star, write them well with a Gi'u wam liquid. If a women who

9 Shes rab skyed pa'i za yig 'di / [image] ga bur kha che sbyar ba la / yar mar tshes bcu bstun te bri / mdze pho yugs sal sogs pa'i / yang zhud pa'i grib la 'dzem / bcu bzhi bco lnga'i tho rangs la / gsang sngags rgyal po bcu bcu bza' / shes rab me ltar 'bar bar 'gyur. Rdo rje gling pa, "Za yig nor bu'i bang mdzod ces bya ba bzhugs so," vol. 67, 627. 10 Gtad dang byad kha zhi ba'i za yig 'di / [image] nus pa byin rlabs ldan pa yi / dri chen gla rtsi gur gum sbyar / skar ma rgyal dang bstun te bri / snying gar rdo rje rtse lnga pa sam / gung gnyis dag tu zos gyur na / pha rol byad du bcug pa ni / gtad dang mnan pa gang gis kyang / Ius ngag yid la tshugs mi 'gyur. Ibid., vol. 67, 628. 11 Rmi lam ngan pa zhi ba'i za yig 'di / [image] gnyid sad Ito ru zos gyur na / rmi lam ngan pa thams cad zhi. Ibid., vol. 67, 630. 12 Mtshon rigs srung pa'i za yig 'di / [image] Ii khri gur gum gla rtsi dang / ga bur dag gi chu yis bri / khra yi khrag dang bya rgod shas / grit la zas su bza' bar bya / mtshon cha'i 'jigs pa kun dang bral / g.yul ngor 'gro dus bdun bdun bza' / rgyun du sngags khar chud na chog. Ibid., vol. 67, 630-31.

90

Frances Garrett is devoted one-pointedly to her faith and vows eats them for a month, she will conceive a son quickly.13 These are the edible letters for rectifying vow infringements. Write them with your own excrement and musk, and if you eat them in the morning, all vow infringements will be rectified. 14

These recipes are of average length and detail - many are much shorter, essentially providing no more than the required letter configuration and the recipe's purpose. Others add to the sorts of detail just given by prescribing a contemplative or visualization practice that should accompany the eating of the rolled-up paper. Rdo rje gling pa's "Za yig nor bu'i bang mdzod," for example, has a companion text, the "Za yig nor bu'i bang mdzod kyi zhal shes," which describes a simple mental exercise for each recipe. For example, commentary on some of the prescriptions above recommends,
With reference to [the edible letters for] pacifying evil spells and curses, eat them having generated thoughts of your lama or personal deity (yi dam), whatever is suitable; then contemplate how, due to the power of the recipe, specialists in black magic will be turned away.1S
If you want to extend your life and merit, [picture] light-rays radiating from a clear physical image of Amitabha in front of you, and imagine [those light rays] gathering in the quintessence of the body, speech and mind of the Victor and his spiritual heirs as well as the longevity and merit of all living beings. Dissolve these into the body of Amitabha, allow his body to melt into light, and dissolve that into

13 Srid 'bebs za yig 'di / [image] skarma rgyal gyi dus tshod du / gi'u waM chu yis legs par bri / dad pa dam tshig ldan pa yi / rtse gcig gus pa'i bud med kyis / zla ba'i bar du zos gyur na / myur du bu dang ldan par gyur. Ibid., vol. 67,632. 14 Mna'i nyes pa zhi ba'i za yig 'di / [image] rang gi dri chen gla rtsis bri / nangs par zhogs pa ltor zos na / mna' yi nyes pa thams cad zhi. Ibid., vol. 67, 634. 15 Gtad dang byad kha zhi ba'i skabs / rang bla ma yi dam gang rung du bsgoms nas zos pas / sngags kyi mthus byad ma mkhan rang dbang med par phar zlog par bsgom. Rdo rje gling pa, "Za yig nor bu'i bang mdzod kyi zhal shes bzhugs so," in Rin chen gter mdzod chen mo ed. 'Jam mgon kong sprul blo gros mtha' yas (Paro: Ngodrup and Sherab Drimay, 1976-1980), vol. 67, 643.

Eating letters in the Tibetan treasure tradition

91

the edible letters. By eating them, your life and merit will expand. Meditate on that, and recite the mantra "OIp. vajra ayu~e pUIfye pUI).ye mahii pUIfye svahii" 100 times, or not less than 21 times. 16

If you want to generate wisdom, [picture] those edible letters as a shining orange dhf/:t [syllable] from which light is radiating, and gather up the total quintessence of the Victor's supreme wisdom [with those light rays]. Dissolve them into the dhf/:t, melt them into light, and merge [that light] into yourself. Imagine yourself as a wise and powerful Lord, and recite the mantra "OIp. vagisvari muIp. prajiia siddhi phala hUIp." 100 timesY

Revelations of edible letters continue

Within a century of Rdo rje gling pa's recipe book, two important Nyingma (rnying ma) Treasure Revealers, Rat na gling pa (14041479) and Pad rna gling pa (1450-1521), discovered several texts on edible letters. One work by Rat na gling pa, for example, focuses on remedies for afflictions caused by malicious spirits called Za (gza'), which are associated with the planets (see Figure 2). Resembling what we might refer to as seizure disorder, stroke or a form of mental illness, in Tibetan medical and religious works Za diseases are considered contagious and often best treated by ritual means. Rat na gling pa's text explains how to diagnose the particular type of Za disease by examining the patient's urine, and it describes three different healing technologies: wearing circular charms ('khor [0), eating edible letters, or writing letters directly on the patient's body.

16 Tshe dang bsod nams rgyas par 'dod na / mdun mkhar tshe dpag med spyi itar gsal ba'i sku las 'od zer 'phros / rgyal ba sras dang bcas pa'i sku gsung thugs kyi bcud dang / skye 'gro'i bsod thams cad bsdus / tshe dpag med pa'i sku la thim / sku 'od du zhu za yig la thim / de zos pas tshe dang bsod nams rgyas par bsam zhing / oM badzra AyuSe puNye puNye ma hA puNye swA hA / shes pa'i sngags 'di brgya 'am nyi shu rtsa gcig las mi nyung bar bzla'o. Ibid., vol. 67, 642. 17 Shes rab skyed par 'dod na / za yig de [image] dmar ser 'od 'bar bar gyur pa las 'od 'phros rgyal ba'i mkhyen rab ma Ius pa'i bcud bsdus Dhi la thim / de 'od du zhu nas rang la thim / rang mkhas pa'i dbang phyug tu gyur par bsams la / oM wA gI swari muM phra dznyA siddhi pha la hUM / shes pa'i sngags 'di brgya rtsa sogs bzla'o. Ibid.

92

Frances Garrett

Each edible letter recipe specifies ingredients to stuff inside or smear onto the paper on which the letters are written; for letters written directly on the body, these ingredients compose the ink that is used to write them. is As an example of a text that provides little instruction, by contrast, Rat na gling pa's "Phan byed zayig skor" is essentially an index of around thirty letter images, including a threesome of edible letters for protection, exorcism and killing, and recipes for various purposes, such as liberating yourself from possession by the eight classes of demons (sde brgyad), malevolent gnomes (the'u rang), or evil spells. 19 Said to be the teaching of Vajrayoginl, Pad rna gling pa's "Sku gsung thugs kyi za yig" also recommends the consumption of letters for repelling malevolent spirits and other personal dangers. 20 Like other such texts, it offers drawings of the letters to be written, recommendations for type of paper, and the ingredients with which to make the ink. Visualization exercises are important in this work. For example, you should recite a particular mantra before eating the roll, imagining that
is Rat na gling pa, "Gza' srung rna bu dgu 'khams kyi 'khor 10 lag len rna bu za yig za sman bdug sman Ius yig btags thabs 10 rgyus dang beas pa bzhugs so," in Rin chen gter mdzod chen mo, ed. 'Jam mgon kong sprul blo gros mtha' yas (Paro: Ngodrup and Sherab Drimay, 1976-1980), vol. 71, 525-35. 19 Rat na gling pa, "Gu ru drag po'i ehos skor las: Phan byed za yig gi skor mams," in Rat na gling pa'i gter chos (Darjeeling: Taklung Tsetrul Perna Wangyal, 1977-1979), vol. 5, 655-64. Another text on edible letters by the same author may be found at Rat na gling pa, "Za yig them med" in Rat na gling pa'i gter chos (Darjeeling: Taklung Tsetrul Perna Wangyal, 19771979), vol. 9, 539-42. 20 Pad rna gling pa, "Khros nag gi las mtha' sku gsung thugs kyi za yig log non las mtha' dang beas pa" in Rig 'dzin padma glin pa yi zab gter chos mdzod rin po che (Thimphu: Kunsang Tobgay, 1975-1976), vol. 2, 653-62.

Eating letters in the Tibetan treasure tradition

93

your body fills with light as the remedy is swallowed. As in Pad ma gling pa's"Phan byed za yig skor" and elsewhere, remedies here are categorized into those of body, of speech, and of mind. Consuming the letters of body, Pad rna gling pa explains, will protect you from earth spirits, niiga spirits, and demons, and froin poisoned food. While undergoing this practice, you should imagine yourself as a red Hayagrlva holding a club in the right hand and a noose in the left, making a threatening hand gesture. Consuming the letters of speech will protect you from dream omens, vow infringements, warfare, weapons, disputes, illness and uncleanness, as well as from various malevolent spirits. In this case, you should picture yourself as a dark blue Vajrapa:t;li, holding a nine-tipped vajra and a bell and grimacing aggressively. Consuming the letters of mind also addresses attack by various malevolent spirits. As you swallow this remedy, Pad rna gling pa instructs, you should imagine that your own body is filled with five lights, and that immeasurable light rays radiate outwards from a five-tipped vajra of gold within you; you will relax into a state of equanimity and bliss. For Pad ma gling pa, eating these letters is equivalent to eating the body, speech and mind of the deity in a process that actually purifies your own body, speech and mind, which in turn has the effect of exorcising malevolent spirits or protecting you from other sorts ofharm.21 The seventeenth century brings us a number of edible letter texts with some new features, indicating, as we should expect, that the practice may have changed in orientation over time, or that there were different schools of practice. The very short "Gos sngon can gyi za yig," by Klong gsal snying po (1625-1692), has the unusual message that the practitioner need not use special ingredients to inscribe the paper, nor recite special mantras, because it is sufficient simply to eat the inscribed paper: "There is no need for ingredients

21 Although the digestive mechanics of swallowing letters is not made as explicit as we might like, it appears that for many authors, the letters themselves act as deities within the body to effect certain aims. At the turn of the twentieth century, Mi pham explained that after you eat the rolled-up paper, it will reach your heart, at which point you should imagine that "all sickness, evil influence, misdeeds and obscurations will be consumed by fire."

94

Frances Garrett

and mantras; the practice of only eating is sufficient," he insists. 22 The text contains ten drawings of letters that are provided as protection against lightning, arrows or malevolent spirits, as well as drawings of the letters of body, speech and mind. If you eat the letters of body, speech and mind, he says, "at the three [times], dawn, daybreak and evening, you will reverse curses, evil spells, and malevolent yantra."23 Another twist is found in the work of the master of occult technologies of this era, Mi 'gyur rdo rje (1645-1667), whose collected works in thirteen volumes contain numerous edible letter recipes, including some quite distinctive ones (about which I will say more below). His "Khyi smyon bcos pa'i thabs lag tu blang pa'i rim pa," a text devoted to the problem of mad dogs (khyi smyon), begins,
Homage to Padmasambhava! In the degenerate age, the energies of the agitated ma mo [demoness], monk-demons (rgyal 'gong) and ghosts ('byung po) have become more active, and they have amplified their magical displays (cho 'phrul). At this time, mad dogs, mad people and mad donkeys increase their attacks on all beings. It is said that at the time of the male-fire-monkey year in particular, mad dogs greatly increased their harming [activities].24

22 Klong gsal snying po, "Ye shes rab 'bar las: Gas sngon can gyi za yig," in Rin chen gter mdzod chen mo ed. 'Jam mgon kong sprul blo gros mtha' yas (Paro: Ngodrup and Sherab Drimay, 1976-1980), vol. 67, 653-54. This text is also found at Klong gsal snying po, "Ye shes rab 'bar las: Gos sngon can gyi za yig," in Klong gsal snying po'i zab gter gsung 'bum ed. Pad rna rgyal mtshan (Darjeeling: Kargyud Sungrab Nyamso Khang, 1997), vol. 6, 482-84. 23 Klong gsal snying po, "Ye shes rab 'bar las: Gos sngon can gyi za yig," vol. 67, 654. 24 Padma 'byung gnas sku la phyag 'tshal lo / snyigs dus ma mo 'khrugs pa'i kha rlangs dang / rgyal 'gong 'byung po kha dar cho 'phrul che / de dus khyi smyon mi smyon bong smyon gyis / 'gro ba kun la gnod 'tshe che ba dang / khyad par me pho sprello tshes pa'i tshe / khyi smyon gnod pa chen po dar bar gsungs. Mi 'gyur rdo rje, "Khyi smyon bcos pa'i thabs lag tu blang pa'i rim pa bzhugs so," in Rin chen gter mdzod chen mo, ed. 'Jam mgon kong sprul blo gros mtha' yas (Paro: Ngodrup and Sherab Drimay, 1976-1980), vol. 73, 394. The fire-male-monkey year is 1556-7, assuming it is meant to be during his lifetime.

Eating letters in the Tibetan treasure tradition

95

Mi 'gyur rdo rje has an answer to this problem. He recommends first that the practitioner-magician purify himself or herself by abandoning meat, garlic and onions, and by drinking pure water collected from a river. Unlike edible letter recipes of an earlier period, this text, like that of Klong gsal snying po, states that ingredients and special preparatory instructions for the paper are unnecessary: the key aspect of the remedy here is the visualization. To protect yourself from attack by mad dogs, Mi 'gyur rdo rje recommends that you recite a mantra, "0Ip. tog ge ya Ie huIp. phar," while imagining yourself as Avalokitesvara with white and red glowing light radiating out from your heart, and in this form you should bestow a blessing on all dogs. On behalf of someone who has been bitten by a mad dog, you should engage in a similar visualization, picturing a stream of ambrosia flowing from your body over that of the victim, thereby neutralizing the dog bite's poisonous effect. While he also offers edible letters to go with these remedies, Mi 'gyur rdo rje argues that while edible letters, protection cords or water may be used, the important activity is the visualization exercise. In a work on veterinary remedies, "Rta phyugs kyi rigs la phan gdags pa'i man ngag nyer mkho bsdus pa," which he is said to have revealed at the age of thirteen, Mi 'gyur rdo rje recommends an edible letter remedy for the problem of contagious throat blockage (yams nad al 'gag), which may refer to the possibly fatal infection in horses that we refer to as "strangles." For this, the prescribed letters should be written with a combination of pen ink and shellac that has been dipped into a paste of specified ingredients such as molasses, sal-ammoniac, nutmeg, clove, cardamon, saffron, bamboo pitch and (possibly) semen (chas sman). The practitioner should then recite a mantra 30,000 times while doing the protection-granting gesture (skyabs sbyin phyag rgya) with the right hand, holding a skull-cup in the left, and meditating on Avalokitesvara. The edible letter preparation is dipped into purified water. A spell is recited, the "Master of contagious throat blockage" is invoked, the tlintrika shouts three times, and the edible letter preparation is fed to the afflicted patient, which in this case, apparently, is a horse. 25
25 Mi 'gyur rdo rje, "Rta phyugs kyi rigs la phan gdags pa'i man ngag nyer mkho bsdus pa bzhugs so," in Rin chen gter mdzod chen rna, ed. Jam

96

Frances Garrett

All the sources I have cited so far are Terma (gter rna), or Revealed Treasure texts. of the Nyingma tradition. Edible letter recipes are found elsewhere too, however, particularly in medical works. A text by the. major medical scholar De'u dmar bstan 'dzin phun tshogs (b. 1672), for example, recommends eating letters for the treatment of phlegm disorders. The work discusses a variety of phlegm disorders, beginning with instructions on the classification and identification of these conditions, followed by a range of treatments. De'u dmar begins with medicinal treatments - that is, the creation of pills, powders or decoctions - but he then describes three kinds of protection rituals: protective charms, mantra recitation, and edible letters. His sources for these therapies are the Revealed Treasures of earlier masters, specifically citing our friends above, Rdo rje gling pa and Mi 'gyur rdo rje. De'u dmar cites the work of Mi 'gyur rdo rje in particular for its expertise in edible letters that can heal phlegm diseases. 26 Finally, we also encounter edible letters practice, or something rather like it, in biographical literature. The autobiography of the Treasure Revealer 'Jigs med gling pa (1730-1798), translated by Janet Gyatso, records the following story. A ~Uikinl appeared before 'Jigs med gling pa and gave him a wooden box containing five rolls of yellowed paper and seven pea-sized crystals. When he unrolled one of the papers he smelled "an aromatic fragrance of camphor and other good medicines" and he found on the paper a stupa-shaped outline filled with "secret ~akinl sign-letters" that he could not read. As he started to roll it up, the letters suddenly transformed into Tibetan, revealing to 'Jigs med gling pa a Mahakarur;tika siidhana cycle. He unrolled another paper, which also turned out to contain an entire Tibetan text. At that moment a girl who looked like his mother came down from the sky and told him to eat the rolls of paper. Obediently he did so, swallowing them
mgon kong sprul blo gros mtha' yas (Paro: Ngodrup and Sherab Drimay, 1976-1980), vol. 71, 579-98. 26 De'u dmar bstan 'dzin phun tshogs, "Bdud rtsi snying po man ngag gi rgyud las nye bar mkho ba'i gces par btus pa bad kan gyi bcos khol bur phyung ba man ngag bdud rtsi 'chi sos," in Gsa rig gees btus rin chen phreng ba bzhugs sa (Zi ling: Mtsho sngon mi rigs dpe skrun khang, 1993), 165-75.

Eating letters in the Tibetan treasure tradition

97

without chewing, whereupon, he reports, "all the words and their meanings were printed on my mind."27 While this story is given as an example of a "mind terma" (dgongs gter), a teaching that has been concealed in the Revealer's own mind, the similarities to an edible letter roll are striking. If nothing else, this example serves to bring even closer the connections between the practice of edible letters and the Treasure tradition.

Contexts and sources for edible letter practices

In the 1ll-volume Rin chen gter mdzod, which organizes hundreds of Revealed texts according to the threefold classification of tantric practice into Mahayoga, Anuyoga and Atiyoga, edible letter practices are considered Mahayoga teachings for the achievement of mundane aims. 28 They are most commonly found in the context of rituals for protection; Rat na gling pa's Za disease text, for example, is found within the Rin chen gter mdzod's section on rites for the prevention of contagious diseases and plagues, which is itself found within the several volumes on protective rituals. As we have seen, however, in addition to these prophylactic or pacifying recipes there are many remedies aimed at enhancement or acquisition (rgyas), subjugation (dbang) and aggression (drag), following the standard fourfold classification of ritual activities (las bzhi).29 With the exception of the few texts that are devoted exclusively to the topic, typically edible letter recipes are presented alongside descriptions of other technologies such as wearing amulets and talismans or the writing of letters on the body, and they are considered useful for a similar range of effects. The recipes we have examined here were found in texts attributed to Treasure Revealers associated with Nyingma or Kagyii (Bka' rgyud) lineages. The fourteenth-century Treasure Revealer Rdo rje
27 Gyatso 1998: 57-58. This episode is also recorded at Khyentsel Padmakara Translation Group 1999: 7. See a similar story at Thondup 1986: 89. 28 For a basic overview of the Rin chen gter mdzod, see Gyatso 1996: 14769. 29 For more on this classification of ritual see Cuevas forthcoming.

98

Frances Garrett

gling pa, for example, is renowned in both Nyingma and Bon traditions. He is an innovative and controversial writer on Nyingma and Bon Great Perfection (Rdzogs chen) teachings,30 and he revealed numerous texts from the age of thirteen onwards, around Tibet and in Bhutan, on a wide range of topics, including medicine, astrology, rejuvenation alchemy (bcud len), and thread-cross (mdos) rituals. 31 The fifteenth-century Treasure Revealer Rat na gling pa is especially known for his compilation of a broad range of early texts now known as the Rnying rna rgyud 'bum, and his Treasures also cover various topics, including medicine empowerment rituals (sman sgrub) and recipes for special pills that bestow superpowers (siddhi). The Nyingma Treasure Revealer Pad ma gling pa is known as the national Treasure Revealer of Bhutan. Several figures important to the Kagyu tradition also wrote about edible letter practices, such as 'Bri gung rin chen phun tshogs (1509-1557), a well-known Kagyu master and Treasure Revealer who was the Abbot of 'Bri gung monastery for several years,32 and 'Bri gung chos kyi grags pa (1595-1659), a renowned Kagyu master and also a physician. 33 De'u dmar bstan 'dzin phun tshogs, author of many works on topics such as medicine, religious practice, drama and art, and founder of a Kagyu monastery in Kham, is most famous as a medical scholar. In the works of all of these scholar-adepts we find edible letter practices to be at once contemplative, devotional, occult, medical, astrological, cryptological and dietetic, crossing 30 Karmay 1988: 186. Also see his biography in Dargyay 1977: 139-43. 31 Karmay 2000: 3. Karmay cites gTer rabs rgya mtsho, sNgags 'chang
pa rdo rje gling pas (pa'i) zab gter rnams dang mjat ba'i gter byang rnam thar gter rabs rgya mtsho (Cover title: rDor rje gling pa'i zab gter rnams dang mjal ba'i rnam thar dang gsung thor bu), Texts of the rDo-rje gling-pa Traditionfrom Bhutan (Thimphu: Kunsang Tobgeyl, Druk Sherig Press, 1981:J.), Vol.21, 1-146. 32 See, for example, a single edible letter recipe said to be "from the terma of 'Bri gung chos rgyal rin chen phun tshogs," in Rin chen gter mdzod chen mo, vol. 71, 518. 33 See, for example, 'Bri gung rig 'dzin chos kyi grags pa, "Drangs srong gza'i bsrun ba za yig bdud rtsi 'chi gso 'am gnam 1cags rdo rje'i gter mdzod gza' yi gnyen po," in Collected works (gsung 'bum) of 'Bri-gung rig-'dzin chos-kyi-grags-pa (Bir, Distt. Kangra: D. Tsondu Senghe), 695-706.

Eating letters in the Tibetan treasure tradition

99

academic and sectarian boundaries but remaining firmly rooted in the Treasure tradition. But where does this practice come from? While edible letter recipes do address a wide range of needs, a significantly large number of them address the healing of illness. Some may wish, therefore, to identify these as a form of "tantric medicine" deriving from Indian tantras and siddha literature, Ayurveda, or even from the Atharvaveda, as indeed Nik Douglas has done. The category of tantric medicine, however, is ill-defined. It is well known that many Indian tantras, Hindu and Buddhist alike, describe methods of healing illness. For some who use the term, it simply refers to cases where a "ritual" or "magical" remedy, by which is often meant mantra, is applied to the healing of illness. Leaving aside the difficulty of determining precisely how "ritual" or "magical" remedies might differ from other sorts of remedies, this simple definition of tantric medicine is clearly unacceptable; healing remedies that use mantra are common even in early Indian Buddhist texts and, in any case, the necessary correlation between tantric Buddhism and mantra or dhiirm;.f use has been widely discredited. 34 Some better developed definitions of "tantric medicine" add principles such as a particular understanding of the power of extracting the essence (rtsi or bcud) of a medicinal substance, the requirement that the "doctor" be a tantric master, and the ability of consecrated medicinal substances to bestow not only health but also supernatural powers (siddhi).35 Indeed, it is true that many therapeutic techniques in Indian esoteric sources are both what some will call "medicinal," in the sense of involving the prescription of organic and non-organic compounds in the form of pills, powders or decoctions, and what some will call "tantric," in that they prescribe recitation of mantras, visualization exercises, participation of invoked deities, or empowerment or consecration of the medicinal substances to be consumed. Does the Tibetan practice of eating letters derive from this (ambiguously defined) area of Indian tradition?
34 For a summary of debates on this issue, see McBride 2005. Also see Davidson forthcoming. 35 See Samuel forthcoming, Stablein 1976, Walter, 1980, Garrett forthcoming.

100

Frances Garrett

As it happens, there is no clear record of a South Asian practice of eating talismans -not in the Atharvaveda, not in the extant Indian tantras or siddha texts, not in Indian medical traditions. The practice of wrapping. medicinal or otherwise special substances in textually-inscribed paper rolls to be eaten has not been found in these South Asian traditions. If edible letters are "tantric medicine," in other words, they do not come to Tibet frok India, at least not directly. Is this therefore an indigenous Tibetan tradition? While it certainly did develop over many hundreds of years in distinctively Tibetan ways, a look toward China suggests a surprising connection.

Sealing with script: a continuity of practice

A chapter in Michel Strickmann's Chinese Magical Medicine discusses "ensigillation," a technology of exorcism and prophylaxis that was widespread across China by the ninth or tenth centuries. Centuries before that, the swallowing of talismans, or the drinking of water in which their ashes had been dissolved, appears to have been one the primary modes of Daoist medical therapy. Strickmann argues that the practice of "sealing" is found first in the hands of early medieval occultists, then by the end of the sixth century it was codified by Daoist priests, and finally it was thoroughly incorporated into Chinese Buddhist circles as well, from which it was subsequently transmitted to Japan and to peoples on the borderlands of China alongside other tantric ritual practices. According to Strickmann's research, the earliest Daoist form of the practice, recorded in a Six Dynasties period (220-598 CE) text, involves a doctor's stamping a square-shaped inked seal onto various parts of the patient's body while reciting spells and engaging in a complex series of visualizations. Mter being "sealed" in this way; the patient was to swallow several talismans. The patient was advised to avoid certain foods before being sealed. A fifth-century text, the Book of Consecration (Kuan-ting ching, T. 1331) demonstrates a Buddhist adaptation of the practice: the Buddha recommends the use of a "seal" (mudrii) on a round piece of wood measuring seven inches by seven tenths of an inch. This seal should be held in the right hand while standing seven paces from the patient;

Eating letters in the Tibetan treasure tradition

101

after performing a visualization exercise, the practitioner should approach the patient and press the seal onto his chest. Vapors of different colors' enter the patient and neutralize the malevolent disease-causing forces, which then escape from the patient's navel in a cloud of smoke. The Buddha says, "By this mudra all demons are crushed and annihilated," and "if you use it to seal any of the body's diseases, all will be cured."36 The sixth-century Dharm}l Book of Atavaka, General of the Demons (T. 1238) also recommends the carrying of seals. Strickmann summarizes,
Even if a monk has broken the rules for monks, he need only rub the seal with cow's bezoar (bezoars being concretions found in the stomachs of cows, oxen, tapirs, and certain other animals) and impress it on a bamboo-membrane. After he has swallowed a hundred thousand such sealed talismans, he will become a bodhisattva of the first or second stage (bhiimi) and all his transgressions will be obliterated. 37

The text describes seals to be used either by pointing them in the direction of harm, or by imprinting them on paper which is then to be swallowed. Other seventh- and eighth-century texts recommend square seals made of wood that are to be either pressed upon the patient, or printed on paper that is swallowed by the patient. Strickmann also describes a seal collection attributed to Nagar-juna, possibly written in North China in the sixth century, which gathers "talismans for treating diverse maladies and for prophylactic use against disease-demons, government officials, and other robbers" plus talismans for childbirth and restoring harmony; these include instructions for their consumption "according to a precise astrological schedule."38 The practitioner should avoid meat and alcohol while engaging in the practice. Paper talismans may be printed using the seal and swallowed. While we cannot be certain of whether Chinese Buddhist texts .with such remedies had Sanskrit originals or whether they are apocryphal, the absence in any existing South Asian literature of 36 Strickmann 2002: 134-35. 37 Ibid., 147. 38 Ibid., 171.

102

Frances Garrett

the particular practice of eating talismans and its earlier presence in Chinese Daoist sources, suggests that this interpretation of the practice, at the very least, is Chinese. 39 Regardless of where r these texts may have originated, ~ however, similarities between the " , practices in Chinese and Tibetan sources should be obvious - the l .... 1!~.. "!"" ~ ._ _ _." '\ .............. eating of talismans for a variety of exorcistic and prophylactic aims, Figure 3: Unusually shaped (stiipa shaped?) edible letter design in a the writing of those talismans on strange script, for protecting against papers using vermillion and their epidemic disease (left) and poison treatment with herbal and other (right) (Gnam chos, vol. 4, 511). substances, the astrological, behavioral and contemplative recommendations - all of these, despite the differences, make it clear, I think, that we are dealing with a continuity of practice, and suggest to me that Tibetans were aware of this Chinese technology. The fact that such a practice does not exist in India makes it all the more evident that this is a case where Tibetans shared knowledge with their neighbors to the east rather than those to the south. Why are edible letters said to be effective? Thus far I have outlined the key features of a typical edible letter prescription. But which of these components is said to account for the remedy's effectiveness? Of course, the answer to this question will vary by practitioner. We have seen, for example, that for Klong gsal snying po, the act of just eating the rolled-up paper is sufficient. Mi 'gyur rdo rje's text on mad dogs states, by contrast, that the important aspect of the remedy is the visualization. Elsewhere
39 That is to say, if some of Strickmann's Buddhist sources did in fact have their origins in India, it could be that in the process of their transmission to China certain practices (if not entire texts) were adapted to suit Chinese traditions.

Eating letters in the Tibetan treasure tradition

103

in his collected works, however, Mi 'gyur rdo rje comments that the important aspect of using edible letters is not the ingredients, recitation of dharaJ}X, nor the manner of rolling or folding the paper. 40 De'u dmar, who gives detailed recipes for medicines for phlegm disease elsewhere in the same text (and is therefore concerned with proper combination of medicinal ingredients in those cases), ignores completely the presence of potentially medicinal ingredients in edible letter recipes, only providing images of the letters and advising one to swallow them. Indeed, while the authors we have discussed here do offer recipes of varying detail, what is by far the most common is the inclusion of only the name of the ailment or desired aim, and an image of letters, without any real recipe at all. It seems clear overall that what is critical to the practice of edible letters is - and this may not be so surprising - the letters! Let us look for a moment, then, at the letters. As we have seen, they are most typically, at least in the editions we have, written in the Tibetan block printing (dbu can) script; they are drawn in this way even when found within texts that are otherwise written in a cursive script. The letters are stacked vertically (Tibetan is not written vertically). They are not words - they have no linguistic meaning - and they are not mantras; that is to say, they are not called words or mantras in Tibetan, but simply "letters" (yig). While I have found not a single comment on why particular letters are used for particular aims, the letters are clearly important, given that we see many instances of identical letter arrangements transmitted across centuries of literature. Indeed, the letters appear to be what Tibetan scholars themselves find most interesting academically, not to mention therapeutically. Mi pham rgya mtsho (18461912), for example, who quotes in full edible letter recipes from various sources, has nothing to add himself to those sources' reci-

40 See Gnam chos, vol. 2, 537; Gnam chos, vol. 1, 94; Gnam chos, vol. 2, 328, where he writes that the choice of "ink, paper, or way of folding makes no difference" and that you should "eat it when you wish; ink and paper make no difference." Also see Gnam chos, vol. 2, 537, where after a list of more than ten edible letter recipes, a concluding statement claims that there is no enumeration of mantras, nor do ingredients and manner of folding matter. The same sentiment is found in Gnam chos, vol. 3, 197 and 198.

104

Frances Garrett

pes. The only thing he does say by way of critically comparing or commenting on his sources is to discuss how the letters are written - that is, does a particular recipe have six letters or seven, is it a "ta" or a "na" that should be written on top, and so forth.41 Mi 'gyur rdo rje, similarly, makes a point of emphasizing that the letters should be written very carefully, without any error.42 Although most edible letters are recorded in texts using dbu can, the script or cryptographic system used to write these letters is also a topic of interest for some Tibetan scholars. Making both functional and paleographic distinctions, Rat na gling pa classifies his remedies (which are not only of the edible letter sort) by letter type: mother letters, son letters, inner letters, Indian letters (rgya yig), ancient Zhang-zhung "Mar letters" (smar yig) , wild letters (rgod yig), edible letters. Mi 'gyur rdo rje refers in passing to "illness letters" (na yig).43 Rdo rje gling pa explicitly refers to his edible letters as "magical letters" ('phrul yig).44 The collection of edible letters by occult master Mi 'gyur rdo rje includes the use of various alternative scripts (see Figure 3, for example), with a series of cryptanalytic deciphering tables (such as in Figure 4) for those needing to read these ciphers. Edible letter recipes are in these contexts placed within a taxonomy

ffl

--it,:-

l;t
~;-: ~
~ ~
Ilif._

l~~
l~ ~~~ if .Iv
fr'irt .:.i1IJ
~~ u..iIr
W-

~f

iji
i ;Jt.:
f;.~ ~r .,... i-~

~I" -go

.1(

ft.~~ Ii [o,e-f

fF.

i"

~;-I/

;- -'>i
'iT

tr~
-~

41 For example, see Mi pham rgya mtsho, Las sna tshogs pa'i sngags kyi be'u bum dgos 'dod kun 'byung gter gyi bum pa bzang bo (Zi ling: Mtsho sngon Zhing chen Zhin hwa dpe khang gis bkram, 1999), 108 and 109. 42 Mi 'gyur rdo rje, Gnam chos vol. 3, 479. 43 This could be a scribal error; Ibid., vol. 2, 436. 44 'Phrul yig is the term for writing that changes form magically when a Treasure Revealer reads it, as sometimes happens in the discovery of Treasure texts. Rdo rje gling pa, "Za yig nor bu'i bang mdzod ces bya ba bzhugs so," voL 67, 637.

Eating letters in the Tibetan treasure tradition

105

of letters, rather than one of conditions to be treated, aims to be achieved, or Buddhas to be visualized. It may be, therefore, that a fruitful context in which to understand the consumption of letters is that of a much broader embodied alchemy of the alphabet. Of course, Buddhism in general has a long history of interest in syllabaries of various types, with elaborate traditions of use and interpretation of dhiirar;z, mantra, blja, and so forth, which have been implicated in various sorts of "letter magic" (despite the many and varied critiques ofthe controversial term "magic").45 What we have in edible letters is something rather unusual, however, in that it is a gastrosemantically significant interaction with the power of letters, where internal reception of the written forms of letters themselves is what is efficacious - an alchemical cryptography, we might say, or a cryptological alchemy. In this case, we can see this practice in another sort of context. The therapeutic or otherwise extraordinary power of written letters in the body is important not only in technologies of the Mahayoga sort, such as those that we have seen above, but it pervades much of Buddhist practice, all the way to the most esoteric and complex. Letters are linked to embodiment in prominent ways in a range of Buddhist tantric traditions. In various Tibetan religious, astrological and medical conceptions, key locations of the body are marked by syllables. 46 Kiilacakra systems of astrology and contemplative practice are particularly rich in phonemic analytical schemes; commenting on the Kiilacakra tantra, the Vimalaprabhii links the consonants and vowels of the Sanskrit alphabet with the joints of the body, for instance. 47 Tibetan Great Perfection writings explain that tiny letters pervade the body from the moment of conception: in the first week of gestation letters appear immediately inside the newly developed circulatory channels; these enduring letters are said to produce Buddha Bodies during contemplative visualization practices later in life. 48 David Germano

45 For example, see Gyatso 1992. For a recent critique of the use of the term "magic" in the context of mantras, see Burchett 2008. 46 Gerke 2008: 155, 355-6l. 47 The Kiilacakratantra 2004: lOI. 48 Garrett 2008.

106

Frances Garrett

has written about the contemplative use of letters in the Eastern Tibetan Peyti (dpaZ yuZ) sub-tradition in particular,49 as well as a Great Perfection contemplative practice that is actually called "sealing" (rgyas gdab pa): in this practice the meditator imagines the form of a blue huYfl syllable at the heart, which then comes out from the nose and pervades her surroundings such that everything in the external environment is sealed by, and in fact therefore becomes, the syllable huYfl. This meditation is followed by an "internal sealing" by those letters, which reenter her through the pores of her skin and fill her entire body. 50 Complicated and extensive, contemplative traditions such as these cannot be adequately addressed in this article. My intention here is simply to recommend that the internal reception of letters, in the case of edible letter remedies, not be seen in isolation from the internal presence or embodied activity of the alphabet, as taught in a wide range of Buddhist traditions in Tibet. For the users of edible letters, there may be a broad understanding that the special letters you eat may, quite unexceptionally, act efficaciously inside your body, as indeed letters are wont to do. By eating these letters, everyone, from the isolated Great Perfection yogin to the dogs and horses of the village lane, can partake of the alphabet's power.

49 Germano 1997: 322.

50 Ibid., 322-23. The multivalent uses of the model of the seal is worth pursuing. Strickmann discusses the use of seals for more than a thousand years in China, typically in the form of official "signature" seals, and the possible links of this to the therapeutic practice of ensigillation, both of which are aimed at controlling the spirit world (Strickmann 2002: 192). One feature therapeutic seals share with signature seals is the use of an "archaizing 'seal script' to whose secrets only officers and gentlemen, priests and administrators, scholars and commanders were privy" (Strickmann 2002: 156): In Tibet too, official seals are used; the seal alphabet was derived from the 'Phags pa script created in the Yuan period (1280-1368 CE), itself based on the Tibetan alphabetbut developed to more accurately represent Sanskrit, Mongolian and Chinese sounds. Note that it, like other seal scripts, is written not horizontally but vertically (like edible letters, that is); the first seal script noted in Tibetan histories, attributed to Sa skya pa1f<;li ta Kun dga' rgyal mtshan (1182-1252), was known as Hor yig, "Mongolian letters," and was also written vertically. See Clauson 1959: 305.

Eating letters in the Tibetan treasure tradition

107

Conclusions We have seen that despite the "Indian-ness" of presentation in edible letter texts, no South Asian precedent exists, and connections to similar Chinese practices are instead more likely. Some in Buddhist Studies have suggested a Chinese derivation to the Treasure tradition as a whole, and many have written specifically about connections between the Tibetan Great Perfection and Chinese Ch'an traditions. 51 While it is not conclusive, I might add the fact that all of the edible letter authors I have mentioned in this article are either from Eastern Tibet or spent large parts of their lives there (Eastern Tibet being, of course, the region of Tibet closest to China). Mi 'gyur rdo rje, the author whose collected works is most thoroughly pervaded by edible letters, is one of the most important scholars of the Nyingma Peyti tradition of Eastern Tibet. Indeed, eating letters may continue to the present day as such a regionally focused tradition, as I have recently been told about the use of edible letters by Eastern Tibetans. 52 Finding aspects of Tibetan Buddhism that are especially Eastern Tibetan is important, given the dominance ofresearch on Central Tibet, and the inaccurate but common perception that Central Tibet is representative of Tibet as a whole. Finding connections to Chinese religious practice is also interesting to those who do Buddhist Studies with a focus on Tibet, given the near erasure of explicit Chinese voices in Tibetan literature. Despite apparent continuities with Chinese practices, however, as Revealed Treasures, edible letters are made entirely Tibetan, deriving extra authority from astral connections to Indian Buddhist masters such as Padmasambhava. Entanglement with Chinese histories and practices is not the only pattern of exchange that I find interesting in the literary corpus on edible letters, however. Firmly rooted in the Treasure tradition,
Gyatso 1992: 153, Ueyama 1983: 327-50, Tucci 1958. Mona Schrempf has met tantrikas in Rebkong, Amdo, who use edible letters (personal communication, 2008). Dan Martin has described the Khampa Dilgo Khyentse Rinpoche passing them out at a ritual in Boudha (personal communication, 2008). Recall too that the nun Ani Panchen, whose biography mentioned edible letters (cited above), grew up in Kham.
51 52

108

Frances Garrett

it is labeled a Mahayoga teaching, though it plays a role in medical traditions too and is certainly quite centrally focused on the healing of illness. It is apparent here that distinguishing between what are often called "religious" and "medical" modes of healing is not so easy. The rubric of "tantric medicine," moreover, is inadequate in its neglect of the contexts in which these remedies actually appear, not to mention its lack of clarity about what the components of the term mean. What we have in this practice is a remarkably cross-disciplinary practice that unites contemplative, devotional, occult, medical, astrological, cryptographic and dietetic realms of knowledge and practice. The patterns of exchange between these realms recommend new ways of reading across disciplines, and new ways of understanding how practices and discourses may wind their ways through and across geographic, sectarian, professional or doctrinal boundaries.

Tibetan sources, arranged chronologically

Authorship uncertain Rims srung gi man ngag zab dgu phyogs gcig tu bsdus ba bzhugs so, In Rin chen gter mdzod chen mo, A reproduction of the Stod-lun Mtshur-phu redaction of 'Jam-mgon Kon-sprul's great work on the unity of the gter-ma traditions of Tibet, 111 vols. (Para, Bhutan: Ngodrup and Sherab Drimay, 1976),Vol. 71, pp. 495-518. This collection will elsewhere be referred to as Rin chen gter mdzod. Gter kha sna tshogs las byung ba gtso bor gyur pa'i srung ba'i man ngag phan bde'i go cha zhes bya ba bzhugs so, in Rin chen gter mdzod, Vol. 70. Slob dpon padma'i phyogs bcu'i za yig shin tu zab pa'i yig chung bzhugs so, In Rin chen gter mdzod, Vol. 67, pp. 647-651. Rdo rje gling pa (1346-1405) Za yig nor bu'i bang mdzod ces bya ba bzhugs so, In Rin chen gter mdzod, Vbl. 67, pp. 625-640. Za yig nor bu'i bang mdzod kyi zhal shes bzhugs so, In Rin chen gter mdzod, Vol. 67, pp. 641-645. Rat na gling pa (1403-1473)

Eating letters in the Tibetan treasure tradition

109

Gza' srung ma bu dgu 'khams kyi 'khotlo lag len ma bu za yig za sman . bdug sman lus yig btags thabs 10 rgyus dang beas pa bzhugs so, In Rin chen gter mdzod, Vol. 71, pp. 525-535. Phan byed za yig skor, In Ratna gling pa'i gter chos, Collected rediscovered teachings of Ratna-glin-pa, 19 vols. (Darjeeling: Taklung. Tsetrul Perna Wangyal, 1977-1979), Vol. 5, pp. 639-647. This collection will elsewhere be referred to as Ratna gling pa'i gter chos. Za yig them med, In Ratna gling pa'i gter chos, Vol. 9, pp. 539-542.
Pad rna gling pa (1450-1521) Bla ma nor bu rgya mtsho las khros nag gi las mtha' sku gsung thugs kyi za yig log non las mtha' dang beas pa bzhugs so, In Rig 'dzin padma glin pa yi zab gter chos mdzod rin po che, the rediscovered teachings of the great padma-glin-pa, 21 vols. (Thimphu, Bhutan: Kunsang Tobgay, 1975-1976), Vol. 2, pp. 653-662. This collection will elsewhere be referred to as Padma gling pa'i gter chos. Gu ru drag po'i skor las them med gud du sbas ba'i za yig, In Padma gling pa'i gter chos, Vol. 3, pp. 503-505. Klong gsal las krod rdom za yig nyin thub rdo rje'i go khrab bzhugs so, In Padma gling pa'i gter skoT, Collected gter-ma rediscoveries of Padma gling pa (A reproduction of a rare manuscript collection from Manang), 7 vols. (New Delhi: Ngawang Topgay, 1975- ), Vol. 3, pp. 333-334. 'Bri gung chos rgya! rin chen phun tshogs (1509-1557) [No title], In Rin chen gter mdzod, Vol. 71, p. 518. 'Bri gung pa mi bskyod rdo rje / 'Bri gung rig 'dzin chos kyi grags pa (15951659) Drangs srong gza'i bsrun ba za yig bdud rtsi 'chi gso 'am gnam lcags rdo rje'i gter mdzod gza' yi gnyen po, In Collected works (Gsun 'bum) of 'Bri-gun Rig-'dzin Chos-kyi-grags-pa, 1 vol. (Bir, Distt. Kangra, H.P.: D. Tsondu Senghe, 1985- ), pp. 695-706. Klong gsa! snying po (1625-1692) Ye shes rab 'bar las gos sngon can gyi za yig, In Rin chen gter mdzod, Vol. 67, pp. 653-654. Ye shes rab 'bar las gos sngon can gyi za yig, In Klong gsal snying po'i zab gter gsung 'bum, Collected revelations and writings of rig-'dzin klongsal-siiin-po (1625-1692) of Ka/:t thog, 12 vols. (Darjeeling: Kargyud sungrab nyamso khang, 1997), Vol. 6, pp. 482-483. Mi 'gyur rdo rje (1645-1667) Khyi smyon beos pa'i thabs lag tu blang pa'i rim pa bzhugs so, In Rin chen gter mdzod, Vol. 73, pp. 393-397.

110

Frances Garrett

Rta phyugs kyi rigs la phan gdags pa'i man ngag nyer mkho bsdus pa bzhugs so, in Rin chen gter mdzod, Vol. 71, pp. 579-598. Collection des tresors reveles par Gnam-chos Mi-'gyur rdo-rje, 13 vols. (Paro Kyichu, Bhutan: Dilgo Khyentsey Rinpoche, 1983). This collection will elsewhere be referred to as Gnam chos. De'u dmar bstan 'dzin phun tshogs (b. 1672) Bdud rtsi snying po man ngag gi rgyud las nye bar mkho pa'i gces par btus pa bad kan gyi bcos khol bur phyung ba man ngag bdud rtsi 'chi sos, In Gso rig gces btus rin chen phreng ba bzhugs so (Zi ling: Mtsho sngon mi rigs dpe skrun khang, 1994), pp. 165-175. Mi pharo rgya mtsho (1846-1912) Las snam tshogs kyi be'u bum bzhugs so (Hong Kong: Zhang kang then ma dpe skrun khang, 1999)

Sources arranged alphabetically

'Bri gung rig 'dzin chos kyi grags pa. "Drangs srong gza'i bsrun ba za yig bdud rtsi 'chi gso 'am gnam 1cags rdo rje'i gter mdzod gza' yi gnyen po." In Collected works (gsun 'bum) of 'bri-gun rig-'dzin chos-kyi-grags-pa, 695-706. Bir, Distt. Kangra: D. Tsondu Senghe. Burchett, Patton E. "The 'magical' language of mantra." Journal of the American Academy of Religion 76, no. 4 (2008): 807-43. Clauson, Gerard. "ThehP'ags-pa alphabet." Bulletin of the School of Oriental and African Studies 22, no. 1/3 (1959): 300-23. Cuevas, Bryan. "Mi pharo's 'calf's nipple' (be'u bum): A handbook of Tibetan ritual magic." In Tibetan Ritual, edited by Jose Cabez6n, Forthcoming. Dargyay, Eva M. The Rise of Esoteric Buddhism in Tibet. Delhi: Motilal Banarsidass, 1977. Davidson, Ronald M. "Studies in dhara literature i: Revisiting the meaning of the term dharaI}.l," Journal of Indian Philosophy (Forthcoming). De'u dmar bstan 'dzin phun tshogs. "Bdud rtsi snying po man ngag gi rgyud las nye bar mkho ba'i gces par btus pa bad kan gyi bcos khol bur phyung ba man ngag bdud rtsi 'chi sos." In Gso rig gces btus rin chen phreng ba bzhugs so, 165-175. Zi ling: Mtsho sngon mi rigs dpe skrun khang, 1993. Dilgo Khyentse, and Padmakara Translation Group. The Wish-fulfilling Jewel. Boston: Sharobhala, 1999. Douglas, Nik. Tibetan Tantric Charms and Amulets. New York: Dover, 1978. Garrett, Frances. Religion, Medicine and the Human Embryo in Tibet. Routledge Critical Studies in Buddhism. Abingdon; New York: Routledge, 2008.

Eating letters in the Tibetan treasure tradition

111

"Tapping the body's nectar: Gastronomy and incorporation in Tibetan literature." (Forthcoming). Gerke, Barbara .. "Time and longevity: Concepts of the life-span among Tibetans in the Darjeeling hills, India." Doctoral dissertation, University of Oxford, 2008. Germano, David. "The elements, insanity, and lettered sUbjectivity." In Religions of Tibet in Practice, edited by Donald Lopez. Princeton: Princeton University Press, 1997. Gyatso, Janet. "Letter magic: Peircean meditations on the semiotics of rdo grub-chen's dharal).I memory." In In the Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism edited by Janet Gyatso, 173-214. Albany: State University of New York Press, 1992. Gyatso, Janet B. Apparitions of the self: The secret autobiographies of a Tibetan visionary. Princeton: Princeton University Press, 1998. - - - . "Drawn from the Tibetan treasury: The gter ma literature." In Tibetan Literature: Studies in Genre, edited by Jose Ignacio Cabezon and Roger R. Jackson, 147-69. Ithaca: Snow Lion Publications, 1996. Hill).<;ia, Omacanda. Buddhist Monasteries in Himachal Pradesh. New Delhi: Indus Pub. Co., 1987. The Kalacakratantra: The Chapter on the Individual together with the Vimalaprabha, translated by Vesna A. Wallace, edited by Robert A.F. Thurman, Treasury of the Buddhist Sciences Series. New York: American Institute of Buddhist Studies, Columbia University, 2004. Karmay, Samten G. "Dorje lingpa and his rediscovery of the 'gold needle' in Bhutan." Journal of Bhutan Studies 2, no. 2 (2000): 1-35. - - - . The Great Perfection: A Philosophical and Meditative Teaching of Tibetan Buddhism. Leiden: Brill, 1988. Klong gsal snying po. "Ye shes rab 'bar las: Gos sngon can gyi za yig." In Rin chen gter mdzod chen mo, edited by 'Jam mgon kong sprul blo gros mtha' yas, vol. 67, 653-54. Paro: Ngodrup and Sherab Drimay, 1976-1980. - - - . "Ye shes rab 'bar las: Gos sngon can gyi za yig." In Klong gsal snying po'i zab gter gsung 'bum, edited by Pad rna rgyal mtshan, vol. 6, 482-83. Darjeeling: Kargyud Sungrab Nyamso Khang, 1997. McBride, Richard D. II "Dharal).I and spells in medieval Sinitic Buddhism." Journal of the International Association of Buddhist Studies 28, no. 1 (2005): 85-114. Merck's archives. Merck & Co., 1904. Mi 'gyur rdo rje. Gnam chos 13 vols. Paro kyichu, Bhoutan: Dilgo Khyentsey Rinpoche, 1983.

112

Frances Garrett

- - - . "Khyi smyon beos pa'i thabs lag iu blang pa'i rim pa bzhugs so." In Rin chen gter mdzod chen mo, edited by 'Jam mgon kong sprul blo gros mtha' yas, vol. 73, 393-97. Paro: Ngodrup and Sherab Drimay, 19761980. - - - . "Rta phyugs kyi rigs la phan gdags pa'i man ngag nyer mkho bsdus pa bzhugs so." In Rin chen gter mdzod chen mo, edited by 'Jam mgon kong sprul blo gros mtha' yas, vol. 71, 579-99. Paro: Ngodrup and Sherab Drimay, 1976-1980. ' Mi pham rgya mtsho. Las sna tshogs pa'i sngags kyi be'u bum dgos 'dod kun 'byung gter gyi bum pa bzang boo Zi ling: Mtsho sngon Zhing chen Zhin hwa dpe khang gis bkram, 1999. Pachen, Ani, and Adelaide Donnelley. Sorrow Mountain: The Journey of a Tibetan Warrior Nun. Kodansha America, 2000. Pad rna gling pa. "Khros nag gi las mtha' sku gsung thugs kyi za yig log non las mtha' dang bcas pa." In Rig 'dzin padma glin pa yi zab gter chos mdzod rin po che, vol. 2, 653-62. Thimphu: Kunsang tobgay, 1975-1976. Rat na gling pa. "Gu ru drag po'i chos skor las: Phan byed za yig gi skor rnams." In Rat na gling pa'i gter chos, vol. 5, 655-64. Darjee1ing: Taklung Tsetrul Perna Wangyal, 1977-1979. - - - . "Gza' srung rna bu dgu 'khams kyi 'khor 10 lag len rna bu za yig za sman bdug sman Ius yig btags thabs 10 rgyus dang bcas pa bzhugs so." In Rin chen gter mdzod chen mo, edited by 'Jam mgon kong sprul blo gros mtha' yas, vol. 71, 525-35. Paro: Ngodrup and Sherab Drimay, 1976-1980. - - - . "Za yig them med." In Rat na gling pa'i gter chos, vol. 9, 539-42. Darjee1ing: Taklung Tsetrul Perna Wangyal, 1977-1979. Rdo rje gling pa. "Za yig nor bu'i bang mdzod kyi zhal shes bzhugs so." In Rin chen gter mdzod chen mo, edited by 'Jam mgon kong sprul blo gros mtha' yas, vol. 67, 641-45. Paro: Ngodrup and Sherab Drimay, 19761980. - - - . "Za yig nor bu'i bang mdzod ces bya ba bzhugs so." In Rin chen gter mdzod chen mo, edited by 'Jam mgon kong sprul blo gros mtha' yas, vol. 67,625-40. Paro: Ngodrup and Sherab Drimay, 1976-1980. Samuel, Geoffrey. "A short history of Indo-Tibetan alchemy." In Studies of Medical Pluralism in Tibetan History and Society, edited by Mingji Cuomu, Sienna Craig, Frances Garrett and Mona Schrempf. Halle: International Institute for Tibetan and Buddhist Studies GmbH, Forthcoming. Stablein, William. "The Mahiikiilatantra: A theory of ritual blessings and tantric medicine." Doctoral dissertation, Columbia University, 1976. Strickmann, Michel. Chinese Magical Medicine, edited by Carl Bielefeldt and Bernard Faure. Stanford: Stanford University Press, 2002.

Eating letters in the Tibetan treasure tradition

113

Thondup, Thlku. Hidden Teachings of Tibet: An Explanation of the Terma . Tradition of the Nyingma School of Buddhism, Buddhayana series 1. London: Wisdom Publications, 1986. Tucci, Giuseppe. Minor Buddhist Texts, part II. Rome: Istituto Italiano per il Medio ed Estremo Oriente, 1958. Ueyama, Daishun. "The study of Tibetan Ch'an manuscripts recovered from Tun-huang: A review of the field and its prospects." In Early Ch'an in China and Tibet, edited by Whalen Lai and Lewis R. Lancaster. Berkeley: Berkeley Buddhist Studies Series, 1983. Waddell, L. A. The Buddhism of Tibet; or, Lamaism. 2nd ed. Cambridge [Eng.]: W. Heifer & sons, limited, 1934. Walter, Michael. "The role of alchemy and medicine in Indo-Tibetan tantrism." Doctoral dissertation, Indiana University, 1980.

"This inferior female body:" Reflections on life as a Tibetan visionary through the autobiographical eyes of Se ra mkha' 'gro (Bde ba'i rdo rje, 1892-1940)1 Sarah H. Jacoby

As Barbara Aziz informed us twenty years ago, the handful of accounts authored by women in pre-1959 Tibet detailing the "fabulous accomplishments of outstanding divine women" cannot provide an adequate basis for a sociology of women in Tibet.2 However, the few extant auto/biographies written by or about women certainly provide a good starting point to gather information about women, their social contexts, and their religious opportunities. 3 One such
1 This research could never have been undertaken without the support of Bya bral sangs rgyas rdo rje rin po che, who made Se ra mkha' 'gro's autobiography available to me, and Rgyal rong mkhan po thub bstan blo gros mtha' yas (Mkhan po sangs rgyas), who tirelessly read it with me. 2 Aziz 1987: 84-85. 3 For research on other early 20 th -century Tibetan autobiographies and biographies about women, see Havnevik 1997, 1999 for a study of the autobiography of Rje btsun 10 chen (also called Rig 'dzin chos nyid bzang mo, 1865-1951) and Rossi 2008a, 2008b for an introduction to the hagiography of Bde chen chos kyi dbang mo (1868-1927?) and a collection of religious biographies exclusively of women (Bde chen 1985) that she revealed. For research on earlier Tibetan-language auto/biographies about women, see Schaeffer 2004, 2005 for a study of the autobiography of the nun from Dolpo named 0 rgyan chos skyid (1675-1729) and Diemberger 2007 for a comprehensive study of the Bsam sding rdo rje phag mo female reincarnation lineage including a translation of the biography of the first in this lineage, Chos kyi sgron rna (1422-1455/6). Other pre-1959 Tibetan women who have biographies or autobiographies include Bsod nams dpal dren (1388?-1432?),

Journal of the International Association of Buddhist Studies Volume 32' Number 1-2 2009 (2010) pp. 115-150.

116

SarahH. Jacoby

previously unexplored autobiography is that of Se ra mkha' 'gro (also called Bde ba'i rdo rje, Kun bzang bde skyong chos nyid dbang mo, and Dbus bza' mkha' 'gro, 1892-1940). Se ra mkha' 'gro is one of the most prolific female authors in Tibetan history. Her writings include four volumes of revealed Treasure teachings (gter chos), a commentary on Buddhahood Without Meditation (Ma bsgom sangs rgyas) by Bdud 'joms gling pa (1835-1904), a biography of her teacher and consort Dri med 'od zer (also known as Gsang sngags glingpa, 1881-1924), and her own autobiography.4 Se ra mkha' 'gro's autobiography is one of the only Tibetan autobiographies written by a woman who was both a Treasure revealer herself and a consort to male Treasure revealers. 5 It is a long work

'Phrin las dbang mo (1585?-1668?), Lha 'dzin dbyangs can sgroI rna (17th century), Smin gling rje btsun mi 'gyur dpal sgron (1699-1769), Smin gling rje btsun 'gyur med phrin las chos sgron (early 19th century), Bsam sding rdo rje phag mo chos dbyings bde chen mtsho mo (19 th century), and Rgya gar ri rdo rje phag mo bde skyong ye shes dbang mo (1886-1909). Additionally, there are biographies of the women who were part of the Gung ri mkha' 'gro female incarnation line from A mdo. There are also several 'das log accounts about women in which they visit and return from the realm of the dead. For English sources, see Epstein 1982; Pommaret 1989; Delog Dawa Drolma 1995; Cuevas 2008. For hagiographies in English of important Tibetan female saints, see Nam-mkha'i snying-po 1983; Allione 1984; Dowman 1984; Edou 1996; Lama Chonam and Sangye Khandro 1998; Gyalwa Changchub and Namkai Nyingpo 2002; Harding 2003. 4 See Bde ba'i rdo rje 1978; Mkha' 'gro bde skyong dbang mo 1981; Dbus bza' mkha' 'gro 2002, 2009. 5 Treasure revelation is a system of on-going Buddhist revelation that traces much of its history to the religious activities of the eighth-century Indian Tantric master Padmasambhava, the figure who is credited with spreading Buddhism in Tibet. Padmasambhava gave teachings to be revealed in future times to incarnations of his twenty-five disciples, including his Tibetan consort Ye shes mtsho rgyal, who transcribed and concealed the Treasures. Treasures can be material or textual objects (sa gter, earth Treasures) discovered in the ground or they can be visionary revelations (dgongs gter, mind Treasures). There is a complex network of stages in the revelation process as well as conditions known as auspicious connections (rten 'brei) that must converge including finding the right time, place, doctrine holder, and consort for the prophesied Treasure revealer to successfully reveal his or her Treasures. See Thlku Thondup 1986; Gyatso 1986, 1993.

This inferior female body

117

_ more than 400 Tibetan folios - replete with Se ra mkha' 'gro's divinely inspired visions as well as her recollections of the difficulties of life as a 'female Treasure revealer who struggled to find her place amidst the religious encampments (chos sgar) of early twentieth-century Mgo log and Khams. 6 Se ra mkha' 'gro's narrative is a first-person account of the life of a woman who renounced her wealthy Lhasa upbringing as the daughter of Mongolian nobility in order to pursue a religious life in Mgo log, fleeing from an imminent forced marriage to a Chinese official at the age of fourteen. 7 En route to Mgo log, she recounts nearly starving and freezing to death in hopes of receiving Great Perfection (rdzogs chen) teachings from Dri med 'od zer, who was one of Bdud 'joms gling pa's eight sons. The text charts Se ra mkha' 'gro's transformation from being a central Tibetan noble girl to becoming a beggar in Mgo log and then a tjiikinf renowned for revealing textual and material Treasures (dgongs gter and sa gter). Not only did she teach her own Treasure teachings to monastics and devout laity in Eastern Tibet, but she also taught those of Bdud 'joms gling pa and Dri med 'od zero Se ra mkha' 'gro was not a nun; her life story includes reflections on the difficulties of living with her "life partner" (tshe grogs) Rgyal sras, son of the late nineteenth-century Treasure revealer Mgar ra gter chen pad ma bdud 'dul dbang phyug gling pa

6 Se ra mkha' 'gro lived the majority of her life in what is today Gser rta County, Dkar mdzes Tibetan Autonomous Prefecture, Sichuan, and Pad rna County, Mgo log Tibetan Autonomous Prefecture, Qinghai. The three main communities in which she lived were 1) Bdud 'joms gling pa's seat Zlar tshang dgon in Gser rta County (GPS N: 32 47.505 ft. E: 100 04.097 ft. elevation 13711 ft.), 2) Mgar ra gter chen pad rna bdud 'dul dbang phyug gling pa's seat Ban nag dgon in Pad rna County (GPS N: 32 54.129 ft. E: 100 46.338 ft. elevation 11,924 ft.), and 3) Se ra Monastery (Se ra theg chen chos 'khor gling) in Gser rta County (GPS N: 32 28.703 ft. E: 99 50.441 ft. elevation 13,407 ft.). 7 Se ra mkha' 'gro recorded her age according to the Tibetan system, which I have modified to accord with the international convention. Hence, when Se ra mkha' 'gro writes that she escaped from her family estate in Lhasa at the age of fifteen, I give the age as fourteen.

118

Sarah H. Jacoby

(1857-1910) from Pad rna County, Mgo log.s But Rgyal sras, with whom she bore two children, was not the most significant man in her life. As is evident in both Se ra mkha' 'gro's autobiography and the 248-foliobiography she wrote about her lama and prophesied consort (thabs grogs) Dri med 'od zer, her devotion and love for he whom she calls "Wish-fulfilling Jewel" (Yid bzhin nor bu) or "the Lord himself" (Rje de nyid) forms the backbone of her auto/ biographical narratives. Se ra mkha' 'gro's long separations from Dri med 'od zer and their ultimate reunion mirror her journey from suffering to realization, a Tantric journey symbolized by the union of method and wisdom in the form of predestined male and female partners. Se ra mkha' 'gro's prolific writings record her extraordinary accomplishments amidst the vicissitudes of life in early twentiethcentury Eastern Tibet, where she struggled to find a place for herself as an outsider from Central Tibet and as an unusual female agent in the largely male domain of being a Treasure revealer. Hindrances such as poverty, disease, and jealousy among competing female consorts threatened the fragile auspicious connections (rten 'breI) that Se ra mkha' 'gro sought in order to fulfil her life's purpose of revealing Treasures. Throughout her sorrowful tale of separation from loved ones, physical illness, and community conflicts, Se ra mkha' 'gro's main sources of unconditional support were the visionary t!akinfs who appeared before her in waking and dreaming life. pakinfs encouraged her to withstand the abuses of her quotidian world by reminding her of her divine origins and promising celestial respite in Ye shes mtsho rgyal's Buddhafield after she fulfilled her earthly duty of taming disciples. This article aims to present some of the key elements of Se ra mkha' 'gro's autobiographical narrative, focusing in particular on her portrayal of the female body. Se ra mkha' 'gro's writing is pervaded by references to her "inferior female body" (skye Ius dman

S Se ra mkha' 'gro never mentions Rgyal sras's full name in her writings, but it was most likely Pad rna mam rgyal (1882/3-?). For references to him, see Mgar gter chen pad rna bdud 'dul dbang phyug gling pa 2005: 77, 103.

This inferior female body

119

pa)9 at the same time as she asserts herself to be an incarnation of Ye shes mtsho rgyal and a divinely prophesied Treasure revealer. I suggest that Se ra mkha' 'gro's repeated negative comments about her female body reflect not only real societal restraints on the possibilities for Tibetan women in her milieu, but also serve the literary purpose of emphasizing her humility. The denigration with which she describes her femaleness neutralizes the radical implications of writing such a long autobiographical accolade, an endeavor typically assumed only by accomplished religious males. Laments about her lowly female form 'put her in her place' even as they allow her to articulate a convincing identity as a religious exemplar that is both sufficiently humble and spiritually superlative. 10 Through the dialogical tone of her writing, Se ra mkha' 'gro transforms her female gender from being a potential problem for her acceptance as an authentic Treasure revealer to a boon. Her continual laments about how a lowly woman like herself couldn't possibly accomplish the Dharma have the subversive effect of eliciting responses from visionary cjtikinfs and siddhas as well as important male religious hierarchs that affirm the great potential of women and proclaim Se ra mkha' 'gro's true identity as a divinely prophesied Treasure revealer. Through these direct speech dialogues, Se ra mkha' 'gro contradicts her own negative valuation of being female. She conveys to her readers that she is Ye shes mtsho rgyal incarnate via the words of her interlocutors while rarely making lofty assertions about herself in her own voice. Visionary cjtikinfs take the lion's share of this supportive role, often appearing immediately after discouraging events in Se ra mkha' 'gro's life to reassure her of her sanctity and to counter the worldly forces that seemed determined to discredit her.
9 One could also translate this as 'inferior body,' but given that its short form 'skye dman,' literally 'low-birth,' is a commonly used Tibetan word for . 'woman' or 'wife,' its gendered connotations are quite overt. For a discussion ofthe term 'skye dman,' see Aziz 1987: 79, Diemberger 2007: 10-13. 10 The dialectic between Se ra mkha' 'gro's gender-based self-denigration and her eloquent and loquacious depictions of her spiritual accomplishments formulate what Janet Gyatso has referred to as a "diffident autobiography," or one that bears a tension between humility and selfaggrandizement. See Gyatso 1998: 105.

120

Sarah H. Jacoby

l)iikinfs' confidence-inspiring words are especially significant given that Se ra mkha' 'gro mentions only one other woman who gave religious teachings in her entire autobiography, an elderly incarnation of Ma cig lab sgron who was a nun named Lama A ne bzang mo. ll With few other examples upon which to model her life as a female Treasure revealer, Se ra mkha' 'gro's.main inspiration was clearly the paradigmatic female saint of Tibet, Ye shes mtsho rgyal. According to Se ra mkha' 'gro's writings, Ye shes mtsho rgyal was not only Padmasambhava's Tibetan consort who wrote down and concealed his religious teachings to be rediscovered in later times as Treasure texts (gter rna), but Se ra mkha' 'gro was also an emanation of her. I suggest that Ye shes mtsho rgyal was a model for Se ra mkha' 'gro in another sense as well: her extensive autobiography bears noticeable resemblances to Ye shes mtsho rgyal's hagiography.12 The two texts also share common narrative elements including mothers from the powerful Tibetan Gnubs clan, fathers who were local political leaders, great beauty in youth, extreme repulsion from impending marriages, consort relationships with important male Buddhist teachers, resistance to unwanted male attention, and explicit use of male consorts to serve their own spiritual needs. Whether one interprets these similarities as indicative of Se ra mkha' 'gro's status as Ye shes mtsho rgyal's incarnation or as signs that Se ra mkha' 'gro shaped elements of her writing and even her life on Ye shes mtsho rgyal's hagiography, it is striking to note how few other precedents for human female sanctity Se ra mkha' 'gro found.13
11 Bde ba'i rdo rje, Dbus mo bde ba'i rdo rje'i rnam par thar pa, fol. 122a3 (henceforth I refer to this text as Se ra mkha' 'gro's Autobiography because she is most widely known to English-language readers by her epithet Se ra mkha' 'gro rather than Bde ba'i rdo rje, which was her Treasure name or gter ming). Se ra mkha' 'gro writes that she never met A ne bzang mo, who died when Se ra mkha' 'gro was nineteen years old. 12 Se ra mkha' 'gro never describes which version of Ye shes mtsho rgyal's biography she read or heard, but I have found at least three passages from her autobiography that are parallel to passages in the 17th -century Bstag sham nus ldan rdo rje edition. 13 Aside from A ne bzang mo and the hosts of celestial fjakinfs who appear before her in her visions, the only other female religious adepts Se ra

This inferior female body

121

Text description

Se ra mkha' 'gro's autobiography is titled The Spiritual Biography ,oj the Central Tibetan Woman Bde ba'i rdo rje: A Reliquary for Fortunate and Faithful Ones [Serving as] a Chariot Leading to Renunciation. 14 In the two-part colophon of the text, Se ra mkha' 'gro specifies that she completed the first segment of the text in 1927 at the age of thirty-five at Se ra Monastery in Gser rta County, and that this portion of the text was transcribed by her male disciple Tshul khrims rdo rje. 15 Then, in the second part of the colophon, Se ra mkha' 'gro documents that she herself wrote down her life story from age thirty-five to forty-two. 16 Se ra mkha' 'gro's disciple Bya bral sangs rgyas rdo rje rin po che describes the long version of her autobiography (rnam thar rgyas pa) quoted in this article as the outer, inner, and secret autobiography combined

mkha' 'gro references are Jo mo sman mo, Kun dga' 'bum rna, Ma gcig lab sgron, and Sa skya'i rta mgrin dbang mo. She also writes of two women she knew personally, Sngags btsun nyi rna sgron rna, partner of Mkhan po nor bu dbang rgyal from A bswe, and the tf,likinffrom Khrom dge named Mkha' 'gro zla ba sgron rna, though she does not indicate whether or not these women had their own disciples. 14 Bde ba'i rdo rje, Dbus mo bde ba'i rdo rje'i rnam par thar pa nges ibyung 'dren pa'i shing rta skal ldan dad pa'i mchod sdong, unpublished , manuscript, ca. 1934. The manuscript I quote in this article is 407 folios long iand was lent to me by Bya bral sangs rgyas rdo rje rin po che in Pharping, Nepal. I have also found another 526-folio version of this manuscript that is privately owned in Gser rta County. The variation in length between these ,two manuscripts appears to be the result of different handwriting sizes as to , date I have not found any differences in content. For a study of this autobi,ography and the biography Se ra mkha' 'gro wrote of Dri med 'od zer, see ;Jacoby 2007. 15 Se ra mkha' 'gro, Autobiography, fol. 406a. The monk Tshul khrims rdo Tje transcribed many of Se ra mkha' 'gro's writings. He was Se ra mkha' 'gro ,and Dri med 'od zer's close disciple and attendant, who travelled with Se ra ',mkha' 'gro after Dri med 'od zer's death along with her other monk attendant Thub bstan bzang po. 16 Se ra mkha' 'gro, Autobiography, fol. 407a. Since she died at fortyeight, all but the final six years of her life are included in the autobiography.

122

Sarah H. Jacoby

into one. 17 Additionally, there is a condensed fourteen-folio version published in Se ra mkha' 'gro's collected Treasure revelations (gter ChOS).1B Until recently, Se ra mkha' 'gro's autobiography was a rare text, unpublished in Tibetan or English.19 Se ra mkha' 'gro's writing style is influenced by Mgo log dialect Tibetan -and the text includes some spelling and grammar variants which. attest to its distance from standard classical literary Tibetan. The colloquial style of the text is highly oral in nature; the text reads like a conversation between Se ra mkha' 'gro and the plethora of divinities and humans who populate her narrative, written as direct speech. Interspersed in the dialogic prose are elaborate verse poems and songs that demonstrate Se ra mkha' 'gro's linguistic prowess and insight into Buddhist philosophy and praxis, insight that repeatedly surprised and inspired great faith in those around her who thought she was uneducated. 20 Approximately half of the text documents

17 Bya bral sangs rgyas rdo rje 1976: 5.

1B The condensed version of Se ra mkha' 'gro's autobiography is titled Ku su lu'i nyams byung gi gnas tshul mdor bsdus rdo rje'i spun gyis dris Ian mos pa'i lam bzang, written in 1929, in Bde ba'i rdo rje 1978, vol. 4:103-129. 19 Reasons for the rarity of Se ra mkha' 'gro's Autobiography in Tibet include: (1) The lack of a blood lineage heir to carryon Se ra mkha' 'gro's teachings. Only one of Se ra mkha' 'gro's children lived beyond childhood, her daughter Chos dbyings sgron rna, who had a son fathered by Dri med 'od zer's son Bsod nam lde'u btsan named Mdo li nyi rna (or Dpal don 'grub 'phrin las mam rgyal mchog), b. 1946, who was killed during the Cultural Revolution (Pad rna 'od gsal mtha' yas 2003: 71). (2) The esoteric content of her narrative, which includes many of her visions and accounts of consort practices. (3) Perhaps the insecure political and economic climate in the early-mid 20 th c. in parts of Mgo log and Gser rta due to incursions by warlords followed by People's Liberation Army soldiers caused her lineage and manuscripts to become relatively obscure in Tibet aside from a few places in which her manuscripts have been preserved, notably Ri bo che in the Chab mdo Prefecture of the Tibetan Autonomous Region and Bkra legs and Se ra Monasteries in Dkar mdzes Tibetan Autonomous Prefecture, Sichuan Province. However, Se ra mkha' 'gro's works will soon circulate more widely due to the recent Tibetan-language publication of her collected works compiled and edited by the Mgo log khul gna' rtsom bya ba'i 'go khrid tshan chung gzhung las khang (Dbus bza' mkha' 'gro 2009). 20 Se ra mkha' 'gro, Autobiography, fols. 79a, 81b-82a, 88b, U1a, 126b.

This inferior female body

123

Se ra mkha' 'gro's meditative experiences (nyams snang) during waking and dreaming life in which she travelled widely through Buddhafields and met many qiikinfs, lamas, siddhas, and other divine beings who lavished prophesy and at times Treasures (gter) upon her. The other half chronicles Se ra mkha' 'gro's travels, teachings, thoughts, and human interactions. 21 Life in Lhasa as a Chinese leader's daughter A fascinating element of Se ra mkha' 'gro's liberation story is the way she describes the ethnicity of her family and the events surrounding her birth. She writes that she was born in an area of Lhasa west of the Jo khang Temple called Klu mo mthil in a palace called Rgya ra ga shar to a mother of the Tibetan Gnups clan named Tshe ring chos 'dzom and a father of royal Mongolian ancestry named Lha sras byams pa mgon pO.22 Interestingly, she refers to her family as being a Chinese leader's family.23 She describes her 21 Se ra mkha' 'gro's Autobiography contains 88 visions, which comprise 46% of the text. Out of the 88 visions she describes, 55% of them are visions of female divinities, 27% of them are male divinities, and 16% of them include both male and female divinities (or neither in the case of formless Visions and visions of animals not distinctly gendered). This breakdown between visionary and worldly life is similar to that of the biography Se ra mkha' 'gro wrote of Dri med 'od zer in which visions comprise 52% of the text. Out of 80 different visions in his biography, 58% of them are of female divinities, 31% of them male divinities, and 10% of them include both male and female divinities (or are genderless, i.e., they are of animals or formless voices). 22 The palace name Rgya ra ga shar is similar to the name of the Lhasa noble family known as Mdo mkhar or Rag ga shag, whose palace was also to the West of the Jo khang. Another similarity between the Rag ga shag family and that of Se ra mkha' 'gro is an affiliation with Stag lung Monastery, in particular with the abbot Ma' rin po che. Despite these overlaps, I am not aware of any evidence suggesting that Se ra mkha' 'gro was connected to the Rag ga shag family. 23 Se ra mkha' 'gro describes her father's race as sog po dbang gi dung rgyud, i.e., "royal Mongolian ancestry," and their family as rgya dpon sang, i.e., "Chinese leader's family." Se ra mkha' 'gro, Autobiography, fols. 7b, 17a. Se ra mkha' 'gro's description of her father as a "Chinese leader" probably means that he was a Qing official, not that he was ethnically Chinese, given

124

Sarah H. Jacoby

father as a man fluent in both Chinese (rgya skad) and Tibetan who wore Chinese attire (cha lugs rgya nag) but was a devout Tibetan Buddhist Seminal Heart practitioner (chos lugs snying thig).24 Given that she mentions that the famous Rnying rna lama and descendent of Gushri Khan named Chos rgyal ngag kyi dbang po (1736-1807) was her father's ancestor, it is likely that Se ra mkha' ?gro's father was descended from the Mongol Khans who settled in Lhasa. Se ra mkha' 'gro recounts a dispute that emerged regarding her birth year between a Chinese group (rgya sde rnam) who argued that she was born in the Chinese male hare year (1891) since there were two 30th days at the end of the last month and a Tibetan group (bod shag pa rnams) who claimed that she was born on the first day of the first month of the Tibetan dragon year (1892).25 While the birth year debate ended in favor of the Tibetan group's perspective, Se ra mkha' 'gro's father insisted that she learn Chinese characters (rgya yig) before she was allowed to learn Tibetan. 26

that there were very few ethnically Han Chinese officials in Tibet until the twentieth century. According to Kolmas 1994: 456-57, out of the 135 Ambans or Assistant Ambans in Tibet from 1727-1912, only 7 were Chinese while the majority were Manchu or Mongolian. In Se ra mkha' 'gro's usage, the term rgya nag, "Chinese," seems to refer to non-Tibetans affiliated with the Qing Empire and not to the Han Chinese ethnicity. In any case, Se ra mkha' 'gro explains that her real father was none other than Gnyan chen thanglha, the pre-Buddhist land deity who Padmasambhava converted to Buddhism (Se ra mkha' 'gro, Autobiography, fols. 17a, 19a). Because her real father was not the "Chinese leader," she later concluded that she was free to pursue her religious aspirations following her true father's footsteps. Incidentally, Gnyan chen thang lha is also said to be the father of Mdo mkhyen brtse ye shes rdo rje (1800-1866), who was from Mgo log. See Thondup 1999: 180. 24 Se ra mkha' 'gro, Autobiography, fols. 27b-28a. 25 Se ra mkha' 'gro, Autobiography, fol. 9a. 26 Se ra mkha' 'gro, Autobiography, fols. 12b, 36b-37a. There is prece~ dence for Chinese language education and usage in Lhasa during the period in which Se ra mkha' 'gro lived there [between 1892-ca.l906]. See Richardus 1998: 165, 168 for an autobiographical reference to Chinese literacy in Lhasa at the time: Ts'an-chih Chen (b. 1893) recounts being educated in a Chinese school in Lhasa in preparation for work in the Amban's office while his sisters received Chinese language education privately at home, as did Se ra mkha' 'gro. Thanks to Gray Tuttle for directing me to this reference.

This inferior female body

125

Despite her father's resistance to her religious yearnings, Se ra mkha' 'gro represents her childhood as one of sanctity in which she continually recited the six-syllable rna lJi prayer instead of playing games with other children. Despairing at the thought of never having a chance to practice Buddhism because she was being groomed to become a leader (dpon rno), she writes that when shewas seven years old, she threw her schoolbooks into the river. At the same time, she contemplated suicide, making the first of many prayers lamenting her female body:
Alas, I pray to the Lotus Born One from 0 rgyan, all refuges condensed into one, To Mother Mtsho rgyal, and to the Venerable Noble Lady TaraLook upon me with compassion! This female body (skye dman) of mine is the foundation of cyclic existence. Having thrown away this contaminated, unclean, negative body, Bless me with attaining a body replete with the freedoms and advantages. As soon as I am born, having met with the dharma, May I have the freedom to practice properly.27

Just as she was about to throw herself into the water along with her books, a terrifyingly ugly woman brandishing a sword appeared before her and threatened to kill her. This r;liikinf vision is the first of many in her autobiography, all of which serve to protect her and encourage her religious yearnings. Premonitions of her future as a Treasure revealer also began at the age of seven when she pulled a ritual dagger (phur bu) part-way out of a rock at Brag yer pa near Lhasa. However, from the beginning, important people in Se ra mkha' 'gro's life rejected the possibility that there could be a female Treasure revealer. When she was eight years old, she and her mother went on pilgrimage to 'On phu stag tshang in Lho kha where Ye shes mtsho rgyal had received . VajrakIlaya empowerments from Padmasambhava. Remembering her former karmic connections to the place, Se ra mkha' 'gro told

27 Se ra mkha' 'gro, Autobiography, fo1. 13bl-3.

126

Sarah H. Jacoby

her mother, "It seems like I have lived in this land in a previous life. I remember my former actions and my share of texts that are here."28 Her mother retorted,
"Don't talk that way. From the time of your exalted paternal ancestors [such as] the Dharma King Ngag gi dbang po until now, there have been many who have gained spiritual accomplishment. Because it is impossible for a girl to take out Treasure texts, the king's punishment will soon come. Be silent."29

Immediately after her mother's harsh words, Se ra mkha' 'gro had her first dream in which a t;liikinf came to her and gave her a share of the t;liikinfs' feast, telling her that she was the t;liikinfs' magical emanation and that they would meet again quickly. In this way, Se ra mkha' 'gro's vivid visionary experience belied the discouragement those in her human world repeatedly projected onto her. Many other visionary experiences soon followed in which she travelled to other realms, including many Buddhafields (zhing khams) and hells. After one such visionary encounter with a t;liikinf that occurred when she was ten years old, she amazed everyone by curing those stricken by a smallpox epidemic that raged through Lhasa just after the turn of the twentieth century by blowing on them and reciting special mantras. 30 Se ra mkha' 'gro's childhood world was shattered when her father began arranging her betrothal. Although requests for Se ra mkha' 'gro's hand in marriage came from Tibetan ministers including those from the Shar khra, Rdo ring, and Zur khang families, she writes that her father refused their offers and returned their gifts because those families were from the Tibetan group (bod sde). Instead, he chose a Chinese leader (rgya sde'i dpon) named Kru'u da 10 yi from Skyid grong, whose clothing was Chinese but whose religious sect was Seminal Heart, just like that of Se ra mkha' 'gro's father.31 Depressed about her upcoming marriage, Se ra mkha' 'gro
28 Se ra mkha' 29 Se ra mkha' 30 Se ra mkha' 31 Se ra mkha' 'gro, Autobiography, fo1. 17b. 'gro, Autobiography, fo1. 17b. 'gro, Autobiography, fo1. 31al-6. 'gro, Autobiography, fo1s. 27b-28a: de dag bod sde yin pas

ma gnang nas rgyu rdzas thams cad phyir sprad / nub phyog skyid lung grong

This inferior female body

127

recalls that at the age of eleven, she attempted suicide by drinking a mixture of opium and alcohol only to be discovered and resuscitated by her father.32 Although her father tried to backtrack on his marriage promise in light of her despondency, the Skyid grong family threatened war if he broke their agreement. 33 Further adding to her sorrows, the following year Se ra mkha' 'gro's mother passed away and her father promptly remarried. . Unhappy at home and dreading her imminent marriage, the twelve-year-old Se ra mkha' 'gro's life transformed when Vajravaram appeared before her in a dream and empowered her in the two Treasure cycles that would be her life's main practice, The Secret Treasury of Reality pakinfs (Chos nyid mkha' 'gro'i gsang mdzod) and The pakinfs' Heart Essence (Mkha' 'gro'i thugs thig).34 Despite VajravarahI's encouraging proclamations, this momentous vision is one more instance in which Se ra mkha' 'gro laments her "inferior .female body." After receiving empowerment and instruction from VajravarahI, she writes,

chen zer ba rgya sde'i dpon kru'u da [0 yi'i zhes pa cha lugs rgya nag dan / chos lugs snying thig pa nga rang tsho dang 'dra ba / bsod nams long spyod [dan pa zhig la bug cig pu las med pa /. Da lo yi (Ch. da lao ye) is a Chinese honorific title meaning "great old gentleman." It was a very minor title meaning "your honor," given to local leaders as a polite way to refer to a Prefect, Department, or District magistrate. Sperling 1998: 331 translates it as "his honour" in the context of the title of a translator of the Qianlong emperor in the late 18th century. 32 Se ra mkha' 'gro, Autobiography, fo1. 33b. Se ra mkha' 'gro writes that her father saved her by pouring seed oil down her throat through a straw and forcing her to regurgitate the poison she drank. 33 Se ra mkha' 'gro, Autobiography, fo1. 33a. 34 This is reminiscent of the life of another female Treasure revealer, the thirteenth-century Jo mo sman mo, who was also an incarnation ofYe shes mtsho rgyal who had a vision of VajravarahI when she was thirteen (i.e., twelve according to international convention) in which she empowered her in Treasures. One of the seven prophecies in the beginning of Se ra mkha' 'gro's autobiography is from Jo mo sman mo, who predicts her future rebirth as Se ra mkha' 'gro (Autobiography, fols. 6a6-6b2). For Jo rno sman mo's hagiography, see 'Jam mgon kong sprul blo gros mtha' yas 2004: 379-380.

128

Sarah H. Jacoby

"You are extremely kind to bestow empowerment and instruction upon me. Because I am one with an inferior female body (skye lus dman pa), it is difficult to accomplish the Dharma and [benefit] beings. However I can do this, how great! Why do I have the karmic connection to be empowered in these two profound Treasures?" The akinfanswered, "Noble woman, you are not one with an inferior female body (lus dman pa). Before, in response to Mtsho rgyal's inquiry, Guru [Padmasambhava] proclaimed: 'The body of a female bodhisattva is a supreme body without [karmic] remainder; It is the mother [who gives birth to] the great assembly of primordially perfected Ones Gone to Bliss (Skt. sugata). Generate the aspiration to attain a supreme body like The Great Vajra Queen with the knowledge of the expanse of emptiness. The beneficial deeds of the Buddhas of the three times, a dance of the moon [reflecting] in water, Arise in physical form on account of disciple's devotion. Their source, the lotus mouth of the Vajra Queen, Is the foundation of the Buddha's teaching, like a ripening bud. This female body (bud med lus) is not inferior; this body is good. Mtsho rgyal, you must generate the pure aspiration Never to part with this [female] body.' Because [the Guru] spoke like this, [your body] is not inferior."35

This is one of the many instances in which Se ra mkha' 'gro interweaves her own biography with the sacred history of the Treasure tradition in the form of the divine couple Ye shes mtsho rgyal and Padmasambhava, thereby emphasizing her genealogical relationship to them. If Padmasambhava proclaimed that Ye shes mtsho rgyal's female body was not inferior, then, as her emissary, Se ra mkha' 'gro's female body must also not be inferior. The dual tone of Se ra mkha' 'gro's gender-based self-denigration emerges in passages such as this; thematizing her "inferior female body" is more than an internalization of the misogynist views of her social milieu because it has the subversive effect of occasioning responses from
35 Se ra mkha' 'gro, Autobiography, fols. 50al-50b2.

This inferior female body

129

others that articulate highly positive views of women. I suggest that this dual tone of Se ra mkha' 'gro's repeated laments about her female body was intentional; by drawing attention to her "inferior" status as a woman, she both mitigated others' potential criticism and defended herself in writing from the charge of overestimating her social status and religious authority. In the midst of these transformative cjClkinf visions and oppressive marriage preparations, the lama who would change Se ra mkha' 'gro's life, Dri med 'od zer, arrived from Mgo log along with his group of travelling pilgrims in search of shelter at her family's Lhasa residence. Instantly, tears flowed from her eyes and goose bumps rose on her flesh as she prayed in all her lifetimes never to separate from him. Intrigued by Dri med 'od zer and drawn to the life of a religious devotee free from the constraints of worldly attachments, Se ra mkha' 'gro caused a sensation when she sneaked out of her Lhasa home at the age of fourteen in about 1906 to follow Dri med 'od zer to Eastern Tibet, where she lived for the rest of her life. A new life unfolds in Mgo log The few snacks Se ra mkha' 'gro carried and the refined Lhasa clothing she wore quickly proved insufficient for the rigors of life as a wandering religious mendicant. Se ra mkha' 'gro recounts the shock she faced acclimating to her new rough circumstances outside of the wealth and privilege of her family's protection. Starving and in search of food, at first she mistook a nomad's tent for a wild yak. When she put the barley flour nomads gave her into her shoe because she didn't have a bag for it, the nomads laughed at her. 36 She had no idea how to make tea, and when she learned, she felt as though her eyes, which had never before felt the sting of wood fire smoke, would go blind. 37 She nearly froze to death on the road

36
37

Se fa mkha' 'gro, Autobiography, fols. 74a-74b. Se ra mkha' 'gro, Autobiography, fols. 76a-76b.

130

Sarah H. Jacoby

due to inadequate clothing, and as a result, sores erupted all over her body.38 When she caught up with Dri med 'od zer's group at Mtshur phu, her presence was met with hostility from Dri med 'od zer's encampment members (sgar pa) who viewed her suspiciously as an outsider to their religious community.39 As a! Central Tibetan amidst Mgo log people whose speech she initially describes as that of "savage barbarians" (mtha' 'khob pa kla klo) and as a young unattached female, Se ra mkha' 'gro's presence threatened many around her who feared she would become their lama's consort.40 Once they arrived at Dri med 'od zer's homeland in the high pasture land of Zlar tshang in Gser rta, Dri med 'od zer's travelling pilgrims all returned to their homes while Se ra mkha' 'gro was left alone in a foreign land where she couldn't understand the dialect, was illiterate in Tibetan, and had no work experience. Initially, she tried to rent a room at Dri med 'od zer's household (bla brang), but others forced her to leave out of concern that Dri med 'od zer's consort A skyong bza' wouldn't return to the household unless Se ra mkha' 'gro left. Expelled from another home, Se ra mkha' 'gro managed to learn how to be a servant in a nearby nomadic household while she began religious practices under the tutelage of Dri med 'od zer and several of his brothers.41 Her devotion to her lamas, ascetic perseverance in completing her preliminary religious practices, and eloquent and poetic religious songs quickly demonstrated that she was no ordinary laywoman.

38 Se ra mkha' 'gro, Autobiography, fol. 78a. 39 Se ra mkha' 'gro's identity as a Central Tibetan woman and not an insider to the Mgo log religious encampments (chos sgar) in which she lived remained salient throughout her life and beyond, as is evident by the name still used in Mgo log today to refer to her, Dbus bza' mkha' 'gro, "The pakinf Lady from Central Tibet." 40 I follow Se ra mkha' 'gro's lead of describing Dri med 'od zer and his companions as people from Mgo log, but more specifically they were people from an area neighboring what is now called Mgo log that during Se ra mkha' 'gro's day was controlled by the Dbal shul tribal group. 41 Se ra mkha' 'gro, Autobiography, fol. 87a.

This inferior female body

131

Despite this, as one with no biological family relations nearby, no monastic vows, and no husband, Se ra mkha' 'gro didn't easily fit into any community in Mgo log. Jealous consorts made life difficult in the non-celibate religious communities (chas sgar) that were often centered around Treasure revealers (gter stan).' When the seventeen-year-old Se ra rokha' 'gro received a letter summoning her to the presence of a prominent Treasure reveaier named Mgar ra gter chen pad ma bdud 'dul dbang phyug gling pa of Ban nag Monastery in what is today Pad ma County, Mgo log Tibetan Autonomous Prefecture, she yearned to leave her post as a servant girl to attend to the lama's needs. However, before she began her journey, she received a message from Mgar ra gter ston's consort Yag shul bza' threatening her with harsh words and warning her hot to come near the lama. Yag shul bza' knew that Mgar ra gter ston had received a prophecy from the rjakinfs indicating that Se ra mkha' 'gro could help remove the obstacles to his life and cure his worsening illness by acting as the lama's consort, but she didn't have any interest in competing for her partner's attentions with the young, beautiful, and increasingly renowned rjakinf from Central Tibet. When Mgar ra gter ston was on the verge of death, Yag shul bza' finally relented. But since Se ra mkha' 'gro arrived too late, the auspicious connection (rten 'breI) her presence could have brought .to preserve the lama's life did not occur and he died. 42

42 In addition to Se ra mkha' 'gro's Autobiography, at least two other publications mention this incident including 1) the Dar thang dgon pa'i gdan rabs and 2) Mgar ra gter ston's autobiography tit1~d Mgar gter chen pad ma bdud 'dul dbang phyug gling pa'i rang rnam. Mgar ra gter ston's short biography in Dar thang dgon pa'i gdan rabs (0 rgyan brtson 'grus 1999: 379) states that Mgar ra gter ston died after "the auspicious connections of the $upreme consort in the tjakinf prophecies named the Central Tibetan Lady mkha' 'gro chos nyid bde skyong dbang mo [i.e., Se ra mkha' 'gro] were confused." Secondly, Mgar ra gter ston's disciple named Thub bstan chos 'phe1 concludes Mgar ra gter ston's autobiography by giving a detailed account of how his consort Vag bza' was a bad person who "wanted him to die" (zhing khams la 'dun pa gtod rgyu yin) and did whatever she could to harm him such as preventing Se ra mkha' 'gro from dispelling the obstacles to his life (Mgar gter chen pad rna bdud 'dul dbang phyug gling pa 2005: 161-170).

132

Sarah H. Jacoby

Immediately afterwards, Se ra mkha' 'gro had a visionary experience in which the newly departed Mgar ra gter ston empowered her in his Treasure cycles and designated her as the appropriate person to maintain and protect his teachings. Se ra mkha' 'gro describes the following interaction she had with his visionary form after he empowered her: "Because of my inferior female body (Ius dman), it is difficult for me to benefit beings in the world. Hence, having abandoned this body, if I try to attain a man's body, I wonder if I will benefit beings?" The lama said, "Don't think like this. This body you have is that of a great female bodhisattva, the mother who gave rise to all the Buddhas. It is the chariot that traverses the grounds and paths, the . foundation that propagates men who adhere to Sutra and Tantra. Like snow comes from water, the expansive space of the greatly secret empty realm of the mother [i.e., the womb], the space (mkha') of reality itself in which the energy of compassion goes Cgro), is the source of both saf!!siira and nirviilJa. Hence, this body of yours is the superior [way to] benefit beings. If you wonder how it is that it is superior, before, during the time of the previous Buddha Kasyapa, when you were the daughter of a Brahmin named Nam mkha' sgron rna, based on your pure aspiration to generate the supreme awakening and by the power of your prayers to the Lady Mtsho rgyal, you dispelled all manifest and unmanifest negative obstacles to the life forces of men who adhered to the teachings. You raised the victory banner of the teachings. Via skillful means (thabs) and auspicious connections (rten 'brei), you propagated the happiness of beings and established them in bliss. In particular, be like a mother to exhausted sentient beings and it is certain that inconceivable benefit for beings will gradually arise. 43 Mgar ra gter ston's response to Se ra mkha' 'gro's despair transforms her negative valuation of her female body by honoring its extraordinary qualities, qualities that reflect a Tantric perception of the female body's virtues centered on the womb. The womb is the source of the Buddhas, which Mgar ra gter ston metaphorically links to the Tibetan word for cjakinf (mkha' 'gro), or sky-goer. He specifies that Se ra mkha' 'gro's previous incarnation Nam mkha' sgron rna dispelled obstacles to the life force of religious men, an Se ra mkha' 'gro, Autobiography, f01s. 105a6-106a2.

43

This inferior female body

133

ability which refers to her capacity to engage in consort practices. This equation between sexuality and longevity that Mgar ra gter ston's comment and the stories surrounding his death evokes persists in Se ra mkha' 'gro's writings as one of the main purposes for consort practices. 44 Despite the unfriendly welcome Se ra mkha' 'gro initially received at Ban nag Monastery, she soon settled with Mgar ra gter ston's son named Rgyal sras because she received a prophecy indicating that she should do so for the sake of preserving Mgar ra gter ston's Treasure teachings. By the time she was nineteen in approximately 1911, Se ra mkha' 'gro's extraordinary abilities as a rJiikinf including curing people of sickness, performing divinations, and experiencing many divine visions earned her increasing acclaim in the wider community around Ban nag Monastery. Nevertheless, her former nemesis Yag shul bza', who was now effectively her mother-in-law, strove to expel her from the community. Rgyal sras was soon influenced by Yag shul bza's antipathy towards Se ra mkha' 'gro. He belittled her religious understanding, saying she was unfit to receive esoteric instructions. He rejected her claim to be a Treasure revealer as her own mother had earlier, stating that,
Before, during the time when there were one hundred great Treasure revealers, two of them actually manifested as female Treasure revealers (gter rna). These days in this Degenerate Age, [we] don't see the existence of female Treasure revealers. 45

After this, Se ra mkha' 'gro writes that, "From that time forward, I wrote my Treasures down secretly.''46 Soon, Se ra mkha' 'gro became pregnant with Rgyal sras's child, who she intuited was an
44 See Chapter Four of Jacoby 2007 for a discussion of consort practices in Se ra mkha' 'gro's auto/biographical writings. I have delineated three major purposes for consort practices including soteriological, pragmatic, and hermeneutical, but pragmatic purposes such as prolonging life and curing illness are the most common. 45 Se ra mkha' 'gro, Autobiography, fo1. 141a3-4. The two female Treasure revealers he refers to here are Jo mo sman mo and Kun dga' 'bum ma, both mentioned in 'Jam mgon kong sprul bio gros mtha' yas's (1813-1899) Gter ston brgya rtsa (The One Hundred Treasure Revealers). 46 Se ra mkha' 'gro, Autobiography, fo1. 141a5.

134

Sarah H. Jacoby

incarnation of his father Mgar ra gter ston. However, Se ra mkha' 'gro records that the child's fate changed when he transformed from a boy into a girl in utero:
That year in the first month of fall when it became time to see the face of my child, I went to collect firewood near a man named Chos lhun who Mgar ra gter chen previously [said was] one who adhered to demonic scriptures. Because he called out in a fierce voice and performed curses and so forth, the protection of my child was reversed and he became a girl. At that time Rgyal sras felt great regret. Saying, "If this girl dies I'll be happy," he treated her really badly. Although I wasn't pleased with this, I didn't say anything. 47

The opinion of Se ra mkha' 'gro's milieu about having a female child is clear, given that Se fa mkha' 'gro attributes the birth of a daughter to demonic intervention. After the child's inauspicious sex transformation, there is no more discussion that the child might be Mgar ra gter ston's incarnation. Despite this, still feeling that her daughter was extraordinary, Se ra mkha' 'gro took the baby to a lama who said the child had great merit and named her Dbyangs can sgron ma. 48 After Se ra mkha' 'gro gave birth to her daughter, she writes that a Ma gcig lab sgron Accomplishment Practice (Skt. sadhana) came to her as a Treasure, but because she secretly taught it to some monks, Rgyal sras scolded her. She writes that she felt extremely sad and then supplicated the heroes and rjakinfs of the three realms with the following prayer:
In the unmanifest land of the Glorious Copper-coloured Mountain, Omniscient Mother Ye shes mtsho rgyal and Heroes and Heroines, consider this! I, this messenger sent by the mother rjiikinfs, Was born in an impure body as a woman (za rna rno). My impure action is that I maintain a household. My impure land is the demonic land of Mgo log. My impure deed is that I continually stay with a negative community. From what causes and conditions did this karma originate? ... 49
47

48

Se ra mkha' 'gro, Autobiography, fols. 146a5-146b2. Se ra mkha' 'gro refers to her daughter more commonly as Chos dbying Se ra mkha' 'gro, Autobiography, fo1s. 147b6-148a2.

sgronrna.
49

This inferior female body

135

In response to her song of lamentation, Se ra mkha' 'gro heard a voice with no physical form coming from the sky saying,
Alas, alas yogin! Is the great self-emergent primordial wisdom Free from exertion? Is the unborn self-emergent spontaneous presence The ground of sarrrsara and nirvalJa? Are the appearances of good and bad, happiness and sorrow The play of reality itself? Is the poisonous water-lily which is the eight worldly dharmas 50 Something to be abandoned quickly? Your wealth of the tjakinf's heart essenceWill you take it out in due time? Will benefit for both the teachings and sentient beings Corne from the south? You don't need to be displeased and weary On account of the illusory emanations of this world. The Lotus [Born One's] testament is changeless; It is certain that in time the aspiration prayers will corne together. 51

In this way divine voices weave their words of encouragement through Se ra mkha' 'gro's narrative; they provided support and consolation just when her mundane circumstances seemed to overwhelm her with obstacles and difficulties. Despite her best efforts . . to "take a low place and write down the profound Treasures that came to me extremely secretly,"52 living with her "life partner" (tshe grogs) Rgyal sras proved to be at loggerheads with her main purpose of revealing Treasures. To Rgyal sras's chagrin, by 1915 the twenty-three-year-old Se ra mkha' 'gro began to receive broader acclaim as a Treasure revealer and as a ejiikinl. The prominent lama from Dpal yul dar thang Monastery in Mgo log named Sgo sprul rin po che ('Jigs bra! chos
50 The eight worldly dharmas refer to ordinary human reactions to 1) gain and loss, 2) fame and infamy, 3) praise and blame, and 4) happiness and sadness. 51 Se ra mkha' 'gro, Autobiography, fols. 148b5-149a3. 52 Se ra mkha' 'gro, Autobiography, fol. 144a4.

136

Sarah H. Jacoby

kyi blo gros) was one of the first people to recognize Se ra mkha' 'gro as an incarnation of Ye shes mtsho rgyal and to encourage her to reveal Treasures. Se ra mkha' 'gro portrays their first interaction in the following passage:
Then, I gave [Sgo sprul rin po che] a high quality turquoise wrapped in five-coloured silk along with a silk scarf with auspicious symbols on it as offerings and I prostrated to him. When I asked him, "How are you?," the incarnate himself got up from the middle of the assembly and offered me silver and a silk scarf. The lamas, monks, and laymen and women gathered there all didn't know what was going on. Each one looking at the other, they didn't say anything. Again, the incarnate said, "You all don't need to have doubt. I think this rjakinf is certainly the speech incarnation of Ye shes mtsho rgyal. You all abandon your wrong views and be faithful. This is not a deception; it is real." I thought, "Before [now] he has never met me or those affiliated with me, so how can he know that I am Mtsho rgyal? Not only that, how is it possible for someone like me - who has an inferior female body (Ius dman), who on the outside has little learning and on the inside is ignorant with regard to generation and perfection stage practice, and who cannot move beyond sa~saric work no matter what activity I begin - to be a rjakinf?" When I wondered this, he said, "Not needing to exert yourself in learning and training, you fully comprehend all the paths of generation and perfection. Not needing to learn how to write and recite texts, [you know] from the awakening of the residual karma of having learned [before]. Nevertheless, not letting your own Treasures go, you must write and propagate them or the rjakinfs' punishment will quickly come." I said, "I am not a Treasure revealer. I have never attained even one type of dharma teaching and I don't know about generation and perfection [stage practices]."53

Se ra mkha' 'gro thus vociferously denies his lauding words, maintaining her characteristic humility through reference to her "inferior body," at the same time as her autobiography proclaims
53 Se ra mkha' 'gro, Autobiography, fols. 162a5-163a4.

This inferior female body

137

her extraordinary status through Sgo sprul rin po che's voice. Nevertheless, after denying her abilities, Se ra mkha' 'gro hints that this was riot the only outcome of their meeting by writing, "Not keeping anything that I knew secret, I spoke to him and we nurtured each other's practices."54 Another turning point occurred in Se ra mkha' 'gro's twentythird year when Rgyal sras, herself, and the first object of her devotion Dri med 'od zer met at a consecration ceremony for a new prayer wheel sponsored by a lama named Hor 'od chung in A mdo. Se ra mkha' 'gro writes that at the ceremony, a beautiful bejewelled woman saying she was Rdo rje gyu sgron rna appeared before her and gave her a prophecy and a white crystal vessel. Before she could hide the Treasure vessel, Dri med 'od zer saw it and insisted that she give it to him. Although she received many signs a_nd prophecies indicating that her choice to maintain a contentious relationship with Rgyal sras instead of living with Dri med 'od zer was like choosing brass over gold, and although she recounts that she "... wanted to see him [Dri med 'od zer] like a person suffering from thirst wants water,"55 it was Rgyal sras who decided her future. Se ra mkha' 'gro explains that in conversation with Dri med 'od zer, Rgyal sras offered her to be his consort, saying:
"... she is in a relationship with me [Rgyal sras] because of my father's former aspiration prayers - she is not one who is prophesied for me. Perhaps she is your prophesied consort. She constantly and singlemindedly prays to you and your father [Bdud 'joms gling pal In particular, when she hears your name tears flow uncontrollably from her eyes and, with goose bumps on her body, she prays in a loud voice and sings hymns of sadness. In any case, if [you] do a retreat [with her], perhaps obstacles [regarding] your retinue and life force will be dispelled. I offer you this consort (mkha' 'gro) of mine - aside from you, she has no husband. 56 She has a gentle disposition, a lucid intellect,

54 Se ra mkha' 'gro, Autobiography, fo1. 163b4-5.

Se ra mkha' 'gro, Autobiography, fo1. 155b3-4. 56 The multiple meanings of the word I translate here as "husband," Tib. bdag po, include "owner, lord, master, boss, husband, and guardian." Hence, the word implies that she is now under Dri med 'ad zer's control, although Se ra mkha' 'gro does not mention any formal marriage rite aside from this
55

138

Sarah H. Jacoby and she doesn't assochite with other men."57

At first, Dri med 'od zer responded to Rgya1 sras's offer hesitantly, mentioning concern over whether his other consorts would accept Se ra mkha' 'gro's presence at his household, which had already been a problem when she first arrived there at the age of fourteen. Additionally, he wondered whether having another consort was in accord with his intention to enter strict retreat for a few years. Rgyal sras insisted, and thus Se ra mkha' 'gro was exchanged between two powerful male lamas without having a say in the matter. Despite Se ra mkha' 'gro's devotion to Dri med 'od zer, she writes of her dismay at being given away. Indeed, she did not leave Rgyal sras for another six years, during which time troubles pervaded her life at Ban nag Monastery. Her rjiikinf visions and several visits on horseback over the mountain passes and valleys separating Ban nag from Zlar tshang to see Dri med 'od zer were her only consolations during this turbulent time. For example, when she was twenty-four in 1916, Se ra mkha' 'gro had a vision in which rjiikinfs transported her to Padmasambhava's Buddhafie1d, the Coppercoloured Mountain. There Se ra mkha' 'gro lamented her fate to be born in a female body again, telling a rjiikinf,
I abide in an inferior corporal body. Although I have good intentions, it is difficult to tame beings in this negative time. I have met with Treasures, yet the auspicious connections of skillful means are bad. Because of my limited knowledge, it is difficult to understand the meaning of the symbols. This inferior female body (skye lus dman pa) cannot get free from the mouth of my husband; When I wander about this country, it can't get free from the mouth of ,dogs. Whatever I do, people don't like me. When things like these occur, I am one who must abandon this body.

agreement between the two men. 57 Se ra mkha' 'gro, Autobiography, fols. 164a5-164b3.

This inferior female body

139

Even though there is no difference in the mind of enlightenment (Skt. bodhicitta) between men and women, In the perspective of disciples, my body is inferior. I think that if I were to transform into a hero with a supreme body, I certainly will accomplish the great purpose of benefiting [all]beings who have been my mother. Without this, how will this body Have a way to accomplish great benefit for the teachings and sentient beings? Lady, give me advice. 58

In response, the fjiikinz gave her a detailed prophecy in which she taught her about types of fjiikinz script and foretold her future doctrine holders, consorts, and teachings. Nevertheless, in her quotidian life, Se ra mkha' 'gro had good reason to perceive her female body to be a source of suffering. In 1918 when she was twenty-six years old and still living with Rgyal sras, community tensions at Ban nag Monastery had escalated. Rgyal sras had taken on a second wife named Gsal sgron, about whom Se ra mkha' 'gro writes nothing positive. Not only were Rgyal sras and Se ra mkha' 'gro quarrelling, but two factions of the religious community at Ban nag were disputing about land rights, which was a common source of strife in the nomadic pasture lands of Mgo log. Se ra mkha' 'gro records that as a result of this social tension, her unborn child, who was an incarnation of Rgya rong gter smyon, died and she gave birth in great pain to a stillborn son. Aside from three members of Rgyal sras's family who liked her, she writes that her detractors headed by Yag bza' were happy and said,
"The Central Tibetan girl's merit is exhausted and her son died. Still what misery doesn't she give to the Mgo log Lady 59 and Rgyal sras? All that she says about her son dying in her stomach and giving birth to the corpse with its head and feet un-reversed (i.e., breech) is certainly lies. Whenever a calf dies in a female yak's stomach [everyone] can see that both mother and baby die. Her son was not a corpse in

58 59

Se ra mkha' 'gro, Autobiography, fols. 174b5-175a4. "Mgo log Lady" probably refers to Rgyal sras's new spouse Gsal sgron.

140

Sarah H. Jacoby his mother's stomach; he didn't die [in the womb]. Because she didn't die, there is absolutely no way she could have gotten the baby's corpse
OUt."60

Yag bza' and her cohorts essentially accused Se ra mkha' 'gro of murdering her son because they could not believe cshe could have given birth to a stillborn child without dying herself. Only one year later, she became pregnant again and gave birth to a son named Rig 'dzin 'gyur med rdo rje, who she describes as an incarnation of Sgo sprul rin po che, the Dpal yul dar thang lama who recognized her a few years earlier as a r;liikinf incarnation of Ye shes mtsho rgyal. But her troubles only grew worse. Her infant son became ill and Se ra mkha' 'gro was increasingly afflicted with an arthritic condition causing swelling in her legs Cbam) that left her with greater and greater difficulty walking. In 1921 when she was twenty-nine years old, Se ra mkha' 'gro's social world was also affected by another crisis: the "northern Chinese" (byang rgya) were creating havoc Cur 'ur zing zing byed), causing both Dri med 'od zer's and Rgyal sras's religious encampments to move repeatedly in order to avoid battles. 51 With all this internal physical and external social commotion swirling around her, Se ra mkha' 'gro still returned to Rgyal sras's community after her pilgrimage journeys to see Dri med 'od zer, even though she knew that Dri med 'od zer was the prophesied consort she needed to reveal her Treasures. piikinfs who manifested before Se ra mkha' 'gro made this imminently clear, such as the r;liikinf who appeared in a dream and pronounced in no uncertain terms that her time with Rgyal sras was over when she was enroute returning to him:
"Why are you going towards those with deteriorated commitment vows? It is as if you have mistaken brass for gold, water for wine. You throwaway your destined bodhisattva as unnecessary. You toss beSe ra mkha' 'gro, Autobiography, fols. 215a4-215b2. The "northern Chinese" (byang rgya) refers to the Hui Chinese Ma family warlords who were active in Northwestern China during the Republic of China era including Ma Qi (1869-1931), Ma Lin (1873-1945), and Ma Bufang (1903-1975). For more detail on this 1921 invasion of Mgo log by the Ma warlords, see Don grub dbang rgyal/Nor sde 1991: 156-158.
50 51

This inferior female body

141

hind you [the responsibility of] maintaining your profound Treasure . . You are distracted holding on to cyclic existence. From the time you were young lintil now, I have given you honest advice. I have given you the quintessence of the profound Dharma Treasures. Although I have reared you like a mother who loves her child, repelling negative conditions, outer and inner obstacles, and so forth, still you are unable to be independent and you need to be under the power of another. What is the meaning of thiS?"62

Although Se ra mkha' 'gro replied that she was only returning to Rgyal sras because she was trying to fulfil her obligations to maintain Mgar ra gter ston's Treasure teachings, the fierce 4akinfwould have nothing of this and pronounced, "From this year forward, you belong to us. The time for you to live with Mgar ra is finished."63 As the 4akinf predicted, after returning to Rgyal sras, the twentynine-year-old Se ra mkha' 'gro's arthritic illness became severe and when she was completely unable to walk and on the verge of death, her disciples could bear it no longer. Everyone, including Rgyal sras, agreed that Se ra mkha' 'gro along with her young daughter and son should be sent to live with Dri med 'od zero Having been carried on horseback by her disciples, Se ra mkha' 'gro was finally reunited with Dri med 'od zer in 1921 at a place called Nyi rna lung in the Zlar tshang region where her life in Mgo log had begun fifteen years earlier. Dri med 'od zer, who himself had become increasingly ill, treated her illness by performing blessing rituals. After several months of strict retreat, she gradually regained her physical strength. Se ra mkha' 'gro describes amazing outer and inner signs of realization that emerged during this time of intense practice in retreat with Dri med 'od zer:
At that time, the signs in the outer environment were that day and night the mountains spoke, the earth quaked, a roaring was heard, the sound of a flute and cymbals came forth, a fragrant smell spread, and so forth. The whole group of fortunate disciples actually saw this. The inner signs that the domains of the channels, winds, and seminal essence were becoming purified were that the knots of my head and

62

Se ra mkha' 'gro, Autobiography, fols. 254a6-254b6.

63 Se ra mkha' 'gro, Autobiography, fol. 255a5.

142

Sarah H. Jacoby

throat became liberated. Not needing to train in and traverse the path, I understood all the Dharmavehic1es completely. The essential meaning of the symbols of all the teachings in accordance with the disposition and devotion of beings from the hidden vessel of the intention of the profound Treasures became self-liberated. The ten w:inds became pliant and our bodies had youthful flesh and were light like cottonwool. When I spoke spontaneously-uttered hymns of vajra words, the seeds of the first concentration were planted in the mind-streams of those who heard them. By many small deeds such as this, benefit for sentient beings arose. In particular, the inexpressible understanding of awareness and emptiness that is the primordial wisdom of great bliss and emptiness without elaboration endowed with the three liberations 64 manifested [in me]. The minds of both of us, method and wisdom, actually merged into one taste. 65

Se ra mkha' 'gro thus describes the moment of her spiritual liberation as one in which method and wisdom, Dri med 'od zer and herself, became utterly inseparable. Not only was her body cured from illness as realization dawned, but she and Dri med 'od zer "became extremely quick in decoding each other's Treasures."55 Thus, Se ra mkha' 'gro portrays her relationship with her lama and prophesied consort Dri med 'od zer as one of mutual benefit in which both male and female partners relied on each other to reveal and decode Treasures, cure illness, and dispel obstructions. Tragedy struck when Se ra mkha' 'gro and Dri med 'od zer had lived happily together for only a few years. A plague swept through Dri med 'od zer's religious encampment in 1924, killing both Dri med 'od zer and Se ra mkha' 'gro's five-year-old son 'Gyur med rdo rje within the same week. She writes that Dri med 'od zer "put his hands on my head, stared into space, and having prayed [for us] to arise as Vajradhara in the [form of] the non-dual union of method and wisdom, his form body departed for the expanse of reality."67 Devastated with grief, Se ra mkha' 'gro was again utterly homeless when Dri med 'od zer's other consorts expelled her from his
The three liberations are emptiness, signlessness, and wishlessness. 55 Se ra mkha' 'gro, Autobiography, f01s. 262b4-263a6. 56 Se ra mkha' 'gro, Autobiography, fol. 264a2. 67 Se ra mkha' 'gro, Autobiography, fo1. 295a2-4.
54

This inferior female body

143

household (bla brang) before the embers of his funeral pyre had grown cold. Dri med 'od zer's close disciple Bsod sprul sna tshogs rang grol, the incarnate lama from Se ra theg chen chos 'khor gling Monastery of Gser rta County, kindly invited Se ra mkha' 'gro and her entourage including her daughter Chos dbyings sgron rna, her attendant Thub bstan bzang po, and her scribe Tshul khrims rdo rje to live at Se ra Monastery. They stayed there for several years, earning Se ra mkha' 'gro her nickname "the /)akinf of Se ra." In the years following Dri med 'od zer's death when Se ra mkha' 'gro was living at Se ra, she wrote first Dri med 'od zer's biography, which she explains came to her from tjakinfs as self-emergent symbols, and then her own long autobiography, which she authored as her own composition. She frequently travelled from Se ra Monastery around Mgo log and Khams with her attendants, empowering monastics and lay people in the Treasure teachings of Bdud 'joms gling pa, Dri med 'od zer, and her own Treasures. During this time, she continued to encounter visionary tjakinfs who urged her on with prophetic encouragement to accomplish her life's purpose of revealing Treasures. When she was forty-one, she spent several months gathering and editing Bdud 'joms gling pa and Dri med 'od zer's entire collection of Treasures (gter chos) that totalled forty volumes by themselves and forty-four volumes with the inclusion of her own revelations. 68 Se ra mkha' 'gro's autobiography ends when she was forty-two in 1934, six years before her death in 1940. She died in Ri bo che at the estate of her disciple, the Ri bo che Zhabs drung Tshe dbang grags pa. 69 Her disciple Bya bral sangs rgyas rdo rje adds' that when they brought her body out to the cremation site, "Her body had dissolved into light and all that remained was about the size of an eight-year-old child."70

Se ra m kha' 'gro, Autobiography, fol. 388b4-6. A first-hand account of her death can be found in Tshul khrims rdo rje 1948: 9-12. See Dbus bza' mkha' 'gro 2009 (vol. 1): 538-547 for a published version of this text. 70 Bya bral sangs rgyas rdo rje 1976: 15a5-15b4. This would be seven years old according to international convention.
68

69

144 Conclusion

Sarah H. Jacoby

That Se ra mkha' 'gro's female gender was a salient issue for her and others in her communities is clear by the frequency with which she mentions her "inferior female body," only a few examples of which I depict in this article. I suggest that we can interpret the more than thirty instances in which Se ra mkha' 'gro describes herself as "an inferior woman" in two ways: both as a very literal internalization of misogynist views regarding women's religious aptitude and as a rhetorical device that enabled her to make extraordinary claims about herself. Through harping on her "inferior female body," Se ra mkha' 'gro successfully composed an autobiography that was both humble enough to be accepted by many in her androcentric religious milieu, yet assertive and confident enough to confirm her status as a woman with religious authority. This feature of Se ra mkha' 'gro's writing bears resemblance to some female Medieval Christian mystics' writings in which female authors had a general tendency to use their femaleness as a way to accentuate their humility. For example, Hildegard of Bingen describes herself as "a weak and fragile rib;" Clare of Siena as "a useless handmaid and unworthy servant;" and Julian of Norwich as "a woman ignorant, weak, and frail."71 Monica Furlong attributes this "ritual obeisance" of gender-based self-denigration in Medieval women mystics' writings as a reflection of a socially imposed sense of inferiority as much as it was a literary device that helped their divine mystical visions to be heard, an interpretation I suggest can be applied to Se ra mkha' 'gro's writings as well. Closer to home in Tibetan Buddhist literature, we can also find precedence for the way in which Se ra mkha' 'gro expressed humility through referring to her female form. For example, Kurtis Schaeffer finds a strong connection between suffering and the female 'body in the autobiography of a nun from Dolpo, Nepal named o rgyan chos skyid (1675-1729). Just as 0 rgyan chos skyid prayed to be reborn as a man and never again as a woman because "This female body is itself sarrzsiira - the round of existence," Se ra

71 Furlong 1996: 14.

This inferior female body

145

mkha' 'gro described her female body as a "contaminated, unclean, negative body" that was "the foundation of cyclic existence."72 This association between suffering, saf[lsara, and the female body has ancient Buddhist roots that hardly served the interests of women, but both 0 rgyan chos skyid and Se ra mkha' 'gro managed to turn this negative valuation of the female body from a vice to a virtue. 73 Orgyan chos skyid's analogy between saf[lSara and her female body put her in a very good position to claim direct experience of the Buddhist truth of suffering. 74 Se ra mkha' 'gro's descriptions of herself as an unlearned woman, as a tramp expelled from empowerment lines, and as an unhappy wife subservient to Rgyal sras had the subversive effect of occasioning others' powerful praise. In particular, the authoritative others who respond to Se ra rokha' 'gro's laments are often r;lakinfs. pakinfs unapologetically assert what Se ra mkha' 'gro herself cannot, that she is an emissary of Ye shes mtsho rgyal working for the benefit of sentient beings by revealing and teaching Padmasambhava's Treasure teachings. However ephemeral their appearances may be, r;lakinfs provide direction and protection for Se ra mkha' 'gro just when she needs it most. pakinfs do not try to destabilize her sense of self, as they do in at least one male-authored Tibetan autobiography,?5 More than being "a powerful representation of the repressed feminine aspects of the male psyche," as some Western scholars interpret them, 72 Schaeffer 2004: 143; Se ra mkha' 'gro, Autobiography, fol. 13b2. It is important to note, however, that not all Tibetan women's autobiographies reference the female body in the same way. For example, the autobiography of the nun and Abbess Rje btsun 10 chen rin po che (Rig 'dzin chos nyid bzang mo, 1865-1951) of Shug gseb Nunnery does not emphasize her humility through repeated references to her female body. See Havnevik 1999 for an English translation and study of this autobiography. 73 See Wilson 1996 for a study on the ways in which the female body has been portrayed as horrific in early Indian Buddhist scriptures written by and for monks. 74 Schaeffer 2004: 103. 75 Gyatso finds that in the context of the 18th -century 'Jigs med gling pa's autobiographical writings, iikinfs' words bolster his sense of himself at the same time as they destabilize it, contributing to the ultimate undecidability of his experience (Gyatso 1998: 254).

146

Sarah H. Jacoby

tjiikinfs act as enlightened female role models whose main func-

o rgyan chos skyid as well, tjiikinfs are a major source of encour-

tion is to encourage Se ra mkha' 'gro's religious yearnings.76 For

agement; tjiikinfs even over-ride her master's refusal to help her write by bestowing literacy upon her.77 Perhaps we can conclude that marshalling the tjiikinfs' support was especially important for female autobiographers in the Tibetan Buddhist world who had so few human female role models from whom to derive inspiration. Despite the potentially subversive elements of Se ra mkha' 'gro's treatment of her "inferior female body," we must ask ourselves what effect if any this subversion had for those who inherited her religious lineage and/or copies of her manuscripts? I suggest that the long list of women in Mgo log and beyond who have been recognized as incarnations of Se ra mkha' 'gro including Rdzong sprul rin po che's daughter Lha leam chos kyi sgron ma, Ta: re lha mo (1938-2002), and Mkha' 'gro rin po che (b. 1954) demonstrates that Se ra mkha' 'gro's example did help pave the way for women who came after her to claim religious authority.7B After all, tjiikinfs supportive words spoke not only to Se ra mkha' 'gro, but to all female and male disciples who read her autobiography, exhorting them to see the female body as "the mother [who gives birth to] the great assembly of primordially perfected Ones Gone to Bliss."79

76 Sponberg 1992: 28. See Herrmann-Pfandt 1990, 1992-93 and SimmerBrown 2001 for comprehensive studies on the significance of ej,akinfs in Tibetan Buddhism as well as interpretations of the diverse ways in which Western scholars have analyzed them. In the case of Se ra mkha' 'gro, her writing affirms that ej,akinfs are far more than symbols serving the needs of the rp.ale yogic subject. On the other hand, the prevalence of ej,akinfs in Se ra mkha' 'gro's writing does not imply that human women had an exalted status in communities practicing Tantric Buddhism. 77 Schaeffer 2004: 132. 7B Pad rna 'ad gsal mtha' yas 2003: 73 and 1997: 110. Outside of Mgo log, the daughter of Bya bral sangs rgas rdo rje rin po che of Pharping, Nepal named SaraswatI is widely thought to be an incarnation of Se ra mkha' 'gro. 79 Se ra mkha' 'gro, Autobiography, fol. 50a4-5.

This inferior female body

147

Bibliography

Tibetan sources
Mkha' 'gro bde skyong dbang mo (1892-1940). Skyabs rje thams cad mkhyen pa grubs pa'i dbang phyug zab gter rgya mtsho'i mnga' bdag rin po che Pad ma 'gro 'dul gsang sngags gling pa'i rnam par thar pa snying gi mun sel dad pa'i shing rta ratna'i chun 'phyangs utpala'i phreng ba. Dalhousie: Damchoe Sangpo, 1981. Mkha' 'gro bde chen dbang mo (b. 1868). Mkha' 'gro rgya mtsho'i rnam thar. Dolanji: Tibetan Bonpo Monastic Community, 1985. Mgar gter chen pad rna bdud 'dul dbang phyug gling pa (1857-1910). Mgar gter chen pad ma bdud 'dul dbang phyug gling pa'i rang rnam, Stag lung dgon gyi sde snod rigs pa'i dpe tshogs (ed.). Si khron: Si khron zhing chen khron lin par 'debs bzo grwa, 2005. 'Jam mgon kong sprul blo gros mtha' yas (1813-1899). Dpal snga 'gyur rnying ma'i gter ston brgya rtsa'i rnam thar bai d'urya'i phreng ba. Odiyan: Dharma Publishing, 2004. Stag sham nus Idan rdo rje (b. 1655). Bod kyi jo mo ye shes mtsho rgyal gyi mdzad tshul rnam par thar pa gab pa mngon byung rgyud mangs dri za'i klu phreng. Kalimpong: Zang mdog dpal ri Monastery, 1972. Bde ba'i rdo rje (1892-1940). Dbus mo bde ba'i rdo rje'i rnam par thar pa nges 'byung 'dren pa'i shing rta skalldan dad pa'i mchod sdong. Unpublished manuscript. ca. 1934. - - Mtsho rgyal dngos sngang bde skyong dbang mo yi zab gter gsang ba'i chos mdzod rin po che. 4 volumes. Kalimpong: Dupjung Lama, 1978. Don grub dbang rgyal & Nor sde (b. 20th century). Mgo log 10 rgyus deb ther. In Mgo log rig gnas 10 rgyus, vol. 1. Xining: Srid gros Mgo log khul u rig gnas 10 rgyus rgyu cha zhib 'jug u yon lhan khang, 1991. Pad rna 'od gsal mtha' yas (b. 1957). Nam sprul 'jigs med phun tshogs dang mkha' 'gro Tii re Iha mo'i rnam thar. Chengdu: Si khron mi rigs dpe skrun khang, 1997. - - Gter chen bdud 'joms yab sras kyi rnam thar. Chengdu: Si khron mi rigs dpe skrun khang, 2000 . .. - - Deb chung a ru ra'i dga' tshal. Chengdu: Si khron mi rigs dpe skrun khang, 2003. Bya bral sang rgyas rdo rje (b. 1913). Dbus bza' mkha' 'gro bde ba'i rdo rje'i rnam thar chen mo'i mjug gi kha skong nyung du gyu yi phra tshom. Unpublished manuscript, 1976. Dbus bza' mkha' 'gro (1892-1940). Snang sbyang dbus 'grel. Beijing: Mi rigs dpe skrun khang, 2002.

148

Sarah H. Jacoby

--Dbus bza' mkha' 'gro'i gsung 'bum. Mgo log khul gna' rtsom bya ba'i gzhung las khang (editor). Si khron: Si khron mi rigs dpe skrung khang,
2009. Tshul khrims rda rje (19 th-20 th century). Rgyal yum ye shes mtsho rgyal mi

mo'i rnam thar bzhengs pa mkha' 'gro rin po che bde ~a'i rdo rje'i 'das rjes kyi rnam thar don bstan su bkod pa. Unpublished Manuscript, 1948. rgyan brtson 'grus (b. 20 th century). Dar thang dgon pa'i gdan rabs. Chengdu: Si khron mi rigs dpe skrun khang, 1999.

Western language sources

Allione, Tsultrim. 1984. Women of Wisdom. London: Routledge & Kegan Paul. Aziz, Barbara Nimri. 1987. Moving towards a sociology of Tibet. In Janice D. Willis (ed.) Feminine Ground: Essays on Women and Tibet. Ithaca: Snow Lion Publications. Cuevas, Bryan. 2008. Travels in the Netherworld: BuddhistPopular Narratives of Death and the Afterlife in Tibet. New York: Oxford University Press. Diemberger, 2007. When a Woman Becomes a Religious Dynasty. New York: Columbia University Press. Dowman, Keith. 1984. Sky Dancer: The Secret Life and Songs of Ye shes mtsho rgyal. London: Routledge & Kegan Paul. Drolma, Delog Dawa. 1995. Delog: Journey to Realms Beyond Death. Richard Barron (trans.). Junction City: Padma Publishing. Edou, Jerome. 1996. Machig Labdron and the Foundations of Chad. Ithaca: Snow Lion Publications. Epstein, Lawrence. 1982. On the history and psychology of the 'das-Iog.' The Tibet Journal 7/4, 20-85. Furlong, Monica. 1996. Visions & Longings: Medieval Women Mystics. London: Mowbray. Gyalwa Changchub and Namkhai Nyingpo. 2002. Lady of the Lotus-Born. The Padmakara Translation Group (trans.). Boston: Shambhala. Gyatso, Janet. 1986. Signs, memory, and history: a Tantric Buddhist theory of scriptural transmission. JIABS 9/2, 7-35. - - 1993. The logic of legitimation in the Tibetan Treasure tradition. History of Religions 33/1, 97-134. --1998. Apparitions of the Self. Princeton: Princeton University Press. Gyatso, Janet and Havnevik, Hanna (eds). 2005. Women in Tibet. New York: Columbia University Press.

This inferior female body

149

Harding, Sarah. 2003. Machik's Complete Explanation: Clarifying the .Meaning of Chad. Ithaca: Snow Lion Publications. Havnevik, Hanna:. 1997. The Autobiography of Jetsun Lochen Rinpoche, a preliminary research report. In Helmut Krasser, Michael Torsten Much, Ernst Steinkellner, Helmut Tauscher (eds). Tibetan Studies, Proceedings ofthe 7th Seminar of the International Association of Tibetan Studies, Graz 1995. Wien: Verlag Der Osterreichischen Akademie Der Wissenschaften. --1999. The Life of Jetsun Lochen Rinpoche (1865-1951) as Told in her Autobiography. PhD thesis, University of Oslo. Herrmann-Pfandt, Adelheid. 1990. pakinfs: Zur Stellung und Symbolik des Weiblichen im tantrischen Buddhismus. Bonn: Indica et Tibetica Verlag. - - 1992-3. pakinfs in Indo-Tibetan Tantric Buddhism: Some Results of Recent Research. Studies in Central and East Asian Religions 5/6, 45-63. Jacoby, Sarah. 2007. Consorts and Revelation in Eastern Tibet: The Auto/biographical Writings of the Treasure Revealer Sera Khandro (1892-1940). PhD thesis, University of Virginia. - - 2009. To be or not to be Celibate: Morality and Consort Practices According to the Treasure Revealer Se ra mkha' 'gro's (1892-1940) Auto/ biographical Writings. In Sarah Jacoby, Antonio Terrone (eds). Buddhism beyond the Monastery: Tantric Practices and their Performers in Modern Tibet. Proceedings of the Tenth Seminar of the International Association for Tibetan Studies, Oxford, 2003. Leiden: Brill. Kolmas, Josef. 2004. The Ambans and Assistant Ambans of Tibet (17271912), some statistical observations. In P. Kvaerne (ed.) Tibetan Studies: Proceedings of the 6th International Seminar of the International Association for Tibetan Studies, Fagernes 1992. Oslo: The Institute for Comparative Research in Human Culture, vol. 1, pp. 454-67. Lama Chonam and Sangye Khandro. 1998. The Lives and Liberation of Princess Mandarava. Boston: Wisdom Publications. Nam-mkha'i snying-po. 1983. Tarthang Tulku (trans.). Jane Wilhelms (ed.). Mother of Knowledge: The Enlightenment of Ye-shes mTsho-rgyal. Berkeley: Dharma Publishing. Pommaret, Frangoise. 1989. Les Revenants de L'au-Dela Dans Ie Monde Tibetain. Paris: Center National de la Recherche Scientifique . . Richardus, Peter (ed.). 1998. Tibetan Lives: Three HimalayanAutobiographies. Richmond: Curzon. Rossi, Donatella. 2008a. mKha' 'gro dBang mo'i rNam thar. The Biography of the gTer-ston-ma bDe chen Chos kyi dBang mo (1868-1927?). Revue d'Etudes Tibetaines 15. Tibetan Studies in Honour of Samten Karmay, Part 2,371-378.

150

Sarah H. Jacoby

- - 2008b. Sacred Female Biographies in the Bon Religion: A 20 th Century gTer-ma. Rivista di Studi Sudasiatici (3), 127-143. Schaeffer, Kurtis. 2004. Himalayan Hermitess. Oxford: Oxford University Press. Simmer-Brown, Judith. 2001. Dakini's Warm Breath. Boston: Shanibhala. Sperling, Elliot. 1998. Awe and submission: a Tibetan aristocrat at the Court of Qianlong. The International History Review, xx(2), 32-35. Sponberg, Alan. 1992. Attitudes toward women and the feminine in early Buddhism. In Jose Ignacio Cabez6n (ed.). Buddhism, Sexuality, and Gender. Albany: State University of New York Press. Tulku Thondup Rinpoche. 1986. Harold Talbott (ed.) Hidden Teachings of Tibet. London: Wisdom Publications. - - 1999. Harold Talbott (ed.). Masters of Meditation and Miracles. Boston: Shambhala. Wilson, Liz. 1996. Charming Cadavers: Horrific Figurations of the Feminine in Indian Buddhist Hagiographic Literature. Chicago: The University of Chicago Press.

Aryadeva's gradual stages Their transmission from India to Tibet!

Andrew McGarrity

Introduction
Indian Madhyamika formulations of the gradual path tend to be most often associated with the Madhyamaka-Yogacara synthesis of Santarak~ita and KamalasIla, as well as the work of Atisa. However, an earlier account of the gradual nature of the Madhyamika critique is found in the CatuJ:tsataka (CS) of Aryadeva. It occurs, unsurprisingly, in the eighth chapter which focuses on preparing the student (slob rna yongs su sbyor ba). At CS 189, Aryadeva explains that the Buddha teaches generosity to the lowest students, morality to intermediate students and pacification to the highest.2 Three distinct stages are then presented in the following verse (CS 190). As recorded in Aryadeva's Root Verses, as found in the Tibetan translation of Pa tshab Nyi rna grags and Si1k~majana, this typology reads:
That one who understands, firstly, the rejection of the non-meritorious, at the intermediate stage, the rejection of the self, and finally the

! This article is based on a paper first presented as part of a Numata .Research Fellowship at Ryukoku University in Kyoto. I am grateful especially for the suggestions of Prof. Shi:iryu Katsura and Prof. Peter Skilling (who were both kind enough to read over earlier versions of this paper in full), as well as the assistance of Prof. Karen Lang, Dr Zhihua Yao, Dr Sara McClintock and Dr Paul Fuller. 2 CS 189 (8.14; p. 82): sbyin pa dman pa la gsungs shing / / 'bring la tshul khyims gsungs pa ste / / mchog la zhi ba gsungs gyur pa / / des na rtag tu mchog tu byos /

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 151-212

152

Andrew McGarrity

rejection of all views is a wise person. (bsod nams min pa dang por bzlog I bar du bdag ni bzlog pa dang I phyi nas Ita ba kun bzlog pa I gang gis shes de mkhas pa yin 1). 3

This verse is also quoted by Candrakirti in his PrasannapadCi (PP) analysis of MUlamadhyamakakCirikCi (MMK) 18.6, allowing it to be preserved in Sanskrit. There, it reads:
That one who understands, firstly, the rejection of the non-meritorious, at the intermediate stage, the rejection of the self, and, finally, the rejection of all is a wise person. viirafla1!l priig apuflyasya madhye viiraflam iitmanaJ:t I sarvasya viirafla1!l pasciid yo jiinfte sa buddhimiin 11. 4

The difference between the two is immediately striking. The root verse in the Tibetan translation takes the climactic final rejection as being of all views (lta ba kun bzlog pa), while the original Sanskrit instead takes it as being simply of "all" (sarvasya vCira!Jaf!l). Now, the Tibetan understanding of "all views" is obviously an amplification of this "all;" "all views" could have simply always been implied and the Sanskrit's sarvasya vCira!Jam could, in principle, simply be an abbreviation, for metrical reasons, of sarvadr~tlnCif!l vCira!Jam. "All views" would obviously be implied in the broader category of "all." However, the Sanskrit's sarvasya vCira!Jam, "rejection of all," I suggest, actually specifically means "rejection of all things" (sarvabhavCinCif!l vCira!Jam) and not views. In the following root verse (CS 191), for instance, Aryadeva, working on the assumption that this rejection of all in CS 190 is tantamount to understanding the emptiness of all, takes up the issue of how, practically, this rejection of the "all" could be carried out. He states that S

3 4

CS 190 (8.15; p. 82).

PP ad MMK 18.6 (359.8-9). CS 191 (p. 82): dngos po gcig gi Zta po gang I I de ni kun gyi lta par bshad I I gcig gi stong nyid gang yin pa I I de ni kun gyi stong pa nyid I Sanskrit quoted at PP ad MMK 4.9 (128.4): bhiivasyaikasya yo dra~tii dra~!ii sarvasya sa smrtaJ:t I ekasya siinyatii yaiva saiva sarvasya siinyatii I I Further on how this "knowing one" to "knowing all" principle operates, see McClintock
5

2000: 225-244.

.Aryadeva's gradual stages

153

That one who is seer of one thing (bha.vasyaikasya) is said to be the seer of all: that which is the emptiness of one thing is the emptiness of all.

Although there is no mention here of all things, the reference to the "one thing" (ekabhiiva) with which the "all" is counter-balanced firmly locates the discussion in the metaphysical terms of things (bhiiva) rather than "views about things" (dr~ti) (although, as we shall see, there is a certain blurring of these two in early Madhyamaka). This would imply that the sarvasya in the previous verse refers to sarviiniif!l bhiiviiniim, not dr~t;:niim. More conclusively, as we shall see in detail, Candraklrti, in his Catu!:z.fatakatlkii (CST) commentary only ever assumes "all things," explicitly understanding Aryadeva's rejection of "all" to be of "all things" (dngos po thams cad, *sarviiniif!l bhiiviiniim) without any mention of "all views."6 So, I suggest that there is a genuine alteration that occurs here between the Sanskrit and Tibetan. Accordingly, this present paper attempts to account for this alteration, and ventures a working hypothesis as to how and why it may have occurred. It will also shed further light on the broader hermeneutic assumptions implicit in this act of translation. In my analysis, I will discuss the reading of "all things" and "all views" as each are understood in two direct commentaries on Aryadeva's CS, one being Candraklrti's CST analysis and the other rGyal tshab's bZhi brgya pa'i rnam bshad (bZhi brgya). rGyal tshab's analysis will afford an insight into how a Tibetan exegete understands Aryadeva's model in the light of the received "all views" reading, and how this is reconciled with Candraklrti's CST analysis, which, by contrast, assumes "all things." I will also set these direct commentaries alongside CandrakIrti's PP analysis of MMK 18.6 in which the Sanskrit of CS 190 is preserved. I will argue that the MMK 18.6 context provides the most direct factor that influences the alteration, while a series of wider background doctrinal
6 CST 139al-2 ad CS 190: sdig pa bzlog pas ni phung po dang khams dang skye mched rnams la bdag rnam pa Ingar yang mi srid pa'i phyir bdag tu Ita ba cha nyi shu dor bar bya'o / / phyi nas bdag la chags pa yang dag par ston pa'i phung po dang khams dang skye mched dag la chags pa yang yongs su gtong bar bya ba'i phyir dngos po thams cad bzlog par bya'o / /

154

Andrew McGarrity

assumptions act in an indirect way to validate this alteration once it has occurred.
Textual consIderations

There does not seem to be any reason to doubt " authenticity the of the Sanskrit reading from the PP. According to J.W. de Jong, it is confirmed by all the manuscripts from which La ValleePoussin constructed his edition, as well as the manuscript found by Giuseppe Tucci from which de Jong then made his emendations? It is of course possible that the Tibetan sources were based on an alternate Sanskrit reading which has been lost. However, it is difficult to see how a Sanskrit reading which corresponds to the Tibetan Ita ba kun bzIog pa, e.g., sarvadr~tfniiY[l viiraTJam, could be made to conform to the meter. Indeed, the pattern of the rest of the verse would perhaps suggest that, if anything, simply dr~tfniiY[l viiraTJam would make for a better balance with the other rejections: the first rejection, for instance, is simply given as apuTJyasya vClraTJam and not sarviipuTJyasya viiraTJam, even though this is implied. However, the context of the verse itself, especially, as mentioned, given the verse which follows, reinforces the fact that it is precisely the "all" (sarvasya), that is, the universal nature of the final rejection, that is being deliberately emphasized. 8 On the other hand, the Tibetan translators could, theoretically, have easily read phyi nas dngos po kun bzIog pa without any metrical difficulties, yet they do not. According to Karen Lang's edition of A.ryadeva's root verses, the reading of Ita ba (views) is confirmed by the sDe dge, Peking, sNar thang and Co ne recensions - again it is possible that all of these could have been based on some lost Sanskrit alDe Jong 1978: 22; see also ibid., p. 25. As already noted, the following verse, CS 191 (p. 82), reads: "The one who is the seer of one thing is said to be the seer of all things: that which is the emptiness of one thing is the emptiness of all things." (dngos po gcig gi Zta po gang / / de ni kun gyi Zta par bshad / / gcig gi stong nyid gang yin pa / / de ni kun gyi stong pa nyid / Sanskrit quoted by Candrilirti, PP ad MMK 4.9 (128.4): bhiivasyaikasya yo dra~rii dra~rii sarvasya sa smrtaJ:! / ekasyasunyatii yaiva saiva sarvasya siinyatii / / This would seem to be more in accord with the final rejection being of all things rather than all views.
7,

Aryadeva's gradual stages

155

ternate reading. Unfortunately, Xuanzang's Chinese translation of Aryadeva's Root Verses contains only the ninth chapter onwards, nor is Dharmapala's commentary available for this section, so we also do not have a Chinese version against which to compare the Sanskrit and Tibetan. 9 More interestingly, the verse is, as in CS, also rendered with "all views" when it occurs as quoted in the Tibetan translation of PP itself, with the slight variation of Zta zhig kun. This is also agreed upon by the sDe dge, Peking, sNar thang and C~ ne recensions of the Tibetan translation of the PP citation of Aryadeva's verse.lO The verse itself is not quoted by either Buddhapalita or Bhiiviveka/Bhavya in their earlier commentaries, suggesting that its direct quotation in the context of Nagarjuna's root verse is a move of Candraklrti's own. In his heavily PP influenced rTsa shes tik chen commentary on the MMK, Tsong kha pa reads the verse ~s Zta zhig kun,11 and in his commentary on Aryadeva's text itself, rGyal tshab, as we shall see, proceeds on the assumption that the verse reads Zta. 12 Likewise, the direct quotation of the verse itself in Candraklrti's CST, which is preserved only in Tibetan, also renders "all views," even though Candraklrti, in his actual CST comments, makes no mention of the final rejection being of "all views" but rather takes it as simply being of "all (things)," in accordance with his PP citation. Red mDa ba who also provides a direct commentary on Aryadeva's verses, and who is probably rGyal tshab's most direct source, also reads the final rejection here as being of "all views."13
9 I am grateful to Dr Zhihua Yao for pointing this out to me (personal communication, April 29, 2009). 10 PPt Peking 133a6-7, sDe dge 116a7. I have consulted sDe dge and Peking directly, and sNar thang and Co ne indirectly through the CPA reprint (2000) of the sDe dge canon which provides textual variations with Peking, sNar thang and Co ne (dBu rna: 'a: vol. 60, 284.11-13). 11 rTsa shes tik chen 285.18. 12 bZhi brgya pa'i rnarn bshad 63bl. 13 dBu rna bzhi brgya pa rtsa 'greI (1974: 95.6-8): phyis nas bdag dang braI ba'i phung po tsarn Ia rnngon par zhen pa spang ba'i phyir rang bzhin stong pa nyid kyi sgo nas mthar 'dzin gyi Ita zhig kun bzIog par bya ste. I am

156

Andrew McGarrity

The distinction between "all things" and "all views" thus conforms neatly to the division between Indian and Tibetan readings of the verse. Or perhaps we ought to say rather that the division conforms neatly to the distinction between Candraklrti's readings, and, presumably, other Indian MMhyamikas before him, on the one hand, and, on the other hand, the readings of the collaborative translation teams working during the later transmission (phyi dar) and the indigenous Tibetan traditions which has then followed them.
Pa tshab Nyi rna grags and the emergence of the SvatantrikaPrasatigika distinction

According to the respective colophons of the CS and CST, Pa tshab Nyi rna grags collaborated with the Indian pa1Jita Siik~majana on their translations ,14 and with Mahasumati on translating

grateful to Professor Karen Lang for passing on the Tibetan of this to me (personal communication, September 18, 2006, May 17, 2009). 14 See especially on this Lang 1990: 275-8 and Ruegg 2000: 41-55. CS colophon D 3846, vol. tsha, 18a5-7: mal 'byor spyod pa bzhi brgya pa las slob dpon dang slob ma rnam par gtan la dbab pa bstan pa ste rab tu byed pa bcu drug pa'o / / / / slob dpon 'phags pa lha'i zhal snga nas mdzad pa / byang chub sems dpa'i mal 'byor spyod pa bzhi brgya pa'i tshig le'ur byas pa rdzogs so / / / / kha che'i grong khyer dpe med kyi dbung rin chen sbas pa'i kun dga' ra bar / rgya gar gyi mkhan po su smra dz'a na* dang / bod kyi lo tsa ba pa tshab nyi ma grags kyis bsgyur cing zhus te gtan la phab pa / / * Read: sukshma dza' na (i.e., Suk~majana). In the CST colophon, we also learn that Suk~majana is the spiritual son of "the great brahmin, Sajjana" whose own teacher, in turn, was another "great brahmin, Ratnavajra (rin chen rdo rje)," as part of a scholarly lineage which go~s back "a countless number of generations." (239a5-7): bzhi brgya pa'i 'grel pa slob dpon zla ba grags pa'i zhal snga nas kyis mdzad pa rdzogs so / / / / dpal grong khyer dpe med kyi dbus rin chen sbas pa'i kun dga' ra bar / rgya gar gyi mkhan po sukshma dz'a na zhes bya ba gdud rabs grangs med par pa I)i ta* brgyud pa'i rigs su sku 'khrungs pa bram ze chen po rin chen rdo rje'i dpon po bram ze chen po sad dz'a na'i sras gcig tu bde bar gshegs pa'i bstan pa la gcef sgras su mdzad pa bdag dang gzhan gyi gzhung lugs rgya mtsho'i pha rol tu son pa'i zhal snga nas dang / bod kyi sgra bsgyur gyi lo ts'a ba dge slong pa chang nyi magrags** kyis bsgyur cing zhus te gtan

Aryadeva's gradual stages

157

the PP, a translation which he then revised with the Kashmiri Kanakavarman. 15 Pa tshab emerges as the common link between the texts in which the alteration from "all things" to "all views" occurs. For this reason it is his role, rather than that of his collaborator, Siik~majana, that is of interest. For our immediate purposes, it is espe,cially significant that neither Aryadeva's CS, nor Candraklrti's CST, nor Candraklrti's PP had been translated at all before this, thus not only inextricably associating Pa tshab with their introduction into Tibet, but also placing these three texts into a sort of symbiotic relationship with each other in the propagation of what comes to be known as the "PrasaJigika" viewpoint (although the PP is obviously the most significant among them in this respect). I suggest that it is this "symbiosis" between these three texts, although with the PP assumed as paramount, that is more directly significant at the level of translation and which becomes especially significant in explaining the discrepancy between the Sanskrit and Tibetan readings of Aryadeva's stages. This "symbiosis" further illustrates both Pa tshab's direct role in reshaping its formulation, as well as his indirect influence upon later Tibetan interpretations which proceed on the assumption of "all views." At the same time though, it is also worth briefly noting some of the doctrinal contexts in which Pa tshab was operating and also
laphab pa / / * Read: pa1J4i ta (i.e., paIJejita). It may be that the middle vertical stroke of -ej- has faded here, as there seem to be two strokes still remaining under -IJi-. ** Read: pa tshab nyi rna grags. 15 PPt colophon sDe dge 200a4-7: gsung rab rgya mtsho bsrubs byung de nyid bdud rtsi'i snod / / shes rab ces bya snang byed tshig gsal rin chen sgron l/ bsgyur las dge ba gang bsags de yis nyi ma grags / / ita thibs spangs .te spros bral mkha' la 'gro bar shag / kha che'i grong khyer dpe med kyi dbus / / rin chen sbas pa'i gtsug lag khang gi 'dabs su* / rgya gar gyi mkhan po rtog ge ba chen po / ma h'a su ma ti'i zhal snga nas dang / bod kyi 10 ts'a ba pa tsab nyi rna grags kyis kha che'i dpe dang mthun pa ltar bsgyur / phyis ra sa ra mo che'i gtsug lag khang du kha che'i mkhan po ka na ka ba ra ma dang / bod kyi 10 ts'a ba de nyid kyis nyi 'og shar phyogs kyi dpe dang gtugs shing legs par bcos te gtan la phab pa'o / / * Read: 'debs suo

158

Andrew McGarrity

the contexts which then informed how his work may have been received. It is well known that Pa tshab was especially significant in the context of the later transmission of Buddhism to Tibet. His impact is found as a teacher in his own right, and in his role in the establishment of the supremacy of Candrakirti's (Prasatigika) Madhyamaka in Tibet. 16 A complete commentary! on MMK written by Pa tshab, which, Kevin Vose notes, "would seem to be the earliest Tibetan commentary on Nagarjuna,"17 is also included in the recently published thirty volume collection of early bKa' gdam pa teachers, the bKa' gdam gsung 'bum phyogs bgrigs.18 Pa tshab may in fact have been the first to actually use the terms "Svatantrika" (rang rgyud pa) and "Prasatigika" (thai 'gyur pa) to denote different branches of the Madhyamika tradition. 19 What can be said for certain, as Karen Lang has pointed out, is that "prior to sPa-tshab Nyi-ma-grag's translation and teaching of the PP and Candraklrti's other major works there was no basis for making these distinctions [i.e., between Svatantrika and Prasatigika]."20 'Gos Lo tsa ba, in the Blue Annals (Deb ther sNgon po) states that along with CS, PP and CST Pa tshab also translated Candrakirti's Yuktisa~tikiivrtti and the Madhyamakiivatiira and Bhii~ya (MAv; MAvBh).21 Although MAvBh had previously been translated by Nag tsho, Helmut Tauscher has shown that "after Tsong kha pa, Nag tsho's translation seems to have been forgotten" and super-

16 'Gos Lo tsa ba, Deb ther sngon po; trans. Roerich 1976: 341-343; Bu ston, Chos 'byung, 93.7-10ff. For further biographical summaries of Pa tshab's career and the contexts of his teachings, see especially Lang 1990 and Vose 2009: 45-56ff. 17 Vose 2009: 7. 18 lac. cit and p. 173, n.21, regarding the publication details of the bKa' gdam gsung 'bum phyogs bgrigs (Lhasa: dPal brstegs bod yig dpe rnying zhib 'jug khang, 2006). 19 Mimaki 1983: 163; Ruegg 2000: 47; Dreyfus and McClintock 2003: 3; Tauscher 2003: 207, 212ff. 20 Lang 1990: 127. 21 'Gos Lo tsa ba, Deb ther sngon po; trans. Roerich 1976: 342. This seems to be the only reference however to his translating Yukti~a~tikavrtti; he is only mentioned in the colophon to Yukti~a~tika and not Yukti~a~tikavrtti.

Aryadeva's gradual stages

159

seded by Pa tshab's later translation. 22 According to 'Go 10 tsa ba, Pa tshab undertook his most important translations (namely of Candraklrti's and Aryadeva's works) having already established the study of Madhyamaka,23 although the colophons would seem to suggest that he may have already completed his translations in Kashmir prior to revising them in Lhasa. By "establishing the study of Madhyamaka" is presumably meant the establishment of the study of Madhyamaka along what comes to be understood as "Prasangika" lines with Candraklrti's interpretation taken as authoritative. Certainly, we would expect Pa tshab's own teachings of Madhyamaka, and most notably the reliance upon Candrakirti's work, including Candraklrti's use of his predecessors' works via citation, to have influenced his translations. The emerging Svatantrika-Prasangika distinction may also have contributed, at least in some indirect way, to Pa tshab's alteration of Aryadeva's final rejection of "all things" to "all views." The interpretation of the rejection of views (dr$ti), and more significantly the closely related rejection of the pratijiiii (thesis), pak$a (position) and pratipak$a (counter-position) in early Madhyamika works, is especially significant in the early emergence of the SvatantrikaPrasangika distinction in Tibet. Of particular importance is of course the interpretation of Nagarjuna's famous disavowal of a thesis (pratijiiii) at Vigrahavyiivartanf (VV) 29 in response to an opponent's charge of self-contradiction against the Madhyamika's claim for the emptiness of all things. Nagarjuna states that "If there were some thesis, then this fault [i.e., self-contradiction] would come about for me; but I do not have a thesis (pratijiiii) , thereTauscher 1983: 30l. 23 'Gos Lo tsa ba, Deb ther sNgon po; Tibetan as quoted in Lang 1990: 140, n. 23: de nas dbu ma'i 'chad nyan legs par tshugs te gra ba rnang du bskyangs / rtsa 'jug bzhi gsum / rigs pa drug bcu rnams kyi slob dpon zla bas mdzad pa'i grel pa rnams legs par bsgyur zhing / Lang (1990: 134) translates: "Then, after he successfully established Madhyamaka studies and supervised many students, he translated well the commentary written by Candraklrti on the Yukti~a~tikii, and three [commentaries on] the MulamadhyamakakiirikiilJ [rtsa], the Madhyamakiivatiira ['jug] and the CatulJsataka [bzhi]." Both Lang (1990: 140, n. 23) and Tauscher (2003: 245, n. 21) have noted that Roerich (1976: 342) misses the reference to CS in his translation.
22

160

Andrew McGarrity

fore I do not have such a fault at all."24 Georges Dreyfus and Sara McClintock have noted that
In eleventh century Tibet, the translator Khu mdo sde 'bar, a student and collaborator of Jayananda, appears to have interpreted this rejection [of the pratijna] as meaning that Madhyamikas should have no thesis at all. Pa tshab is said to have understood Nagarjuna's statement to rule out only positive theses but not exclude negative ones.25

This conception of "thesislessness" as characterizing "Prasailgika" Madhyamika is associated especially with Jayananda who, along withPa tshab, played akeyrole in promoting CandrakIrti's works, and with whom Pa tshab co-translated Atisa's Mahiisutrasamuccaya. 26 Kevin Vose has noted, however, that Jayananda was Pa tshab's junior and that when he was active Pa tshab's reputation had already been made, thus suggesting that his association was with Pa tshab's students rather than directly with Pa tshab himself. 27 Jayananda is famously criticized by Tsong kha pa in his Lam rim chen mo on two counts especially:28 firstly, for arguing that Prasailgikas do not rely upon pramiiIJas themselves but only insofar as they are agreed upon by their opponent. 29 Secondly, and more signifi24 VV 29 (61.12-15): yadi ka cana pratijiiii syiin me tata niisti ca mama pratijiiii tasmiin naiviisti me do~alJ. II
26

e~a

me bhaved

do~alJ. I
25

Dreyfus and McClintock 2003: 21. Following Vase 2009: 196, n.81 ad Mahiisutrasamuccaya, D 3961. According to the colophon, Jayiinanda, Pa tshab and Khu mdo sde 'bar translated this work. 27 Vase 2009: 54. 28 See for an overview ofthis, Yoshimizu 1993: 207-13. 29 Lam rim chen mo 673.8-674.6: 'di la dza ya a nanda ni 'di ltar smra ste I 'jug pa'i 'grel bshad las I de la 'ga' zhig na re gal te thaI 'gyur gtan tshigs su 'dod na de tshad mas grub pa yin nam 'on te ma grub pa yin I de la gal te phyogs dang po ltar na de'i tshe gnyis ka la 'ang grub pa yin pas gzhan gyis khas blangs pa zhes ci ltar brjod I phyogs gnyis pa ltar na gzhan gyis khas len par mi 'os pa nyid yin pas gzhan gyis khas blangs pa zhes ji ltar brjod ce na I de la Ian ni tshad mas grub pa gang yin pa de gnyis ka la grub pa yin no I I zhes pa de nyid kho bos mi shes te 'di ltar rgol bas sgrub byed bkod pa'i dus na gang gtan tshigs 'god pa de la tshad mas grub pa yin grang gzhan la tshad mas grub par des ji ltar shes te I gzhan gyi sems kyi khyad par mngon sum dang I rjes

Aryadeva's gradual stages

161

cantly for the issue of "views," he is indirectly criticized by Tsong kha pa for holding that PrasaIigika Madhyamikas hold no thesis (pratijfiii) whatsoever, but merely refute the assertions of others
(dbu ma ba la gzhan gyi 'dod pa 'gog pa tsam ma gtogs pa'i rang gi dam bca' ba med).30 A significant forerunner of Tsong kha pa's critique of thesislessness is found in the dBu ma shar gsum gyi stong thun of Phya pa chos kyi seng ge. 31 However, unlike Tsong

su dpag pa'i yul ma yin pa'i phyir TO / / rang nyid la yang tshad mas grub par ci ltar shes te /* **gzhan gyi sems kyi khyad par mngon sum dang rjes su dpag pa'i yul rna yin pa'i phyir ro / / rang nyid la 'ang tshad mas grub par ji Itar shes te 'khrul ba'i rgyu mtshan gyis rgol ba dag gis tshad ma nyid du khas blangs pa'i stobs kyis dngos po rnams kyi rang bzhin khas len pa yin no / des na gzhan gyis khas blangs pa'i sgo nas gzhan gyi phyogs sun 'byin pa yin no** zhes zer te / phyi rgol la rtags tshad mas grub ma grub snga rtol gyis mi shes te de'i bsam pa ni snga rgol gyi tshad ma gnyis mi 'grub pas so / / rang gis kyang rtags tshad mas grub par mi shes te rang gis tshad mas grub par thag bcad kyang bslu ba srid pas so / / de'i phyir gnyis ka la tshad mas grub pa'i rtags med pas tshad mar khas blangs pa la thug pas na tshad masma grub kyang khas blangs pa'i sgo nas sun 'byin 'thad pa yin no / / zhes 'chad do / * Jayananda, Madhyamakiivatiirabhii~yatfkii (120a6-120bl). ** Madhyarnakiivatiirabhii~yatfkii (120bl-3): khrul pa'i rgyu rntshan gyis dus ring po nas bzung ba'i phyir slu ba 'drid pas so / / de phyir rgol ba dang / phyir rgol ba dag gis tshad rna nyid du khas blangs pa'i stobs kyis dngos po rnams kyi rang bzhin khas len pa yin no / / des na gzhan gyis khas blangs pa'i sgo nas gzhan gyi phyogs sun 'byin par rigs pa yin no / 30 Lam rim chen mo 675.6-12: pa1J4i fa de'i slob ma* 10 tsa ba dag** kyang 'di skad du / dbu ma ba la gzhan gyi 'dod pa 'gog pa tsam ma gtogs pa'i rang gi dam bca' ba med cing chos can la sogs pa gnyis ka la grags pa'i thun mong ba ma grub pas rang rgyud mi 'thad do / / rigs pas rnams par dpyad pa'i 'brad bu yang gzhan gyi grub mtha' 'dor ba tsam zhig yin la / de las gzhan pa'i rang 'dod med pas rang rgyud kyi gtan tshigs rnam pa thams cad du brjod par mi bya'o / / des na thai 'gyur kho na yin la de la 'ang sgrub pa'i thai 'gyur ni rang rgyud kyi mthar thug pas sun 'byin gyi thai ba kho na'o / * That is, Jayananda. ** For example, the aforementioned eleventh century lotsawa, Khu mdo sde 'bar, according to 'Jam dbyangs bzhad pa, (cf. Cutler et a12002, vol. 3, p. 413, n. 474). 31 dEu rna shar gsum gyi stong thun; ed. Tauscher 1999: 66.7-67.17; trans. Vose 2009: 151-153.

162

Andrew McGarrity

ka pa, who critiques thesislessness while still taking CandrakIrti's reading of Madhyamaka. to be authoritative, Phya pa simply critiques thesislessness and Candrakirti's standpoint along with it, upholding instead the interpretation of Madhyamaka championed by Santarak~ita, KamalasIla and Jfianagarbha (i.e., "the three Madhyamikas from the East," dbu ma shar gsum). On the other hand, Tsong kha pa carefully distances CandrakIrti himself from "thesislessness" (by shifting the discussion from epistemological to ontological grounds, as we will consider in a moment), preferring instead to associate the charge with Jayananda's "misrepresentation" of PrasaIigika. Jose Ignacio Cabez6nhas noted, however, thatJayanandahimself, in his Madhyamakiivatiirabhii~yatfkii (MAvBhT), only refers to the Svatantrika position and only ever opposes it to the Madhyamaka position, and not specifically "PrasaIigika Madhyamaka."32 From the work of Georges Dreyfus and Drongbu Tsering on Pa tshab's own writings, it appears as though his position does not differ markedly from the sort of standpoint associated with Jayananda. Pa tshab advocates a form of PrasaIigika that Dreyfus typifies as being akin to a form of radical skepticism, allowing only a "negative thesis."33 Due to only ruling out a "positive thesis" and not all theses per se, Pa tshab does not meet with the sort of criticism at the hands of Tsong kha pa that Jayananda does, and in general he seems to be viewed favorably. He is, for example, included, via the Sa skya teacher, Red mDa ba, in Tsong kha pa's own lineage of correct Madhyamika teachers. 34 Hinting at further differences between the two that may in some way account for these differing reputations, Kevin Vose has pointed out that Jayananda's MAvBhT contains citations from MAv and PP which differ from Pa tshab's translations of these works, suggesting also that Jayananda did not utilize Pa tshab's translations of Candraklrti's works in his own translation of his MAvBh'f.35 32 33 34 35

Cabez6n 2003: 292. Dreyfus and Tsering 2010. Lang 1990: 137. Vose 2009: 54.

Aryadeva's gradual stages

163

In any case, there is certainly considerable change between Pa tshab's time and Tsong kha pa in views over what exactly constitutes Pdisarigika Madhyamika. Following Tani Tadashi, Helmut Tauscher has noted an "old Svatantrika-Prasarigika distinction" which "centers on the question of whether an own thesis is denied or maintained," associated with Jayananda et aI, in contrast to "a 'new,' i.e., dGe lugs pa distinction,"36 while "Phya pa chos kyi seng ge appears to be somewhere in between these."37 Most significant about this "new" (dGe lugs pa) distinction, championed by Tsong kha pa and his followers, is that it reconceptualizes the epistemological issues of the Svatantrika-Prasarigika distinction in an ontological light. Rather than the issue simply being about the use of pramiil;as and holding a pratijfiii per se, for Tsong kha pa, the Svatantrika postulation of svatantra (autonomous) inferences still involves some subtle reification at a metaphysical level. As Cabez6n puts it, "Tsong kha pa embarks on a project of trying to show that svatantras are syllogisms that contain the assumption of independent existence ... [they] are indicative of a kind of 'crypto-realism'."38 Or, in Dreyfus' and McClintock's summation, for Tsong kha pa, "an argument is autonomous if, and only if, its terms refer to objectively or intrinsically existent phenomena."39 Thus, what appears on the surface to be merely an epistemological concern carries, on Tsong kha pa's reading, deeper ontological implications, and these implications, and not simply the mere usage of pramiilJas and holding a thesis, are what distinguishes Svatantrikas from Prasarigikas. Contrary to Jayananda (and to a lesser extent Pa tshab) the issue ceases to simply be about the use of pramiilJas and the pratijfiii - and perhaps also views, dr~ti - per se, but rather pramiilJas and pratijfiiis only insofar as they implicate inherent existence (svabhiiva). From this perspective, as Tauscher notes, our transitional figure, Phya pa, could be considered a "Svatantrika" according to the "old distinction" and a "Prasarigika" according

36 Tauscher 2003: 211.

37

ibid, p. 238. 38 Cabez6n 2003: 297. 39 Dreyfus and McClintock 2003: 23. See also Yoshimizu 2003.

164

Andrew McGarrity

to the "new (dGe lugs pa) distinction" insofar as he also rejects the idea of thesislessness. 4o However, the key difference is that the "new (ontological) Svatantrika-Prasangika distinction" has the effect of still leaving open for Tsong kha pa the possibility of a view, and the legitimate use of pramii1)a methodology to the extent that inherent existence is not implicated in their usage. This means that, for Tsong kha pa, the legitimacy of pramii1)a is reconcilable with Candraklrti's critique, whereas for Phya pa it was not. So, the context of the "old Svatantrika-PrasaIigika distinction" prevalent in Pa tshab's time that emphasizes an outright rejection of the thesis (or at least just a "positive thesis"), and perhaps also views, can, in general, provide a snapshot of an earlier climate in which there might be some sense in moving the Madhyamika's realization of ultimate emptiness from being a rejection of all things to being a rejection of all views, hence the alteration of Aryadeva's verse. On the other hand, the context of the "new" SvatantrikaPrasangika distinction from Tsong kha pa onwards provides a climate in which any discrepancy between a reading of Aryadeva's verse that assumes "all things" (such as Candraklrti's CST analysis) and the received "all views" reading in the CS Kiirik~s could be reconciled quite unproblematically, and in fact largely go unnoticed. We will see this when we examine rGyal tshab's direct analysis of the verse. The "new" Svatantrika-PrasaIigika distinction provides a climate in which the Madhyamika's rejection of all things at the ontological level and the rejection of all views at the epistemological level could be readily accepted as synonymous, inasmuch as it could be assumed that both are referring to inherently existing views and things. From the perspective of the "new" Svatantrika-Prasangika distinction, the rejection of svabhiiva operates as the crucial determinant that implicates both epistemological views and metaphysical things, inasmuch as it implicates a particular type of misunderstanding relevant for both. However, it is the "old" distinction that offers a general backdrop of why such an alteration might have been made in the first place.

40

Tauscher 2003: 238.

.Aryadeva's gradual stages

165

Now, while this cursory sketch provides the general context within which Pa tshab was operating and how his alteration may have been understood, for our immediate purposes we do not need to dwell much further on this. I suggest that the alteration from things to views can be understood through reference to Indian textual materials, rather than through reference to developm(;!nts actually occurring in Pa tshab's own time, although these developments may provide a background context in which such a move may have made sense. Tibetan readings that directly engage with Aryadeva's verse - such as rGyal tshab's which we will focus on in particular '-' would seem to reinforce this.
dr~ti, pratijfiii, pak~a

and pratipak~a in early Madhyamaka

Certainly, there is considerable precedent in early MMhyamika materials themselves for viewing metaphysical "things" in the same light as epistemological "views" (most notably, problematic views that implicate truly existent or non-existent things). Both views and things are equally subsumed into the same general critique of metaphysics. Starting from the original exchange in Candraklrti's PP that eventually gives rise to the Svatantrika-Prasangika distinction, Nagarjuna's VV 29 disavowal of a pratijfiii comes to be frequently cited alongside Aryadeva's CS 400 in which it is said that "Although [they might try] for a long time, [an opponent] is not able to utter a criticism against [the MMhyamika] who does not hold a position (pakija) of 'existence,' 'nonexistence' or 'existence and non-existence'."41 The term used here is pakija ("position") but the citation of this verse alongside the disavowal of the pratijfiii would suggest that from Candrakirti onwards, it is assumed that the same point is being made in both instances. Aryadeva's argument at CS 400 may have influenced Pa tshab's alteration to "all views" since the association of the pakija that is rejected is one that postulates "existence," thus neatly implicating in a single move both a view 41 CS 400 (16.25; p. 150); yod dang med dang yod med ces I gang la phyogs ni yod min pa I I de la yun ni ring po na 'ang I I klan ka brjod par nus ma yin I Sanskrit cited in PP ad MMK 1.1 (16.4-5): sad asat sadasac ceti yasya pak~o na vidyate I upalambhas cirelJapi tasya vaktufJl na sakyate II

166

Andrew McGarrity

about - or position regarding - things and existent things themselves. As David Seyfort Ruegg has summed it up, "the term pratijna very clearly refers to a propositional thesis postulating an entity (bhi'iva). And it is just this that Nagarjuna and Aryadeva have disowned in VV and CS."42 Indeed, if anything, the move to views (if views can be taken as synonymous with this sort of pak~a) becomes wider in scope than just the original "things," since the pak~a that is rejected is also, in addition, of "non-existence" and "existence and non-existence" as well. Thus, the move from "all things" to "all views" potentially expands the Madhyamika's climactic rejection, implicating what we shall especially see the Abhidharmakosa (AK) denote as "extreme views" (antadr~ti), especially eternalism and nihilism. Further on, we will also refer to RatnavaIf(RA) 2.3-4, which CandrakIrti cites alongside CS 190, in which Nagarjuna also suggests that neither a pak~a nor a pratipak~a are in accordance with Reality, in accordance with the standard Madhyamika principle of what Ruegg has termed "the complementarity of binary concepts and terms.''''3 Candraklrti also draws together the specific terms of dr~ti and bhi'iva following this same "complementarity of binary concepts and terms" principle in relation to Yukti~a~tika (Y~) 23 where Nagarjuna uses the term dr~tikrta, "constructed view"44 to denote the views that are overcome when one no longer equates dependent origination with metaphysically grounded arising and cessation.45 Specifically, Candraklrti states that while the absence of cessation precludes the view of nihilism (ucchedadr~ti), the absence of arising precludes dngos por Ita ba46 - which, as Cristina
42

Ruegg 1983: 213.

43 Ruegg 1977: 56.

44 'Edgerton (1970, vol. 2, p. 269) suggests a more overtly pejorative sense of dr~tikrta as "false view"/"heresy." I have preferred the less loaded and literal "constructed view,' but this pejorative sense is worth bearing in mind. 45 y~ 23 (62.3-6): gang dag rten cing 'brei par 'byung / / skye dang 'jig pa rnams spangs pa / / shes par gyur pa de dag ni / / Ita gyur srid pa'i rgya mtsho brgal / / 46 y~V ad y~ 23 (62.7-14): skye ba med pas dngos por Ita ba med 'jig pa med pas chad par Ita ba med do / / skye ba dang 'jig pa med na rang gi mt-

Aryadeva's gradual stages

167

Scherrer-Schaub suggests, would seem to denote *bhiivadr:fti.47 This usage of the term bhiiva when associated with dr:fti, specifically as an instance of dr~tikrta, suggests not just a "view of things" but a "view of being" or "realism" more broadly, corresponding to the aforementioned pak:fa postulating "existence" (sat). Here, at y~ 23, Niigiirjuna also refers to the "ocean of becoming [i.e., sa~siira] consisting of constructed views" (Ita gyur srid pa'i rgya mtsho, dr:ftikrtabhaviirl;ava). In a clever intertwining of both sets of metaphors used, Candraklrti states that the "vision of emptiness" (stong pa nyid mthong ba, sunyatiidarsana) acts as the vessel (gru, nau) that bears one across the "ocean of becoming;" this "ocean of becoming" in turn itself consists of "rivers of views" (Ita ba'i chu bo), which are precisely "constructed views" (dr~tikrta). "Views," when understood as dr~tikrta - which are inclusive of the view of bhiiva(s) - are directly juxtaposed against the "vision" of emptiness. The clearest discussion of the question of all dr~tis in the wake of emptiness occurs at MMK 13.8a, providing a significant blueprint for how the "views" reading of Aryadeva's final rejection might be understood. While at MMK 13.8a the term "rejection" (viirw;am) is not used, emptiness is associated with the "relinquishment of (or perhaps "release from") all views" (sarvadr:ftfnii~ niftsara1}am):48
The conquerors proclaim emptiness to be the relinquishment of all views (sarvadr:ftfnii1!L nilJsarwJam). But those for whom emptiness is a view, they are said to be unable to be rectified.

Given his influence on Pa tshab, it is especially worth considering Candraklrti's PP analysis of this verse. In keeping with Yukti:fa:ftikii 23 above, Candraklrti reads the "views" (dr:fti) in question as
shan nyid med pas rtag par Ita ba mi 'byung ngo / / de Itar rten cing 'brei par 'byung ba'i skye ba dang 'jig pa mi dmigs na Ita par gyur pa'i srid pa'i rgya mtsho las brgal ba yin no / / Ita bar gyur pa nyid srid pa'i rgya mtsho'o / I Ita ba'i chu bo nyid Ita bar gyur ba'i srid pa'i rgya mtsho yin no I I stong pa nyid la brtson pa dag ni stong pa nyid mthong ba'i gru chen pos gdon mi za bar de las brgal bar 'gyur ro I I 47 Scherrer-Schaub 1991: 207, n. 36l.
48 MMK 13.8 (247.1-2): sunyata sarvadr-!tfnar[! prokta niJ:tsaralJar[! jinaiJ:t I yqa1!L tu sunyattidmis tan asadhyan babha-!ire II

168

Andrew McGarrity

"constructed views" (dr~tikrta), and they are false precisely inasmuch as they consist of "adherence to (existent) thingsj being"
(bhiiviibhiniveSa):49
Here, emptiness is' that which is the relinquishment, i.e., the nonproceeding, of all constructed views (dr~tikrta), i.e., all adherence to grasping (sarvagrahiibhinive.sa). But it is not the case that the mere cessation (nivrttimiitra) of constructed views constitutes a [positively defined] existent thing (bhava). Rather, we do not respond to those who, even with respect to emptiness, maintain adherence (abhinive.siina) to an existent thing (bhiiva). So how, on the basis of our teaching, will there be liberation without the turning away from the accumulation of conceptual elaboration? [Answer:] Were one, being told "I will give you nothing, no commodity at all," to respond "Well then! give me that very commodity called nothing!," how would they be made to apprehend the non-existence (abhiiva) of the commodity?50 Thus, for one for whom there persists the adherence to an existent thing even in regard to emptiness, how might their adherence to an existent thing (bhaviibhinive.sa) in regard to that [i.e., emptiness] now be refuted? Therefore, due to their erroneous conception, even in regard to the great medicine [emptiness], they are simply rejected by the Tathagatas, the great physicians, whose treatment is most supreme.

Candraklrti's use of the term nivrttimiitra, "mere cessation" in regard to emptiness suggests that he is drawing upon his earlier insistence upon the Madhyamika's critique being a non-implicative negation (prasajyaprati~edha) to explain why emptiness itself cannot constitute any positively defined entity (bhiiva). That is, empti49 PP ad MMK 13.8 (247.3-248.3): iha sarve~iim eva dr~tikrtiinarrz sarvagrahabhiniveSiinarrz yanni/:!saralJ-am apravrtti/:! sii sunyata I na ca dr~tikrtanarrz nivrttimatrarrz bhiiva/:!, ye tu tasyam api unyatayarrz bhiivabhiniveSinas tan praty avacaka vayam iti kuto 'smadupadesat sakalakalpanavyavrttya mok~o bhavi~yati I yo na kirrz cid api te palJ-yarrz dasyiimfty ukta/:! sa c'ed dehi bhos* tad eva mahyarrz na kirrz cin nama palJ-yam iti bruyat sa kenopayena sakya/:! palJ-yiibhiivarrz grahayiturrz I evarrz ye~arrz sunyatayam api bhiivabhiniveSa/:! kenedanfrrz sa te~arrz tasyarrz bhiivabhiniveSo ni~idhyatam iti I ato mahiibhai~ajye 'pi do~asarrzjiiitvat paramacikitsakai/:! mahiivaidyais tathiigatai/:! pratyakhyata eva te II * Read ukta/:!, san dehi bhos following de Jong 1978: 56.

50 Lit. "Being told '1 will give you no commodity at all,' how is one who would respond 'Hey you, give me that very commodity called nothing! ' ... "

Aryadeva's gradual stages

169

ness does not indicate some positive thing which happens to lack inherent nature, but instead applies simply to the mere absence of inherent nature'. 51 On the other hand, if the Madhyamika's negation of inherent nature were implicative (i.e., a paryudasaprati~edha), then it could be seen as applying to a p<?sitive entity which is negative only by implication, such as an actual positive thing, or commodity, called "nothing" (na kirrz cin nama pa1Jyam). Thus, the principle of niJ:tsvabhavatva being non-implicative provides, for CandrakIrti, the means of reconciling the rejection of all things , (i.e., in the wake of universal emptiness) with the rejection of (or, 'here, release from) all views, inasmuch as views are equated with "constructed views" (dr~tikrta) which are, in tum, synonymous with "grasping at things" (bhavabhiniveSa). Given the shared assumption of dr~tikrta, the term bhavabhiniveSa as used here would

51 Elsewhere, CandrakIrti more explicitly states that the Madhyamika's "critique of svabhiiva is a prasajya-prati~edha, as well as drawing out the specific doctrinal and methodological implications from this principle of emptiness as being non-implicative. See for example, PP ad MMKl.1 (13.4-6): .naiva svata utpannii, ity avadhiiryamiiTJo parata utpannii ity ani~ta7!l priipnoti / na priipnoti, prasajyaprati~edhasya vivak#tatviit parato 'py utpiidasya .prati~etsyamiinatviit / At CST 234b3-5 ad CS 388, CandrakIrti illustrates the non-implicative 'nature of the refutation of svabhiiva by way of the non-implicative nature . of space (iikasa) as the "the mere absence of form" (rupiibhiivamiitra): rgyu dang rkyen las skyes pa'i phyir rang bzhin byas pa can nyid du thaI ba'i phyir . dngos po rnams kyi rang bzhin gang yin pa de ni rgyu med pa can kho na yin la rgyu med pa la yang yod pa nyid mi srid pas med par dgag pa la brten ;nas dngos po med pa kho na rang bzhin yin te dngos po ma yin pa'i phyir i.~o / / de'i phyir rang bzhin 'di thams cad la dbyer med pa'i ngo bo yin pa'i :phyir dngos po thams cad rang bzhin gyis ma skyes par rang gi ngo bo gcig 'pa ste / 'di Ita ste dngos po med pa'i rang bzhin gyi ngo bo'o / / CandrakIrti ;'specifica1ly defines space as rupiibhiivamiitra, following the Sautrantika formulation, at CST ad coS 205; Sanskrit fragment ed. Suzuki 1994: 180.3-8: lrupiibhiivamiitra eviikiiSavyavahiiriin na ki7!l caniikiiSa7!l niima vasturupam asti / rupantariibhave tu rupiTJam utpattipratibandhiibhiiviit sa eva rupiini tariibhiivo bhrsam asyiintalJ. kiisante bhiivii ity iikiiSam ity iikhyiitalJ. / tad ,. asyiivastusato 'ki7!l canasya niimadheyamiitropade1avyiimu4hair abhidhar~ masiistre vaibhii~ikair yad vastutvam iiropita7!l na tat pramiiTJam iti na tenii~smiikam abhyupagamabiidhiicodana7!l nyiiyyam /

170

Andrew McGarrity

seem equivalent to what is denoted by *bhavadr~ti as used in relation to y~ 23. So certainly there is a tendency in early Madhyamika works to associate the Madhyamika's rejection of all views with a rejection of all things, and in fact to treat the two as synonymous,"especially given the understanding of bhava to denote "being" or "real existence," hence a view of "realism." No doubt this synonymity influenced Pa tshab's rendering of Aryadeva's verse. It may be that the reading of "all views" simply draws out an implication originally there in "all things" all along, and that it was simply understood in sarvasya Vara1Jam, with "views" simply being included within the larger category of "things." There may indeed have also been a tradition of understanding "all things" to specifically imply "all views." However, this in itself does not tell us why "views" became worth explicitly drawing out in this particular instance in the Tibetan translation. After all, if views and things are generally accepted as being perfectly, and in all instances, syn'onymous in the context of Aryadeva's, Candraklrti's and Nagarjuna's works (i.e., in the sense of "views of being"), then it could equally be said that no change even needs to be made; one would naturally be assumed to implicate the other. So, I would suggest that this general assumption in early Madhyamaka of metaphysically loaded views (i.e., especially of views that presuppose bhava) does not in itself provide the direct reason why this change itself took place. However, this general assumption does provide a context in which the rejection of all things and all views can so seamlessly be understood as virtually interchangeable, as well as a context within which the change itself might have made hermeneutic sense. The MMK 18.6 context Rather, in order to explore the direct background behind Pa tshab's alteration from "all things" to "all views," it is worth comparing CandrakIrti's CST and rGyal tshab's bZhi brgya analysis of CS 190 with the overall PP analysis of MMK. 18.6, at the culmination of which the Sanskrit original CS 190 is quoted. It is most important

A.ryadeva's gradual stages

171

to note MMK 18.6 and the context it provides for Candraklrti's citation. Nagarjuna's root verse itself reads: 52
"Self" is conventionally designated (prajfiapitam) by the Buddhas and "no-self" is taught (deitam) [by them]; [yet] neither is there any self or non-self taught.

Candraklrti quotes CS 190 at the culmination of his analysis of this verse. I suggest that Nagarjuna's typology here, and especially Candraklrti's PP analysis of it, is the most direct factor for understanding Pa tshab and Siik~majana's alteration of Aryadeva's final rejection from "all things" to "all views." Such a hierarchy is at least implicit in Nagarjuna's contrasting use of the terms prajftapitam ("conventionally designated") and deiitam ("taught").53 The provisionally expedient "designation" (prajftapitam) of the self is subordinated to the actual "teaching" (deiitam) of non-self, which is in turn subordinated to the superior understanding of neither self nor non-self. However, this hierarchy, and the sequential gradual stages it suggests, while only implicit in Nagarjuna's root verse, is made explicit by CandrakIrti in his commentary, as it is in the work of Buddhapalita and Bhaviveka/ Bhavya before him (although they do not cite Aryadeva's verse). This provides the context in which Candraklrti then subsequently cites Aryadeva's gradual rejections. For Aryadeva's verse, with its initial, intermediate and final stages,does indeed explicitly provide a sequential hierarchy that can be grafted onto Nagarjuna's typology. Pa tshab and Siik~majana's translation of "all things" with "all views" then subsequently makes the correlation between Nagarjuna's typology 52 MMK 18.6 (355.11-12): tumety api prajiiapitam antumety api deitaf[l / buddhair natma na canatma kat cid ity api desitaf[l / / 53 The weaker sense of prajiiapitam in contrast to deitam also of course connotes the idea of prajiiapti ("designation'); atman is only prajfiapitam whereas anatman is actually deitam. Suggesting this, Candraklrti (La Vallee-Poussin, p. 357.4) takes the self that is initially posited (prajiiapitam) as being "the self as based on worldly convention" (tumety api prajiiapitam, loke vyavasthapitaf[l). Reflecting this distinction, Ruegg (1983: 212) paraphrases: "while the designation atman has been used and anatman has been taught, the Buddhas have also taught neither atman nor its opposite (anatman)." (my underlining)

172

Andrew McGarrity

and Aryadeva's gradual rejections even more explicit, although of course such a correlation with Nagarjuna's schema here may have already been implied by Candraklrti's citation of Aryadeva's verse in the first place. The alteration also allows for a more subtle and refined set of alignments to be drawn than just the more generally applicable correlation between views and things. The MMK 18.6 "stages" provide the direct reason for the alteration, while the general Madhyamika attitude to views and things provides an indirect and general context in which the alteration then makes sense. The rejection of the non-meritorious and the conventional designation of a self In his CST analysis of CS 190, Candraklrti immediately places Aryadeva's rejections within the context of a graduated approach. He specifies that "here, this teaching ought to be understood as the gradual rejection of all conditioned phenomena in stages."54 His reference to "all conditioned phenomena" ('dus byas thams cad) places the rejections firmly in an ontological setting, foreshadowing that it is simply the rejection of all things rather than views that represents the culmination of Aryadeva's three rejections. In commenting on the initial rejection in which Aryadeva prescribes the rejection of the "non-meritorious" (bsod nams min pa, apu1}ya), Candraklrti explains that 55
firstly, one ought to carry out the rejection of all non-meritorious positions (Phyogs thams cad), for they are the cause of many hundreds of evils seen and unseen. One ought to renounce them; they are to be rejected first of all.

While Candraklrti does initially refer here to "the rejection of all positions" (phyogs thams cad bzlog pa), this does not seem related to P!l tshab's "all views." The specific qualifier that these positions 54 CST 138b7: bstan pa 'dir ni rim gyis 'dus byas thams cad bzlog par bya'o / 55 ibid (138b7-139a): de la 'dir re zhig bsod nams ma yin pa'i phyogs thams cad bzlog pa ni dang por bya ste / mthong ba dang ma mthong ba'i nyes pa brgya phrag du ma'i rgyu yin pa'i phyir te / spang bar bya ba dang por bzlog parbya'o /

Aryadeva's gradual stages

173

are "what are not of merit" (bsod nams ma yin pa) indicates that this is limited only to Aryadeva's initial rejection. rGyal tshab elaborates on this initial rejection by more fully relating the "turning back of the non-meritorious" to the ethical implications of karmic causality, as well as showing why it provides the necessary starting point. He explains that 56
if Reality is taught to [a student] incapable of acting as a vessel for such a teaching, then, all actions and agents having been denied, [the student] may suppose that there is no distinction between good and evil. Having then committed evil [on this basis of this misconception], they will go to a bad rebirth. So, initially, if it is taught that good and evil actions have causes and effects, this is easy to understand. Engagement with evil demeritorious [deeds] ought to first of all be rejected.

rGyal tshab locates the need to turn back from the non-meritorious in terms of avoiding the negative outcome (i.e. a "bad rebirth," ngan 'gro) which results from a certain moral relativism in which there is no distinction made between good and evil (dge sdig la khyad par med pa). This may result from the student's premature exposure . to non-dualistic Reality. This nihilistic view involves the denial of the conventionally valid causal relation between actions and their results, suggesting the views of the MaterialistsY While at an explicit level rGyal tshab deviates from Candraklrti's CST analysis of the "rejection of the non-meritorious" by drawing out this specific nihilistic outcome, his analysis has strong parallels
56 bZhi brgya pa'i rnam bshad 63a5-6: de nyid bstan pa'i snod du mi rung . ba La bstan na, bya byed mtha' dag La skur ba btab nas dge sdig La khyad par med par shes te I sdig pa byas nas ngan 'gror 'gro bas I thog mar dge mi dge'i las rgyu dang 'bras bur bcas pa bstan na rtogs sla bas, bsod nams min pa sdig pa La 'jug pa dang por ldog par bya ba yin I 57 Although such an outcome is not explicitly spelt out here as being nihilism (nlistikaviida, uccedavlida) in a formal sense, it is not hard to infer that this is the case based on similarities with such descriptions as found in Nagarjuna's Ratnlivalfv1.18 and 1.57 (8.9-17, 24.5-8): krodhlid durvaTlJatli maurkhyam aprasnena vipascitlim I phalam etan manu~yatve sarvebhyaf:t priik tu durgatif:t II and nlistiko durgatirrz yliti sugatirrz yliti clistikaf:t I yathlibhataparijftlin mok~am advayanisritaf:t II as well as the first part of CandrakIrti's PP account of MMK 18.6 to which we shall tum in a moment.

174

Andrew McGarrity

with Candrakirti's PP analysis of MMK 18.6a, i.e. of Nagarjuna's

atmeti prajiiapitam. There, Candraldrti also explains that the rea-

son for making the self known (as a designation) is to preclude the Materialist extreme. In direct keeping with Aryadeva's rejection of the non-meritorious (varar.zam apur.zyasya), Candraldrti first of all makes it clear that "in order to turn away from evil (akusalan nivartayitum) the doers of evil action among inferior followers, "the self," as based upon convention in the world, is sometimes posited by the Buddhas, the Lords."58 He then explains that59
There are those who do not see the entirety of things, even though it is not beyond the scope of the clear view of an ordinary person in the world. [Such fault occurs] by dint of the eye of their mind being completely obscured by the cataract whose dimness is none other than the distorted false view of the non-existence of the self (atmabhava). While based only on the truth of conventional practice, their utmost allegiance is to the mere reality of what are assign~d to be earth, water, fire and air. . ..They claim the mind has only come together based on the mere gestation of the elements, such as the fetus does. Being actively engaged in denying any beginning or endpoint [of personal identity], they deny both the self and the next world with such claims as "This world does not exist; the next world does not exist; the ripening of the fruits of actions of good and evil deeds does not exist; a

58 PP adMMK 18.6 (357.3-4): hfnanaT{l vineyanam akusalakarmaktiri1J.iim akusaliidi* nivartayituT{l buddhair bhagavadbhiJ:z kvacid iitmety api prajfiapitaT{lloke vyavasthiipitaT{l I * Read akusalad following de Jong 1978: 227. 59 ibid. p356.1-3 ... 5-9: iha ye atmiibhiiva*viparyiisakudarsanaghana timirapataliivacchiiditiise~abuddhinayanatayii laukikavadiitadarsanavi~ayiinatikriintam api bhiivajiitam apa.yanto vyavahiirasatyiivasthitii eva santl(lJ:z k~itisalilajvalanapavaniibhidhiinatattvamiitriinuvar1J.anaparii ... kalaliidimahiibhiUaparipiikamiitrasaT{lbhutii eva buddhfr anuvar1J.ayantaf:t purviintiipariintiipaviidapravrttiiJ:z santaJ:z paralokam iitmiinaT{l ciipavadante niisty ayaT{l loko niisti paraloko niisti sukrtadu~krtiiniiT{l karma1J.iiT{l phalavipiiko niisti sattva upapiiduka ityiidina I tadapaviidiic ca svargapavargavisi~!e~!aphalavise~iik~epaparii1imukhiif:t satatasamitam akusalakarmiibhisaT{lskara1J.apravrttii narakiidimahiiprapiitapataniibhimukh
iiMI

* Read iitmfyiisad following de Jong 1978: 226.

Aryadeva's gradual stages

175

sentient being is not brought into being.'>6O And from denying those [assumptions], their face turned away, due to their various scorns, from the distinguished and desired rewards of heaven and ultimate salvation, they are constantly, continuously engaged in the formative innate tendencies of evil actions and face a fall, a mighty plunge into hell etc. Like rGyal tshab's reading of the sort of position to be avoided through the rejection of the non-meritorious, CandrakIrti similarly takes the self that is designated (prajfiapitam) i.e., "the self as based on worldly convention" (iitmety api prajfiapita1!l "loke vyavasthiipitam), as providing a necessary counter to the Materialist-style standpoint that leads to moral relativism. That is, initially positing a self, like initially rejecting the non-meritorious, wards off the materialist view that reality is merely reducible to .the four elements, thus undermining the relation between actions and their results from one life to the next. Specifically, CandrakIrti names the Lokayatikas as holding this viewpoint, contrasting them with such views as the Sarp.khyas' who see a self as necessary to relate actions and results: 61 Having assumed that composite factors perishing from moment to moment do not possess any connection between actions and their results, the S~khyas etc designate the "self" [as the enduring possessor ensuring a connection between actions and their results]. "Nonself," on the other hand, is designated by the Lokayatikas who do not see as necessary a self as the wanderer in sarrtsiira . . The clear similarity with rGyal tshab's treatment of the first part of CS 190, i.e., bsod nams min pa dangpor bzlog, suggests that rGyal

60 This standard summation of the Materialist viewpoint is also found, for example, in Arthaviniscayasutra 315.26-28: mithyiidr~ti/:t katamii? niisty ayarrz loka/:t, niisti paraloka/:t ityiidi purvavat I etan miinasarrz trividham I iyam ucyate mithyiidr~ti/:t II I am grateful to Prof. Peter Skiling for pointing this Siitra out to me. For the Pali antecedents of this summation, see Fuller

2005: 16-18. 61 PP ad MMK 18.6 (358.3-4): iitmety api prajiiapitarrt siirrtkhyiidibhi/:t, pratik~a1J-avinasvarii1J-iirrt sarrzskiirii1J-arrz karmaphalasarrzbandhiibhiivam utprek~ya I aniitmety api prajiiapitarrz lokiiyatikair upapattyiitmiinarrz sarrzsartiiram apasyadbhi/:t I

176

Andrew McGarrity

tshab is actually more influenced by the PP context in which CS 190 is quoted than he is by Candraklrti's direct CST analysis of CS 190 itself. Although the "rejection of the non-meritorious" (viiral}am apul}yasya) is not mentioned in PP ad MMK 18.6a, there is mention of the near-identical need "to turn away from evil" (akusaliid nivartayitum). Likewise, although the need to provisionally posit a self is not mentioned in Candraklrti's CST ad CS 190, there is direct agreement between what rejecting the non-meritorious and provisionally positing a self are meant to counter, namely the materialist viewpoint, thus averting a bad rebirth. rGyal tshab's analysis makes clear the implicit correspondence between Aryadeva's initial rejection of the non-meritorious and the provisional positing of the self which Nagarjuna advocates. Moreover, it also indicates that this correspondence seems to have been assumed in the Tibetan tradition which rGyal tshab inherits. The assumption is that the initial positing (prajfiapitam) of the self amounts to a means of turning back the non-meritorious, seeing as both have a shared outcome (avoiding a hell rebirth). Following his explanation of the initial provisional positing of the self in PP ad MMK 18.6, Candraklrti introduces the threefold division of inferior, intermediate and advanced followers (hfnamadhyotkr~tavineya)62 to whom the Buddha addresses his teaching, in accordance with his wisdom, skillful means and great compassion (prajfiopiiyamahiikarul}ii).63 This would also seem to correspond to Aryadeva's delineation of three types of students at CS 189 before he sets out the three rejections at CS 190. 64 Thus, in his PP ad MMK 18.6 Candraklrti makes explicit the hierarchy implicit in Nagarjuna's verse, and brings us a step closer to the threefold typology of Aryadeva's gradual rejections.

62 PP ad MMK 18.6 (357.3), 63 ibid p. 357.2. 64 CS 189 (8.14; p. 82): sbyin pa dman pa la gsungs shing / / 'bring la tshul khryims gsungs pa ste / / mchog la zhi ba gsungs gyur pa / / des na rtag tu mchog tu byos /

Aryadeva's gradual stages

177

The rejection of the self and the teaching of non-self

1. The five types of relation between the self and the skandhas
The most obvious correspondence between MMK 18.6 and CS 190 is between Niigiirjuna's "teaching of no-self" (anatmety api desitam) and .Aryadeva's intermediate "rejection of the self" (varal}am atmanaJ:t). In Candraldrti's CST analysis, the analysis of .Aryadeva's second rejection is virtually inseparable from the analysis of the final rejection. CandrakIrti sets out a clear transition from the initial rejection of the non-meritorious to the rejection of the self, and then on to the rejection of all things, which is considered necessary for, and is directly aligned with, complete liberation without any grasping remaining (i.e., the non-rejection of even a single thing would be tantamount to there still being some grasping left over). Candraklrti explains that65
By rejecting such faults [as those previously set out as leading to the non-meritorious], the five types [of relation between] the self and (i) the skandhas, (ii) dhatus, (iii) iiyatanas are rendered impossible, meaning that the twenty-part view of the self may be cast aside. Finally, attachment to the skandhas, dhiitus and iiyatanas correctly shown [to stem from] attachment to the self may also be completely eliminated, hence all things may be rejected (dngos po thams cad). So it is in such a way that this wise person is completely liberated (yongs su mya ngan las 'das = parinirvrta) without any grasping remaining. In this way, one who understands the attainment of the rejection of all, having entered upon these gradual stages, is explained, according to the Buddha's teaching, as a wise person.

. 65 CST 139al-3 ad CS 190: sdig pa bzlog pas ni phung po dang khams :dang skye mched roams la bdag roam pa Ingar yang mi srid pa'i phyir bdag :tu Ita ba cha nyi shu dor bar bya'o / / phyi nas bdag la chags pa yang dag par ,ston pa'i phung po dang khams dang skye mched dag la chags pa yang yongs 'su gtong bar bya ba'i phyir dngos po thams cad bzlog par bya'o / / de'i phyir de Itar na blo dang ldan pa 'di len pa med par yongs su mya ngan las 'das te / 'de Itar na rim pa'i sgo nas gang zhig gis thams cad bzlog pa'i thabs shes pa / ;de Ia ni sangs rgyas kyi bstan par mkhas pa zhes brjod do /

178

Andrew McGarrity

There is no suggestion that Candraklrti sees the final rejection itself as being of all views, but simply all things, as per the most straightforward reading of "sarvasya" in his PP citation. Candraklrti's primary concern is with providing a step by step metaphysical critique. If a certain type of attachment can be first of all undone, then the next, and then the rest can be unraveled too. Hence, the order in which they are approached is important: the rejection of the self naturally leads on to the rejection of all things. CandrakIrti's main point is to lay bare a set of necessary metaphysical entailments that necessitate the rejection of all things, which amounts to complete liberation without any remnant grasping. This rejection of all things is seen as tantamount to realizing the emptiness (of inherent nature) of all things, which is then, in Aryadeva's following verse, said to be possible by correctly seeing the emptiness of just one thing. 66 However, in order to understand how Tibetan readings, such as rGyal tshab's, which simply proceed on the basis of the assumption of Pa tshab's "all views" translation, are reconciled so unproblematically with Candraklrti's analysis here of the "all things" reading, we also need to focus upon the strategy that Candraklrti employs, and not just the state of affairs he wishes to expose. CandrakIrti understands the rejection of the self in terms of a rejection of possible ways that the self might be conceived. The first of these mentioned ad CS 190 is the five types of possible relation it may have with the skandhas, dhiitus or ayatanas (phung po dang khams dang skye mched rnams la bdag rnam pa Inga). Now, in regard to the self and the skandhas, five possible types of relation are found in MMK 10 (agnfndhanaparfk~ii) using the template of the critique of possible relations between "fire and kindling." The relations themselves are specifically set out at MMK 10.14 before their applicability to the self and skandhas is made explicit at MMK 10.15. CandrakIrti ar66 CS 191 (8.16; p. 82): dngos po gcig gi ita po gang II de ni kun gyi ita par bshad II gcig gi stong nyid gang yin pa II de ni kun gyi stong pa nyid I ("The one who is the seer of one thing is said to be the seer of all things: that which is the emptiness of one thing is the emptiness of all things.'). Sanskrit quoted by CandrakIrti, PP ad MMK 4.9 (128.4): bhavasyaikasya yo dra~tii dra-?!ii sarvasya sa smrtalJ I ekasyasunyatii yaiva saiva sarvasya sunyatii II

Aryadeva's gradual stages

179

gues in his PP analysis of MMK lO.14 that Nagarjuna's arguments in the root verse are simply an amplification of the basic identity and difference refutations already applied at the beginning of the agnlndhanaparlk~ii (MMK lO.1):67
(MMK 10.14) Fire, moreover, is (i) not kindling, (ii) nor is fire elsewhere apart from kindling. Fire is (iii) not a possessor of kindling, (iv) neither is kindling in fire, (v) nor is it [fire] in that [kindling]. (PP) There, [at MMK lO.14a, it is as already set out at MMK lO.1ab, viz.,] ...
"If what is kindling is fire, then identity between the actor and the action ... "

.,. would ensue. It is by this rationale [at MMK lO.lab] that [now at lO.l4a] fire, moreover, is not kindling, on account of the negation of identity between fire and kindling. Along such lines as [MMK lO.lcd viz.,] .. . " ... if fire were other than kindling, it would come about even without kindling ... " ." [it is now explained at MMK lO.l4b that] nor is fire elsewhere apart from kindling, on account of the negation of difference. The alternatives of what depends and what it is dependent upon, i.e., the alternatives of possession, are in fact also negated on account of

67 MMK 10.14 with PP (211.14-212.12):

(10.14) indhanaJ?'! punar agnir na niignir anyatra cendhaniit I niignir indhanaviin niigniiv indhaniini na te~u sal;t II tatra I

(1O.1ab) yad indhanaJ?'! sa ced agnir ekatvaJ?'! kartrkarmalJol;t I

bhaved ity aneniignfndhanayor ekatvaprati~edhiid indhanaJ?'! punar agnir nail (lO.1cd) anyas ced indhaniid agnir indhaniid apy rte bhavet I ityiidiniinyatvasya prati~edhiin niignir anyatra cendhaniit II tattviinyatvob hayapak~aprati~edhiid eva tadvatpak~iidhiiriidheyapak~iilJiim apy arthatal;t prati~iddhatviit tan api nigamayann iiha I niignir indhanaviin nagniiv indhaniini na te~u sal;t II iti I tatriignir indhanaviin na bhavati I indhanam asyiismin vii vidyata iti vyatirekelJa vii vyutpiideniivyatirekelJa va I tatra vyatirekelJa tadyathii gomiin devadattal;t, avyatirekelJa buddhimiin devadatto rupaviin ityadi I agnfndhanayos ca pak~advayasyiipi prati~iddhatviid indhanaviinagnir iti prati~edho vihital;t II

180

Andrew McGarrity

the very negation of both alternatives of either (i) identity [between fire and kindling] or (ii) difference [between them].58 Hence, understanding any of those, it is then said [at MMK lO.14cd], "fire does not possess kindling, neither is kindling in fire, nor is it [fire] in that [kindling]." There, "fire is not a possessor of kindling" is that "there is kindling either belonging to [fire] or in it," whether [possession] in terms of separation (vyatireka) or in terms of non-separation with distinction (vyutpiideniivyatireka). There [possession] with separation is along such lines as "Devadatta possesses a cow," and [possession] with non-separation is along the lines of "Devadatta possesses intelligence or possesses physical form."59 And so the negation of "fire as 58 tattviinyatvobhayapak~a. tattva here is equivalent to ekatva. That is, tattva (= ekatva) is in opposition to anyatva, given the reference to ubhayapaksa. Cf. also La ValIee-Poussin 1911: 311 on the compound tattviinyatve ad MAv 6.143d when he renders "l'litman n'est ni la meme chose, ni autre chose que Ie rupa" for bdag ni gzugs las de nyid gzhan nyid med I tattviinyatve rupato niitmanalJ stalJ (Sanskrit cf. PP ad MMK 22.1 (435.2)). 59 Candraklrti'svyatireka(e.g.gomiindevadattalJ)andvyutpiideniivyatirekal avyatireka (e.g. buddhimiin devadatto rupaviin) seem equivalent to the early Vaie~ika notions of "conjunction" (samyoga) and "inherence" (samaviiya) respectively. In the Vaise~ika Sutras (VS), saY(!yoga (conjunction) between one thing and another is subject to, or born from, action (e.g., VS 7.2.9: anyatarakarmaja ubhayakarmajalJ saY(!yogajas ca saY(!yogalJ). On the other hand, samaviiya (inherence) between two things is independent of action (like quality and substance etc) (cf. VS 5.2.25: ni~kriyiil}-iiY(! samaviiyalJ karmabhya~ prati~iddhalJ). SaY(!yoga is temporary while samaviiya is not. At VS 1.1.1519, samyoga and vibhaga ("disjunction') are pointedly distinguished from the characteristic of quality (gul}-alak~al}-am) in relation to substance, due to their relation with their "possessor" (saY(!yogin) being merely characterized by action: dravyiiSrayy agul}-aviin saY(!yogavibhiige~v akiiral}-am anapek~a
iti gUl}-alak~al}-am II ekadravyam agul}-aY(! saY(!yogavibhiige~v anapek~aT(! kiiral}-am iti karmalak~al}-am II dravyagul}-akarmal}-iiT(! dravyaY(! kiiral}-aY(! siimiinyam II tatha gUl}-a~ II saY(!yogavibhagiiniiY(! karma II As in the relation betvyeen Devadatta and possessing a cow (gomiin) in Candraklrti's example of vyatireka, at VS 7.2.19 saY(!yoga is compared to the relation between a stick and the person holding it (= the saY(!yogin). Similarly, like the posesessive relation between Devadatta and his physical form (rupaviin) or mind (buddhimiin) in Candraklrti's example of vyutpiideniivyatirekalavyatireka, at VS 9.1.15 qualities of the soul (litmagul}-a) are related to the soul itself by way of samaviiya: iitmasamaviiyiid litmagul}-e~u II. See also VS 5.1.1-2 in which saY(!yoga denotes the relation between a hand and the pestle (mu~ala) it touches (i.e., a temporary relation determined by action), while at VS 8.9 samaviiya

Aryadeva's gradual stages

181

the possessor of kindling" is prescribed from the negation of either of the two alternatives of fire and kindling.

Explicitly drawing this typology back to the self and the skandhas, Nagarjuna himself states in the following root verse (MMKlO.15) thafD
By means of fire and kindling is explained the method for the self and the [skandhas subject to] appropriation, along with pots and cloths etc', all without exception.

The fivefold critique of the relation between fire and kindling is directly transferable to the critique of the self and the upadanaskandhas ("the skandhas subject to appropriation") and then to the refutation of all. In his move in this verse from the self and skandhas onto "all without exception" (sarvo niravase~et:za) Nagarjuna foreshadows the final two rejections (varat:zam) of the self and then all (things) in (the Sanskrit of) Aryadeva's three rejections. This suggests not only that this verse indicates a set of assumptions in common with Candraklrti's CST analysis - perhaps even providing Candraklrti with his most immediate MMK source for his reference to the five possible relations in relation to CS 190 .,- but it also indirectly confirms "all" as found in the Sanskrit. In his PP analysis of MMK 10.15, CandrakIrti assumes the intertwined associations of upa + ..;aa as "clinging/appropriation" and, epistemologically, the "dependent basis of a designation" following the general Buddhist assumption that designation is itself a type of clinging or appropriation. CandrakIrti assumes that the five upadanaskandhas, i.e., the five skandhas subject to appropriation (hence mistakenly appropriated as a self), must necessarily entail, by definition, the self (atma) which is designated in the appropriation (upadaya prajfiapyate) of them. The self is precisely a "dependent designation" (upadaya prajfiapti)71 imputed upon the
describes the quality-substance relation between whiteness (vaitya) and a white object (veta) (= the samavayin). 70 MMK 10.15 (212.16-17, 213.14): agnfndhanabhyaTJl vyakhyata atmopadanayoJ:! kramaJ:! I sarvo niravae~e1Ja sardhaTJl ghatapatadibhiJ:! II 71 Cf. Nagarjuna's famous terminology elsewhere at MMK 24.18 (503.1013): yaJ:! pratftyasamutpadaJ:! iinyataTJl taTJl pracak~mahe I sa prajiiaptir

182

Andrew McGarrity

appropriated skandhas (upadanaskandha). In the very act of such subtle clinging to them, the skandhas as "subject to appropriation" (upadana) are necessarily the imputandum (upadana) of the "self"imputation. As Jikid6 Takasaki has neatly put it, "the aggregates of five elements 'taken' as the stuff for the concept of iitman are called 'paficopadanaskandhas ... in contrast to paficaskandhas as the elements of dharma in general."72 Candraklrti explains that73
There, what is subject to appropriation (upiidiina), the "appropriated" (upiidfyata), refers to the five skandhas subject to appropriation (paiicopiidiinaskandhiilJ'). That which is designated in the appropriation of (upiidiiya prajiiapyate) those [five upiidiinaskandhas], i.e., what appropriates (upiidiitii), the grasper, the accomplisher, is referred to as the "self." Due to its taking the I-notion as its object, egotism is generated in dependence upon this [the "self"]. The method of proof pertaining to the self and what is subject to appropriation [i.e., the five upiidiinaskandhas] as though in a relation of possession (tadasya) is to be understood as applying to all things without exception, as exupiidiiya pratipat saiva madhyamii II 72 Takasaki 1988: 1453[3]. See also on this, Gethin (1986: 35-53) who notes in relation to the Pali Nikiiyas that the simple term khandha acts as a "neutral term" (p. 39) in contrast to the defiled "upiidiinakkhandha." Hence, if a "self" were not mistakenly imputed, then the term upiidiinaskandha must necessarily be used as opposed to the "neutral" "skandha." Gethin explains (p.39): The term upiidiinakkhandha signifies the general way in which the khandhas are bound up with upiidiina; the simple khandha, universally applicable, is used in the Nikiiyas and especially the Abhidhamma texts as a neutral term, allowing the specific aspects of, for example, upiidiina's relationship to the khandhas to be elaborated. 73 PP ad MMK 1O.15abc (212.18-213.2 ... 213.7-11): tatropiidfyata ity upiidiinaTfl paiicopadiinaskandhii~ I yas tiin upiidiiya prajiiapyate sa upiidiitii gralfftii ni~piidaka iitmety ucyate I ahaTflkiiravi~ayatviid iihita utpiidito 'haTflmiino 'sminn iti II tadasyiitmana upiidiinasya ca ya~ kramaf:t siddh*, sa sarvo 'gnfndhaniibhyiiTfl vyiikhyiito 'vagatantavyo niravase~elJa II ... tatra yad evopiidiinaTfl sa eviitmety evaTfl kartrkarmalJor ekatvaprasaftgiin na yujyate I niipy anyad upiidiinam anya upiidiitii skandhavyatirekelJiipy iitmopalabdhiprasaftgiit, paratra nirapek~atviid ityiidiprasaftgiic ca I ekatviinyatvaprati~edhiic ca skandhaviin apy iitmii na bhavati I anyatviibhiiviic ca niitmani skandhii na skandhe~v iitmii I yata evaTfl paiica~u prakiire~v iitmano na sattvaTfl

Aryadeva's gradual stages

183

plained by means of fire and kindling ... In that case, it is not fitting for the self to be exactly what is subject to appropriation (upadanam), for otherwise there would be identity between the actor and the action. Nor also is what is subject to appropriation (up adana) other than what appropriates (upadata), for otherwise the ascertainment of the self would take place even in separation from the skandhas, and elsewhere it is said to be because otherwise they would be unrelated etc. And due to the negation of identity and difference, the self is not a possessor of the skandhas either; and from the non-existence of difference, it cannot be the case that the skandhas are in the self, nor is the self in the skandhas, in which case there is no existence of the self in the five types of [relation with the upadanaskandhas].

Thus, following the fire and kindling model, the five types of relation critiqued between the imputed self (as the upiidiiya prajfiapti) and the (upiidiina)skandhas (as the upiidiina) are:
(i) (ii) (iii) (iv) (v) the self is the skandha; the self is different tol other than the skandha; the self possesses the skandha; the skandha is in the self; the self is in the skandha.

2. The twenty-fold view of the self

Candraklrti's specific reference to "the twenty-fold view of the self" (bdag tu Ita ba cha nyi shu, perhaps more literally, "the twenty-part view of what is conceived as the self) in relation to CS 190 suggests the older tradition of formulating the view of selfhood in terms of the "satkiiyadr~ti" (Pali: sakkiiyadi!thi). The satkiiyadr~ti / sakkiiyadi!thi indicates the general view of existent personality or selfhood, involving, as Paul Fuller puts it, an "identification with the khandhas."74 For reasons that will become apparent, I am hesitant to attempt to provide a direct translation of satkiiyadr~ti for now. 75 In the Pali sources, we find in the Ciilavedalla Sutta an exFuller 2005: 26. Fuller (loc.cit) provides a summary of various English translations of the Pali sakkayaditthi, noting that "Collins translates sakkaya-ditthi as 'per74

75

184

Andrew McGarrity

change between the bhikkhunl Dhammadinna and the lay follower Visakha, in which the sakkiiya itself is defined as the five upiidiinakkhandas (the five khandhas subject to appropriation).76 Further on, Visakha then asks Dhammadinna specifically "How then, 0 Lady, does the sakkiiyaditthi come about?" to which she responds:??
Here, a friend Visakha, an ignorant ordinary person, unhearing of the Noble Ones, without a correct understanding with respect to the Dhamma of the Noble Ones, untrained with respect to the Dhamma of the Noble Ones, unhearing of good men, without a correct understanding with respect to the Dhamma of good men, untrained with respect to the Dhamma of good men, regards (i) rapa as the self, or (ii) the self as possessing rapa, or (iii) rupa as in the self, or (iv) the self as in rupa [and so on for the other four khandhas] ... in this way, a friend Visakha, does the sakkayaditthi come about.

Stripped of this narrative context, an identical description also occurs in the DhammasangalJ-i,78 while in the Pafifiiibhuminiddeso of the Visuddhimagga, Buddhaghosa cites the same formulation and actually describes the sakkiiyaditthi as being "twentyfold" (literally, as "having a twenty-fold basis," vlsativatthukii).79 The twenty

sonality belief,' Gethin as 'the view that the individual exists' while Bikkhu Bodhi takes the term as 'identity-view'."
76 Cu!avedallasutta (121; p. 299.7-9): sakkiiyo sakkiiyo ti ayye vuccati. Katamo nu kho ayye sakkiiyo vutto Bhagavatii ti. - Pafica kho ime iivuso Visiikha upiidiinakkhandhii sakkiiyo vutto Bhagavatii. 77 ibid, 300.4-10): Katharh pan' ayye sakkiiyaditthi hotfti. - Idh' iivuso Visiikha assutavii puthujjano ariyiinarh adassiivf ariyadhammassa akovido ariyadhamme avinlto, sappurisiinarh adassiivf sappurisadhammassa akovido sappurisadhamme avinlto, ruparh attato samanupassati, rupavantarh vii attiinarh, attani vii ruparh, rupasmirh vii attiinarh ... Evarh kho iivuso Visiikha sakkiiyadiuhi hotfti. 78 'Dhammasaliga]Ji 1003 (182.30-35): Tattha katamii sakkiiyadiuhi? Idha assutavii puthujjano ariyiinarh adassiiv! ariyadhamassa akovido ariyadhamme avinfto sappurisiinarh adassiivi sappurisadhammassa akovido sappurisadhamme avinfto ruparh attato samanupassati. 79 Visuddhimagga 27: Pafifiiibhuminiddeso (570.2-6): Vfsativatthukii sakkiiyadiuhi attaviidupiidiinarh. Yath' iiha.' - tattha katamarh attaviidapiidiinarh? Idha assutavii pathujjano ... pe ... sapurisadhamme avinfto ruparh attato samanupassati ... pe ... vipariyesagiiho: idarh vuccati attaviidupiidiinan ti.

Aryadeva's gradual stages

185

permutations of the sakkayaditthi are arrived at by way ofJour possibilities for the relation between the imputed self and each of the five khandhas. Specifically, in relation to each respective khandha, it consists of the mistaken views that
(i) the self is the khandha; (ii) the self possesses the khandha (iii) the khandha is in! dependent upon the self; (iv) the self is in! dependent upon the khandha

- thus adding up to twenty (4 permutations x each of the 5 khan"dhas).BO Although not doing so ad CS 190, CandrakIrti elsewhere refers to the satkayadr$ti by name, describing it as "expansive, rising up to twenty peaks" (vif!Lsatisikharasamunnatataratiprthusatkayadr$ti)B1 which abbreviates the fuller image of "the satkayadr$ti-mountain reaching up to twenty peaks [= the twenty permutations]" (vif!LsatiiikharasamudgataJ:t satkayadr$tisaila) found, for example, at Mahavyutpatti 208 and recurrently in Avadanasataka 53-63. B2 This draws upon an older Buddhist tradition of appropriating the cosmological imagery of Mount (Su)Meru to denote the view of selfhood, suggesting that just as the cosmos is supported by Meru, so all afflictions are derived from the view of the self. In relation to MMK 13.8, CandrakIrti cites the Kasyapaparivarta of the Ratnakuta Sutra which refers to the pudgaladr$ti as being the size of Sumeru (sumerumatra), B3 thus suggesting that this sutra is CandrakIrti's immediate source for his "mountain-imagery." Suggesting his audience's familiarity with the image, CandrakIrti in fact uses it to conclude his most extensive analysis of the

BO For a further summary of sakkayaditthi from the Pali sources, see Fuller 2005: 26-28. B1 PP ad MMK 16.8 (294.7-8): viTfLsatisikharasamunnatataratiprthusat. kayadr~ti B2 Mahiivyutpatti 208 as cited by Louis de la Vallee-Poussin ([1992] 190313: 294, n. 2); Avadanasataka 68, 125ff - see further note below. B3 Kasyapaparivarta from the Ratnakilta cited at PP ad MMK 13.8 (248. 9-10): varaTfL khalu kasyapa sumerumiitra pudgaladr~tir asrita na tv evabhavabhiniveSikasya silnyatadr~ti/:l /

186

Andrew McGarrity

satkiiyadr~ti at MAv and MAvBh 6.143-5. Drawing upon a motif found especially in Avadiinasataka 53-63, CandrakIrti embellishes the association by colorfully elaborating that84

Having conquered, through diamond-like wisdom, the heights of the twenty peaks of the satkayadr,v{i-mountain, the fruits of 'stream-entry are manifested. 85 As is explained in Agama: (6.145) The self and what accompanies it [i.e., "mine," atmiya] are destroyed by conquering, through the diamond-like understanding of selflessness, the [satkaya-]dmi mountain, [specifically,] the heights of these peaks situated upon the massive satkliyadmimountain. Untouched by the diamond-like wisdom of the Arya, the satkliyadr~{i mountain is built up day by day from the rocks of defilements. Arisen from beginningless sa'!lslira, its height elevated into the three realms (tridhlitu), it extends in all directions without exception, rising up out of the golden base of ignorance. 86 It should be understood then that
84 MAv 6.145 with Bhii,vya (267.14-268.7): de la jig tshogs la Ita ba'i ri'i rtse mo nyi shu mtho ba ye shes kyi rdo rjes beom nas rgyun du zhugs pa'i 'bras bu mngon du byas so zhes gang lung las gsungs pa ni / Ita ri bdag med rtogs pa'i rdo rje yis / / beom bdag gang dang lhan cig jig 'gyur ba / / jig tshogs Ita ri lhun stug la gnas pa'i / / rtse mo mtho bar gyurpa 'di dag go / / jig tshogs la Ita ba'i ri bo 'phags pa'i ye shes kyi rdo rje ma babs pas nyin re bzhin nyon mongs pa'i brag rdo 'phel ba / 'khor ba thog ma med pa nas byung ba / dpangs khams gsum du mtho ba / phyogs kyi ngos ma Ius par khyab pa / ma rig pa'i gser gyi sa gzhi las 'thon pa / bdag med pa khong du ehud pa'i rdo rjes bshig pa na / rtse mo ehes mtho ba gang dag dang lhan eig jig par 'gyur ba de dag ni rtse mo yin par shes par bya'o / / 85 Avadlinasataka53-63 (135.14-15,143.18-19,146.18-19,148.30,150.2122" 158.6-7, 160.3-4): vi'!lsatisikharasamudgata'!l satkiiyadr,v{isaila'!l jiiiinavajre1J.a bhitvii srotaiipattiphala'!l siik~iitkrtam. Again, I am grateful to

Prof. Peter Skilling for pointing me in the direction of this work. 86 The cosmic Mt (Su)Meru is said to arise out of a base of gold. For a Buddhist account of this, see AK and Bhii~ya 3.45-51b. There, the "receptacle-world" (bhiijanaloka) rests on space (iikiisaprati~tha). Within the bhiijanaloka is first of all the "circle of wind" (viiyuma1J.q.ala), upon which rests the "circle of water" (apama1J.q.ala), upon which, in tum, then rests the "circle of gold" (kiiiieanama1J.q.ala). Resting upon this are the eight

.Aryadeva's gradual stages

187

upon the destruction [of the satkayadr~ti-mountain] by the diamondlike comprehension of selflessness, [the respective "self-skandha" alternatives] are referred to as "peaks," inasmuch as the peaks are, at their height, those along with which [the mountain itself] would be destroyed.

Candraklrti's technical analysis of the satkiiyadr'i!i over the previous two verses accords with the four permutations found in the Pali tradition. He explains that87
(6.143) The self is not regarded as possessing form, since the self does not exist, hence the meaning denoted by possession (mattvartha) is not applicable, in the case of either [possession] in separation, e.g., possessing a cow, or in non-separation, e.g.,

great mountains, with Mt Sumeru at their centre (AKBh ad AK 3.48b-49c: mahaparvatiilJ / madhye sumerulJ I). There are some difficulties with this account however and I have drawn on Siistri's edition as Pradhan lacks 3.48b-49c as cited above. 87 MAv 6.143-144 with Bhana (266.1.2-19): bdag ni gzugs ldan mi 'dod gang phyir bdag / yod min de phyir ldan don sbyor ba med / / gzhan ni gnag ldan gzhan min gzugs ldan na / / bdag ni gzugs las de nyid gzhan nyid med / / bdag phung po rnams de nyid dang gzhan nyid yin pa ni sngar bkag zin la / ldan pa'i skyon yang lha sbyin gzugs dang ldan no zhes bya ba de ni tha mi dad pa nyid la dmigs la / ba lang dang ldan no zhes bya ba ni tha dad pa yin na / de nyid dang gzhan nyid de dag ni gzugs bdag la yod pa ma yin pas bdag gzugs dang ldan no zhes bya ba 'di yang mi srid do / / da ni bkag zin pa'i phyogs rnams bsdams te bstan pas 'jig tshogs la lta ba dmigs pa dang rnam pa log pa la 'jug par grangs kyi sgo nas bstan pa'i phyir bshad pa / gzugs bdag ma yin bdag ni gzugs ldan min / / gzugs la bdag med bdag la'ang gzugs yod min / / de ltar rnam bzhir phung kun shes bya ste / / de dag bdag tu lta ba nyi shur 'dod / / phung po lnga po bdag dang bral dag* la 'jig tshogs la lta bas rnam pa bzhi'i sgo nas bdag nyid du hzung ba na 'jig tshogs fa lta ba'i cha nyi shur 'gyur ro / * Read bdag following MAvBht 302bl.
kiiftcanamaf}alaprati~tha a~tau

188

Andrew McGarrity possessing physical form. The self has neither identity with, nor difference from, form.88 The self as either identical to or different from the skandhas has been refuted earlier: Moreover, the fault of possession is in the conception of non-separation, e.g., Devadatta possesses physical form, or in separation, e.g., possessing a bullock. Yet by virtue of the fact that identity and difference do not pertain to the self, this notion that "the self possesses form" is also impossible. Now, such an explanation acts as a summation of the positions already refuted. Hence, for the sake of an explanation in terms of an enumeration actually introducing the mistaken modes as covered in the notion of satkiiyadr~ti, it is explained:

(6.144) (i) Form is not the self; (ii) the self is not a possessor of form; (iii) the self does not exist in form nor (iv) does form exist in the self. In such a way according to these four modes are all the skandhas to be understood; they are regarded as the twenty views of the self. The satkiiyadr~ti consisting of either the self apart from or possessing the five skandhas apprehends the self according to four modes, hence the twenty-part satkiiyadr~ti.

Identity, dependence (either of the self on/in the skandhas or the skandhas on/in the self) and possession (including the further subdivision into "separation" and "non-separation") are all in keeping with the account of fire and kindling and the self and the upiidiinaskandhas at MMK 10.14-15 and its PP analysis. The overlap with the PP account is apparent right down to the same examples for vyatireka- and avyatireka-possession. The glaring discrepancy, however, is that the original sakkiiyadinhi/satkiiyadr/fti arrangement as found in the Pali and in Candraklrti's MAv 6.143-144 and Bhii/fya is only fourfold, lacking the explicit delineation of difference between the self and the skandhas, whereas the arrangement in regard to fire and fuel is fivefold, with difference also being included. In regard to a fourfold
88 Quoted at PP ad MMK 22.1 (434.19-435.2): i~to niitmii rupaviin niisti yasmiid iitmii mattviirthopayogo hi niitaJ:t I bhede gomiin rupaviin apy abhede tattviinyatve rupato niitmanaJ:t staJ:t II

Aryadeva's gradual stages

189

arrangement, the twentyfold view of the self is straightforward (4 permutations x 5 skandhas), however the problem remains of how the fivefold arrangement is reconciled with the "twenty peaks," as is the case in the reference CST ad CS 190 where the five modes and twentyfold view of the self are mentioned together. Candrakirti is aware of this and in concluding his Bhii~ya ad MAv 6.144 explains that 89
If [a query is raised to the effect that] the investigation [into the self and the skandhas] is on occasion acknowledged as five modes - after all, is it not said in Sastra that
(MMK 22.1) "The Tathiigata is (i) not the skandhas, nor (ii) other than (anya) the skandhas; the skandhas are not (iii) in Him, nor (iv) is He in them; nor (v) does the Tathiigata possess the skandhas - who then is the Tathiigata?,,90- thus, this would be five parts; why then is it said to be twenty? [Then our response is;] Those parts of the satkayadr~ti being respectively determined from scripture, one is unable to adhere to a self as it is ungrasped by any of those [modes, i.e., identity, possession and 2 x dependence], thus providing the conception of the skandhas according to the four modes. In this way, there is no adherence to the self anywhere apart from the skandhas [in any of the four modes], hence a fifth mode of the satkiiyadr~ti does not apply [for Buddhists].
89 Bhii~ya ad MAv 6.144 (266.19-267.14): gal te rnam par dpyad pa rnam pa lnga la skabs su bab la I bstan beos las kyang I phung min phung po las gzhan min I I de la phung med der de med I I de bzhin gshegs pa phung ldan min I I de bzhin gshegs pa gang zhig yin I I zhes gsungs pa ma yin nam I de'i phyir 'di eha lngar 'gyur na I eha nyi shur ji skad brjod ee na I 'jig tshogs la lta ba'i eha de dag ni mdo sde las rnam par gzhag la des kyang phung po rnams ma bzung bar bdag tu mngon par zhen par mi nus pas rnam pa bzhi'i sgo nas phung po rnams la dmigs shing 'jug go I I 'di ltar phung po ma gtogs par ni 'ga' la yang bdag tu mngon par zhen pa med pas 'jig tshogs la lta ba'i rnam pa lnga pa mi srid do I I de'i phyir 'jig tshogs la lta ba ni eha nyi shu kho na'o I I bstan beos las lnga pa gzhan pa nyid kyi phyogs su gsungs pa ni mu stegs pa'i lugs dgag pa'i phyir ro zhes bya bar shes par bya'o I I 90 MMK 22.1 (432.12-13); skandhii na niinyaJ:! skandebhyo niismin skandhii na te~u saJ:! I tathiigataJ:! skandhaviin na katamo 'tra tathiigataJ:! II

190

Andrew McGarrity

Therefore, the satkiiyadr~ti is only twenty. In scripture, the explication of the position of difference (anyatva) as a fifth mode is in order to refute non-Buddhists [who might indeed posit a self that is beyond the four possible relations with the skandhas].

By this rationale, Candraklrti's reference to five types of relation in CST ad CS 190 must be directed at non-Buddhists, while the reference to the twenty-fold view also leaves it open to his fellow Buddhists, thus covering both audiences. Another way of dealing with the five types of relation between the self and the skandhas is found in Candraklrti's comments at the beginning of Nagarjuna's iitmaparfkifa (MMK ch.18). There, Candraklrti refers to the satkiiyadrifti by name and sees it as the basis of all defilements. At a more subtle level, the self is described as the object-condition (iilambana(-pratyaya)) of the satkiiyadrifti, while at a more overt level the self is defined as whatever is the object of the I-notion (ahaf(lkiira). Candraklrti begins by first of all citing his earlier MAv (6.120), using this as a platform for his further explanation of the five relations: 91
It is pointed out at length in the MAv with such verses as:

91 PP ad MMK 18.1 (340.8-341.6): satkiiyadr~[iprabhaviin ase~iin klesiil'{ls ca dO:jiil'{ls ca dhiyii vipasyan I iitmiinam asyii vi:jayal'{l ca buddhvii yogf karoty iitmani:jedham eva II ityiidinii madhyamakiivatiiriid vistare1Jiivaseya/:t II kiiyadr:j[imulakam eva sal'{lsiiram anupasyan, tasyiis ca satkiiyadr:j[er iilambanam iitmiinam eva samanupasyann iitmiinupalambhiic ca satkiiyadr~tiprahii1Jal'{l tatprahii1Jiic ca sarvakleiavyiivrttil'{l samanupasyan prathamataram iitmiinam evopaparfk~ate ko 'yam iitmii niimeti yo 'hal'{lkiiravi:jaya/:t sa ciiyam ahal'{lkiirasya vi~aya/:t parikalpyamiina/:t skandhasvabhiivo vii bhavet skandhavyatirikto vii II iidhiiradheyatadvatpaksii1Jiim apy ekatviinyatvapak:ja eviintarbhiiviit sal'{lk~epe1Jaiva ca* vivak:jitatviid ekatviinyatvapak~advayaprati:jedhenaiviitma ni:jedham iirabdhukiima iiciirya iiha I (18.1) iitmii skandhii yadi bhaved udayavyayabhiig bhavet I skandhebhyo 'nyo yadi bhaved bhaved askandhalak~a1Ja/:t I kim arthal'{l punar anyatra tathiigataparfk:jiiyiim agnfndhanaparfk~iiyiil'{l ca paiica pafica pak:jii upanyastii iha tu puna/:t pak:jadvayam eveti II * Read sal'(lk:jepasyaiva ca following de Jong 1978: 224.

Aryadeva's gradual stages

191

(MAv 6.120) Discerning with their mind defilements and faults, without exception, arising from the satkiiyadr$ti, and having understood the self to be its [satkiiyadr$ti's] object, the yogin carries out the refutation of the self.92 Understanding sa'!lsiira to have as its very root the kiiyadr$ti, and seeing the very self to be the object condition (iilambana) of that satkiiyadr$ti,93 one understands that from the non-ascertainment of the self occurs the abandonment of the satkiiyadr$ti, and from the abandonment of that [satkiiyadr$ti], one realizes the rejection of all defilements. One first of all conducts an investigation into the self itself [in this way:] What is this that is referred to as the self? [Answer:] Whatever is the object of the I-notion (aha'!lkiira). And this object of the I-notion being envisaged, it would be [taken as] either the nature of the skandhas or separate from the skandhas. Due to the positions [of the self as] (i) the basis of [the skandhas], (ii) based upon [them] or (iii) possessing them being included in just the position of identity and difference, and out of a desire for conciseness of expression, the teacher's [Nagarjuna's] wish is to begin the refutation of the self with just the negation of the two positions of identity and difference. He states that (MMK 18.1) If the self were the skandhas, it would participate in their arising and passing away; if it were other than the skandhas, it would not have the characteristic of the skandhas. What is the point moreover [of the fuller explanation] as each of the five alternatives are set out elsewhere in the Tathagata chapter and the Fire and Kindling chapter?94 Here, on the other hand, there are just two alternatives.

Here, Candraklrti considers the rejection of the five types of relation between the self and the skandhas as reducible to the extremes of identity and difference. Or, more specifically, the two types of "basis"/ dependence relation and possession are reducible
92 MAv 6.120 (233.16-19): nyon mongs skyon rnams ma Ius 'jig tshogs la / / ita las byung bar blo yis mthong gyur zhing / / bdag ni 'di yi yul du rtogs byas nas / / mal 'byor pa yis bdag ni 'gog par byed / 93 Tibetan (PPt D 110b6): bdag ni 'jigs tshogs la Ita ba de'i dmigs pa yin par mthong lao 94 That is, MMK 22.1, 10.14-15, which have both been cited above. This confirms that it is indeed the satkiiyadr$ti that Candraklrti has in mind in the alternatives we noted earlier in the agnfndhanaparfk$ii.

192

Andrew McGarrity

to the basic a priori assumption of difference, meaning that if difference is rejected then these further three are automatically ruled out too. This was also implied in the fivefold formulation in the agnfndhanaparfk$ti at PP ad MMK 10.14-15 with Candraklrti's immediate direct reference back to the simple twofold identity and difference refutation at the beginning of the chapter, suggesting that the specific relations were mere amplifications of identity and difference. The framing of the alternatives in terms of pak$as, as Candraklrti does here ad MMK 18.1, would also seem to make better sense when reduced to two alternatives, implying a pak$a and pratipak$a. Candraklrti's comments ad MMK 18.1 also provide an extra dimension vis-a-vis the issue of how the fivefold Madhyamika formulation is reconcilable with the earlier fourfold arrangement. At the most immediate level, his suggestion that the refutations of possession and dependence are covered in the refutation of difference is intended to show that Nagarjuna's explicit reference to "difference" automatically covers the relations of both "dependence" and "possession," meaning that they need not be explicitly spelt out. By the same token, if dependence and possession are explicitly spelt out, then difference need not be, since it is then implied. This suggests that difference should not be counted as being in addition to dependence and possession but implied in them. Hence, what might appear as five relations (when difference is mentioned) can actually be seen as four (focusing on dependence and possession explicitly, with difference merely implicit in them). Alternately, what could be counted as four relations (identity, possession and two types of dependence) can be abbreviated to two relations (identity and difference) depending on context. So, although not as eplicit as the strategy at Bhii$ya ad MAv 6.144, Candraklrti's explanation ad MMK 18.1 of the critique of identity and difference as being shorthand for the more extensive critique provides a further means of reconciling the fivefold critique with the earlier fourfold satktiyadr$ti/sakktiyaditthi tradition. Candraklrti sees no problem in understanding the fivefold relation in terms of the satktiyadr$ti although it is noticeable that he pointedly does not refer to it as being twentyfold here (ad MMK 18.1). In principle, this strategy (i.e., not needing to mention difference when possession and de-

Aryadeva's gradual stages

193

pendence are stated and not needing to mention possession and dependence when difference is stated) would seem to be implied in Candraklrti's MAvBh distinction between Buddhists (who accept and are trained to recognize the satkiiyadrifti) and non-Buddhists (who do not). However, it also raises another issue. The assumption that difference (anyatva) is automatically entailed in possession and dependence suggests that here at PP ad MMK 18.6 (and for that matter at PP ad MMK 10.14-15) a different conception of anyatva is envisaged to that at the Bhiiifya ad MAv 6.144 cited above. In regard to MAv 6.144, it was precisely the fact that anyatva is taken as entailing a self elsewhere apart from the skandhas and as additional to possession and dependence as covered in the standard four alternatives that is important, hence it is referred to as a "fifth" mode, over and above the standard accounts of relation. We might term this a "more extreme" conception of anyatva. On the other hand, ad MMK 18.1 and 10.14-15, anyatva in terms of the self and skandhas covers simply any basic common-sense form of difference which necessarily acts as the a priori condition presupposed in any account of possession and dependence, or indeed relation in general,95 We might term this the "less extreme" conception of anyatva. It is implicit in any form of relation itself whereas the more radical form of anyatva mentioned in relation to MAv 6.144 indicates something beyond or apart from these relations (hence a "fifth" alternative).96 The ambiguity may reflect the ongoing dif95 In regard to PP ad MMK 10.14 (212.5-6, cited above), Nagarjuna's reference to difference as "fire elsewhere apart from kindling" (agnir anyatra cendhanat) is treated by Candraldrti as simply implied in possession and dependence, inasmuch as the negation of difference automatically accounts for the negation of possession and dependence. Hence, difference here in PP ad MMK 10.14 is not a "fifth" mode additional to dependence and possession as at MAvBh 6.144. 96 Further muddying the waters though, in regard to MAv 6.143, immediately before characterizing difference as a "fifth" mode ad MAv 6.144, Candraklrti actually still refers to the fact that "the self has neither identity with, nor difference from form" (bdag ni gzugs las de nyid gzhan nyid med, tattvanyatve rupato natmanaf:t staf:t) by way of concluding the refutation of the two types of possession. In his Bhasya on this, Candraklrti glosses that

194

Andrew McGarrity

ficulty of how to reconcile the five modes of relation with the original four. This may come to the fore for MMhyamikas in particular due to their especially rigorous and widespread insistence on the basic identity and difference critique, hence giving rise to the need to reconcile this more streamlined MMhyamika critique with the earlier tradition. The other possibility is that refuting a self beyond these relations was simply not such an issue at an earlier stage but became one later on. In any case it is clear that for Miidhyamikas the significance of the satkiiyadr~ti lies in its providing a methodological template for the wider critique of substantialismY

"by virtue of the fact that identity and difference do not pertain to the self, this notion that 'the self possesses form' is also impossible" (de nyid dang gzhan nyid de dag ni gzugs bdag la yod pa ma yin pas bdag gzugs dang ldan no zhes bya ba 'di yang mi srid). This would seem to suggest the less extreme conception of difference is envisaged here inasmuch as the refutation of difference necessarily entails the refutation of possession and dependence, echoing MMK 18.1 and 10.14. If anything, it is in the subsequent analysis ad MAv 6.144, in which anyatva is taken as additional to possession and dependence that stands out as unique. This uniqueness is understandable given the unique hermeneutic context of MAvBh 6.143 in actually consciously reconciling the twenty-fold satkiiyadr~ti with the fivefold arrangement as found in Siltra; it is understandable that it is here that a certain stretch would have to be made. Reconciling both accounts, it is perhaps the case that non-Buddhists, untrained in the satkiiyadr~ri, must have the critique explicitly framed in terms of difference, which automatically rules out both possession and dependence (d la the original fourfold satkiiyadr~ti), as well as anything beyond these as well (i.e. a "fifth" alternative). 97 By way of contrast, in a Yogacara context, Schmithausen (1979) originally suggested that the satkiiyadr~ti, specifically the satkiiya (= upiidiinaskandhas) to which we turn in a moment, may have been a direct antecedent of 'the notion of the kli~!amanas. However, he has later modified this somewhat (1987, vol. 1, pp. 146-150158-164, 167-169), suggesting that the satkiiyadr~ri may simply playa role in the development of kli~!amanas. Be that as it may, it does seem as though in a Yogacara context, the satkiiyadr~ri plays more of a descriptive role in providing a phenomenological account of subjectivity, as opposed to its use in a Madhyamika setting in which its significance lies in its providing a methodological template, thus indicating more of a prescriptive orientation (i.e., in how a critique ought to be framed).

Aryadeva's gradual stages

195

3. From the self to views of extremes

(antadr~ti) via the Abhi-

dharmako,abha~ya's account of satkayadr~ti

Returning to our concern for the correspondences between the PP analysis ofMMK 18.6 and Aryadeva's three rejections, Candraklrti refers to the satkayadarsana in relation to Nagarjuna's (intermediate) "teaching of no-self" (anatmety api deiitam) at MMK 18.6b, suggesting a further parallel with Aryadeva's intermediate varalJam atmanaJ;.. Satkayadarsana seems equivalent to satkayadr~ti. In regard to MMK 18.6b, having just described the needs of the lowest of followers for whom the provisional self is posited, Candraklrti then describes the intermediate followers to whom no-self is taught
(anatmeti deiitam):98

But on the other hand, there are those who, like birds, are tied to the long, firm cord of attachment to the existence of "me" and "mine"(iitmiitmfyabhiivasneha), owing to their view of the self as actually real (sadbhatiitmadr~ti). Although they have gone very far, not even by their turning back from the path of evil deeds are they able to approach the peaceful, ageless and deathless city of nirviilJa once they have passed into the occurrence of birth in the threefold realm. In order to loosen the adherence to satkiiyadarSana among those followers at the intermediate level, "no-self" is also taught by the Buddhas, the Lords, in their desire to provide assistance to those people who are such followers, with the aim of bringing about the longing for
nirvii1J.a.

CandrakIrti's satkayadarsana (=satkayadr~ti) is associated with which is in turn associated with "attachment to the existence of 'me' and 'mine'" (atmatmfyabhavasneha). This closely reflects the standard Vaibha~ika and Sautrantika definitions of the satkayadr~ti. In the Mahavibha~aJ it is more specifically exsadbhatatmadr~ti
98 PP ad MMK 18.6 (357.7-358.3): ye tu sadbhutiitmadr~tikathiniitidfr ghataratmatmfyabhavasnehasutrakarabaddha* vihaT(lgama iva sudaram api gata!:t kusalakarmakari1J.o 'kusalakarmapathavyavrttyapi na saknuvanti traidhatukabhavopapattim ativahya sivam ajaram amara1J.aT(l nirvii1J.apuram abhigantuT(l I te:jaT(l madhyanaT(l vineyanaT(l satkayadarsanabhiniveSasithiIfkara1J.iiya nirvii1J.abhila:jasaT(ljananiirthaT(l buddhair bhagavadbhir vineyajananugrahacikfr:jubhir anatmety api dditaT(l II * Read satrakiivabaddha following de Jong 1978: 227.

196

Andrew McGarrity

plained that five of the twenty permutations are in fact iitmadr~ti (= the actual identification of the self with each of the five skandhas) and the remaining fifteen are iitmfyadr~ti (= the three relations of possession and dependence x each of the five skandhas).99 At Abhidharmakosa (AK) 5.3, the satkiiyadr~ti is first of all named as one of the five wrong views.lOO Subsequently, in his Bhii~ya ad AK 5.7 (AKBh) when he expands on the "nature" (svabhiiva) of the five views,101 Vasubandhu follows the Vaibha~ika identification of the satkiiyadr~ti with the view of self and what pertains to self (iitmiitmfyadr~ti) with which Candraklrti's iitmiitmfyabhiivasneha accords. Candraklrti, as we have noted, reflects the Vaibha~ika and Sautrantika tradition of satkiiyadr~ti as iitmiitmfyadr~ti by treating Nagarjuna's aniitmety api deiitam in MMK 18.6b as directed towards the satkiiyadarsana, which is in turn associated with sadbhiitiitmadr~ti and iitmiitmfyabhiivasneha. This treatment would indicate that Candraklrti assumes the satkiiyadr~ti here, just as he does in his explanation of Aryadeva's "rejection of the self" being in accordance with understanding the fivefold relations and twenty-fold view. Thus, although the satkiiyadr~ti is mentioned by name in neither the CST nor PP passages - although satkiiyadarsana is mentioned in the latter - its implicit presence in both lends support to viewing Nagarjuna's teaching of non-self at MMK 18.6 and Aryadeva's rejection of the self at CS 190b in tandem. The set of assumptions shared between the two, i.e., the assumption of the satkiiyadr~ti in Candraklrti's reading of both, provides further indication of the influence of PP ad MMK 18.6 on Pa tshab's alteration of CS 190. Now, like the Pali sakkiiyaditthi as a general "view of personality/identity," the AKBh's conception of the satkiiyadr~ti covers the view of existent selfhood (i.e., "me" and "mine"), following iitmiitmfyadr~ti. The Pali tradition seems to straightforwardly take the sa~ in satkiiya as something like "existent" and kiiya as someRahder 1931-32: 229. 100 AK 5.3 (p. 279): dmayal:z paftca satkiiyamithyiintagriihadr~tayal:z / dr~ta.f[lavratapariimarsiiv iti punar dasa ;.-'/-99
101 Cf. lead in to AK 5.7 (p. 281): dr~tayal:z paftca niimato nirdi~tii na tu svabhiivatas tat kas tiisiil'J1. svabhiival:z /

Aryadeva's gradual stages

197

thing like "body," rendering something like "body in being, the existing body" for sakkiiya. 102 However, Vasubandhu, in the key definitional passage explaining satkiiyadr~ti in relation to AK 5.7, instead reads sat as "what passes away" (sfdati) and kiiya as referring to the skandha as an "accumulation, the multiple collection" (cayaJ:z kiiyaJ:z sa1!lghiitaJ:z). This unusual reading has led Edward Conze to remark that "[t]he etymological derivation of this very archaic term [satkiiyadr~ti] presents great and almost insuperable difficulties."103 Similarly, Etienne Lamotte has remarked that "[l]a satkiiyadr~ti, d'etymologie obscure ... designe la croyance au moi et du mien (iitmiitmfyagriiha) ... Le systeme de l'Abhidharma ... est plus complique."104 Forewarned by these notes of caution, we could perhaps, following Vasubandhu's comments overall, render the satkiiyadr~ti in the Sautrantika context with something like "view of (or "view attributing existence (sat) to?") the passing (sat == sfdati) relational assemblage (kiiya)." I offer the following tentative translation of the AKBh passage in question with a reading of satkiiyadr~ti that is sympathetic to Vasubandhu's explanation, whilst still trying to preserve the sense of sat as "existence:"105
The view attributing existence to (?) the passing relational assemblage (satkiiyadr~ti) is the view of the self or the view of what pertains to the self (iitmadr~tir iitmfyadr~tir va). "Existent" (sat) indicates what "passes away" (sfdatfti); the "relational assemblage" (kiiya) is the accumulation, the multiple collection, meaning the skandha (cayaf:t kayaf:t saf!!ghataf:t skandha ity arthaf:t). Comprising both [what is taken as] existent (sat) and this relational assemblage (kiiya), it is hence the (attribution of) existence [of selfhood, i.e., iitmiitmfya] to the passing relational assemblage, i.e., to the five skandhas subject to appropriation (pancopiidiinaskandha). [Satkayadr~ti] is explained in order to dispel (i) the conception of a permanent [entity] [by showing sat

102 T.w. Rhys Davids and William Stede 2001: 660-661. 103 Conze 1962: 276, n. 16. 104 Lamotte 1981: 737, n. 3. 105 AKBh 281f. ad AK 5.7: iitmadr>!tir iitmfyadmir vii satkiiyadmif:t / sfdatfti sat / cayaf:t kiiyaf:t saf!!ghiitaJ:! skandha ity arthaJ:! / sac ciiyaY[l kiiyas ceti satkiiyaJ:! paficopiidiinaskandhiiJ:! / nityasaY[ljfiiiY[l pi1JasaY[ljfiiiY[l ca tyiijayitum evaY[l dyotitii / etatpurvako hi te~v iitmagrahaJ:! /

198

Andrew McGarrity as "passing away"] and (ii) the conception of a single whole [by emphasizing kiiya as an "accumulation," "multiple collection" etc]. This [was covered] earlier, viz., that grasping at a self is [directed toward] those [upiidiinaskandhas].

Speculating on Vasubandhu's reading, it is almost as though he is suggesting that the satkaya, which one originally mistakenly reifies as "existent" (sat), is subsequently found to in fact be nothing more than the set of "passing (sat, i.e., szdatf) conglomerations/ accumulations/ assemblages (of the upadanaskandhas)," thus making the "double" meaning of satkaya deliberate, or at least making some sense of it. To put this another way, the satkaya as the passing assemblage is revealed as the object of the mistaken view of the kaya as existent (sat) or as a self. Accordingly, there seems to be some ambiguity as to whether satkayadr~ti is positive or negative: it is negative when held unwittingly, but it is positive when one understands the technical formulation of what it actually is, thus enabling it to be redressed. La Vallee-Poussin has noted that Vasubandhu here provides not only the unique Sautrantika reading of satkaya but also the basis of the Tibetan translation, remarking that "c'est cette etymologie que repose la versione tibetaine: 'jig tshags la Eta ba = vue sur la collection de ce qui perit."106 The Tibetan 'jig tshags la Eta ba closely follows Vasubandhu's reading and exports it beyond just the immediate Sautrantika context. When confronted with the problem of how to render the term sat and its multiple connotations, the Tibetan translators chose to drop any sense of "existent" altogether and simply render the "view of the passing assemblage," suggesting that they too found Vasubandhu's reading problematic. Significantly for our purposes, in concluding the passage, Vasubandhu then associates the satkayadr~ti with the (grasping at) extreme views (anta(graha)dr~ti) of eternalism and nihilism, thus enabling the satkayadr~ti formulation to provide a template for a

106 Louis de la Vallee-Poussin 1980: 15. See also Lokesh Chandra 2001:
805.

Aryadeva's gradual stages

199

progression from the view of the self to views more broadly. He explains thatl07
The view [directed] towards the passing relational assemblage (as existent) is explained as the view of the passing relational assemblage (satkayadr~ti)/view attributing existent [selfhood] to the reiational assemblage. The view which has as its object-condition the defiled (siisravalambanii dr~tif:t), i.e., [the defiled as] any [of the five types of upiidiinaskandhas] at all (sarvaiva),108 is directed towards the passing relational assemblage (as existent) (satkiiye). But the very view of the self and what pertains to the self (iitmiitmfyadr~tir eva) is explained as the view (attributing existence to) the passing relational assemblage (satkayadr~ti), in such a way that (yatM) when there is neither self nor what pertains to the self, then this view attributing existence to the passing relational assemblage (satkiiyadr~ti) would be understood. As it is said, 'Whoever, whether bhik~us, sramar.zas or briihmar.zas observing, observe a self, all of them [are actually observing] only these: the five skandhas subject to appropriation (paficopiidiinaskandha).'109 Whether the view of permanence or the view of annihilation, [any such view] is only of that, i.e., an entity (vastu) regarded as the self (iitmiibhimata); such a view grasps at extremes (antagrahadr~ti), since it constitutes the grasping at the extremes of eternalism or annihilation.

Vasubandhu suggests that while all views which have the defiled (sasrava) in terms of the upadanasakandhas as their object condi107 AKBh 281-2 ad AK 5.7: satkiiye dr~ril:t satkiiyadr~fil:t / sarvaiva siisravaZambana dr~til:t satkaye / atmatmfyadr~tir eva tu satkayadr~tir ukta / yatM gamyeta satkayadNtir iyal'{l ntitmani natmfye veti / yathoktal'{l ye kecid bhik~aval:t sramar.za va brahmar.za va atmeti samanupasyantal:t samanupasyanti, sarve ta imiln eva paficopadanaskandhan iti / tasyaivtitmiibhimatasya vastuno dhruvadr~tir ucchedadr~tir viintagrahadr~til:t / sasvatocchedantagraha1Jat / 108 Cf. AKV p. 607 ad AKBh 5.7: sarvaiveti paficaprakarapi / satkaya iti / paficasllpadanaskandhe~u / ---

109 Khandha-Vagga XXII.47.3 (p. 46): ye hi keci bhikkhave samar.za vii brilhmar.za va anekavidham attanal'{l samanupassamana samanupassanti sabbe te paficllpadanakkhandhe samanupassanti etesal'{l va afifiataram.

200

Andrew McGarrity

tion (iilambana) may happen to take the satkiiya as their object, iitmiitmfyadr$ti itself constitutes the actual satkiiyadr$ti as such (cf. iitmiitmfyadr$tir eva tu satkiiyadr$tir uktii). Likewise, the views of permanence and annihilation are said to have as their object precisely the entity conceived of in terms of a self (iitmiibhimatavastu), with such views constituting grasping at the extremes of eternalism and nihilism respectively. The vastu which comes to be "regarded as a self" (iitmiibhimata), i.e., the object of iitmiibhimata, seems equivalent to the satkiiya (i.e., the paficopiidiinasakandhas) as the object of satkiiyadr$ti. There is at least the potential here for the conceptual expansion from the critique of the self to things here via the critique of extreme views inasmuch as Vasubandhu seems to use the terminology of iitmiibhimatavastu here as a stepping-stone between satkiiyadr$ti and the extreme views of eternalism/permanence and nihilism/annihilation. A little further on at AK and AKBh 5.32e33a, the false views (mithyiidr$ti) of satkiiyadr$ti and anta(graha) dr$ti are related via the more straightforward intermediary of the extremes in question (i.e., eternalism and annihilation) being specifically applicable to the grasping at the extremes of the eternalism or annihilation of the self (iitmanaJ:t sasvatocchediintagrahalJa), without reference to iitmiibhimatavastu.uo On the other hand, at AKBh 5.7, with the use of the term iitmiibhimatavastu, Vasubandhu here provides an alternate way of framing this relation between satkiiyaand antagraha-dr$ti, although of course both AKBh 5.32e-33a and AKBh 5.7 really amount to a rephrasing of the same dynamic. At AKBh 5.7, Vasubandhu's implication that the template applicable to how the satkiiya is taken as the object of the view of selfhood (iitmiitmfyadr$ti) is also applicable to the iitmiibhimatavastu has the potential to apply the critique more generally, to any entity regarded in the reified substantial terms of a self (i.e., iitmiibhimatavastu);
AK and AKBh 5.32e-33a (pp. 304-305): (32e) satkiiyadrk tataJ:! II mithyiidr~teJ:! kila satkiiyadr~riJ:! pravartate duJ:!khataJ:! skandhiin apohyiit mato 'bhiniveSiit I (33a) tato 'ntagraha1}am satkiiyadr~ter antagrahadr~tiJ:! pravartate iitmanaJ:! sasvatocchediintagraha1}iit I
110

Aryadeva's gradual stages

201

any such object is then in turn what is grasped at by extreme views (arttagrahadr~ti). It is this strategy, more broadly applying the critique of the self to things generally, that is more applicable in a Madhyamika setting and its Tibetan interpretation. The rejection of all (extreme) views and the teaching of neither self nor non-self The association of the satkiiyadr~ti with extreme views (antadr~ti) resurfaces when we return to rGyal tshab's analysis of CS 190. Recalling that rGyal tshab inherits Pa tshab's translation of "all views" as Aryadeva's climactic rejection, he explains the transition from the rejection of the self to the final rejection thus: 111
At the intermediary stage, on the basis of an analysis of the five types [of relation] with the skandhas one takes as their deliberate object of refutation (zhen yul bkag) the twenty-fold satkiiyadr~ti. From this, the coarse self is rejected. Finally, when the [student's] mental continuum comes to be a vessel capable [of receiving such a teaching], the selflessness of the person is taught. Once it is has been taught that even selflessness is not [the final] Truth, all views which grasp at extremes (mthar 'dzin gyi lta ba kun :::; *sarvantagrahadr~ti) are rejected. One who understands the stages of this teaching is a wise person in regard to the order of the stages in which one ought to be trained.

rGyal tshab's most immediate source here is probably his former teacher, Red mDa ba. Red mDa also takes the views which are refuted, and which lead on from the refutation of the self, as being "all extreme views" (phyis nas bdag dang bral ba'i phung po tsam la mngon par zhen pa spang ba'i phyir rang bzhin stong pa nyid kyi sgo nas mthar 'dzin gyi Ita zhig kun bzlog par bya ste).112 However, for both rGyal tshab and Red mDa ba, the satkiiyadr~ti
111 bZhi brgya pa'i mam bshad 63a5-63b2: bar du phung po dang mam pa Ingar dpyad pa'i sgo nas 'jig tshogs la Ita ba nyi shu'i zhen yul bkag nas rags pa'i bdag ni ldog pa dang phyis nas rgyud snod rung du song ba la gang zag gi bdag med bstan / bdag med kyang bden par med par bstan nas mthar 'dzin gyi lta ba kun ldog pa ston pa'i rim pa gang gis shes pa de gdul bya bkri ba'i rim pa de la mkhas pa yin no / 112 dBu ma bzhi brgya pa rtsa 'greI95.6-8.

202

Andrew McGarrity

and its Abhidharma association with the anta(graha)dr~ti provide the key point of departure for how the rejection of the self can, at a methodological or pedagogical level, lead on to the rejection of all (extreme) views. This is despite the fact that, at the ontological level, with which Candraklrti and probably Aryadeva were originally directly concerned, the rejection of the self flows on to the rejection of all things. Even though Candraklrti has the satkiiyadr~ti and its association with iitmiitmlyadr~ti in mind, he does not take the AKBh's lead in associating it explicitly with extreme views in his CST ad CS 190 as he has no need to move toward a climactic rejection of all views. By contrast, rGyal tshab, like the Tibetan rendering of Jigs tshogs lta ba in general, is more led by Vasubandhu's reading of satkiiyadr~ti, to the extent that he sees it in terms of the basis of extreme views, thus enabling it to be reconciled with the "rejection of all views" reading that he inherited. rGyal tshab's readings of the progression through Aryadeva's three rejections thus represents a certain shift in the approach to gradualism. The factual set of metaphysical entailments that Candraklrti emphasizes in his CST analysis in which the rejection moves from the self and on to all things later becomes, for rGyal tshab, indistinguishable from the correct approaches that ought to be taken so that these entailments can be rightly understood and the rejections carried out. For rGyal tshab, the rejection of the self almost becomes more important at a normative level as a methodological template: it is how the self is analyzed that is significant for the subsequent rejection, and not just the bare fact of its rejection (i.e., the propositional fact that the self is non-existent), although in terms of content these may amount to the same thing. This methodological focus does not usurp Candraklrti's more ontologically inclined "progressive entailments" reading and in fact it is clearly implied in Candraklrti's work itself as shown in his overall treatmentof the satkiiyadr~ti; the normative methodological and metaphysical entailments readings go hand in hand. However, rGyal tshab's reading represents a slight shift in emphasis in exactly what allows one stage to follow on from another, and makes explicit a further dimension, only implicit in the earlier reading, of what exactly makes the approach itself "GraduaL" It is not simply what was ascertained at a previous stage that allows for progression to the

.Aryadeva's gradual stages

203

next stage, but also, in addition to this, how it was ascertained at the previous stage that counts. The background assumption of the satktiyadr.Jti as indicative of extreme views provides the link which allows for the move from the view of the rejection of the self to the rejection of all views to occur so seamlessly. The background assumption of the satkiiyadr.Jti as indicative of extreme views also allows for transition from the metaphysical/ontological to the normative/methodological readings to take place to such an extent that in the indigenous Tibetan understanding illustrated by rGyal tshab's analysis, these two aspects are virtually fused together. Now, this provides not so much a direct reason for Pa tshab's alteration in itself, but more for how it could then be so readily accepted and adapted in the Tibetan tradition. rGyal tshab's reconciliation of "all views" with Candraklrti's "all things" of course operates only after the fact and on the assumption of "all views." While the Sautrantika treatment of satkiiyadr.Jti provides an explanation for the alternate transition from the self to views rather than to things, it does not completely explain why Pa tshab may have rendered "all views" in the first place. Rather, I suggest that the direct reason lies in the context in which CS 190 is quoted in PP ad MMK 18.6, specifically, in relation to 18.6cd in which neither self nor non-self is taught (niitmii na ciiniitmii kascid ity api deiitarrt). We have noted that the audience for Aryadeva's rejection of the non-meritorious corresponds to those for whom the self is postulated at MMK 18.6a while the rejection of the self via the satktiyadr.Jti formulation corresponds to MMK 18.6b with the teaching of nonself to intermediate followers. Candraklrti then outlines the third group, the superior followers, to whom neither self nor non-self are taught: 113

PP ad MMK 18.6 (358.4-9): ye tu parviibhyiisavise~iinugatagambhfra pratyiisannavartini* nirviirze te~iim utkr~tiiniif!Lvineyiiniif!L vigatatmasnehiiniif!L paramagambhframaunfndrapravacaniirthatattviivagiihanasamarthiiniim adhimuktivise~am avadhiirya I buddhair atmii na ciiniitmii kascid ity api deSitaf!L II yathaiva hy iitmadarsanam atattvam evaf!L tatpratipak~abhatam apy aniitmadarsanaf!L naiva tattvam iti I evaf!L niisty atmii kascin na ciipy anatmii kascid astfti deitaf!L II * Read pratyiisannavartino following de Jong 1978: 227.
113

dharmiidhimok~alabdhabfjaparipiikaJ:t

204

Andrew McGarrity

And yet there are those who have obtained the ripening which follows their exceptional practice at an earlier stage, [the ripening of] the seed which is their commitment to the profound Dharma.114 Their abiding in nirviilJa near at hand, those followers at an advanced level, for whom attachment to the self has ceased, are capable of immersion in the real meaning (tattaviirtha) of the pronouncements of the Great Sage in all their [the pronouncements'] extreme depth. The commitment of [such advanced followers] having been ascertained, "neither any self or non-self" is taught by the Buddhas. For in just the same way that the view of the self (iitmadarsanam) is not in accordance with reality, so in this way, neither is the view of non-self (aniitmadarsanam), which occurs as its opposite, in accordance with reality. Thus, it is taught that there is not any self, nor any non-self. Candraklrti treats "self" and "non-self," neither of which are in accordance with reality (tattva) and which are thus left behind, in terms of the views of self and non-self (atmadarsanam and anatmadarsanam). Before finally quoting Aryadeva's verse itself, Candraklrti also cites Ratniivalf (RA) 2.3-4, where, in a similar vein, Nagarjuna explains ll5 In this way neither self nor non-self are found with respect to the actual state of things as they are, hence the Great Sage rejected views composed of self and non-self. What is seen, heard, etc, is said by the Sage to be neither true nor false; for from a position (pak:fa) there would be a counter-position (pratipak:fa), but neither of them are in accordance with reality. Aryadeva's verse, with its original climactic rejection of all things, is then cited after this. The overall context here, with the reference to self and nonself as darsanas and the citation of the RA reference in which self
114 This follows La Vallee-Poussin ([1992] 1903-13: 358, n. 2) who explains that "Ie bfja est l'adhimok:fa; il est paripakva en raison de P[=purva] abhyiisa" and also the Tibetan sngon goms pa'i khyad par gyis zab mo'i ehos

la lhag par mos pa rjes su 'bras bu 'bral ba'i sa bon yongs su smin par gyur pa thob pa. llS RA 2.3-4 (40.9-16): (2.3) naivam atmii na eanatma yathabhutyena labhyate I atmanatmakrte dr~tfvavarasman mahamuniJ:t II (2.4) dr:ftasrutadyaY[l munina na satyaY[l na mr~oditam I pak~iid dhi pratipak:faJ:t syad ubhayaY[l tae ea narthataJ:t II Cited at PP ad MMK 18.6 (359.1-4).

A.ryadeva's gradual stages

205

and non-self are associated with a pakifa and pratipakifa respectively would seem more conducive to a reading of "all views" as Aryadeva's fin'al rejection. On the other hand, the final rejection being of "all things" as it is here in the Sanskrit original stands out somewhat and is more difficult to reconcile within the overall context of the rest of Candraklrti's analysis. So whereas the first two of Aryadeva's rejections could be more simply reconciled with Nagarjuna's initial positing of the self and the teaching of nonself, there is more difficulty in harmonizing Nagarjuna's teaching of "neither self nor non-self" with Aryadeva's original rejection of all (things). This, I suggest, provides the immediate reason for the alteration to "all views." Pa tshab's translation of "all views" actually seems to make more sense in the PP context in which the verse is quoted than does Aryadeva's original reading of "all (things)" which is found there, even if it then is not as in accord with Candraklrti's CST analysis. Moreover, as implied in the treatment of the non-existence of the self in terms of nihilism and then the existence of the self as permanence, self and non-self can also be seen to themselves correspond to the extreme views (antadrifti) which are subsequently rejected. This allows CS to also be reconciled with the AKBh's account of extreme views via the shared reference to the teaching of non-self and rejection of the self in terms of the structure of the satkayadrifti that is implicit (or explicit) in all three (CST, PP and AKBh). That PP ad MMK 18.6 is the immediate influence that draws together these shared assumptions, thus prompting Pa tshab, can then, as we have seen, be illustrated most explicitly in the terms in which rGyal tshab's later analysis is framed.
Conclusions

While Candraklrti may not necessarily intend a neat set of correspondences between Nagarjuna's and Aryadeva's stages, this may be what Pa tshab means to bring out with his translation, and this is brought into sharper relief when one also examines rGyal tshab's analysis and the set of assumptions he simply takes as given. Pa tshab's "all views" would provide a means of harmonizing both accounts in a single coherent framework. Certainly, such correspond-

206

Andrew McGarrity

ence is at the very least Implicit in CandrakIrti's analysis given his citation of the verse in this context in the first place, so Pa tshab is merely drawing on an implication that is already there. In sum, the set of assumptions behind Pa tshab's alteration are first of all that: MMK 18.6ab
(i) atmety api prajfiapitam (ii) anatmety api desitam

CS 190ab
apulJyasya varalJam varalJ-am atmanaJ:t

However, when it comes to pushing this unified framework through to the final of the three stages in each case, it is more difficult to force a clear-cut correspondence between the teaching of neither self nor non-self and the rejection of all things. On the other hand, the teaching of neither self nor non-self can be more easily made to correspond to a "rejection of all views" owing to the assumption of self and non-self as mutually opposing darsanas or a pak~a and pratipak~a respectively along the lines of the aforementioned Madhyamika "principle of the complementarity of binary concepts and terms."116 Thus, Pa tshab's alteration may perhaps be summarized: MMK 18.6cd
(iii) natma na canatma kascid ity api desitam

CS 190c
sarvasya VaralJ-am
:;:117

Zta ba kun bzlog (*sarvadr:jtfna1!l varalJ-am)

Pa tshab places more emphasis upon interpreting CS 190 in the light of the surrounding context in which Candraklrti quotes the verse in his PP than within the specific context of the CS itself. The verse as quoted in PP is effectively read "back into" its original CS rather than "out of" it, with the result that it becomes altered to fit in more fluidly with the context of its citation in PP. Clearly, 116 Ruegg 1977: 56.
117 The symbol ""," here denotes something more akin to "is made to correspond to" rather than simply "corresponds to" or the standard "is approximately equal to."

Aryadeva's gradual stages

207

sacrificing some of the consistency between CS 190 and its direct CST analysis was considered worth it for the sake of increasing the con~istency between CS 190 and the PP context ofMMK 18.6, and, indirectly, MMK 18.6 itself. There are broader implications to be gleaned from this for how we understand the Tibetan interpretation of Indian philosophical materials, and the process by which ideas were transmitted and understood. Certainly, for Pa tshab, and for the Tibetans who followed him, the prime concern is with the significance of Aryadeva's verse within the context of the overall Madhyamaka project, for which Nagfujuna's root verses and CandrakIrti's PP are obviously considered more significant. This concern overrides any need to preserve the verse in what, from a modern point of view, would be valued as an objectively accurate form that is true to the original textual context within which it was embedded. Pa tshab and the Tibetan translators he worked with were more concerned with doctrinal consistency and systematization than they were with semantic accuracy. Somewhat ironically, if the verse itself had not been quoted by CandrakIrti in his PP, thus preserving its Sanskrit original, then we would not have even been aware of the discrepancy between the Sanskrit and its Tibetan translation and that the alteration had taken place; and yet, if it were not quoted in the PP then, as I have argued, there would probably not have been any direct reason for the verse to have been amended by Pa tshab to "all views" in the first place. No doubt as more and more Sanskrit originals emerge of materials previously only preserved in Tibetan and Chinese, further instances of this sort will come to light, allowing us to glimpse more of the hermeneutic presuppositions at work in the process of the translation of the Dharma.

Abbreviations and bibliography

Abbreviations
P - D.T. Suzuki, ed. The Tibetan Tripitaka: Peking Edition. Tokyo-Kyoto: Tibetan Tripitaka Research Institute, 1957.

208

Andrew McGarrity

D - Sera (South India) reprint of the sDe dge bsTan 'gyur made from the Derge Printing House (Tibet) printing, 1992-93.

Primary sources
Pali
Khandha-Vagga - M. Leon Feer ed. Sar!1yutta Nikaya Part III: KhandhaVagga. London: The Pali Text Society, [1975] 1890. Cii!avedallasutta - V. Trenckner ed. The Majjhima-Nikaya. Vol. 1. Oxford: The Pali Text Society, [1993] 1888. Dhammasarigar;i - Edward MUller, ed. The Dhammasarigini. London: The Pa1i Text Society, [1978] 1885. Visuddhimagga - C.A.F Rhys Davids ed. The Visuddhi-Magga of Buddhaghosa. London: The Pali Text Society, [1975] 1920.

Indian sources in Sanskrit and Tibetan translation

AK, AKBh - Abhidharmakosa with Bha.yya; P. Pradhan, ed. Abhidharmakosabha.yyam of Vasubandhu. Patna: K.P. Jayaswal Research Institute, [1975] 1967. AKV - Sphutarthabhidharmakosavyakhya; SvamY Dvarikadasa SastrY, ed. The Abhidharmakosa and Bha.yya ofAcarya Vasubandhu with Sphutartha Commentary of Acarya Yasomitra. 2 vols. Varanasi: Bauddha Bharati, 1987. Arthaviniscayasiitra - ed. P.L. Vaidya, Mahayana-Sutra-Sa1J1-grahal;, Part 1 Buddhist Sanskrit Texts Series No. 17, Darbhanga: The Mithila Institute, 1961. Avadanasataka - ed. P.L. Vaidya, Avadana-Sataka, Buddhist Sanskrit Texts Series No. 19, Darbhanga: The Mithila Institute, 1958. CS - Catul;sataka(karika); Karen Lang, ed. Aryadeva's Catul;Sataka: On the Bodhisattva's Cultivation of Merit and Knowledge. Copenhagen: Akademisk Fodag, 1986. CST - Catul;Satakatfka; D 3865: vol. ya. PP - Prasannapada; Louis de la Vallee-Poussin, ed. M uZamadhyamakakarikas de Nagarjuna avec la Prasannapada commentaire de Candrakfrti. Delhi: Motilal Banarsidass, [1992] 1903-13. PPt - Prasannapada; D 3860: vol. 'a, sDe dge bsTan 'gyur, vol. 60. Beijing: CPA Publishing, 2000; P 3860: vol. 'a.

A.ryadeva's gradual stages

209

MAv, MAvBh - Madhyamakavatara with Bhao!ya; Louis de la ValleePoussin, ed. Madhyamakavatara par Candrakfrti: Traduction tibetaine. Delhi: Motilal Banarsidass, [1992] 1907-12. MAvBht - Madhyamakavatarabhao!ya; D 3862: vol. 'a. MAvBhT - Madhyamakavatarabhao!yatfka; D 3870: vol. ra. MMK - Mulamadhyamakakarika; Louis de La Vallee-Poussin, ed. Mulamadhyamakakarikas de Nagarjuna avec la Prasannapada commentaire de Candrakfrti. Delhi: Motiial Banarsidass, [1992] 1903-13. VV - Vigrahavyavartanf; E.H. Johnston and Arnold Kunst ed. (1949) in Kamaleswar Bhattacharya, The Dialectical Method of Nagarjuna: Vigrahavyavartanf. New Delhi: Motilal Banarsidass, 1998. Vaiseo!ikasutras - Nandaial Sinha, ed. The Vaiseo!ika Sutras of Kanada with the Commentary of Smikara Misra and Extracts from the Gloss of JayanarayalJa. [Allahabad] New York: AMS Press, [1974] 1911. Y~(V) - Yuktio!ao!tika; Cristina Anna Scherrer-Schaub, ed. Yuktio!ao!tikavftti: Commentaire ala soixiante sur le raisonnement au du vrai enseignement de la causa lite par le maftre indien Candrakfrti. Bruxelles: Institut BeIge des Haute Etudes Chinoises, 1991. RA. - Ratnavalf; Michael Hahn, ed. Nagarjuna's RatnavalfVol. I: The Basic Texts (Sanskrit, Tibetan, Chinese). Bonn: Indica and Tibetica, 1982.

Indigenous Tibetan sources

Chos 'byung - Chos 'byung (Bu ston). Janos Szerb ed. Bu stan's History of Buddhism in Tibet. Wien: Verlag der Osterreichischen Akademie der Wissenschaften 1990. Lam rim chen mo (Tsong kha pa). Taipei: The Corporate Body of the Buddha Educational Foundation (Reprint vol. TI002), 2000. dBu ma bzhi brgya pa rtsa 'grel (Red mDa ba). Sarnath, Varanasi: Sakya Students' Union and Pleasure of Elegant Sayings Printing Press, 1974. dBu ma shar gsum gyi stong thun - Helmut Tauscher, ed. Phya pa chos kyi seng ge: dBu ma shar gsum gyi stong thun. Wien: Arbeitskreis flir Tibetische und Buddhistische Studien, Universitat Wien, 1999. rTsa shes tik chen (Tsong kha pa). Mundgod and Taipei: Drepung Gomang Monastery reprint (distributed by The Corporate Body of the Buddha Educational Foundation; reprint vol. TI066). 2002. bZhi brgya pa'i rnam legs bshad snying po (rGyal tshab). Lhasa ZhoI: vol. ka.

210

Andrew McGarrity

Secondary sources
Cabez6n 2003 - Cabez6n, Jose Ignacio. "Two Views of the SvatantrikaPrasangika Distinction in Fourteenth Century Tibet" In Dreyfus and McClintock 2003,289-315. Lokesh Chandra 2001- Chandra, Lokesh. Tibetan-Sanskrit Dictionary. New Delhi: International Academy of Indian Culture and Aditya Prakashan, 2001 [1959-61]. Conze 1962 - Conze, Edward. Buddhist Thought in India: Three Phases of Buddhist Philosophy. London: George Allen & Unwin, 1962. Cutler et al 2002 - Cutler, Joshua W.C. et aI, The Lamrim Chenmo Translation Committee, The Great Treatise on the Stages of the Path to Enlightenment. Ithaca and Boulder: Snowlion Publications, 2002. De Jong 1978 - De Jong, IW. "Textcritical Notes on the Prasannapadii" Indo-Iranian Journal 20 (1978): 25-59 and 217-52. Dreyfus and McClintock 2003 - Dreyfus, Georges B.I and Sara L. McClintock. "Introduction" In The Sviitantrika-Priisarigika Distinction: What Difference Does a Difference Make?, edited by Georges B.I Dreyfus and Sara L. McClintock. Somerville: Wisdom Publications, 2003, 1-37. Dreyfus and Tsering 2010 - Georges Dreyfus and Drongbu Tsering, "Pa tshab and the origin of PrasaIigika" In this JIABS issue pp. 387-417. Edgerton 1970 - Edgerton, Franklin. Buddhist Hybrid Sanskrit Grammar and Dictionary. 2 vols. Delhi: Motilal Banarsidass, 1970 [1953]. Fuller 2005 - Fuller, Paul. The Notion of ditthi in Theraviida Buddhism: The Point of View. London and New York: Routledge Curzon, 2005. Gethin 1986 - Gethin, Rupert. "The Five Khandhas: Their Treatment in the Nikayas and Early Abhidhamma" Journal of Indian Philosophy, 14 (1986): 35-53. La Vallee-Poussin 1910 - La Vallee-Poussin, Louis de. "Madhyamakiivatiira: Introduction au traite du milieu de l'Acarya Candrakirti (1)" Museon 11 (1910): 271-358. La Vallee-Poussin 1911- La Vallee-Poussin, Louis de. "Madhyamakiivatiira: Introduction au traite du milieu de l'Acarya Candraklrti (2)" Museon 12 (1911): 235-328. La Vallee~Poussin 1980 - La Vallee-Poussin, Louis de. L'Abhidharmakosa de Vasubandhu: Traduction et annotations. Edited by Etienne Lamotte. Vol. 16, Melanges Chinois et Bouddhiques. Bruxelles: Institut BeIge des Hautes Etudes Chinoises, 1980 [1923-31]. Lamotte 1981 - Lamotte, Etienne. La Traite de la Grande Vertu de Sagesse de Niigiirjuna (Mahiiprajiiiiparamitiisiistra). 5 vols. Louvain-Ia-Neuve: Universite de Louvain, Institut Orientaliste, 1981.

Aryadeva's gradual stages

211

Lang 1990 - Lang, Karen Christina. "sPa-tshab Nyi-ma-Grags and the Introduction of Prasailgika Madhyamaka into Tibet" In Reflections on Tibetan Culture: Essays in Memory ofTurrell v: Wylie, edited by Lawrence Epstein and Richard F. Sherburne. Lewiston, Queenston, Lampeter: The Edwin Mellen Press, 1990, 127-141. McClintock 2000 - McClintock, Sara L. "Knowing All through One: Mystical Communion or Logical Trick in the Tattvasa'!1-graha and Tattvasa'!1-grahapafzjikii" Journal of the International Association of Buddhist Studies 23/2 (2000): 225-44. Mimaki 1983 - Mimaki, K. "The Blo gSal Grub mtha', and the Madhyamika Classification in Tibetan Grub Mtha' Literature" In Proceedings of the Csoma de Koras Symposium: Held at Velm-Vienna, Austria, 13-19 September 1981, edited by Ernst Steinkellner and Helmut Tauscher. New Delhi: Motilal Barnarsidass reprint, 1983, 161-67. Rahder 1931-32 - Rahder, 1. "La satkiiyadr~ti d'apres Vibha~a, 8" Melanges Chino is et Bouddhiques 1 (1931-32): 227-39. Rhys Davids and Stede 2001 - Rhys Davids, T.W. and William Stede. PaliEnglish Dictionary. New Delhi: Munshiram Manoharlal, 2001 [1921]. Roerich 1976 - Roerich, George N. (trans.) The Blue Annals. Delhi: Motilal Banarsidass, 1976 [1949]. Ruegg 1977 - Ruegg, David Seyfort. "The Uses of the Four Positions of the Catu~koti and the Problem of the Description of Reality in Mahayana Buddhism" Journal of Indian Philosophy 5 (1977): 1-71. Ruegg 1983 - Ruegg, David Seyfort. "On the Thesis and Assertion in the Madhyamaka/dBu-ma" In Contributions on Tibetan and Buddhist Religion and Philosophy: Proceedings of the Csoma de Karas Symposium Held at Velm-Vienna, Austria 13-19 Sept 1981, Vol 2., edited by Ernst Steinkellner and Helmut Tauscher. Wien: Arbeitskreis fUr Tibetische und Buddhistische Studien Universitat Wien, 1983, 205-41. Ruegg 2000 - Ruegg, David Seyfort. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy. Wien: Arbeitskreis fUr Tibetische und Buddhistische Studien, Universitat, 2000. Scherrer-Schaub 1991. - Scherrer-Schaub, Cristina Anna. Yukti~a~tikiivrtti: Commentaire a la soixiante sur le raisonnement ou du vrai enseignement de la causa lite par le maitre indien Candrakfrti. Bruxelles: Insitut BeIge des Haute Etudes Chinoises, 1991. Schmithausen 1979 - Schmithausen, Lambert. "Some Aspects of the Conception of Ego in Buddhism: Satkiiyadr~ti, Asmimiina and Kli~tamanas" Translated into Japanese by K.Yokoyama. Bukkyogaku 7 (1979): 1-18. Schmithausen 1987 - Schmithausen, Lambert. A.layavijfziina: On the Origin and the Early Development of a Central Concept ofYogiiciira Philosophy. 2 Vols. Tokyo: The International Institute for Buddhist Studies, 1987.

212

Andrew McGarrity

Suzuki 1997 - Suzuki, K6shin ed. Sanskrit Fragments and Tibetan Translation of Candrakfrti's Bodhisattvayogacara-CatuJ:tsataka{fka. Tokyo: The Sankibo Press, 1994. Takasaki 1988 - Takasaki, Iikid6. "On Upadana/Upadayaprajftapti" In Orientalia Iosephi Tucci Memoriae Dicata, vol. 3., edited by Gherardo Gnoli and Lionello Lanciotti. Roma: Istituto Italiano per II Medio et Estremo Oriente, 1988, 1451 [1]-1464 [14]. Tauscher 1983 - Tauscher, Helmut. "Some Problems of Textual History in Connection with the Tibetan Translations of the MadhyamakavataraJ:t and its Commentary" In Contributions on Tibetan and Buddhist Religion and Philosophy: Proceedings of the Csoma de Karas Symposium Held at Velm-Vienna, Austria13-19 Sept 1981, Vol 2., edited by Ernst Steinkellner and Helmut Tauscher. Wien: Arbeitskreis flir Tibetische und Buddhistische Studien Universitat Wien, 1983, 293-303. Tauscher 2003 - Tauscher, Helmut. "Phya pa chos kyi seng ge as a Svatantrika" In The Svatantrika-PrasaJigika Distinction: What Difference Does a Difference Make?, edited by Georges B.I. Dreyfus and Sara L. McClintock. Somerville: Wisdom Publications, 2003, 207-55. Vose 2009 - Vose, Kevin A. Resurrecting Candrakfrti: Disputes in the Tibetan Creation of Prasarigika. Boston: Wisdom Publications, 2009. Yoshimizu 1993 - Yoshimizu, Chizuko. "The Madhyamaka Theories Regarded as False by the dGe lugs pas" Wiener Zeitschrift fur die Kunde Sudasiens und Archiv fur Indische Philosophie 37 (1993): 201-27. Yoshimizu 2003 - Yoshimizu, Chizuko. "Tsong kha pa's Re-evaluation of CandrakIrti's Criticism of Autonomous Inference" In The SvatantrikaPrasarigika Distinction: What Difference Does a Difference Make?, edited by Georges B.I. Dreyfus and Sara L. McClintock. Somerville: Wisdom Publications, 2003, 257-88.

Interpreting the tantras A Tibetan debate on the numbers of adepts admissible to tantric consecration Jan-Ulrich Sobisch
It is a commonplace to call Tibetan Buddhism (a form of) "tantric

Buddhism" and to view it as a continuation (in some way) of Indian tantric Buddhism. Yet it becomes increasingly clear these days that Tibetan tantric Buddhism was not a slavish copy of Indian tantric Buddhism. Many experts will agree that Tibetan tantric masters developed their own particular styles, to say the least. The Tibetan "hand-writing" is above all conspicuous in these masters' attempts to systematize the tantric teachings of India, which on occasion have toned down some of the more eccentric aspects and promoted some minor cults to mainstream, but overall have caused what once might have appeared as a conglomerate of cults to be more coherent and, in fact, more uniform. At the same time, each Tibetan tradition will claim, following its most prominent masters, that they preserve faithfully the Indian Buddhist tradition, "without decrease and increase and without insertions and deletions" ('phri snan dang lhag chad med par).l As historians, we deal not only with "a (single) Tibetan approach" to Indo-Buddhist tantrism, but also with how the Tibetan traditions debated their own diverse ways to go about the matter. Inevitably, we also have to consider the way we ourselves deal with this complex of problems. I have discussed elsewhere the ap1 See for example, for the claim of Sa-skya Pm;u;lita (1182-1251, henceforth Sa-pa1).) to continue the authentic Indian tradition of scholarship, Gold 2007, and the present author's review of it (Sobsich 2009), where some of the tensions between the Sa-skya-pas and bKa'-brgyud-pas approaches are briefly discussed together with the academic approach to such problems.

Journal of the International Association of Buddhist Studies Volume 32' Number 1-2 2009 (2010) pp. 213-234

214

Ian-Ulrich Sobisch

pro aches in particular that the Sa-skya-pa and the 'Bri-gung bKa'brgyud-pa traditions took to the coexistence of tantric vows with the priitimok~a and bodhisattva vows in a single mental continuum, and to the tantric practice. of mahiimudrii. 2 With the present paper and a planned series of future articles I would like to deepen this line of investigation, focusing on a variety of aspects of Sa-skyapa and the 'Bri-gung bKa' -brgyud-pa approaches to Indian tantric Buddhism, searching for contested areas and trying to describe and analyze the different views and resulting programs, and the transformations that Indian tantric Buddhism has undergone on its way to becoming the tantric traditions of Tibet. It is quite obvious that much promising material can be found in those works of Tibetan Buddhism where the authors identify and attack the position of opponents and establish their own views. Among such works, two stand out with their commentaries as the most outspoken ones in these respects, namely Sa-pal).'s sDom gsum rab dbye, which is by now well known to interested audiences in the West,3 and the as yet much lesser known dGongs gcig of the 'Bri-gung-pa sKyobpa 'Iig-rten-mgon-po (1143-1217), a chief disciple of sGam-po-pa's (1079-1153) master disciple Phag-mo-gru-pa (1110-1170).4 The problem that I would like to discuss in the present paper may seem to be a trifle one, yet I believe that by investigating the different procedures of the parties involved in the debate, we may be able to glean some particulars of what may be a Sa-skya-pa and a 'Bri-gung-pa approach - not, however, in the sense of a uniform method that serves as a singular key to the understanding of a tradition, but rather as parts of thought patterns that are characteristic for them. We can certainly also not expect total uniformity within these two traditions. In fact by contrasting Sa-pal).'s approach with that of gSer-mdog Pal).-chen Shiikya mChog-ldan (1428-1507),

Cf. Sobisch 2002 and Sobisch, forthcoming. This Tibetan classic was translated by Rhoton (2002). 4 For a few remarks on the dGongs gcig see Martin 1997, van der Kuijp 1987, Liu 2002, and chapter 14 of Sobisch 2002. The dGongs gcig is now in the focus of a research project by Khenpo Rangdrol and myself, see http://freenet-homepage.de/jigten-sumgon/.
2

Interpreting the tantras

215

we already realize some of the tensions that exist within the Saskya tradition. The subject of the debate under investigation, i.e. the number of adepts to be admitted to tantric consecration (Skr. abhi$eka, Tib. dbang bskur), is mentioned by Rig-'dzin Chos-kyigrags-pa (1595-1659) in his commentary of the second topic of the fifth chapter of the dGongs gcig. 5 This particular passage starts with a quotation from the sDom gsum rab dbye of Sa-skya Pal).<;lita, chapter three, verses 21c-25b (in Rhoton's edition of the text), where Sa-pal! says:
When they perform tantric consecration they bestow it on an unspecified number of [of adepts]. This was prohibited by Vajradhara. With regard to the tantric consecration of the caryi'i tantras an unspecified number of disciples was taught. [But] there exists [the teaching] of a specified number with regard to the specifically selected disciples [of] the remaining [tantras]. The following is taught in the Si'imi'inyavidhiguhyatantra: 'The skilled [master] admits odd numbers of disciples such as one, three, five, seven [and so forth] up to twenty-five. It is not auspicious to admit a larger number of disciples [for tantric consecration].'
sngags kyi dbang skur byed pa na II grangs nges med par dbang skur byed II 'di ni rdo rje 'chang gis bkag II spyod pa'i rgyud kyi dbang bskur la II slob ma grangs nges med par gsungs II lhag ma dmigs bsal mdzad pa yi II slob ma la ni grangs nges yod II 'di ni gsang ba spyi rgyud las II mkhas pas slob ma gcig gam gsum II lnga'amyang na bdun dag gam II ni shu rtsa a ni lnga yi bar II
5 'Bri-gung Dharma Klrti, dGongs gcig 'grel ba nyi ma'i snang ba, Maryland: Drikung Kagyu Meditation Center, pp. 246-48; and The Collected Works (gSung 'bum) oj Kun-mkhyen Rig-pa-'dzin-pa Chen-po Chos-kyi-grags-pa, Dehra Dun: Drikung Kagyu Institute 1999, vol. 3, pp. 207.

216

Ian-Ulrich Sobisch

zung du ma gyur slob ma gzunl II de bas" [hag pa'i slob ma ni d II yongs su gzung" bar mi shis so II
e

Variant readings in P (Peking Tripitaka): a dag, bzung

bzung, C las, d dag,

First of all, neither the sDom gsum rab dbye nor the commentaries by Go-rams-pa (1429-1489)6 and Shakya mChog-ldan7 identify Sapa!}.'s opponents in this instance. There are, however, some indications that the practice that is criticized by him was widespread not only amongbKa' -brgyud-pa and rNying-ma-pa lamas, but probably also among the Sa-skya-pas themselves. In this passage, Sapa!}. makes three claims:
1. In the earya tantras it is (under certain preconditions) permitted to bestow tantric consecration to an unspecified number of disciples. 2. In the tantras of all other tantra classes, lower and higher than earya, this is prohibited. 3. The reason is that in the non-earya tantras one has to follow the Samanyavidhiguhyatantra (a kriya tantra), 8 which states that tantric consecration should only be bestowed on odd numbers of disciples such as five or seven and so forth, and that the maximum number of disciples is twenty-five.

6 Go-rams-pa bSod-nams-seng-ge authored six commentaries of the sDom gsum rab dbye, which are identified and briefly described in Sobisch 2002: 28-30. For the present article I have used his sDom pa gsum gyi rab tu dbye ba'i rnam bshad rgyal ba'i gsung rab kyi dgongs pa gsal ba, Sa skya pa'i bka' 'bum, The Complete Works of the Great Masters of the Sa-skya-pa Sect of the Tibetan Buddhism, vol. 14/4, pp. 119-199 (short: sDom gsum rnam bshad). This work was written in 1463 at 'Bras-yul rDzong-dkar sKyed-mo-tshal. It is a detailed exposition that follows Sa-paI:I's sDom gsum rab dbye verse by verse. For a detailed outline of its subject-headings (sa bead), see Appendix B of Sobisch 2002. 7 For some details on SMkya mChog-ldan's replies to (his own) 108 questions on the sDom gsum rab dbye, see below. 8 dKyil 'khor thams ead kyi spyi'i eho ga gsang ba'i rgyud (Sarvamal}r;lalasamanyavidhiguhyatantra), P 9/429, 42-5-4 152-5-5.

Interpreting the tantras

217

As is well known, the carya tantra class is, according to the most wide-spread classification system of the later translation period, the second lowest of the four tantra classes. 9 Why is the practice of bestowing tantric consecration on any possible number of disciples permitted in the carya tantras? In his sDom gsum rnambshad (fol. 90r), the commentator Go-rams-pa bSod-nams Seng~ge quotes from the Vairocanabhisaf!lbodhitantra (P vol. 5, no. 126, 245-1-1):10
More [adepts] than one, two, four [etc.] are possible, sincell one does not need to investigate [them], admit [all disciples]!

And (fol. 90rf., P vol. 5, no. 126, 245-2-2):12

9 Different Indian masters such as Abhayakaragupta, Buddhaguhya, ViHisavajra, Santipa, Nagarjuna, Vajrapa:Q.i, and Atisa have suggested classification systems with two up to seven tantra classes. For a brief overview according to Kong-sprul Blo-gros-mtha' -yas' Shes bya kun khyab mdzod, see Sobisch 2002: 57 ff. For a recent translation of that passage, see Lodro Taye 2005: 89-93. For further remarks on the classification of the Vairocaniibhisa1!lbodhitantra, see also below. 10 For the Tibetan text (in context), see the bold passage two notes further down. For the Vairocanabhisa1!lbodhitantra, see Sobisch 2002: n. 518 and Hodge 2003 (this passage on p. 93); and Wayman and Tajima 1992. 11 I translate here Go-rams-pa's variant reading (dpyad mi dgos pas). The tantra reads: dpyad mi dgos par. 12 In Hodge (2003: 96); Wayman and Tajima (1998, repr.: 120). The relevant passage in P looks like this (P vol. 5, no. 126,244-5-8): slob ma dad cing rigs btsun pa II de bzhin dkon mchog gsum (245-1-1) la dad II zab mo yi ni blo dang ldan II spro ba chen zhing tshul khrims ldan II bzod dang ldan zhing ser sna med II dpa' la yid dam brtan pa ni II bcu'am brgyad dam bdun nam lnga II gcig gnyis bzhi las lhag kyang rung II dpyad mi dgos par gzung bar bya II ... (245-1-3): gsang ba'i bdag po I 'on kyang dkyil 'khor rgyas pa 'di ni I sems can gyi khams ma nges pa yongs su bskyab pa'i phyir snying rje chen po'i snying po 'byung ba zhes bya ste I de ni de bzhin gshegs pa rnams kyis bskal pa grangs med par bla na med pa yang dag par rdzogs pa'i byang chub yang dag par bsgrubs pa byin gyis brlabs so I gsang ba'i bdag po I de Ita bas na rnam grangs 'dis kyang I khyod kyis 'di Itar de bzhin gshegs pa rnams ni sems can gcig gi phyir byang chub mngon par rdzogs par 'tshang rgyab ma yin I gnyis kyi

218

Jan-Ulrich Sobisch

In order to take [the bestowing of tantric consecration] as a cause for the resolve for awakening (bodhicitta), they should admit limitless sentient beings!

The Indian master Buddhaguhya, who lived in the 8th century and visited Tibet, taught that the Vairocanabhisarrtbodhitantra does two things:
1. It occasionally teaches means that accord with the kriya tantras (which are directed outward to an "objective support"), but 2. is actually a yoga tantra that teaches the "profound and vast."

Therefore, according to Buddhaguhya's system of tantra classifications (that included two classes, and according to some exegetes a third that appears to be a combination of kriya and yoga), the Vairocanabhisarrtbodhitantra belongs to the yoga or to the ubhaya tantra class (which combines the two).B According to the system of Bu-ston (with four categories), it belongs to the carya tantra class (Wayman and Tajima 1992: 97 f.) and is its prime example. Gorams-pa quotes this tantra here to show that the chief tantra of the carya class teaches that one does not need to investigate the disciples when bestowing tantric consecration and that one makes the admittance to the tantric consecration of limitless sentient beings the cause for the production of the resolve for awakening (bodhi~ citta). For some reason that is not made explicit by either Sa-pal).
phyir ma yin I gsum gyi phyir ma yin I 'on kyang thugs rje chen po'i dbang du gyur pa dag sems can gyi khams ma lus pa yongs su bskyab pa'i phyir byang chub tu mngon par rdzogs par 'tshang rgya ste I de dag gis sems can gyi khams rnam pa du ma la bsam pa ci lta ba bzhin du I sems can rnams la cho 'chad par rig par bya'o II gsang ba (/) ba'i bdag po I theg pa chen po la goms par ma byas pa rnams ni gsang sngags spyod pa'i tshul mthong ngam thos na sems la dga' ba 'am I dad pa cung zad kyang yang dag par mi skye yi I gsang ba'i bdag po I gang dag gis sngon theg pa chen po gsang'sngags spyad pa'i (245-2-1) tshul sgo mtha' yas bsgrubs pa la goms par byas pa de dag ni rdo rje sems dpa' yin te I de rnams nyid kyi don gyi phyir grangs kyi tshad 'di byas so II 'on kyang slob dpon snying rje chen po dang ldan pas sems can gyi khams ma lus pa bsgral par yi dam bca' ba kho nar bya ste des byang chub kyi sems kyi rgyur 'gyur bar bya ba'j phyir sems can tshad med pa rnams yongs su gzung bar bya'o II 13 Cf. Hodge 2003: 43 and Dalton 2005: 122f.

Interpreting the tantras

219

or Go-rams-pa, this rule only applies to tantras of the caryii class, but not to the (lower) kriyii or the (higher) yoga. The rule of the Siimiinyavidhiguhyatantra, however, which is quoted in the sDom gsum rab dbye, is according to the Sa-pal). to be applied to all other tantras (3.25c: 'di ni kun la 'jug), except, however, those Of the caryii class. But why is the rule of the caryii tantra not to be applied to the other classes as well? After all the Vairocaniibhisarrzbodhitantra refers to bodhicitta, which is one of the most important principles of mahiiyiina practice in the yoga and highest yoga tantras. And why is this principle of bodhicitta overruled by an instruction from a kriyii tantra, which is the lowest of the four tantra classes, into which the tantras are most usually divided in Tibetan Buddhism? Both Sa-pal). and Go-rams-pa refer to this problem only indirectly. In order to establish that the rule of the kriyii tantra has to be applied to all other tantras (except caryii), the sDom gsum rab dbye (3.28a-30b) and Go-rams-pa's sDom gsum rnam bshad (fol. 90v) argue by quoting another passage from the same Siimiinyavidhiguhyatantra (P 9/429, 51-5-4):
If you think: 'Since this is a kriyii tantra, [the rille] doesn't [apply to] rituals [of] other [tantra classes],' [we reply]: In the Siimiinyavidhiguhyatantra itself it is taught that it applies also to all others: 'In those [tantras] where there are activities, but activity rituals are lacking, there the skilled master relies on the ritual that was taught in the Siimiinya-tantra(s?).' Since this has been taught, this ritual is applied to all tantras. [P: Where there are activities, but own activity-rituals are lacking, there the skilled master should perform those activities that are taught in the Siimiinya-tantra(s?).] 'di ni bya ba'i rgyud yin pas II gzhan gyi cho ga min snyam na II gzhan rnams kun la'ang 'di 'jug par II spyi rgyud nyid las 'di skad gsungs II gang du las ni yod gyur la II

220

lan-Ulrich Sobisch

las kyi cho ga mams med pa II der ni spyi yi rgyud dag las II . gsungs pa'i cho ga mkhas pas bsten II de skad gsungs phyir cho ga 'di II rgyud mams kun la 'jug pa yin II [P: gang du las ni yod 'gyur la II rang gi las kyi cho ga med II der ni mkhas pas spyi rgyud las II gsungs pa'i las mams bya ba yin]

The passage from the Siimiinyavidhiguhyatantra is a paraphrase, not an exact quote. It poses as it is a number of problems that cannot be addressed here. Given the frequency with which the tantra is quoted in Tibetan sources and the importance that is attached to it, a detailed study of this text would certainly be warranted. Of considerable interest in our context, however, is Sa-pal).'s remark Ci.e. the last two lines of the above quote), according to which the ritual of this tantra is to be applied to all tantras. The above cited tantra passage "In those [tantras] where there are activities, but activity rituals are lacking, C )," is understood by Sa-pal). as meaning that ... in those tantras that require activities, but are lacking proper rituals of tantric consecration of their own, the tantric consecration ritual of the Siimiinyavidhiguhyatantra is to be used. The word "tantra" is supplied in my translation according to Go-rams-pa's further paraphrase, which also adds the specifications "lacking clarity" and "tantric consecration" CsDom gsum rnam bshad, 90v):
In those tantras possessing activities such as tantric consecration, but the activity rituals are lacking clarity, there ... rgyud gang du dbang bskur ba la sogs pa'i las [nil yod par gyur la I las kyi cho ga mams gsal po med pa der ni ...

The tantra itself, far less explicitly, merely seems to state that activity rituals missing in other tantras should be supplied from "the siimiinya tantra(s?)" (here referring to itself?),14 and in this context 14 One question is indeed, what is meant by the tantra's statement "spyi rgyud las gsungs pa'i las mams." In its fuller version, the name of the tantra is dKyil 'khor thams cad kyi spyi'i cho ga gsang ba'i rgyud, where spyi is related to cho ga, and not to rgyud. The short version of the name, however, is (as found in Tibetan texts) gSang ba spyi rgyud, but the short

Interpreting the tantras

221

it is taken for granted by Sa-paJ;l and Go-rams-pa that this tantra's restrictions concerning the numbers of disciples admitted to the ritual of tantric consecration should be applied to tantras of other classes. By merely quoting this very brief and indeed cryptic passage - even with Sa-paJ;l's and Go-rams-pa's creative interpretation - the doubts whether rituals from a kriyti tantra are applicable to other tantra classes do not seem to have been solved. And furthermore, if Sa-paJ;l's statement "is applied to all tantras" allows for' the exception of the whole class of caryti tantras from this rule, then why can the tantras of higher classes, such as of the yoga and highest yoga tantra class, which certainly employ the principle of

form spyi rgyud is in this sense, as far as I can see, nowhere else attested in the tantra. The question is therefore, whether spyi rgyud las gsungs is an instance of self-reference ("taught in the siimiinya tantra"), or whether it is to be understood as "taught in the general tantras" (understanding spyi'i rgyud). Curiously there exists in Shakya mChog-ldan's gSer gyi thur ma a query that refers to the Siimiinyavidhiguhyatantra and ends with the words (12v3): "(...) [then] it can hardly be the 'general tantra' for all [classes)" (kun gyi spyi rgyud yin par dka'), to which he replies (12v4): Even though it is not the general tantra for all, its specifying of the number of disciples has been applied by many authentic authorities to the other [tantras] (kun gyi spyi rgyud ma yin kyang II de yi slob ma'i grangs nges ni II tshad ldan mang pos gzhan la'ang sbyar II - understanding "{mkhas grub] tshad ldan"). Shiikya mChog-ldan, sDom gsum gyi rab tu dbye ba'i bstan beos kyi 'bel gtam rnam par nges pa legs bshad gser gyi thur ma, The Complete Works (gSung 'bum), Thimphu, Bhutan, Kunzang Tobgey, 1975, vol. 6, pp. 439-648, (=chapters 1-2), and vol. 7, pp. 1-230, (::::chapter 3). Thus, if this should not be a self-reference, but a reference to the "general tantras" (whatever that means), the key passage of Sa-pru;t's argumentation collapses. The immediate context of our passage in the Siimiinyavidhiguhyatantra (51-5-4) does, in my understanding, not seem to discuss a tantric consecration ritual of the present tantra that could be applied to other tantras. It rather leaves the impression that this passage teaches that if rituals for any of the activities are missing somewhere (including in the present tantra), they should be carried out by the skillful master as taught in the general tantras (whatever they are). This is, however, only a preliminary impression.

222

Ian-Ulrich Sobisch

bodhicitta as much as the caryii tantras, not be part of the exception, tOO?15 To sum up so far, we found that Sa-pa:Q. and his commentator Go-rams-pa argue that the rule of a key kriyii tantra, namely the Siimiinyavidhiguhyatantra, that prohibits the bestowing oftantric consecration to even numbers of disciples and to groups larger than twenty-five participants must be applied to all tantras, except those of the caryii class. But while their explanation of the exception for caryii tantras is, due to the explicit statement from the Vairocaniibhisa"f!lbodhitantra, comprehensible, their argument for a transference of the restrictions to the yoga and highest yoga tantras is hardly convincing, since the passage that they quote from the Siimiinyavidhiguhyatantra in this context to support their claim does not explicitly mention these restrictions, nor the need to apply them to tantras other than those of the kriyii class. In fact, it may be a general statement to the effect that rituals missing anywhere can be carried out with the help of rituals found in other general tantras.

* * *
Let us now turn to the question how the Siimiinyavidhiguhyatantra itself justifies its restrictions concerning tantric consecration. As , we have already seen, Sa-pa:Q. quotes from the tantra (sDom gsum rab dbye 3.24bcde-25ab, P 9/429, 44-4-5):
The skilled [master] admits odd numbers of disciples such as one, three, five, seven [and so forth] up to twenty-five. It is not auspicious to admit

15 I am not discussing here the problems of the gradual emergence of the tantras in history, since this would neither be Sa-paIf's nor Rig-'dzin Chos-kyi-grags-pa's (or in other words an ernie) way of thinking. In an historical perspective, the kriyii tantras certainly appeared earlier than most earyii, yoga, or highest yoga tantras, and it would require a different type of argument to show that rituals of the (historically) earlier kriyii tantra are applicable to the (historically) later tantras. Yet, since, as far as I can see, such a gradual emergence is not an issue in Tibetan tantric literature (at least not in a sense as it is discussed by philologists), we cannot expect arguments of this sort in these works.

Interpreting the tantras a larger number of disciples [for tantric consecration].

223

Sa-pal}. explains that the restriction is necessary because the ritual of tantric consecration is to be performed within a single night and has to be completed before dawn (sDom gsum rab dbye 3.2Q): For a larger number of disciples, a complete ritual cannot be concluded in a single night. If it is not concluded in the same night, it is taught that the ritual is defective.
de bas lhag pa'i slob ma la II cho ga yongs su rdzogs pa ni /I mtshan mo gcig la tshar mi nus /I de yi mtshan mor ma tshar na /I cho ga nyams par 'gyur bar gsungs /I

And he quotes again the Siimiinyavidhiguhyatantra, which says (sDom gsum rab dbye 3.27bcde, [a] = P 9/429, 43-5-4; [b] = 43-5-2): (a) When the sun has set, the gods certainly gather through the blessing. (b) It is propitious to worship [them] and then to request their departure until the sun has risen.
(a) lha yang nyi ma nub pa na It nges par byin gyi brlabs kyis 'du /I (...) (b) nyi ma shar bar ma gyur bar II' mchod nas gshegs su gsol ba shis /I

Variant readings in P: a nyi ma nub pa lha rnams ni /I, b nyi ma shar

ba las gyur bar /I

Here, the two passages (a and b) quoted from the tantra actually occur in the tantra in reverse order and two lines apart from one another. With the above statements from the sDom gsum rab dbye Sa-pal}. offers a reason why only a maximum number of twenty-five disciples are admitted, namely that the ritual cannot be completed within a single night if more disciples are present (the necessity of odd numbers is explained by its auspiciousness). Again, the quote from the tantra hardly supports such an interpretation. In fact, in order to create the sequence 'sunset to sunrise' the two passages a

224

Ian-Ulrich Sobisch

and b had to be presented in reverse order and the text in between these lines had to be ignored,

* * *
It is well known that the teachings of the sDom gsum rab dbye

were also discussed critically within the Sa-skya-pa tradition, in particular by Go-rams-pa's contemporary Shakya mChog-ldan. He formulated 108 questions and challenged the Sa-skya-pa masters of his time for answers. Among others, Go-rams-pa provided replies (in 1476), but Shakya mChog-ldan (somehow unsurprisingly) remained dissatisfied and composed his own answers in 1481. To start off the discussion (in our present topic), he first cites a passage from the Samanyavidhiguhyatantra that appears to be in conflict with Sa-paI).'s point of view (gSer kyi thur ma, vol. 7, fol. 9vl, P 9/429, 52-4-5/7):16
The guru should consecrate a single disciple into the maI:u;lalas. It should not happen that the skilled master consecrates two disciples simultaneously. Two, three, or four [disciples] are also possible, [but] regarding the activities of the tantric consecration the guru should bestow [tantric consecration] to all [disciples] individually with a new different set of requisites [for consecrating].

bla mas slob ma gcig pu ni II dkyil 'khar dag tu dbang bskur bya II mkhas pas cig char" slob ma gnyis II dbang bskur ba ni yangs b mi bya II gnyis sam gsum mame bzhi yang rung II bla mas dbang bskur bya ba ni II yo byad gsard pa gzhan rnams kyis II thams cad so so so sar bya II ' V anant read P: a b e ,am, d sar mgs m car, yang,

16 Shakya mChog-ldan, sDam gsum gyi rab tu dbye ba'i bstan bcas kyi 'bel gtam gser kyi thur ma, The Complete Works of gSer-mdog PaI?--chen Shakya-mchog-ldan, voL 6-7.

Interpreting the tantras

225

For the sake of his argument, Shiikya mChog-ldan introduces three sections of the ritual of tantric consecration: (1) Admitting adepts and guiding them, (2) preparations, and (3) the actual consecration. In his actual answer he then ascertains (fol. 10v):
Regarding the intention of the Stimtinyavidhiguhyatantra it is taught that after completely performing the rituals of 'admitting' and 'preparation' at the time of 'admittance of disciples' and 'preparation' for however many [disciples] up to a number of at the most twenty-five there may be assembled into a single group, the ritual of the actual tantric consecration is to be done for each disciple individually. gsang ba spyi rgyud kyi dgongs pa ni I slob rna rjes 'dzin dang sta gon gyi dus su I nyer lnga man chad kyi grangs kha yar ba ji tsam yod pa de thabs gcig tu tshogs pa la I rjes 'dzin dang sta gon gyi cho ga yongs su rdzogs par byas nas I dbang skur dngos gzhi'i cho ga ni slob rna re re nas bya bar gsungs pa yin te I

He furthermore shows that according to the tantra, the master calls up each adept individually in the tantric consecration,17 and that a skilled master never bestows tantric consecration on two (or more) adepts simultaneously, unless he has different new utensils for each ofthem (fol. llr).ls Shiikya mChog-ldan concludes (fol. llr):
If we infer based on that meaning, if not more than one set of tantric consecration tools that are necessary for the tantric consecration is available, the disciples are consecrated successively, and if there are tantric consecration implements available matching the number of [present] disciples, it is apparently so that it is acceptable to consecrate [them] simultaneously.

don de la rjes su dpag na I dbang de la dgos pa'i dbang gi yo byad cha gcig las med na I slob rna rnams rim gyis skur ba dang I slob ma'i grangs dang mnyam pa'i dbang rdzas yod na I cig char du dbang bskur bas chog pa Ita bur mngon no II

17 gSer kyi thur rna, fo1. lOv, P 9/429, 50-1-4/5. blo dang ldan pas cho ga 'dis II slob rna rnams ni legs par bzung (P: gzung) II slob rna re re nas bkug (P: dgug) ste II gsang gtor snga rna bzhin byas fa (P: nas) II. lS See the above quote (fo1. 9vl) from mkhas pas cig car ... until ... so so so sor bya 1/. Shakya mChog-ldan cites here only the first two lines (fo1. lOv).

226

Ian-Ulrich Sobisch

He explains the difference regarding the total numbers of adepts allowed in a tantric consecration in the Samanyavidhiguhyatantra and other tantras (max. twenty-five) and in carya tantras such as the Vairocanabhisambhodhi (unlimited) through the fact that in the former the disciples are first examined and then admitted, while in the latter they are first admitted and then examined (fo1. llv).19 Unfortunately he does not explain this topic further and thus the question remains, why not a number larger than twenty-five adepts can be first examined and then admitted. 20 Of great interest is Shakya mChog-ldan's following remark concerning the maximum of adepts in the tantric consecration of other tantra classes (fo1. 12r):
The fixation 'up to twenty-five' with regard to the number of disciples in the 'admittance of disciples' in the Siimiinyavidhiguhyatantra is without doubt the intention of the dBang gi chu bo, but those gurus who have shaped that [scripture] into a ritual practice - having in mind the special necessity of taking [the admittance of many disciples] as the cause for the resolve for awakening as taught in the Vairoeaniibhisambodhi - have not fixed the number [of disciples] (...)
gsang ba spyi rgyud du I slob ma rjes 'dzin gyi slob ma'i grangs la nyi shu rtsa lnga man chad kyi grangs nges mdzad pa de I dbang gi ehu ba'i dgangs pa yin par gdon mi za yang I bla ma rnams phyag len du

19 It should also be mentioned here that Shakya mChog-ldan (fol. llv) points out that the number of adepts admitted to tantric consecration even in a earyii tantra is not unlimited by default, but only through the special qualities of the master: "With regard to the tantric consecration of the earyii tantra there is no certainty of an upper limit that agrees with the [number of present] disciples. It is not so that they [can be] unlimited, since more than ten are prohibited. Even though it is so, there is no certainty with regard to the number of admitted disciples, since it is taught that 'a great compassionate Acarya should admit [them] without limits!'" (spyod pa'i rgyud kyi dbang bskur la II slob ma kha 'cham kha yar gyi II nges pa med mod grangs med min II beu las lhag pa bkag phyir ro II de Ita na yang rjes bzung gi II slob ma'i grangs la nges pa med II slob dpon snying rje chen po yis II tshad med bzung zhes gsungs phyir ro II). 20 The reason that not a larger number of disciples can be examined is probably that, according to Sa-palf's interpretation of the kriyii tantra, time is restricted, for which see above, sDom gsLtm rab dbye 3.26.

Interpreting the tantras

227

mdzad pa de ni / rnam snang mngon byang nas gsungs pa ltar / byang chub kyi sems kyi rgyur 'gyur ba'i dgos pa khyad par can la dgongs nas grangs nges ma mdzad pa (...)

This concerns the Sa-skya-pa practice of Hevajra, whose tantric consecration is in the focus of the dBang gi chu bo, a text of enormous importance composed by bSod-nams-rtse-mo in the second half of the 12th century.21 According to Shakya mChog-ldan, bSod-nams-rtse-mo intended in his text a limitation of twenty-five adepts, but the ritual works derived from it followed the lead of the Vairocaniibhisa1'Jlbodhitantra. If this is indeed true, we would have a case where eminent Sa-skya-pa authors applied the rule of the ciirya tantra to a tantra of the highest yoga class, placing it above the rule of the kriyii class. It would certainly be worthwhile to study derived ritual works such as those by Grags-pa-rgyal-mtshan of the 12th century (also because of their important remarks with regard to the role and function of the four trantric consecrations of the highest yoga class).22 Shiikya mChog-ldan introduced further distinctions and clarifications, some of them on the very difficult terrain of ritual practice, and I am quite aware that more research is necessary to elucidate the fine points of his argumentation, but I nevertheless hope to have identified the major points of his reply (to his own question) here.

Let me now return to Rig-'dzin Chos-kyi-grags-pa's commentary on 'Jig-rten-mgon-po's dGongs gcig, where, in the second topic of the fifth chapter and in reaction to Sa-pal).'s criticism, he deals with this problem. Not surprisingly, Rig-'dzin Chos-kyi-grags-pa attacks the supposed authority of the kriyii tantra over the yoga and higher 21 bSod-nams-rtse-mo (dPal kyai rdo rje'i dkyil 'khor du slob ma smin par byed pa'i cho ga dbang gi chu bo chen mo, Sa-skya-pa bKa' 'bum 2/20, 35r-89r). 22 Grags-pa-rgyal-mtshan (dBang gi chu bo ma nub par bskur ba, Sa skya Lam 'bras Literature Series 11, 122r-122v), and perhaps also rDzong-pa Kun-dga'-rgyal-mtshan (Kyai rdo rje'i dkyil 'khor du dbang gi chu bo bskur ba'i zhal gdams gnad kyi gsal byed kyi dka' 'grel bdud rtsi'i bum pa). See also Stearns 2001: 13.

228

Jan-Ulrich Sobisch

yoga tantras. 23 He argues in fact that the Samanyavidhiguhyatantra is only a kriya tantra (p. 247: spyi rgyud ni bya rgyud kho na yin) and that what is taught in it is valid only for tantras of the kriya class. He quotes the same tantra (P 9/429, 42-5-6):24
Out of love for the worldlings, for certain, I have taught this general ritual of those maJ;l<;ialas, such as that of VajravidaraJ;la(?), that I have taught as the 3.500 maJ;l<;ialas. 'jig rten dag la brtse phyir nges II rnam par gnon pa la sogs pa II dkyil 'khor sum stong dag dang ni II lnga brgyar bshad pa gang yin pa II dkyil 'khor de dag rnams kyi spyi'i II cho ga 'di ni ngas bshad do II

Thus the Vajradhara's announcement that he has taught a general ritual- which then supposedly contains restrictions with regard to the number of adepts in a tantric consecration - concerns according to Rig-'dzin Chos-kyi-grags-pa only the maJ.l<:ialas of the kriya class, of which there seem to be 3,500, but not the ma:r:t<:ialas of the yoga and highest yoga tantras. Since this is Rig-'dzin Chos-kyigrags-pa's main thrust, he has to deal with the passage of the tantra (already quoted above through sDom gsum rab dbye 3.29), through which Sa-pal). intended to show that the tantra's ritual was valid for all other tantras, too:
In those [tantras] where there are activities, but activity rituals are lacking,
23 Already Shakya mChog-ldan was well aware of this type of criticism. He quotes an unspecified opponent with these words (fo1. 12v): "If the specified number of disciples is established for the tantric consecration of the higher tantras through the reasoning of the kriya tantra, how can the unspecified number not be established through the reasoning of the carya tantra?" (bya ba'i rgyud kyi rigs pa yis II rgyud sde gong ma'i dbang skur gyi II slob ma'i grangs nges 'grub 'gyur na II spyod pa'i rgyud kyi rigs pa yis II grangs nges med pa cis mi 'grub II). 24 This passages stems right from the beginning lines of the tantra, where the purpose of the teaching is stated.

Interpreting the tantras there the skilled master relies on the ritual that was taught in the Samanya-tantra(s?).

229

According to Sa-pal)., this means that wherever a ritual for tantric consecration is missing (or according to Go-rams-pa: "lacks clarity"), the ritual of the Samanya-tantra(s?) is to be applied. Rlg-'dzin Chos-kyi-grags-pa now points out that "nowadays making counts [of disciples] etc. appears in your [i.e. the Sa-skya-pa] system also with regard to the tantric consecration of Hevajra."25 But since there exist a great number of Hevajra tantric consecration rituals by both Indian and Tibetan masters, there occurs to be an inconsistency between the sDom gsum rab dbye and the actual practice in the Sa-skya-pa tradition (bstan beos su bris pa dang lag len 'gal ba, p. 248). The strongest argument, I believe, on the side of Sa-pa1).'s opponents is their questioning of the kriya tantra's authority over the higher tantra classes. This has the twin aspects of the kriya tantras' inferior position within the tantric system of Tibetan Buddhism and the great significance of bodhieitta in Mahayana Buddhism, of which tantra is a part, which, as is expressed through the passage of the VairoeanabhisalJ'lbodhitantra, may serve on the side of the vajraearya as a motivation to bestow tantric consecration on large(r) numbers of disciples. Obviously the different parties engaged in the debate have come to different decisions. Sa-pa1). and some of his followers decided to follow the rules of the Samanyavidhiguhyatantra, whereas his opponents, such as 'Bri-gung Rig'dzin Chos-kyi-grags-pa, have decided to take the formulation of the VairoeanabhisalJ'lbodhitantra as the lead. Here we may also find the key to the understanding of these positions. Did Sa-pa1). perhaps find it most appropriate to make the least common denominator the basis for his doctrinal formulation? Is this perhaps an attempt to preserve the basic principles of the foundations of the practice also on higher levels? And is this perhaps somehow in the same vein as the Sa-skya-pa doctrine of the three vows (of pratimok~a, of the bodhisattvas, and of mantra), where they hold 25 P. 247: deng sang dgyes rdor kyi dbang bskur la yang grangs sogs byed pa khyod kyi lugs la snang mod.

230

lan-Ulrich Sobisch

that vows are transformed while their essence remains the same (sdom gsum gnas 'gyur ngo bo gcig)? In the case of a person, for instance, who maintains the pratimok~a vows and then takes the bodhisattva vows, this means that the main element for taking up pratimok~a, namely renunciation, is first limited to the desire to attain peace and happiness for oneself. When the resolve for awakening (bodhicitta) is produced, the inferior intention is abandoned, but the rules of pratimok~a (such as abandoning killing) remain as an element of the wish to benefit sentient beings. It is thus the case that although the pratimok~a of the auditors has been transformed into the "bodhisattva (or mahayana) pratimok~a," its rules are still valid, and although the desire to attain peace and happiness for oneself is augmented with bodhicitta, replacing the inferior motivation by the vaster intention, renunciation is still at the core. In other words, although the practice has been lifted up to a higher level, certain principles remain intact. 26 This doctrine was, according to Go-rams-pa, formulated in parts as a reply to the teaching of the Indian pm;H;lita VibhUticandra, who had formulated (perhaps while in Tibet during the early 1200s) the doctrine of the vows that are "outshone" (ziZ gyis gnon pa) by the next higher set of vows. Accordingly, the pratimok~a vows, for instance, "remain dormant in the basic consciousness" (kun gzhi la bag la nyal ba'i tshul du gnas) if one forms the resolve for supreme awakeningY In other words, the rules of the lower level remain latent, but unmanifested, when the higher level is practiced. Now, even though this might look like a convenient background for the teaching that the practice of the higher carya tantra overrides the practice of the lower kriya tantra, I think that this is not the case, at least not in the context of the 'Bri-gung-pa Rig-'dzin Chos-kyi-grags-pa. For if "outshining of the lower" would be the principle behind the permission of bestowing tantric consecration
26 For the Sa-skya-pa view of the simultaneous possession of the three vows in a single mental continuum, see Sobisch 2002, esp. pp. 89-96. I have slightly refined my interpretation of the Sa-skya-pa view on this matter in the present article. 27 For VibhUticandra's view, see Sobisch 2002: 35-88 and esp. p. 115.

Interpreting the tantras

231

to groups of disciples with more than twenty-five participants, this would mean (in unison with Vibhuticandra?)28 that the lower system is perceived as something that is more rigid, while the higher system is more lenient. But this is not the attitude of the 'Bri-gung-pas. In fact, a clear statement to that end can be found (in the context of the three vow doctrine) in the dGongs gcig (Addendum 13/14):29
[Others] hold that the three vows are ascendingly more lenient [while 'Jig-rten-mgon-po] maintained that the three vows are ascendingly more rigid. sdom pa gsum gong nas gong du yangs su 'gro bar 'dod I sdom pa gsum gong nas gong du dog tu 'gro bar bzhed do I

According to this view, which is said to have been introduced in Tibet by Atisa,3o the priitimok~a abandons the coarse mental afflictions (nyon mongs pa rags pa) with the help of a small number (grangs nyung) of rules that are guarded through ordinary (tha mal 28 In his Ri chos, Karma Chags-med (1613-1678) says that VibhUticandra's position was practiced by the bKa'-brgyud-pas (citing the 'Brigung-pa 'Jig-rten-mgon-po as an exception). In his words, "the bKa'brgyud-pas proceed a bit lenient with the rules of the vinaya." He contrasts this with the statement from the addendum of the dGongs gcig, for which see below. For the Ri chos, see rTsib ri spar ma, collected and arranged by La-dwags Khrid-dpon 'Khrul-zhig Padma-chos-rgyal, Darjeeling: Kagyu Sungrab Nyamso Khang 1978-1985, 31 vols., vol. 5, pp. 77 ff. 29 The vajra-statements of the founder of the 'Bri-gung bKa'-brgyudpas together with the formulations of what is the "general opinion" or the "opinion of others" (gzhan lugs) can be found in rDo-rje-shes-rab's commentary (Dam chos dgongs pa gcig pa'i 'grel chen snang mdzad ye shes sgron me, Dehra Dun: Drikung Kagyu Institute 1990), pp. 153-187, the present one on p. 178). In the commentary of rDo-rje-shes-rab, this is the thirteenth addendum, in Rig-'dzin Chos-kyi-grags-pa's commentary (Nyi ma'i snang ba) it is the fourteenth, his commentary is on p. 392 f. (Maryland edition) and on p. 338 (Dehra Dun edition). 30 That this view goes back to Atisa is mentioned in a brief versified commentary by Rig-'dzin Chos-kyi-grags-pa, Dam pa'i chos dgongs pa gcig pa'i dka' 'grel tshigs su bead pa mun sel sgron me, The Collected Works (gSung 'bum) of Kun-mkhyen Rig-pa-'dzin-pa Chen-po Chos-kyigrags-pa, Dehra Dun: Drikung Kagyu Institite 1999, vol. 3, p. 463.

232

Jan-Ulrich Sobisch

pa) body, speech, and mind. The bodhisattva vows do the same, but solely for the benefit of others (rnam pa thams cad du gzhan don kho na las) and through a greater number (mang bas) of rules. They are thus considered to be slightly more rigid than the previous set of vows (cung zad snga ma las dog). The mantra vows abandon the very subtle mental afflictions of all three gates (sgo gsum ga'i nyon mongs shin tu phra ba), and this is done, having abandoned attachment to ordinary appearances, through the body of the deity, i.e. the illusion-like samadhi (lha sku sgyu ma lta bu'i ting nge 'dzin las). While the pratimok~a does not need the two higher sets of vows as a foundation stone (rmang rdo), bodhicitta needs at least the basic vows of pratimok~a, such as the refuge vows, and mantra needs the support of one of the seven pratimok~a vows and mahayana bodhicitta. It is therefore considered the most difficult (ches dka') set of vows. Thus Atisa is quoted with the following words (p. 392 f.):
There never occurred a transgression in the priitimok~a vows to me. Occasionally there occurred trifle [transgressions] in the bodhisattva vows. In the mantra vows the transgressions occurred as a stream. kho bo fa so thar gyi sdom pa fa ni nyes ltung gtan mi 'ong / byang chub sems dpa'i sdom pa fa skabs skabs su tsag tsig tsam 'ong / gsang sngags kyi sdom pa fa ni ltung ba char gcig nyar gcig tu 'ong /

This clearly shows that the view of VibhUticandra, according to which "when the sun (= mantra vows) shines forth, the moonlight (= the bodhisattva vows) vanishes, [yet] the world is luminous" (Sobisch 2002: 117), is not accepted by the 'Bri-gung-pas. In other words, the dGongs gcig does not maintain a view according to which the lower is rigid and the higher is lenient - the opposite is the case: the higher vows are seen as more rigid and the higher practices as more difficult to perform. If we apply this thought to the problem under discussion in this article, the procedure of the carya tantra, that allows for a large number of adepts in a tantric consecration, is seen as the more rigid rule, because it presupposes the motivation of mahayana bodhicitta, and it is therefore also the more difficult practice. If this analysis through the somewhat parallel case of the three vow doctrines is correct, the Sa-skya-pas did not simply hold on to the lower principle while the 'Bri-gung-pas have advanced to the

Interpreting the tantras

233

higher, but rather tried to preserve elements of the practice of a lower tantra on a higher level and the 'Bri-gung-pas' aim for achieving the same results (i.e. tantric consecration), but for a vaster group of participants through a vaster motivation, presupposing a greater qualification of the guru who bestows the higher tantric consecrations, as this is a more difficult practice than the performance of a tantric consecration on the lower level. I hope that future research into this and related material will show whether such underlying trends and dispositions are indeed identifiable in the approaches to tantric practice of Sa-skya-pa and 'Bri-gung-pa masters of the past.

Bibliography and abbreviations

Dalton 2005 - Jacob Dalton: "A Crisis of Doxography: How Tibetans Organized Tantra During the 8th-12th Centuries." Journal of the International Association of Buddhist Studies 28/1, 115-179. Gold 2007 - Jonathan Gold: The Dharma's Gatekeepers: Sakya Pm:uJita on Buddhist Scholarship in Tibet. Albany, N.Y.: State University of New York Press. Hodge 2003 - Stephan Hodge (trl.) The Mahii-Vairocana-Abhisarrrbodhi Tantra, With Buddhaglfhya's Commentary. London: Routledge Curzon. Liu 2002 - Kuo-wei Liu: 'Jig-rten-mgon-po and the 'Single Intention' (Dgongs gcig): His view on Bodhisattva vows and its influence on Medieval Tibetan Buddhism. PhD dissertation, Harvard University. Lodro Taye 2005 - Jamgon Kongtrul Lodro Taye: The Treasury ofKnowledge, Book Six, Part Four: Systems of Buddhist Tantra. Elio Guarisco and Ingrid McLeod (trls.), Ithaca: Snow Lion Publications. Martin 1997 - Dan Martin: "Beyond Acceptance and Rejection? The Anti-Bon Polemic included in the Thirteenth-Century Single Intention (Dgongs-gcig Yig-cha) and Its Background in Tibetan Religious History." Journal of Indian Philosophy 25/3, 263-305. P - The Tibetan Tripitaka, Peking Edition (repr.), ed. Daisetz T. Suzuki. Tokyo/Kyoto 1955-1961. Rhoton 2002 - Jarred Rhoton (trl.): A Clear Differentiation of the Three Codes: Essential Distinctions Among the Individual Liberation, Great Vehicle, and Tantric Systems: the Sdom Gsum Rab Dbye and Six Letters. Albany, N.Y.: State University of New York Press.

234

Ian-Ulrich Sobisch

Sobisch 2002 - Jan-Ulrich Sobisch: Three-Vow Theories in Tibetan Buddhism: A Comparative Study of Major Traditions from the Twelfth Through Nineteenth Centuries. (Contributions to Tibetan Studies 1). Wiesbaden: Dr. Ludwig Reichert Verlag. Sobisch 2009 - Id.: Review of The Dharma's Gatekeepers: Sakya Parpfita on Buddhist Scholarship in Tibet. By Jonathan C. Gold. Albany, NY.: State University of New York Press 2007, pp. xii + 267. Journal of the American Oriental Society 129/3, 512-517. Sobisch, forthcoming - Id.: "Guru Devotion in the Bka' -brgyud-pa Tradition: Its Functioning as the Single Means for the Arising of Realization." In: Tibetan Studies 10, Proceedings of the Seminar of the International Associationfor Tibetan Studies, Bonn, 2006, 26 pp. Stearns 2001- Cyrus Stearns: Luminous Lives. Boston: Wisdom Publications. van der Kuijp 1987 - Leonard van der Kuijp: "An Early Tibetan View of the Soteriology of Buddhist Epistemology: The Case of 'Bri-gung 'Jig-rtenmgon-po." Journal of Indian Philosophy 15/1,57-70. Wayman and Tajima 1992 - Alex Wayman, Ryujun Tajima: The Enlightenment ofVairocana. (Buddhist Traditions 18). New Delhi: Motilal.

Tibetan scholasticism in the 11th and 12th centuries

Contributions to a panel at the XVh Congress of the International Association of Buddhist Studies Atlanta, 23-28 June 2008

Guest editors

Pascale Hugon and Kevin Vose

Introduction

Unearthing the foundations of Tibetan Buddhist philosophy Pascale Hugon and Kevin Vose

The contributions to this volume are the result of a panel on the theme "Tibetan Scholasticism in the 11th and 12th centuries" organized at the 15th Conference of the International Association of Buddhist Studies held at Emory University, Atlanta, in June 2008. 1 Our motivation for this panel came in great part from the recent surfacing of new material pertaining to this period, which opens the way to novel research on the development of Tibetan Buddhism in the early part of the Second Diffusion (phyi dar). In particular, the 2006 publication of the first 30-volume set of the bKa' gdams gSUli 'bum by the dPal brtsegs Institute for Ancient Tibetan Manuscripts (dpaZ brtsegs bod yig dpe rfiin f.ib jug khan)2

1 Three further papers were presented in the lABS panel at Emory: "Ron zorn Chos kyi bzan po's Subclassification of the Madhyamaka School" by Orna Almogi, "Ron zorn Chos kyi bzan po on 'Dialectics' (mtshan fiid), 'Secret Mantra' (gsan snags), and the 'View' (Zta ba)" by Nathaniel Rich, and "Challenging Candraklrti: Phywa pa Chos kyi sen ge's Criticism of Prasangika-Madhyamaka" by Dorji Wangchuk. Owing to other commitments, these could not be included in the present volume. Thomas Doctor's paper was originally presented in the panel "Madhyamaka and Yogacara Models of Truth or Reality in Indo-Tibetan Buddhism."
2 The first 30 volumes appeared in 2006 and a second set of 30 volumes in 2007. A third set (vols. 61-90) has been published in 2009. An improved table of contents of the first 30 volumes appeared in appendix A of Kano (2007), and one of volumes 31-60 appeared in appendix A of

Journal of the International Association of Buddhist Studies Volume 32 Number 1-22009 (2010) pp. 237-248

238

Pascale Rugon and Kevin Vose

gives us access to a rich collection of texts stemming from authors belonging to the bKa' gdams pa school of Tibetan Buddhism, dating from the 11th to the 15th century - a mix of famous figures and so far unknown thinkers, whose works were for the most part hitherto unavailable, either considered lost or simply undiscovered. While the bKa' gdams pa scholars are remembered as the primary importers of philosophical or scholastic traditions into Tibet, their activities were complemented by rNin rna pa scholars who approached similar philosophical issues from their commitments to a specific body of tantric literature. Chief among these rNin rna pa scholars was Ron zorn Chos kyi bzan po, a selection of whose writings were published in 1974, but whose "Collected Works" (gswi 'bum) were made widely available only in 1999, sparking renewed interest in this important figure. 3 The lABS panel, organized by Pascale Rugon, Kazuo Kano, and Kevin Vose, was the first occasion to gather scholars investigating this newly recovered material and share our results on this key, but so far uncharted period of early Tibetan scholasticism. In adopting the term "scholasticism," we focus attention on a series of features characteristic of non-tantric Buddhism in Tibet, features that began to take shape during the first decades of the Second Diffusion, a period foundational to the development of Tibet's Buddhist orders. Previous scholarship on Tibetan scholasticism and scholasticism as a comparative category highlights the central importance of both adherence to tradition and rationality: scholastics treat philosophical issues within the bounds of authoritative texts, frequently in the guise of scriptural exegesis, and always through reasoned analysis. 4 The study of dGe lugs pa

Kana (2909). 3 For a discussion of Ron zorn's Collected Works, see Alrnogi (2002). The most widely available edition of Ron zorn Chos kyi bzan po's works is the Ron zorn Chos bzan gi gsun 'bum, while two earlier editions were published in Khams by Padma kun grol and in California by the Yeshe De Project. As Alrnogi (2002: 78) reports, all three editions contain the same thirty-two works of Ron zorn. 4 Cabez6n (1994) and Cabez6n (1998).

Introduction

239

and rN-in rna pa scholastic traditions has given us a sense of the textual, philosophical, and didactic practices these Tibetan schools developed around Indian texts from the fifteenth century into the present, practices that contributed to the enduring success of these orders. 5 The work of Sa skya Pru:t<;lita Kun dga' rgyal mtshan (11821251) to develop an Indian-oriented scholarly method, consisting of exposition, composition, and debate, has also been explored. 6 What remains unclear is the early history of Tibetan scholasticism. With the wealth of newly available bKa' gdams pa material, together with resources from early rN-in rna pa, bKa' brgyud pa, and Sa skya pa scholars, we can now investigate this crucial period in order both to understand its dynamics and to provide a sense of its contributions to later developments in the history of Buddhism in Tibet. A perusal of the bKa' gdams gsuft 'bum reveals that early bKa' gdams pa authors, like their successors, concerned themselves with a limited range of Indian Buddhist literature: the "Perfection of Wisdom" slitras, particularly as systematized by Haribhadra's Abhisamayiilarrtkiiriilokii; the pramii'(ta literature of Dignaga, Dharmaklrti, and their commentators; the "Maitreya" texts, in particular the Mahiiyiinasiitriilarrtkiira and Ratnagotravibhiiga; Madhyamaka treatises; and Vinaya texts. Patterns of explication, too, are familiar from later scholastic literature; in some cases, we can discover the creation of those patterns - regarding either their form or contents (or both) - in this early literature. One of the most striking features, though, is that we see several familiar themes treated to lively philosophical discussion, with unique and frequently competing conclusions. Among the reasons surely is the newness of much of this Indian literature. As Tibetans turned again to India just prior to the year 1000 to revive their Buddhist traditions, they discovered a wealth of literature either not in circulation or simply not yet created during their previous contacts with the subcontinent. A number of Tibetans' scholarly reputations were made on the basis of their translations of this new materiIn addition to Cabez6n's work, see Dreyfus (2003) and (2005). See notably Jackson (1987) and Gold (2007).

5
6

240

Pascale Hugon and Kevin Vase

al; the bKa' gdams gSUli 'bum volumes show us additionally that translators and their followers immediately set about explicating, analyzing, and classifying these texts. The variety of conflicting interpretations that this new literature inspired bears witness to the creativity (sometimes at the expense of fidelity to their Indian sources) and vitality of these early Tibetan authors. A number of texts that would become centrally authoritative for later generations of Tibetan scholastics entered Tibet only in the eleventh and twelfth centuries; among these are Candraklrti's Madhyamakiivatiira and Prasannapadii. While the later Tibetan analysis of the superiority of CandrakIrti's Madhyamaka is familiar to many, in this formative period, Candraklrti's authority was by no means universally accepted. As the contributions to this volume from Thomas Doctor, Georges Dreyfus and Drongbu Tsering, Kevin Vose, and Chizuko Yoshimizu show, Candraklrti's early Tibetan supporters portray his philosophy in quite distinctive terms, emphasizing the incompatibility of his thought with that of Dharmaklrti; at the same time, we see several bKa' gdams pa authors remaining unconvinced by Candraklrti due to this perceived incompatibility. Cabez6n notes that the scholastic method is marked in part by the attempt to integrate rational and experiential religious domains.? As can be witnessed in this volume, such ten,sion came in this period to be embodied in the juxtaposition of Dharmaklrti and CandrakIrti; the marriage of the two, a central characteristic of later Tibetan scholasticism, had yet to be achieved. While Dignaga's and Dharmaklrti's main treatises entered Tibet in the imperial period, eleventh century retranslations together with translations of crucial commentarial materials on these treatises revitalized Tibetan study of them, sparking a "new epistemology" (tshad ma gsar ma) in Tibet, centered at gSaIi phu Ne'u thog Monastery (founded in 1073 by rNog Legs pa'i ses rab). As will be seen 'in the contributions to this volume from Pascale Hugon and Jonathan Stoltz, this revival led to broad Tibetan explorations of the nature of knowledge, including the role of reasoning in religious pursuits. In addition to giving us a first-hand look at the
Cabezon (1994: 19-20, 190-191).

Introduction

241

"new epistemology," the newly recovered materials will allow us to assess the degree to which later generations of Tibetan epistemologists are indebted to gSaIi. phu's innovations. The success of the "new epistemology" indeed marked gSaIi. phu as the pre-eminent training ground for generations of Tibetan scholars. In addition to the "new epistemology" and Can,draklrtian Madhyamaka, a third crucial textual stream that entered Tibet during this same period was the Ratnagotravibhiiga (Uttaratantra) attributed to Maitreya. Reputed to have been rediscovered, along with the Dharmadharmatiivibhiiga, by MaitrIpada around the year 1000, this text became available in Tibet shortly after through rNog Bio ldan ses rab's translation. In Tibet, the Ratnagotravibhiiga became the centerpiece of competing understandings of "Buddhanature;" several enduring interpretations emerged within the early bKa' gdams pa tradition, as seen in Kazuo Kano's contribution to this volume. As Kano discusses, bKa' gdams pa authors understood Buddha-nature to be in perfect harmony with Madhyamaka explanations of emptiness, while some early Tibetan interpreters saw a Buddha-nature endowed with the qualities of a Buddha to be an exception to the rule: Buddha-nature was empty only of defilements. These competing views in time would fuel the debate over "intrinsic emptiness" (ran ston) and "extrinsic emptiness" (gzan ston) in Tibet. The degree of the impact of "new translations" on the forms of Buddhism surviving in Tibet from the imperial period - that would coalesce into the rN-iIi. rna school- remains a conundrum. As Heidi Koppl's contribution to this volume suggests, one pressing concern among defenders of "old" Buddhism was the relationship between Madhyamaka thought, particularly the "Yogacara-Madhyamaka" that integrated the pramiilJa tradition, and the Mahayoga tantras that are central to rN-iIi. rna. At least in the treatment of RoIi. zorn Chos kyi bzaIi. po, this relationship is perhaps cognate to bKa' gdams pa concern with the works of Dharmaklrti and Candraklrti: for RoIi. zorn, pramiilJa-infused Madhyamaka represents an undue preoccupation with reasoning, while Mahayoga alone yields transformative religious experience.

242

Pascale Hugan and Kevin Vase

The selection of papers presented here illustrates well the range of problems figuring among the core preoccupations of this timeframe, and hints also at their interconnection. A large place is devoted in this volume to thinkers belonging to the bKa' gdams pa school - many of them linked to gSan phu Ne'u thog Monastery. The first of them is rNog BIo Idan ses rab (1059-1109), the nephew of the monastery's founder and its second abbot, who embodies severallevels of the scholarly enterprise: besides being renown for his translations - he earned the title of"Lo chen" or "Lo tsa ba chen po" (i.e., "great translator") - he also authored an impressive number of exegeses, summaries and commentaries, and established lineages in several areas of Buddhist learning. Kazuo Kano's paper brings to the fore rNog's contribution to the Buddha-nature teaching in his interpretation of the Ratnagotravibhiiga. Kano provides a firsthand account of rNog's view on the subject by offering a survey of key themes in rNog's commentary, and brings to light the exegetical problems faced by this scholar in his attempt to reconcile the source-text with his own philosophical orientation. This involves, among the core issues, reading the Ratnagotravibhiiga's teaching of Buddha-nature as an ultimate compatible with a Madhyamakaoriented view of emptiness. rNog's exegetical strategy and interpretation paved the way for later generations of scholars. Kano . sketches in particular rNog's impact on the bKa' gdams pa tradition by comparing a commentary on the same text by a scholar active a generation later, Phya pa Chos kyi sen ge (1109-1169), a comparison that, without undermining rNog's indubitable influence, bears witness to a tradition in continuous evolution. Although not all the material we could wish for has become available from rNog's disciples - in particular, no epistemological work by these authors has yet surfaced - we are now in a position to examine their views on several topics, notably Madhyamaka. The massive bsTan rim chen mo (Great Exposition of the Stages of the Doctrine) of rNog's student, Gro lun pa BIo gros 'byun gnas, has been available for some time and includes lengthy discussions of Madhyamaka issues. To this, the bKa' gdams gSUli 'bum adds several works of rGya dmar pa Byan chub grags, including (in volume 31) an independent Madhyamaka composition, dBu ma'i de kho na fiid gtan la dbab pa. Exploring the possibilities for tracing

Introduction

243

the development of Madhyamaka thinking among early gSan phu authors, Kevin Vose's paper guides us through several interpretations of Santideva's Bodhicaryiivatiira, touching upon the readings of four generations of commentators: rNog, rGya dmar pa, his student Phya pa Chos kyi sen ge, and Phya pa's student gTsannag pa brTson 'grus sen ge. Vose shows how interpretations of Santideva's depiction of the ultimate were instrumental in early Tibetan divisions of Madhyamaka, with some authors aligning Santideva with Candraklrti's views, while others chose to read him as agreeing with a logico-epistemologically oriented Madhyamaka. By showing the evolution of efforts to reconcile Santideva's views of the ultimate with the emphasis on logic characteristic of the gSan phu lineage, Vose brings to the fore key aspects of the early development of Madhyamaka categories and shows their relevance to later discussions on the nature and origin of the Svatantrika-Prasangika divide. Kano's and Vose's papers both take into account the views of Phya pa Chos kyi sen ge, the 6th abbot of gSan phu Ne'u thog, who pursued his predecessors' endeavors at exegeses and commentaries on the texts whose lineages were initiated by rNog. While he emerges as an original interpreter of Madhyamaka, Phya pa is especially famous for his contribution to epistemology. The papers by Pascale Hugon and by Jonathan Stoltz both focus on Phya pa the logician. The surfacing of Phya pa's works on the subject offers a welcome opportunity to revise a number of "myths" propagated in the absence of first-hand material. We are now in a position to reassess his relation with the Indian tradition and the extent and orientation of his innovative input, which is to be understood in the more general framework of his Madhyamaka system. Early epistemological treatises like those of Phya pa demonstrate (along with variant interpretations of Indian sources - a number of which can be traced back, let us note, to the influence of Indian commentators) new themes of inquiry that would come to occupy a significant place in the work of later Tibetan scholars. Hugon's paper deals with one such innovation, the so-called "theory of definition," of which the earliest comprehensive account presently available is found in Phya pa's two main works of epis-

244

Pascale Hugan and Kevin Vase

temology. Hugon leaves aside the intricate details of this theory to adopt a historical perspective aiming at identifying possible sources and influences that may have led to the constitution and elaboration of this theory. She first puts to the test some traditional claims as to the arising of this theory - translation issues; crossdomain discussions - and explores the possibility of an influence not acknowledged in the tradition, located in Dharmottara's commentarial works. Although the theory of definition likely originated in epistemological issues and is treated extensively in epistemological treatises, Hugon's paper hints to its interconnection with Phya pa's Madhyamaka agenda. Stoltz takes up Phya pa from a philosophical perspective - and a critical one - throwing light on aspects of Phya pa's typology of mental states and his understanding of dependence. Starting from a single analogical argument between Phya pa and a hypothetical opponent, Stoltz weighs the implications behind the analogy that involves, on the one side, conceptual cognition and its objects, concepts, and, on the other, a specific type of cognition termed "factive assessment" (yid dpyod), which brings about a correct cognition, although not by the recognized means that would qualify it as a "valid cognition" (tshad rna). Analyzing the different models of dependence that could be applied between the respective cognitions and their objects, Stoltz shows Phya pa's difficulty in justifying his stance on the way factive assessment works. Although Stoltz takes a philosophical approach, his discussion of this argument reveals the impact of Phya pa's historical context. He shows that Phya pa's philosophical difficulties come from his reluctance to dismiss traditional categories (in this case, the accepted types of natural dependence, causality and identity), which leaves him in an awkward - and perhaps indefensible - position regarding his new typology of cognitions. Whereas Phya pa's influence in the domain of epistemology was long lasting, as can be seen from the works of his disciple gTsan nag pa and later epistemological works attached to gSan phu Monastery, many of his students turned their backs on his Madhyamaka views. This turnaround comes as a consequence of the Candraklrtian "resurrection" brought about by the activities of

Introduction

245

Pa tshab Ni rna grags (1055-1145?). Pa tshab was not only the translator of Candraklrti's major works, he was also active as a teacher, and founded a'Madhyamaka lineage which came to dominate the Tibetan landscape. The paper written by Georges Dreyfus in collaboration with Drongbu Tsering constitutes a pioneer study of Pa tshab's thought based on three works attributed to Pa tshab published in the bKa' gdams gSUli 'bum and a significant contribution to our knowledge of the early development of Madhyamaka thought in Tibet. Dreyfus and Tsering seek first to establish the authorship of the three works and examine their historical significance, particularly in view of the origin of the Svatantrika-Prasangika divide, which, they suggest, might have its source in the late Indian tradition. The authors' examination of the contents of these texts focuses on two themes, namely, can Madhyamikas have any thesis and can they include valid cognition and related theories within their tradition? The paper portrays Pa tshab's views on these issues as a kind of skepticism, and deals with the probable reasons why, albeit a pivotal figure, Pa tshab was not an influential interpreter of Madhyamaka, particularly when compared with his students, rMa bya Byan chub brtson 'grus and Zan Than sag pa Byun gnas ye ses. The papers by Thomas Doctor and Chizuko Yoshimizu bring into light the respective contributions of these two disciples of Pa tshab and their views on the controversial issues mentioned above. rMa bya came to play a pivotal role in the development of the Candraklrti-inspired interpretation of Madhyamaka in Tibet. He was initially a student of Phya pa Chos kyi sen ge and as such stood in the circle of the gSan phu epistemological school. His turn from his teacher Phya pa's prama~a-infused Madhyamaka to embrace the new Candraklrti movement represents a major development in Tibetan Madhyamaka: based on his epistemological training, rMa bya was the first to begin integrating the Dignaga-DharmakIrti tradition and Candraklrti's Madhyamaka, a process that continues to occupy Tibetan Madhyamaka exegetes. Doctor's paper, based on a newly recovered work by rMa bya and supplemented with his only previously known text, situates the author in view of his hermeneutical framework regarding Yogacara and Madhyamaka and takes up his explanation of the two truths. He demonstrates rMa bya's effort to reconcile the application of valid cognition with his Madhyamaka

246

Pascale Hugan and Kevin Vase

anti-realist stance, in particular in his account of mere appearances (snan tsam), Doctor shows that rMa bya, anticipating Tson kha pa's views two centuries prior to that extraordinarily influential figure, sees Phya pa's brand of "Svatantrika"-Madhyamaka as holding to the conventional existence of "particular characteristics" (ran gi mtshan fiid), while Candraklrti's Madhyamaka avoids this conventional reification and so offers a superior account of "worldly" valid cognition within an empty world. Also a (presumably direct) disciple of Pa tshab, Zan Than sag pa's exact affiliation and lineage remain uncertain. The founder of Than sag Monastery in 'Phan yul, a renowned center of Madhyamaka studies, Zan Than sag pa contributed to the advance of Candraklrti-inspired Madhyamaka in Tibet. While we do not yet know whether he had any schooling at gSan phu or perhaps learned logic from Pa tshab or Kanakavarman, Zan Than sag pa displays a wide and deep knowledge of the Buddhist logico-epistemological system. As with rMa bya, this epistemological background plays a significant role in his reading of Madhyamaka. Yo shimizu's paper focuses on Zan Than sag pa's approach to one of the most prominent questions that Madhyamaka must face, namely, that of the possibility of holding a thesis. This is linked, moreover, with the commentarial difficulty of accounting for Candraklrti's use of the term "thesis" when referring to several of Nagarjuna's statements. On the basis of his only available work, a newly recovered and complete commentary on Candraklrti's Prasannapadii, Yoshimizu examines how Zan Than sag pa reconciles a commitment to logical argumentation with his denial that a Madhyamika holds any philosophical thesis. Zan Than sag pa, then, treats the question of philosophical thesis with regard to both the logicians' - Dignaga's and Dharmaklrti's - discussions and those of Nagarjuna's commentators. The particular status of negation is taken up in relation to this issue, as Yoshimizu examines Zan Than sag pa's refusal to admit even negation as a thesis. Clearly, a central concern for Madhyamaka scholars of this period was the place and role of epistemology: whether supportive of or antagonistic to the new Candraklrti movement, bKa' gdams pa authors sought to utilize valid means of knowledge within re-

Introduction

247

ligious pursuits, while still preserving the unique perspective of Madhyamakaemptiness. Heidi KoppI's examination of Ron zorn Chos kyi bzail po's Madhyamaka presents a quite different concern, the superiority of Tantra to any "sutra"-based viewpoint. Koppl investigates how Ron zorn's commitments to rNinma pa tantras lead him to consider Madhyamaka as a view inferior to that of Mahayoga. Starting with Ron zorn's criticism of Madhyamaka, Koppl inquires into the author's singular view of relative truth as pure appearance and his description of appearances as "divine." She shows that Ron zorn's critique of Madhyamaka assumes the prama~a-infused variety, leading to the possibility that he would have endorsed CandrakIrti's Madhyamaka, with its utter rejection of any reification of ordinary appearances, which Ron zorn understood to be a precondition of seeing those appearances as "divine." While we have no evidence that Ron zorn knew CandrakIrti's work - he may have written prior to Pa tshab's promulgation of Candraklrti in Tibet - KoppI's work suggests some of the larger forces at work in the development of scholasticism in Tibet. The interplay of rNin rna pa tantric commitments and the nontantric literature entering Tibet in this period is certainly a fertile field for future research. Likewise, there is yet more to be discovered about the interaction of Madhyamaka and epistemology, dbu ma and tshad ma, that in later Tibetan scholasticism would come to be fused as "dbu tshad." The newly published texts provide a wonderful opportunity for further research. It is a desideratum that international collaboration continues on these abundant new materials so that individual studies may be woven into a broader and more precise picture of the intricate issues of this early period. The research presented here merely scratches the surface of this foundational stage in Tibet's development of Buddhist scholasticism, but we hope that it gives a sense of its vitality and far-reaching importance.

248

Pascale Hugon and Kevin Vose

Bibliography
bKa' gdams gSUli 'bum phyogs bsgrigs. Chengdu: dPal brtsegs bod yig dpe rfiin :lib 'jug khan, 2006-2009. Ron zom Chos bzan gi gsun 'hum. Chengdu: Si khron mi rigs dpe skrun khan, 1999. . Almogi, Orna. 2002. "Sources on the Life and Works of the EleventhCentury Tibetan Scholar Rong zorn Chos kyi bzang po: A Brief Survey." In Tibet, Past and Present, ed. Henk Blezer. Leiden: Brill, 67-80. Cabezon, Jose 1. 1994. Buddhism and Language: A Study of Indo-Tibetan Scholasticism. Albany, NY: State University of New York Press. Cabezon, Jose 1. (ed.). 1998. Scholastisicm: Cross-Cultural and Comparative Perspectives. Albany, NY: State University of New York Press. Dreyfus, Georges. 2003. The Sound of Two Hands Clapping: The Education of a Tibetan Buddhist Monk. Berkeley, CA: University of California Press. - . 2005. "Where Do Commentarial Schools Come From? Reflections on the History of Tibetan Scholasticism." Journal of the International Association of Buddhist Studies 28.2: 273-297. Gold, Jonathan C. 2007. The Dharma's Gatekeepers: Sakya PalJita on Buddhist Scholarship in Tibet. Albany, NY: State University of New York Press. Jackson, David P. 1987. The Entrance Gate for the Wise (Section III): Sa-skya PalJita on Indian and Tibetan Traditions of Prami.ilJa and Philosophical Debate. 2 volumes. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien. Kano, Kazuo. 2007. "A Critical Edition of rNgog Blo ldan shes rab's sPring yig bdud rtsi'i thig Ie." KiJyasan daigaku mikkyiJbunka kenkyujo kiyiJ 20: 1-58. - . 2009. "An Annotated Translation of rNgog Blo ldan shes rab's sPring yig bdud rtsi'i thig Ie." KiJyasan daigaku mikkyiJbunka kenkyujo kiyiJ 22: 121-178.

rNog Bio Idan ses rab's position on the Buddha-nature doctrine and its influence on the early gSaIi phu tradition' Kazuo Kana

Introduction
The teaching that all sentient beings have Buddha-nature (tathiigatagarbha) was first proclaimed in the Tathdgatagarbhasutra. Developed in a series of Mahayana sutras, such as the Srfmdlddevfsutra and Anunatvdpun;atvanirdeiasutra, it was then systematized in the Ratnagotravibhdga (abbr. RGV), alias Mahdydnottaratantradstra. The core idea of the RGV's teaching is that everyone possesses Buddha-nature. The latter does not change throughout the progression from the level of ordinary beings to that of a Buddha, it is merely purified through the separation from adventitious defilements. Once this purification is complete, awakening is accomplished. Both Indian and Tibetan traditions struggled with the question of the ontological status of Buddha-nature. One finds indeed in some slitras descriptions of Buddha-nature as permanent and pervading every sentient being, which are also characteristics ascribed by non-Buddhists to the Self (atman). But if Buddha-nature were to be understood as a permanent entity akin to a Self, how could this

, I wish to thank Pascale Rugon and Kevin Vose for their very valuable comments on this paper. Work on this paper has been generously supported by the Grant-in-Aid for Young Scientists (Start-up) program of the Japan Society for the Promotion of Science (JSPS).
Journal of the International Association of Buddhist Studies Volume 32 Number 1-22009 (2010) pp. 249-283

250

KazuoKano

teaching be compatible with the standard Buddhist doctrine that everything is impermanent and selfless? Some Mahayana slitras, such as the Mahayana MahaparinirVa1}asutra, would offer support for the assimilation of Buddhanature with a Self. The Mahayana Mahaparinirva1}asutra is quite explicit in associating the two notions, characterizing in particular the dharmakaya in terms of "perfection of Self" (atmaparamita), but warns about the confusion of the "correct" atman, which is Buddha-nature, with atman taken in its ordinary sense. 1 RGV 1.37 and RGVV also speak of the "perfection of Self" as an epithet of the dharmakaya, interpreting however this notion of "Self" (atman) in the sense of selflessness (nairatmya) or quiescence of conceptual proliferations (prapaiica), thus distinguishing Buddha-nature from the notion of a personal, permanent Self (atman).2
1 The Mahiiyiina MahiiparinirviilJasatra equates iitman with Buddhanature (see P 788 tu 105b5 [",T vol. 12, 407b; 883b]: bdag ces bya ba ni de biin gsegs pa'i sfiin po'i don to II) and characterizes the dharmakiiya (that is, the resultant aspect of Buddha-nature; see below [iD in terms of "perfection of permanence" (nityapiiramitii), "perfection of bliss" (sukhapiiramitii), "perfection of Self (iitmapiiramitii), and "perfection of purity" (subhapiiramitii) (see P 788 tu 33b3-34a2 [",T vol. 12, 377c-378a; 862bD. 2 RGVV 31,13-16: tathiigatas tu a punar yathiibhatajfiiinena sarvadharmanairiitmyaparapiiramipriipta/:tb I tac ciisya nairiitmyam aniitmalak~alJena yathiidarsanam avisa1!Lviiditatviit" sarvakiilam iitmiibhipreto nairiitmyam eviitmetid krtvii I yathokta1!L sthito 'sthiinayogeneti I (a Schmithausen [1971: 143] corrected tathiigata/:t to tathiigatas tu; b Johnston xvi; Schmithausen [1971: 143] corrected avisa1!Lviiditviit to avisa1!Lviiditatviit; d Schmithausen [1971: 143] corrected eviitmani to eviitmeti) RGVV. 32,9-10: prajfiiipiiramitiibhiivanayiikiisopamasattvabhiijanaloC

kanairiitmyani~thiigamaniid.

See also RGVV 33,8-10: tiim eva ciividyiiviisabhami1!L pratftya sak~ma nimittaprapaficasamudiiciirayogiid atyantam anabhisa1!Lskiiriim iitmapiiramitii1!L niidhigacchanti. Schmithausen (1971: 143-144 and 1973: 135) links this sentence to the Madhyamaka view. For instance, the Madhyamakahrdaya (III.284cd) similarly defines dharmakiiya as quiescence of conceptual proliferations

 rNog BIo Idan ses rab on the Buddha-nature doctrine

251

Nevertheless, the RGV does not promote the doctrine of emptiness in the sense that everything is ultimately empty of intrinsic nature. Quite on the contrary, the RGV stresses the real existence of Buddha-nature, and proclaims the superiority of the Buddha-nature doctrine to the emptiness doctrine of the Prajfiiipiiramitiisiitras. 3 The RGV thus on the one hand distinguishes Buddha-nature from the disapproved view of a Self, while on the other hand it admits Buddha-nature as ultimately existent4 - an ambiguous viewpoint, and a challenging one for its interpreters. The RGV in Tibet The RGV was first translated into Tibetan by Atisa, working in collaboration with Nag tsho Tshul khrims rgyal ba. 5 A new translation was carried out by rNog 10 tsa ba BIo Idan ses rab (1059-1109)6 in collaboration with Sajjana. 7
(buddhiiniif!l dharmakiiyo 'yaf!l prapaficopasamaJ:t sivaJ:t). 3 The alternative title of the RGV, mahiiyanottaratantra "supreme doctrine of the Mahayana," hints to the superiority of the Buddhanature doctrine to the emptiness doctrine. Cf. RGV L160: purvam evaf!l vyavasthapya tantre punar ihottare / paficado~aprahii1Jiiya dhiitvastitvaf!l prakiisitam // 4 Cf. RGVI.53, L165; RGVV2,1l-13. 5 This translation is not available but is sometimes quoted in gZon nu dpal's RGV commentary. See Kano 2006, chapter 2 ("Six Tibetan Translations of the Ratnagotravibhiiga"). Atisa himself quotes the RGV in his own works; for instance, we find RGV 1.86 quoted in his Dharmadhiitudarsanagfti (D 2314 = D 4475; P 3153 = P 5388; modern edition in Lobsang Dorjee 1999), a work translated by himself and Nag tsho. Note that the translation of this work slightly differs from rNog's translation, namely, the third line has the reading "brdzun med bslu med 'tshe med pas," whereas rNog's translation reads "brdzun med bslu med chos fiid dan" for the Sanskrit amr~iimo~adharmitvam. 6 On rNog's life, see Kano 2006 and Kramer 2007. 7 This is the translation currently found in D 4024/ P 5525. In total six Tibetan translations of the RGV are accounted for: (1) Atisa (982-1054) and Nag tsho Tshul khrims rgyal ba (1011-1064); (2) rNog BIo ldan ses

252

Kazuo Kana

Although there is some evidence of discussions on Buddhanature in pre-11th-century Tibetan works,8 as well as in 11th-century works,9 rNog was the first to compose an extensive commentary on the RGV, as well asa topical outline (bsdus don) of the same text.lO rNog's RGV commentary has been available since 1996 in the form of a facsimile reprint from a xylograph version, and an old manuscript was recently published in the bKa' gdams gsun 'bumY
rab (1059-1109) and Sajjana (late 11th century); (3) Pa tshab Ni rna grags (b. 1055); (4) Mar pa Do pa Chos kyi dban phyug (1042-1136); (5) Jo nan Lo tsa ba BIo gros dpal (1299-1353 or 1300-1355); (6) Yar kluns Lo tsa ba Grags pa rgyal mtshan (1242-1346). No. (2) is the only one presently available. Fragments from (1), (3), and (5) are found in quotations in Tibetan RGV commentaries. For the details, see Kano 2006, chapter 2. 8 See Wangchuk 2005: 178-180. 9 Sajjana's collaborative translator gZu dGa' ba'i rdo rje composed a "brief note" (zin bris) on the RGY. See gZon nu dpal, rGyud bla me lOri, 4,17-18 (Pwujita sadzjana'i gsuri la 10 tsa ba gzu dga' rdor gyis zin bris byas ba'i rgyud bla ma'i rnam Mad) and A khu Chin Ses rab rgya mtsho, Tho yig, no. 11338. Sajjana's disciple bTsan Kha bo che composed the Pad ma lcags kyu that discusses his gian stori viewpoints (see Zab khrid yi ge, 205,2-206,4). According to Zu chen Tshul khrims rin chen, bTsan composed a brief note (zin tho) on the text (Zu chen thob yig, Kha 72a3-4: 'khor 10 mtha' ma'i dgons don dan mtha' phye chos dan chos fiid phye ba'i rjes su 'bran ba 'byed gfiis / rgyud bla rnams dpe tsam yan dka' bar byun tshul btsan kha bo che'i zin thor gsal ba biin rgya gar du'an yod par grags pa'i gian ston gi brgyud pa ni ...). The 'Bras spuns dkar chag mentions a Tibetan manuscript of a RGV commentary by Mar pa (see 'Bras spuns dkar chag, p. 1411: rGyud bla'i bsdus don, 31 folios; 'di mar pa 10 tsa ba'i yin nam brtag), who might be the same Mar pa Do pa Chos kyi dban phyug (1042-1146) to whom a translation of the RGV is also attributed. See above, note 7. 10 An incomplete manuscript of the topical outline found in Kharakhoto, preserved at the British Museum, was edited in Kano 2008. 11 The facsimile comes from an unrevised xylograph version (abbr. A) carved in the twentieth century, which was in possession of Dwags po Rin po che in Paris (see Jackson 1993). A copy of a revised blockprint is preserved in Tohoku University (Tohoku Zogai, no. 6798); another copy

-rNog BIo Idan ses rab on the Buddha-nature doctrine

253

In this commentary entitled "The Essential Meaning afthe Uttaratantra" (Theg pa chen po rgyud bla ma'i don bsdus pa, abbr. rGyud bla don' bsdus), rNog not only glosses on the RGV but presents his own interpretation of Buddha-nature, an interpretation that had a significant influence especially in the early bKa' gdams pa tradition (more precisely, the tradition of gSaIi phu Ne'u thog Monastery) and a long lasting impact in Tibet in generalP bTsan Kha bo che (b. 1021), was also instrumental in introducing the teaching of the RGV in Tibet.13 A lack of materials prevents us from determining his position on Buddha-nature. However, he does not appear to have been excessively influential. According to gZon nu dpal, the transmission of bTsan's teaching tradition had been broken by the 14th century.14 As pointed out above, the interpretation of the RGV presented obvious difficulties in view of its ambiguity. In addition, doctrinal divergences between the RGV and Madhyamaka had to be dealt with by interpreters adopting the latter system. These were handled variously. One attitude consisted in accepting the teaching of the RGV literally and regarding it as a "profound" Madhyamaka position to be distinguished from the standard Madhyamaka position
of the same revised blockprint was microfilmed by the Nepal German Manuscripts Preservation Project (Reel no. L 519/4). All these xylograph versions contain a number of serious errors and lacunas. These can now be corrected thanks to the old dbu med manuscript (abbr. B) found at gNas bcu lha khan temple of 'Bras spuns Monastery, which was included in the bKa' gdams gsuri 'bum. (Citations from rGyud bla don bsdus in this paper are based on B and retain all of its orthographical particularities.) A unique folio from another old manuscript (the last folio of the work) was found by myself in Lhasa in 2007. A critical edition of the first chapter of the rGyud bla don bsdus on the basis of A, along with an English translation, are to be found in Kano 2006 (an improved version collated with B is currently under preparation for publication). 12 rNog is time and again quoted by later masters. See Kano 2006, chapters 5 and 6. 13 See, for instance, Mathes 2008: 32-34, 46-47. 14 Deb ther sri on po, 424.

254

KazuoKano

of the Mulamadhyamakakarika, etc. Proponents of this position, such as Dol po pa Ses rab rgyal mtshan (1292-1361), are called "gian stan pas" because they admit that Buddha-nature is empty of "other," i.e., external defilements, but not empty of its own pure nature, i.e., Buddha-qualities.1s In this paper we will consider another strategy, typical of rNog and his followers, which consists in reformulating the teaching of the RGV so as to make it compatible with their Madhyamaka view. Most previous studies of the Buddha-nature teaching in the Tibetan tradition have focused on post-14th-century works, such as those of Bu ston Rin chen grub (1290-1364), Dol po pa Ses rab rgyal mtshan, rGyal tshab Dar rna rin chen (1364-1432), etc.,16 while only a few pioneer studies address earlier developments. 17 The broader corpus now available puts us in a position to shed more light on the earliest phase of the Tibetan tradition of teachings on Buddha-nature, and thereby to clarify the doctrinal basis for later developments. is The present paper deals with a selection of rNog's most significant views on the doctrine of Buddha-nature and considers some reactions to his interpretations in the works of his followers. Since the RGV commentaries attributed to two of rNog's "four main [spiritual] sons" (sras kyi thu ba bii), Zan Tshes spon ba Chos kyi bla rna and Gro lun pa BIo gros byun gnas,19 as yet remain to be

1S For the details of the gian stan pa position, see for instance Mathes 2008: 45-48. 16 The most relevant studies include Seyfort Ruegg 1969 and 1973, Ogawa 1969, and Hookham 1991. 17 See Kano 2006 (chapter 4) for a discussion of the state of research. Significant contributions include Seyfort Ruegg 1969, Jackson 1993, and Mathes 2008, who made use of rNog's RGV commentary. 18 The present paper elaborates on my earlier research on the subject that can be found in Kano 2001, 2003 and 2006. 19 The other two are Khyuri rin chen grags and 'Bre ses rab 'bar. Cf. bKa' gdams chas 'byun gsal ba'i sgron me, 151.

 rNog BIo Idan ses rab on the Buddha-nature doctrine

255

found,20 we will concentrate on the next-earliest available work, a commentary by Phywa pa Chos kyi sen ge (1109-1169).21 rN"og's position In rNog's rGyud bla don bsdus, the most significant work for his position on the Buddha-nature doctrine, the following discussions can be singled out: (i) The three aspects of Buddha-nature (dharmakaya, tathata, gotra) , (ii) Buddha-nature as a cause,
(iii) Buddha-nature and alayavijiiana,

20 Both A khu Chin Ses rab rgya mtsho and gZon nu dpal ascribe RGV commentaries to these two authors. (Cf. respectively Tho yig, nos. 11333 and 11339, and rGyud bla me lon, 4,23, 574,5.) gZon nu dpal also lists RGV commentaries by Chos kyi bla rna's disciple Nan bran pa Chos kyi ye ses (12th century); Phywa pa's disciple gTsan nag pa brTson 'grus sen ge (12th century); and Dan 'bag sMra ba'i sen ge (12th century), See gZon nu dpal, rGyud bia me lon, 4,23-24. A khu Chin Ses rab rgya mtsho most likely copied gZon nu dpal's references (see Tho yig nos. 11331, 11334, 11335). 21 Phywa pa authored a RGV commentary and a topical outline of RGV (based on rNog's topical outline), both of which were found recently at gNas bcu lha khan of 'Bras spuns Monastery and published in 2006 in the bKa' gdams gsun 'bum. See Theg pa chen po rgyud bla ma'i bstan bcos kyi tshig dan don gyi cha rgya cher bsfiad pa phra ba'i don gsal ba (abbr. rGyud bla don gsal), vol. 7, pp. 163-345; Theg pa chen po rgyud bla ma'i bsdus pa'i don, vol. 7, pp. 145-156. The authorship of the rGyud bla don gsal is confirmed on the basis of (a) the colophon of the manuscript that states "composed by the monk Chos kyi sen ge" (see Phywa pa, rGyud bla don gsal, 93a3: theg pa chen po rgyud bla ma'i bstan bcos kyi tshig dan don gyi cha rgya cher bsfiad pa phra ba'i don gsal ba f.es bya ba II sa kya'i dge slon chos kyi bla ma'i f.abs kyi rdul phyi bos nod pa dge sion chos kyi sen ges fie bar sbyar ba rdzogs s.hyo II) and (b) a citation nominally attributed to Phywa pa in BIo gros mtshuns med's RGV commentary (see Kano 2007: 109, n. 75).

256

Kazuo Kano

(iv) the ontological status of the Buddha-qualities (interpretation ofRGV 1.154-155), and (v) the relation between the emptiness doctrine and the Buddhanature doctrine (interpretation of RGVI.156-167).
(i) The three aspects of Buddha-nature (dharmakiiya, tathata,

gotra) The three characteristic aspects22 of Buddha-nature, namely the dharmakaya, tathata, and gotra, play a central role in the philosophical system of the RGV. They are introduced in RGV 1.27-28 and are explained in more detail inRGVI.130-152. 23 RGV1.27-28 state as follows:
Because Buddha's wisdom [i.e., dharmakaya] pervades [all] kinds of sentient beings, because their [i.e., sentient beings'] immaculateness [i.e., tathata] is by nature non-dual [i.e., is not different from Buddha's immaculateness], and because one metaphorically designates the Buddha-gotra by [the name of] its effect [i.e., Buddha's enlightenment], [the Buddha] taught that all sentient beings have Buddha-nature
(RGV1.27).

Because the body of the Perfect Buddha [i.e., the dharmakaya] is allpervading, because tathata cannot be differentiated [i.e., it pervades everything], and because gotra exists [in sentient beings], the Buddha taught that all sentient beings always have Buddha-nature (RGV1.28).24

The three notions of dharmakaya, tathata and gotra, enumerated in the RGV as constitutive of the "intrinsic nature" (svabhava) of

22 The RGVV (69,17) calls them the "three-fold intrinsic nature" (trividha"f'!'L svabhiiva"f'!'L).
23

rNog's discussion of these notions is found in his commentary on

RGVI.130-152. See rNog, rGyud bla don bsdus, A 4Ia2-42a6; B 25b6-

26b3.
24 RGV 1.27: buddhajiianantargamat sattvarases tannairmalyasyadvayatvat prakrtya I bauddhe gotre tatphalasyopacarad uktaf:z sarve dehino buddhagarbhiif:z II; RGV 1.28: sa"f'!'LbuddhakayaspharafJat tathatavyatibhedataf:z I gotratas ca sada sarve buddhagarbhaf:z sarfrifJaf:z II

rNog Blo Idan ses rab on the Buddha-nature doctrine

257

Buddha-nature,25 are identified by rNog as being, respectively, the resultant Cbrasbu), intrinsic (rali biin), and causal (rgyu) aspects.26 All three are to be equated with emptiness. rNog's commentary on the first states that:
[Dharmakaya pervades all sentient beings] because dharmakiiya is emptiness, and because emptiness, again, abides in [all] sentient beings. Therefore, it was taught [in the RGVV],27 "[There are no] sentient beings who stand outside the dharmakiiya of the tathiigata."28

Note that RGVI.145 uses the term dharmadhiitu to explain dharmakiiya 29 - dharmadhiitu being the sphere of reality that is realized by non-conceptual knowledge and Buddha's instruction (deianii), and associated with emptiness in works such as the Madhyiintavibhiiga. 30 rNog's commentary on the idea that every sentient being possesses tathatii, an intrinsic nature shared by ordinary beings and Buddhas, brings out a further equation between tathatii and emptiness. It spells out also the difference between two "levels" of emptiness; whereas dharmakiiya denotes emptiness only in the

25 RGVI.144ab: svabhiivo dharmakiiyo 'sya tathatii gotram ity api I

26 rNog, rGyud bla don bsdus, A 29a3-4; B 18b5: rdzogs sans sku ni
'phro phyir dan lies bya ba la sogs pa la I 'bras bu dan I ran biin dan I rgyu'i sfiin po can yin pa'i phyir I de biin gsegs pa'i sfiin po can du dgO/is pa'o II 27 RGVV70,18-19: sa kascit sattvaJ:! ... yas tathiigatadharmakiiyiid bahir. 28 rNog, rGyud bla don bsdus, A 41b1; B 26b2: chos sku de ni (B: fiid) ston pa fiid yin la I ston pa fiid kymi sems can thams cad (thams cad om. A) la yod pa'i phyir ro II des na gan iig de biin gsegs pa'i chos kyi sku las phyi rol tu gyur pa'i sems can de dag (am. A) ni ies gsuns so II Cf. Kano

2006: 157 and Mathes 2007: 29. 29 RGV I.145: dharmakiiyo dvidhii jfieyo dharmadhiituJ:! sunirmalaJ:! I
tanni~yandas
30

ca giimbhfryavaicitryanayadesanii II

See in Madhyiintavibhiiga I.l4: tathatii bhiitakotis ciinimittarrz paramiirthatii / dharmadhiitus ca paryiiyiiJ:! sunyatiiyiiJ:! samiisataJ:! II The equation of the dharmadhiitu with emptiness is not explicit in the RGV.

258

KazuoKano

completely pure state (i.e., emptiness realized by Buddhas), tathatii brings emptiness into the scope of both pure and impure states:31
As for the previous case [i.e., dharmakiiya] , since the intrinsic nature of the pure state itself exists [in Buddhas], the pure state is also present in sentient beings [for Buddhas' dharmakiiya pervades all sentient beings]. Therefore, this very purity [i.e.; emptiness]32 was called "the essence" (siUn po). As for the present case [i.e., tathatii], though [ordinary beings] have no [such final resultant state] which is accomplished by accumulation of purifications,33 only the emptiness of intrinsic nature is called "essence."34

As for the gotra, rNog states:

This cause [i.e., gotra] , again, is both emptiness and wholesome seeds. 35 31 rNog's understanding of tathatii as pertaining to both Buddhas and ordinary beings might find its source in the statement in RGVV ad I.148 that tathatii is luminous by nature and undifferentiated throughout both Buddha's and ordinary beings' states. 32 See rNog, rGyud bla don bsdus, A 41b1; B 26b2. 33 In his gloss on the dharmakiiya, rNog previously explained that the extremely pure tathatii has been accomplished on account of the immeasurable accumulations consisting in the Dharma of realization of the previous tathiigatas (rNog, rGyud bla don bsdus, A 41a6; B 26a2). 34 rNog, rGyud bla don bsdus, A 41b2-4; B 26a3-4: snar ni rnam par dag pa'i gnas skabs de fiid kyi ran biin yod pas I rnam par dag pa'i gnas skabs kyan sems can la yod pa yin la I des na rnam par dag pa de fiid silin por brjod pa yin no II da Ita ni rnam par dag pa'i tshogs las yan dag par grub pa de med kyan I ran biin gyis ston pa ilid kho na silin por brjod pa yin no II 35 rNog, rGyud bla don bsdus, A 42a1; B 26a7: rgyu yan ston pa ilid dan (om. B) I dge ba'i sa bon te I See also ibid., A 48a6-b2; B 29b8-30a2: rigs ni (om. A) gilis te I ran biin du gnas pa dan I bsgrubs pa las gyur pa'o II de la ran biin du gnas pa ni ston pa ilid yin te I de ni skyes bu'i byed pas ma bsgrubs par chos ilid du grub pa'i phyir ro II de'an (B: de la) sans rgyas kyi chos skyed (B: bskyed) pa'i nus pa dan ldan pas rigs su giag kyan de biin ilid (A: ilin) kho na yin no II des na bsgrubs pa las gyur pa dge ba'i bag chags ilid sans rgyas pa na yons su gyur nas gnas pa ni I Ions spyod rdzogs pa dan sprul pa yin pas na gter dan ljon sin gi dpes bstan te (A:

-rNog BIo ldan ses rab on the Buddha-nature doctrine

259

rNog sums up:

[Buddha-nature is called] gotra in view of its causal character, tathatii in view of its 'pervading character, and dharmakiiya in view of its resultant character. Mere emptiness pertains to all three. 36

Buddha-nature constituted by the three aspects is, in the end, nothing but emptiness - a conclusion that concords with rNog's definition of Buddha-nature in terms of "the mental continuum which has emptiness as its nature" (sems kyi rgyud ston pa fiid kyi ran
biin)Y
to) I yon tan chos kyis gian gyi dbul pa sel ba dan I rnam par grol pa'i 'bras bu gian la smin par byed pa'i phyir ro II "The gotra is two-fold: the prakrtistha ("inherent spiritual potential") and the samudiinfta ("generated spiritual potential"). Among them, the prakrtistha means emptiness, since it is not brought forth through the efforts of persons and is established by nature. Although it is regarded as a spiritual potential (gotra), since it is endowed with the potentiality (nus pa) to generate the Buddha's qualities, it is nothing but the tathatii [itself, i.e., emptiness]. Therefore, the samudiinfta, namely wholesome impressions which abide in the form of a transformed state after one has become awakened, is [called] the sambhogakiiya and the nirmiilJakiiya [in verse I.150]. Therefore, the two types [of spiritual potential] are illustrated through the analogies of a treasure and a [seed of a] tree, for the former dispels the poverty of others by means of its wholesome qualities, namely teaching, and the latter causes the maturation of fruitions, namely the release [from sal!lsiira], for others." In another passage (A 40b6-41al; B 25b5-6), rNog also refers to the two-fold gotra on the basis of RGVI.149. 36 rNog, rGyud bla don bsdus, A 42a2; B 26a7: des na rgyu'i mtshan fiid nas ni rigs yin la I khyab byed kyi mtshan fiid nas ni de biin fiid I 'bras bu'i mtshan fiid nas ni chos kyi sku ste I ston pa fiid 'ba' Zig rnam pa gsum gar yan 'gyur ro II 37 In the introductory part of his rGyud bla don bsdus, rNog explains the soteriological path according to two systems of the seven vajrapadas (see Kano 2006, chapter 4 for the details). In this context, he redefines Buddha-nature as the mental continuum that has emptiness as its nature. See rNog, rGyud bla don bsdus, A 4a5-6; B 2b5-6: yan dag pa'i Ita ba snon du btan nas thar pa'i don du (B: phyir) brtson pa tshogs kyi lam nas brtsams te (A: nas) byan chub sems dpa'i sa bdun pa'i bar du thob pa'i ran biin gi (am. A) dge ba'i tshogs kyis legs par bsgrubs pa'i sems kyi

260

KazuoKano

Another discussion pertaining to these three aspects concerns the question of the nominal (btags pa) versus the actual (drios) status of dharmakiiya, tathatii, and gotra, depending on whether ordinary beings or Buddhas are being considered. The nominal! actual distinction stems from the mention of "metaphorical designation" (upaciira, ner btags) in RGV I. 27c. The idea behind this is that gotra, insofar as it potentially allows sentient beings to become a Buddha, is a cause of Buddhahood. It can thus be metaphorically designated by the name of its effect. rNog adopts this distinction with regard to dharmakiiya and tathatii as well: the term dharmakiiya directly refers to the result - Buddhahood - and indirectly alludes to the causal faculty present in ordinary beings,38 while the term tathatii directly refers to both the resultant and causal aspects of Buddhas and ordinary beings. This distinction also applies in the context of identifying the "essence" (sniri po), a term already met with in the above passage on tathatii. 39 "Buddha's essence" (that is, the result itself) directly refers to the dharmakiiya (the resultant aspect of Buddha-nature), and indirectly alludes to gotra (the causal aspect).40 On the other hand, "sentient beings' essence" (sems can gyi sniri po, that is, the cause itself) directly refers to gotra. 41
rgyud ston pa nyid kyi ran biin ni I fie bar len pa yin te I khams so II 38 rNog previously restricted the presence of dharmakaya to Buddhas; however, so as not to contradict the teaching of RGV 1.27-28, according to which Buddha's dharmakiiya pervades every sentient being, rNog admits the nominal existence of dharmakiiya in ordinary beings. 39 Although Sanskrit dictionaries do not define garbha as "essence" (sfiin po), some Indian sources refer to such a meaning (Zimmermann 2002: 41). For instance, Ratnakarasanti equates vajragarbha with hrdaya and Jayaratha interprets garbha as sara (see Zimmermann 2002: 41, n. 58). Furthermore, gZon nu dpal refers to Sanskrit equivalents of sfiin po as sara, hrdaya, garbha, and mal}a (see his rGyud bla me lon, 262,24-25).

40 rNog, rGyud bla don bsdus, A 29a4-5; B 18b5-6.

41

rNog, ibid., A 29b1; B 18b7-19al.

rNog BIo ldan ses rab on the Buddha-nature doctrine

261

(ii) Emptiness as a cause

Among the three aspects of Buddha-nature dealt with above, the causal aspect (i.e., gotra), that is, the spiritual potential to become a Buddha, plays the most significant role. rNog's predicament in this regard is a Madhyamaka one: everything is empty by nature and there is no real entity on the ultimate level; how then can Buddha-nature be called a cause and how can it produce a result? Indeed, if we understand Buddha-nature as a real, causally efficient entity, then it could not be emptiness. rNog's solution is to accept it as an efficient cause, which is however a "conventional object" (tha sfiad kyi yul) devoid of real existence:
Therefore, the dhiitu [i.e., Buddha-nature], which is a conventional object pertaining to a non-affirming negation (prasajyaprati~edha), is called the efficient cause (upiidiina) of dharmakiiya, which is the conventional object pertaining to the non-affirming negation. It is, however, not the case that objects which must be produced and which produce [something] really exist. The term "conventional object" means something that is established as having a nature of a non-affirming negation, [namely,] something that does not exist in reality.42 42 rNog, rGyud bla don bsdus, A 4a2-3; B 2b3-4: des na med par dgag pa'i tha sfiad kyi yul du gyur pa'i khams ni med par dgag pa'i tha sfiad kyi yul du gyur pa'i chos kyi sku'i (chos kyi sku'i om. A) ner len du brjod kyi bskyed par bya ba dan skyed par byed pa 'i* don nid ni dnos su yod pa ma yin no II tha sfiad kyi yul ies bya ba ni I med par dgag pa ran biin du grub pa de kho nar med pa'i don to (B: te) II * bskyed par bya ba dan skyed par byed pa'i is my conjecture. Breads: bskyed par bya ba dan bskyed par byed pa'i; A reads: skyes bu byed pa'i. The reading of B, "it is not that objects which must be produced and which produce [something] really exist," is more likely than that of A, "it is not that the object of human activity really exists," since rNog asserts here that Buddha-nature or dhiitu (that is, emptiness) does not have an ontologically grounded causation but a merely nominal causation. Furthermore, this reading is supported by a similar expression found in Phywa pa's rGyud bla don gsal, 3b2: dnos myed dan dkon mchog gsum la rjes su 'gro ba'i ran biin skyed byed dan skyed bya myed kyan ldog pa tsam las rgyur tha sfiad gdags pa ste II

262

Kazuo Kana

rNog thereby accepts neither that dhatu and dharmakiiya should be entities in order to be causes or that the causation in question is a causation of the type that occurs between a seed and a sprout. Even though all phenomena are empty by nature, everything, is able to exist and function conventionally. In this sense Buddha-nature can fulfill a causal function. 43 rNog probably premises his position here on Mulamadhyamakakiirikii XXIV.l4:
Only when emptiness is applicable [to all phenomena], everything becomes applicable; if, on the other hand, emptiness were not applicable [to all phenomena], nothing would be applicable. 44

From another point of view, emptiness is a cause because one gradually attains enlightenment by meditating on emptiness:
Emptiness is the cause for the origination of the [resultant] pure state [i.e., enlightenment] if one takes it as an objective support [in a meditative state]. 45

(iii) Buddha-nature and iilayavijfiiina

When explaining gatra, the causal aspect of Buddha-nature, the RGVV quotes a verse from the Abhidharmasutra:
The beginningless dhiitu is the universal basis of all phenomena. Due to its presence, there arise all forms of existence as well as the attainment of nirviilJa.46
43 The RGV teaches Buddha-nature not only as a cause to become a Buddha but also as the universal foundation of sarrtsiira and nirviilJa (see below [iii]), and this latter aspect is relevant to emptiness's function that enables all phenomena (which include sarrtsiira and nirviilJa) to exist on the conventional level. 44 Millamadhyamakakiirikii XXIV.l4: sarvarrt ca yujyate tasya silnyatii yasya yujyate I sarvarrt na yujyate tasya silnyarrt na yasya yujyate II Ron ston associates rNog's position with Millamadhyamakakiirikii XXIV.l4 when explaining rGyud bla don bsdus A 4al; B 2b2-3. See Ron ston, rGyud bla legs Mad, 79,13-20 (he also refers to Vigrahavyiivartanf). 45 rNog, rGyud bla don bsdus, A 42a1-2; B 26a7: stan pa fiid kyan de la dmigs nas (B: pas) rnam par dag pa'i gnas skabs skye ba'i rgyu yin no II 46 RGVV 72,12-13: aniidikiiliko dhiituJ:t sarvadharmasamiiSrayaJ:t I

rNog BIo ldan ses rab on the Buddha-nature doctrine

263

After quoting this verse that teaches. that dhatu is the universal basis of saf!1-sara and nirvar.za, the RGVV identifies this dhiitu with Buddha-nature;47 One will remain in saf!1-sara as long as Buddhanature is covered with defilements, and will reach nirvalJa when Buddha-nature is purified. This same verse is quoted in Yogacara works, such as Asanga's Mahiiyanasaf!1-graha and Sthiramati's Trif!1-'ikabhii~ya. Those works identify dhatu as alayavijfiana: 48 seeds stored in alayavijfiana cause further existences in saf!1-sara, and the transformation of alayavijfiana (i.e., asrayaparivrtti) causes nirvalJa. rNog endorses the RGVV's identification of dhatu with Buddhanature, as well as the Yogacara identification with alayavijfiana as he states:
Such [dhatu] , which is endowed with various faculties [to generate everything that belongs to saJ'[lSara and nirvar.za], is also called alayavijfiana. 49

The identification of Buddha-nature with alayavijfiana also echoes what we find in the Lankavatarasutra and Ghanavyuhasutra, two sutras that were most probably composed after the RGVV: 50
tasmin sati gatiIJ sarva nirvar.zadhigamo 'pi ca // 47 RGVV 72,16-73,1: dhatur iti / yad aha / yo 'yal'{l bhagaval'{ls tathagatagarbho lokottaragarbhaJ:t prakrtiparisuddhagarbha iti / 48 Mahayanasal'{lgraha (ed. Nagao), I.1, and Tril'{lsikabha~ya (ed. Buescher), p. 116,1-2. 49 rNog, rGyud bla don bsdus, A 42a6; B 26b2-3: de 'dra ba nus pa sna tshogs pa can de la ni kun gii'i rnam par ses pa ies kyan brjod do. In the gloss on RGV I.1, too (rGyud bla don bsdus, A 7bl-4; B 4b5-7), rNog presupposes that dhatu is a foundation of sal'{lSara and nirvar.za and explains why the conventions/expressions of sal'{lsara and nirvar.za are applicable to dhatu which belongs to the ultimate level. This idea can be linked to rNog's understanding of Buddha-nature as a foundation that includes all the seven vajrapadas. See rNog, rGyud bla don bsdus, A 7a6; B 4b4-5: rdo rje'i gnas bdun po 'di mtha' (om. A) dag khams kyi ran biin du bsdu ba ni snar bSad zin pa fiid do // 50 For the source of this equation of Buddha-nature with alayavijfiana found in the Lankavatarasiltra, see Kano 2006, introduction. E.g.,

264

Kazuo Kano

rNog leaves unsolved the problem of how this identification is to be reconciled with his definition of Buddha-nature as emptiness. The notion of iilayavijfiiina indeed involves a level of reality which is hardly compatible with complete emptiness. This difficulty is probably the reason why rNog's successors, for instance, as will be discussed below, Phywa pa ehos kyi sen ge, rejected the identification with iilayavijfiiina.

(iv) The ontological status of the Buddha-qualities (interpretation of RGV 1.154-155)

Verses 1.154-155 of the RGV introduce the notion of emptiness within the system of the Buddha-nature doctrine as follows:
There is nothing to be removed from it [i.e., from Buddha-nature] and nothing to be added. The real should be seen as real, and seeing the real, one becomes liberated. (RGV I.l54)51 The dhiitu is empty of adventitious [stains], which have the defining characteristic of being separable; but it is not empty (asunya) of unsurpassable qualities, which have the defining characteristic of not being separable. (RGV I.l55)52

Lankiivatiirasutra VI (Takasaki ed. p. 6,1-2; Nanjio ed. p. 221,12-13): apariivrtte ca tathiigatagarbhasabdasaY[lsabdita iilayavijiiiine niisti saptiiniiY[l pravrttivijiiiiniiY[l nirodhaJ:t I The same notion is found in the Ghanavyuhasutra: P 778 cu 62bl-2: sa rnams sna tshogs kun gii ste II bde gsegs siiin po dge ba'an II siiin po de la kun gii'i sgras II de biin gsegs pa ston pa mdzad II siiin po kun giir bsgrags par yan II blo gian rnams kyis mi ses so II; T vol. 16, 747a; 775a. For the doctrinal background and development of the equation of Buddha-nature with iilayavijiiiina, see Takasaki 1982. Mathes (2007: 416-417) summarizes gZon nu dpal's and others' positions with regard to the equation of Buddha-nature and iilayavijiiiina. 51 RGVI.l54 parallels AbhisamayiilaY[lkiira Y.21. rNog, in his AbhisamayiilaY[lkiira commentary, explains that verses V.l8-21 teach the fruit of the darsanamiirga-stage and that Y.21 teaches the view of emptiness as a reasonable way for liberation. See rNog, mNon rtogs sgron me, A 86b3-5; B 34a8-9. 52 RGV I.l54-155: niipaneyam ataJ:t kiY[lcid upaneyarrz na kiY[lcana I

- rNog Blo ldan ses rab on the Buddha-nature doctrine

265

The RGVV explains:

There is no defiling factor (saJ'!!kle.animitta) to be removed from it, i.e., dhatu, which is naturally pure, because it is naturally devoid of adventitious stains. Nor does anything need to be added to it aS,a purifying factor (vyavadananimitta), because it is naturally endowed with pure qualities which are inseparable [from it].53

In sum, Buddha-nature is primordially void of adventitious stains but not void of Buddha-qualities. The RGV hence teaches that Buddha-nature is "not empty" (asiinya), insofar as it is not empty of Buddha-qualities. The question that is likely to arise for a Madhyamika interpreter - and is actually already prefigured in RGV I.156 (see below) - is how this affirmation of Buddha-qualities in the RGV can be reconciled with the Prajfiiipiiramitiisiitras' stance that "everything is empty by nature?" Let us consider rNog's interpretation of these two verses:
Neither superimposing the [unreal] ultimate existence of the objective support (dmigs pa) for defilements, nor denying the conventional existence of the objective support of mind and mental factors for purification, [the reality] remains in accordance with the two truths as they are .... In the way of neither superimposing [unreal] ultimate existence (paramiirthasat) nor denying conventional existence (saJ'!!vrtisat), the correct reality, namely the nature of the integration of the two truths, is thus presented [in verse 154].54
dra:jtavyaJ'!! bhatato bhataJ'!! bhatadarsf vimucyate II sanya agantukair dhatu/:! savinirbhagalak~alJai/:! I asiinyo 'nuttarair dharmair avinirbhagalak:jalJai/:! II 53 RGVV 76,5-7: na kiJ'!!cid apaneyam asty ata/:! prakrtiparisuddhiit tathagatadhiito/:! saJ'!!kle.animittam agantukamalasanyataprakrtitvad asya I napy atra kiJ'!!cid upaneyam asti vyavadananimittam avinirbhagasuddhadharmaprakrtitvat I 54 rNog, rGyud bla don bsdus, A 42b3-6; B 24b5-7: kun nas non mons pa'i dmigs pa don dam par yod par sgro mi 'dogs pa dan I rnam par byan ba'i (B: bar) sems dan sems las byun ba'i dmigs pa kun rdzob tu yod pa la skur ba mi 'debs pas I bden pa gnis ji Ita ba biin gnas pa ni / ... de ltar don dam par yod par sgro mi 'dogs pa dan / kun rdzob tu yod pa la skur ba mi 'debs pa'i tshul gyis yan dag pa nid phyin ci ma log pa bden pa gnis zun du 'breI ba'i ran biin bstan to II Cf. Kano 2006: 160 and Mathes 2007: 31.

266

Kazuo Kano

rNog's strategy is here to appeal to the distinction between the two truths. Insofar as perception of reality involves avoiding the denial of conventional existence (kun rdzob tu yod pa), the possibility remains for something to be "not empty." Where the RGVV explains that after removing wrong assertion and denial,55 there remains Buddha-nature which is empty of defilements and there is thus "something remaining" (avasi.Jta), rNog interprets this passage as saying that on the conventional level there remains the objective support for purification which is empty of the objective support for defilement on the ultimate level,56 rNog is thus resorting to a strategy often met with in Madhyamaka, namely, taking the view that everything is empty by nature to refer to the ultimate level, while admitting that existence on the conventional level is not empty. rNog's interpretation along this line thus succeeds in reconciling the view of Buddha-nature being "not empty" of Buddhaqualities with the "everything is empty"-standpoint, but leaves the problem of the status of the Buddha-qualities unresolved: in rNog's framework, the Buddha-qualities would have to be conventionally existent. However, the RGV rather teaches the Buddha-qualities as really existent (bhuta),57 and admits Buddha-nature endowed with the Buddha-qualities as unconditioned or as the ultimate. 58
55 RGVV76,11: samiiropiipaviidiintaparivarjaniid aviparftaY[La silnyatiianena slokadvayena paridfpitam / (a Schmithausen [1971: 159] corrected aparyantaY[L to aviparftaY[L) 56 See rNog, rGyud bla don bsdus, A 43b5-6; B 27a8-b1. 57 As pointed out by Schmithausen (1973: 134), the RGV restricts this "emptiness" to mundane factors which are accidental and unreal (RGV I.158), while claiming the true existence of the reality (bhilta) which is void of faults (RGV I.164cd). See also Schmithausen 1971: 160-161. Furthermore, RGV III.1-3 teach the Buddha-qualities, such as the Ten Powers, as belonging to Buddhas' ultimate body (paramiirthakiiya), and the characteristics of the Great One (mahiipuru~alak~a1Ja) as belonging to Buddhas' conventional body (saY[Lvrtikiiya). However, Phywa pa explicitly understands the Ten Powers, etc., as belonging to conventional truth. See Phywa pa, rGyud bla don gsal, 76a7. 58 RGV 1.53 admits Buddha-nature as unconditioned (asaY[Lskrta), and the RGVV (2,11-13) teaches that the expressions, "Buddha-nature," "the
lak~m:tam

 rNog Blo ldan ses rab on the Buddha-nature doctrine

267

rNog appears not to subscribe completely to the existence of the Buddha-qualities in Buddha-nature, for he says that all Buddha-qualities are "summoned as if called when one realizes the dharmadhiitu."59 What rNog implies here is that dharmadhiitu (i.e., Buddha-nature) is the cause for the attainment of the Buddhaqualities; the Buddha-qualities are thus not strictly speaking present in the Buddha-nature, but generated by it. In this way, rNog strives to avoid the non-buddhistic view of satkaryaviida (the position that admits the result existing in the cause).
(v) Relation between the emptiness doctrine and the Buddhanature doctrine (interpretation of RGV 1.156-167) RGV I.156-167, the last twelve verses of the first chapter, discuss the purpose of teaching Buddha-nature. RGV I.156 first introduces an objection:
Why did Buddhas teach here [i.e., in the Tathiigatagarbhasutra, etc.] that the dhatu of Buddha exists in every sentient being, after teaching here and there (tatra tatra) [i.e., in most sutras such as the Prajfiiipiiramitiisutras]60 that everything should be known to be empty in every respect, like clouds, [visions in a] dream and illusions?61

Replying to this objection, RGV I.157 teaches that the purpose in teaching Buddha-nature is for the sake of abandoning five faults:

ultimate (paramiirtha)," and "the sphere of sentient beings (sattvadhiitu)" are synonyms. 59 See rNog, rGyud bla don bsdus, A 33b2-3; B 21bl: don dam pa rtogs pa ni yon tan kun gyi rgyu yin te I chos kyi dbyins rtogs na sans rgyas kyi yon tan thams cad bos pa biin du 'du ba'i phyir ro II. Cf. Mathes 2007: 31. 60 As for the understanding of tatra tatra, I follow Vairocanarak~ita's gloss, tatra tatreti sutre (see Vairocanarak~ita, Mahiiyiinottaratantratippa~r, 14r1) and also a suggestion by Schmithausen (1971: 169): 'Menfalis ware es denkbar, daB tatra tatra zu ity uktvaiva'!t zu ziehen ist und bedeutet: in manchen [Mahayanasiitren]." 61 RGV 1.156: unya'!t sarva'!t sarvathii tatra tatra jfieya'!t meghasvapnamiiyiikrtiibham I ity uktvaiva'!t buddhadhatuJ:t punaJ:t ki'!t sattve sattve 'strti buddhair ihoktam II

268

KazuoKano

the depressed mind, contempt against those who are inferior, clinging to things unreal, speaking ill of truth, and affection for one's self.62 Further, RGV 1.158-159 teach that the claim "everything is empty in every respect" implies the absence of the defilements, etc., with regard to the real existence of Buddha-nature. Moreover, RGVI.160 hints to the superiority ofthe Buddha-nature doctrine to the emptiness doctrine. 63 rNog comments on RGVI.157 as follows:
[The reality which has] the characteristic [of emptiness] expressed there [in the Prajfiaparamitasutras] is regarded here [in the RGV] as a cause; by focusing on the characteristic [i.e., emptiness] taught there, "it [i.e., emptiness] is the cause for the arising of the Buddhaqualities." The intention [of teaching Buddha-nature] is "in order to eliminate five faults."64 In this way, the contradiction [between the teaching of Buddha-nature and that of the Prajfiaparamitasatras] is eliminated. 65

In order to bridge the gap between the emptiness teaching of the Prajfiiipiiramitiisutras and the "pragmatic function" ascribed to Buddha-nature in the RGV, rNog utilizes his understanding of

emptiness as a cause (see [ii]). Accordingly, rNog does not agree with the superiority ofthe Buddha-nature doctrine to the emptiness doctrine, as taught in RGV 1.157ff. Rather, he is seeking to adapt the Buddha-nature doctrine of the RGV to fit with the emptiness doctrine of the Madhyamaka system.

62 See RGV I.157: lfnal'{l cittal'{l hfnasattve~v avajfiiibhutagraho bhatadharmapavadalJ I atmasnehas cadhikalJ paficado~ii ye~al'{l te~al'{l tatprahalJartham uktam II 63 See note 3.
64 RGV I.160cd (commentarial verse on I.157): paficado~aprahalJaya dhatvastitval'{l prakasitam II 65 rNog, rGyud bla don bsdus, A 45a5-6; B 28a6: 'dir ni der bstan pa'i mtshan fiid de la dmigs nas [B: pas] sans rgyas kyi chos 'grub pa'i rgyu yin no ies der brjod pa'i mtshan fiid de fiid rgyur giag [B: btag] pa yin no II de'i dgons pa yan fies pa Ina span ba'i [B: span bar bya ba'i] phyir ro ies de ltar 'gal ba spon ba'o II

rNog BIo Idan ses rab on the Buddha-nature doctrine

269

The equation of Buddha-nature with emptiness stands at the core of rNog's' interpretation of the RGV. One can wonder what the background could be for this idea; could it have its origin in Indian sources? Two among the three aspects of Buddha-nature, dharmakaya and tathata, are identified with emptiness already in early works, such as the Madhyantavibhaga. However, these works do not discuss Buddha-nature, and do not seem to have had an impact on rNog on this account. The writings of Indian Madhyamikas tum out more likely to provide a source of influence. The first Madhyamaka master who discussed Buddha-nature was Bhaviveka (assuming he was indeed the author of the Tarkajvala), who combined Buddha-nature with the Three Deliverance-doors (trimok~amukha). He asserts that "the teaching that all sentient beings have Buddha-nature implies that emptiness, signlessness, and wishlessness, etc., abide in the mind of all sentient beings; this teaching does not refer to the inherent (anta/:tkaraT}a), eternal puru~a (i.e., atman) pervading everything.''66 KamalasUa further develops Bhaviveka's idea in his Madhyamakaloka:
[The teaching that] "all sentient beings have Buddha-nature" refers to the fact that all [sentient beings] are able to attain the state of perfect awakening; this is because it is held that the expression "tathiigata" [being the first member of the compound tathiigatagarbha] refers to the sphere of reality (dharmadhiitu) which is characterized by the

66 TarkajviiIii VI, D 69a1-3: de biin gsegs pa'i sfiin po can ies bya ba yan ston pa iiid dan I mtshan ma med pa dan I smon pa med pa Ia sogs pa rnams sems can thams cad kyi rgyud Ia yod pa'i phyir yin gyi I nan gi byed pa'i skyes bu rtag pa thams cad du khyab pa Ita bu ni ma yin te I ji skad du I chos thams cad ni ston pa fiid dan I mtshan ma med pa dan I smon pa med pa'i no bo I gan ston pa fiid dan I mtshan ma dan I smon pa med pa de ni de biin gsegs pa'o ies bya ba Ia sogs pa 'byun bas so II Bhaviveka probably premises his interpretation on a passage in the Lankiivatiirasiitra (Nanjio ed., pp. 77,13-79,9). Cf. Takasaki 1980: 232-236 and Eckel 2008: 154-155.

270

KazuoKano

selflessness of person and phenomena, namely natural luminosity (Prakrtiprabhiisvara).67

In this statement, KamalaSIla proposes to take the word tathagatagarbha in the sense of *dharmadhatugarbha 68 and interprets the teaching that "all sentient beings have Buddha-nature" in the sense that all sentient beings are pervaded by dharmadhatu, which is characterized by selflessness. In other words, he understands Buddha-nature as selflessness. Even though rNog does not explicitly acknowledge KamalasI1a's pioneer integration, it is likely that such an integration of the Buddha-nature teaching into the Madhyamaka position had an impact on rNog's interpretation. 69 Precise statements of the equation of Buddha-nature with emptiness are found in works by Jayananda and Abhayakaragupta; whether, being younger contemporaries of rNog, they could have had an impact on the latter's position is difficult to ascertain.7
67 Kamalasila, Madhyamakaloka, D 242b4-7: sems can thams cad ni de biin giegs pa'i sfiin po can no ies bya ba 'dis kyan I thams cad bla na med pa yan dag par rdzogs pa'i byan chub kyi go 'phan thob par run ba fiid du yons su bstan te I de biin giegs pa'i sgra ni chos kyi dbyins gan zag dan chos la bdag med pa'i mtshan fiid ran biin gyis 'od gsal ba yin par brjod par bied pa'i phyir ro II 68 This term is found in the Srfmiiliidevfsutra, and quoted in the RGVV. See RGVV 76,16-18: sa khalv e~a tathiigatagarbho yatha dharmadhatugarbhas tatha satkiiyadr~tipatitiiniim agocara ity uktal!! dr~ti pratipak~atviid dharmadhatof; I As regards the historical and doctrinal background of Kamalasila's position on Buddha-nature, see Kano 2006, introduction. 69 According to Seyfort Ruegg (1981: 95, n. 308), KamalaSila was "perhaps the first of the leading Madhyamaka masters to incorporate the theory of the tathiigatagarbha into the Yogacara-Madhyamaka school." 70 As regards Abhayakaragupta's position, Seyfort Ruegg (1977: 302) noted that: "In Abhayakaragupta's comments it is especially interesting to find the assimilation of the tathiigatagarbha with the prakrtisthagotra whose nature is dharmadhiitu, in other words non-substantiality (nairiitmya) of both pudgala and dharmas, absence of own being (nif;svabhiivatii) and Emptiness." For the details of Abhayakaragupta's discussion, see Kano 2006.

rNog BIo ldan ses rab on the Buddha-nature doctrine

271

rNog's followers in the early bKa' gdams pa tradition We do not know much about the views of the two of rNog's "four main [spiritual] sons" (sras kyi thu bo bii), Zan tshes spon ba and Gro lun pa, who authored lost commentaries on RGY. We can get a glimpse of Gro lun pa's position on Buddha-nature from some discussions occurring in his voluminous bsTan rim 71 where, in addition to an extensive general discussion on gotra in the second chapter,72
Jayananda equates Buddha-nature with emptiness in Madhyamakilvatilratfkil, D 211b5: sems can thams cad kyi lus kyi nan na mchis par brjod de ies bya ba ni ston pa fiid kyis thams cad la khyab pa'i phyir ro II; ibid., D 212b2-3; ibid., D 213a4-5. Whether Jayananda had direct contact with rNog is unclear (Jayananda flourished slightly later, for he is said to have debated with Phywa pa, who was born the year of rNog's death, 1109), but it is most likely that rNog received a Kashmirian tradition of Madhyamaka during his stay there (sometime between 1076-1092) and integrated that system with his RGV commentary. On the other hand, it is noteworthy that we cannot trace rNog's interpretation of emptiness in the Mahilyilnottaratantrasilstropadea of Sajjana, who directly transmitted the teaching of the RGV to rNog. In the marginal note on verse 28 by an anonymous hand found in the manuscript of the Mahilyilnottaratantrasilstropadea, it is said that the conditioned (sa1!lskrta) is empty, but the luminous mind is not empty. This is incompatible with rNog's interpretation. See Kano 2006, chapter 4. Note also that JfianasrYmitra alludes to the identity of Buddha-nature and emptiness in the Silkilrasiddhi (ed. Thakur, p. 487,22-23). 71 The current paper does not take into consideration other works by Gro lun pa which became available recently in the bKa' gdams gsun 'bum: 'Phags pa yon tan rin po che sdud pa tshigs su bead pa rnam par Mad pa (vol. 3, pp. 339-574); brGyad ston 'grel chen gyi Mad pa (vol. 3, pp. 579-746); Rin po che'i phren ba lria bcu pa (vol. 4, pp. 3-6; I am currently preparing a critical edition and annotated translation of this work); Sans rgyas kyi bstan pa la rim gyis 'jug pa'i tshul (vol. 3, pp. 243-321); rGyal ba'i yum bar ma le'u brgyad pa iii khri lria stori pa mdo' 'grel rnam par Mad pa (vol. 4, pp. 331-602). For a preliminary study of the bsTan rim, see Jackson 1996. 72 This chapter is entitled "How to cultivate an awareness of the value of a human life that is free from the conditions that obscure the practice of religion" (37a-47a). Gro lun pa does not focus on the gotra of the

272

KazuoKano

a small section of the eighth chapter is devoted to Buddha-nature.?3 In this context, Gro lUll pa follows faithfully rNog's interpretation as found in the latter's gloss on RGV I.27-28 - the two verses that teach dharmakiiya, tathatii and gotra as being three reasons why all sentient beings possess Buddha-nature.?4 Gro lUll paappears elsewhere in the same text to endorse rNog's idea of tathatii as emptiness,75 and follows rNog's position with regard to the ineffability of the ultimate.?6 Although Gro lUll pa would thus appear to stand as a faithful heir of his teacher, a conclusive determination will have to await further studies and the hopeful surfacing of his RGV commentary. The earliest post-rNog RGV commentary currently available is that by Phywa pa Chos kyi sell ge, better known to most modern scholars for his contribution to the Tibetan tshad ma tradition. According to the colophon of this text, Phywa pa received the teaching of the RGV from Chos kyi bla ma, who is probably to be
RGV's system, but deals with gotra in general. After presenting some general remarks on gotra, Gro lun pa explains the two-fold gotra of the Mahayana. He takes the gotra of the ultimate level as a synonym for dharmata and that of the conventional (vyavahara) level as being constituted of both the prakrtisthagotra (as in the one-vehicle [ekayana] theory) and the samudanftagotra. The latter is the basis for practices and is activated by the wholesome roots at the mok~abhagfya-stage. See Gro lun pa, bsTan rim, 39bl-4. 73 This section is entitled: "The teaching of the [nature of the mind] as being Buddha-nature" (de fiid de biin gsegs pa'i sfiin por bstan pa) (Gro lun pa, bsTan rim, 345a-447a). It is contained in the eighth chapter, "How to cultivate meditatively a realization of ultimate reality." 74 Gro lun pa does not refer to rNog by name, but the source of his explan<j.tions can be identified as rNog, rGyud bla don bsdus, A 29a3-b2; B 18b5-19al and A 41a3-42a2; B 25b7-26a7. Gro lun pa adopts rNog's understanding of dharmakaya, tathatii, and gotra as nominal (btags pa) or actual (dnos) with regard to ordinary beings and Buddhas, respectively (see above riD. See Gro lun pa, bsTan rim, 371a3-5. 75 Gro lun pa, bsTan rim, 68b3: chos thams cad kyi de biin fiid ni ran biin med pas ci'an mi dmigs pa'o II 76 See Cabezon (forthcoming).

rNog BIo ldan ses rab on the Buddha-nature doctrine

273

identified as rNog's "spiritual son" Zan tshes spon ba Chos kyi bla rna. Phywa pa's debt to rNog's (and/or maybe to Chos kyi bla rna's) RGV commentary is clearly visible, as Phywa pa's text mostly follows the outline found in rNog's commentary and adoptsrNog's interpretation of the major issues that were dealt with above, such as the equation of Buddha-nature with emptiness (cf. [i]),77 the causal function of emptiness (cf. [ii]),78 the three aspects of Buddha-nature (cf. [i])79 and the reconciliation of the teachings in the Prajfiaparamitasutras and the Tathagatagarbhasutra (cf. [V]).80 77 Phywa pa defines Buddha-nature as "the reality of the mind, that is, emptiness" (rGyud bla don gsal, 3b8: gfiis pa ni tshogs kyi lam nas brtsams te sa bdun pa'i bar gyi dge ba'i chos kyis rigs kyi dri ma bsal ba'i sems kyi chos fiid ston pa fiid ni khams te fie bar len pa'o II). See also Phywa pa, rGyud bla don gsal, 62al-2: des na 'bras bu'i sfiin po fie bar btags nas yod pa'am thams cad kyi khyab byed ston pa fiid kyi sfiin po dnos su yod pa'am rgyu'i sfiin po ma btags par yod pa la dgons nas rnam pa gsum du bsfiad pa yin no II "Therefore, [sentient beings] metaphorically possess the resultant essence [i.e., dharmakiiya]; they really possess the essence of emptiness [i.e., tathatii] that pervades everything; or they possess the causal essence [Le., gotra], not metaphorically. Having these in mind, [the RGy] explained that [Buddha-nature] has three aspects." 78 See Phywa pa, rGyud bla don gsal, 3b5, 61b4, and 61b8-62al. Phywa pa also discusses the question of how emptiness can be a cause. His initial answer is that emptiness is conventionally called "cause" from the viewpoint of "mere exclusion" (ldog pa tsam), although it does not have a real efficiency: "just as the absence of food is a cause for hunger, so the absence of impurities that have ultimate nature (don dam pa'i ran btin gyi dri ma) is the cause for Buddhahood, and this very absence is called 'cause'." (rGyud bla don gsal, 3b2-3). In another context (Phywa pa, rGyud bla don gsal, 61b4-5), Phywa pa describes emptiness as a metaphorical cause to attain Buddha's wisdom (or the nirvikalpajfiiina) insofar as it is the object of the meditation leading to the attainment of that wisdom; it is, however, the meditating mind of the practitioner that is the "real cause." 79 Phywa pa, rGyud bla don gsal, 34a7 and 61a3 (on dharmakiiya), 34a7-8 and 61a7-8 (on tathatii), 34bl and 62al (on gotra). 80 Phywa pa, rGyud bla don gsal, 67a7-8. There, Phywa pa explicitly

274

Kazuo Kano

Phywa pa's agreement with rNog's views is, however, not complete. A noticable point of divergence occurs, for instance, with regard to the relationship between Buddha-nature and iilayavijiiiina. When commenting on the notion of "dhiitu" occurring in the sense of "Buddha-nature" in the verse from the Abhidharmasutra quoted in the RGVV on RGV 1.149-152, rNog identified this dhiitu with iilayavijiiiina (see above [iii]). Phywa pa, on the other hand, claims that it consists of both (a) emptiness and (b) wholesome seeds, which respectively correspond to prakrtisthagotra and samudiinftagotra. Buddha-nature is thus the foundation of nirviilJa, for nirviilJa is caused (a) by correct understanding of emptiness and (b) by wholesome seeds. On the other hand, Buddha-nature is also the foundation of sa1'(lsiira, for sa1'(lsiira is caused (a) by incorrect understanding of emptiness and (b) by defilements. The latter do not arise from the wholesome seeds themselves but from seeds of defilements stored in the iilayavijiiiina. 81 By avoiding the exclusive identification of Buddha-nature with iilayavijiiiina, Phywa pa seems to be aiming at avoiding both the confiation of iilayavijiiiina with emptiness and the attribution of an ultimate status to iilayavijiiiina. Phywa pa's originality as a thinker is also notable in his discussion of issues not adressed in detail by rNog. A noteworthy discussion in this regard concerns the Bodhisattva's understanding of emptiness. The RGVV divides Bodhisattvas who newly departed on the Mahayana into two categories: (a) those who misunderstand emptiness as mere cessation and (b) those who misunderstand emptiness as something existing (as a real entity).82 While rNog
states that the Buddha-nature doctrine taught in the Tathagatagarbhasutra is provisional (dgons pa can) and is not to be taken literally (sgra ji biin rna yin). Phywa pa thus considers the doctrine of emptiness as primary, leaving ,a secondary role to the Buddha-nature doctrine. 81 See Phywa pa, rGyud bla don gsal, 62a8-b4, and in particular 62a8b1: bsgrubs pa las byun pa'i rigs thar pa'i cha 'thun gyi dge ba'i sa bon dnos las kun nas non mons mi 'byun yan de'i rten kun gii la non mons pa'i sa bon bsgoms las kun nas non mons 'byun bas de'i rten du brjod de kun gii ni nus pa sna tshogs pa'i phyir TO II 82 The RGVV presents the former (a) as: "[Bodhisattvas] who assume that the deliverance-door consisting in emptiness leads to the de-

rNog BIo Idan ses rab on the Buddha-nature doctrine

275

merely offers a summary of this passage,83 Phywa pa extensively discusses both positions. He identifies the first as the attitude of Mind-only adherents, Yogacara-oriented Bodhisattvas, for whom "cessation" is the cessation of paratantra, a process that is also identified with "nirvii"(ta," conceived of as "immaculate purity," i.e., parini~panna. 84 As for the second wrong position (b), he considers it
struction of something existing, saying that parinirvaf.Za is a cessation or destruction throughout future time of a dharma that exists." (RGVV 75,14-15: ye bhavavinasaya siinyatavimok~amukham icchanti sata eva dharmasyottarakalam ucchedo vinasaJ:t parinirvalJ-am iti I) The RGVV presents the latter (b) as: "[Bodhisattvas] who hang on to emptiness by taking emptiness as an object saying, 'we shall attain and realize a certain entity called emptiness existing separately from visible matter (riipa), etc.''' (RGVV 75,15-17: ye va punaJ:t siinyatopalambhena siinyataY(l pratisaranti siinyata nama riipadivyatirekef.Za kascid bhavo 'sti yam adhigami~yamo bhavayi~yama iti I) According to RGVV 74,5-6, they are those whose minds have deviated from emptiness (siinyatavik~iptacittanam) and cannot perceive Buddha-nature. 83 See r:Nog, rGyud bla don bsdus, A 42bl-2; B 26b4. 84 Phywa pa, rGyud bla don gsal, 63b5-7: ston pa fiid kyi tshul de biin gsegs pa'i sfiin po ni dgag bya rnam par gcad pa tsam gyis ston pa'o II de las fiams pa yO/is gcod du Ita ba'o II gan dag ies pa ni sems tsam pa gan dag go II yod pa'i chos fiid dus phyis rgyun chad pa ni ma dag pa'i gian dban don dam (don dam: in the bottom margin) du yod pa phyis ste gsal bas (em.; MS: ba'i) iig nas myed pa'i ston fiid do II de mya nan las 'das par 'dod pa ni dri ma myed pa'i rnam dag du khas len pa'o II de rnam par thar pa'i sgor 'dod pa don dam par khas len pa'o II "Buddha-nature which has emptiness as its mode means being empty in the sense of mere elimination, which eliminates objects which are to be negated. Those who have fallen from this [correct view] are [called] those who [incorrectly] see [emptiness] as elimination [which eliminates anything whatsoever]. '[Bodhisattvas] who .. .' means [Bodhisattvas] who belong to the Mindonly school (sems tsam pa). [The words] 'destruction throughout future time of a dharma that exists' mean emptiness, i.e., the impure paratantra that exists on the ultimate level ceases by means of rejection and becomes a non-entity. [The words] 'they admit this as nirvalJ-a' mean: [the Mindonly school] accepts [this] as immaculate purity [i.e.,parini~panna]. [The words] 'to assume this as the deliverance-door' mean: to accept [this] as ultimate." Note that rGyal tshab (rGyud bla ([ka, 160a1-3) identifies in-

276

Kazuo Kano

to be the position of those who identify emptiness as self-awareness (rali rig) and who accept emptiness from the viewpoint of affirming negation. 85 The identification of the first category of mistaken Bodhisattvas as Yogacara is significant, as it implies that the RGVV is a work that rejects the Yogacara position. As in the previous discussion, here again Phywa pa's preference for the Madhyamaka view appears to direct his choice of interpretation. Early commentaries such as those of rNog and Phywa pa were certainly influential on later compositions of the gSan phu tradition, although the few available sources display abundant disagreement on the key issues dealt with above. For instance, in the RGV commentary of gSan phu pa BIo gros mtshuns med,86 written in the 14th century, Phywa pa (referred to as "some great teacher"81) is repeatedly criticized. 88
stead the second wrong view as that of the Yogacara. 85 Phywa pa, rGyud bla don gsal, 63b7-8: gcig la gcig myed pa khas len pa ni ymi gmi dag ies pa ste I rtogs par bya ba ni 'bras bu'i dus na'o II bsgom par bya ba ni lam gyi dus na'o II gzugs la sogs pa ni gZUli 'dzin giiis so II de las tha dad pa'i dfws po ni des dben pa'i sems ran rig rna yin (rna yin in the bottom margin) dgag gian dban don dam pa'i ston pa'o II ston pa iiid la rten ies pa ni iii tshe ba'i ston par khas len pa'o II 86 gSan phu pa Blo gros mtshuns med (ca. 14th century) was an assistant teacher (zur chos pa) of gSan phu Glin smad bGros riiin monastic college (grwa tshang). His teacher was 'Jam dbyans Sakya gion nu and his disciples included Karma pa Ran 'byun rdo rje (1284-1338), Grags pa sen ge (1283-1364) and g.Yag sde Pal) chen (1299-1378). He composed extensive commentaries on the RGV and Bodhisattvacaryavatara, and independent works, such as the bsNo ba yon bad kyi ston thun and Theg pa rin po che gsal byed rin po che'i rgyan snan (all included in the bKa' gdamsgsuns 'bum, vols. 15-17). See also Kano 2006, chapter 5 and appendix E; Mathes 2008: 91-98. 87 Blo gros mtshuns med calls Phywa pa by name (rigs pa'i dban phyug phywa pa'i ial nas) on a single occasion: rGyud bla sgron me, A 201,5; B 342. See Kano 2003. 88 Blo gros mtshungs med's criticism of Phywa pa's position is found for instance with regard to the understanding of RGV 1.154 (Blo gros mtshuns med, rGyud bla sgron me, A 369,1-6; B 421-422), the causal

rNog BIo Idan ses rab on the Buddha-nature doctrine

277

Conclusion
The RGV itself-neither equates Buddha-nature (or dharmakiiya, the resultant aspect of Buddha-nature) with iitman nor with emptiness. It thereby implies neither that Buddha-nature is an entity nor a nonentity. rNog for his part considers the RGV as a Madhyamaka work, and its teaching as definitive. This position is made clear also in the introductory passage of the rGyud bla don bsdus, where the RGV is given as a treatise that explains sutras of definitive meaning (nftiirtha), whereas the other four treatises of Maitreya (i.e., Abhisamayiilarrtkiira, Mahiiyiinasutriilarrtkiira, Madhyiintavibhiiga, and Dharmadharmatiivibhiiga) are listed as treatises that explain sutras of provisional meaning (neyiirtha).89 Such a reading, however, creates a tension with regard to the RGV's teaching found in verses I.154-155 that stresses the real existence of the Buddhaqualities which abide in Buddha-nature: Buddha-nature is only "empty of defilements," but is not empty of the Buddha-qualities. As we have seen, rNog's strategy to solve the contradiction between the existence of Buddha-qualities taught in the RGV and the emptiness of the Prajiiiipiiramitiisutras by appealing to the Two Truths (satyadvaya) fails to account for the status of "real existence" that the RGV ascribes to Buddha-qualities. Thus, albeit taking the RGV as a teaching of definitive meaning, rNog fails to take such statements of the RGV literally, in order to remain faithful to Madhyamaka. One can see here a local illustration of his generalized policy, which aims at systematizing the whole of

nature of the prakrtisthagotra (ibid., A 358,6-360,2; B 416-417), and the provisional meaning of the Tathiigatagarbhasutra teaching (ibid., A 378,6-382,2; B 426-428. Cf. Phywa pa, rGyud bla don gsal, 67a7-bl). 89 See rNog, rGyud bla don bsdus, A 1b2-4; Blbl-3. It is surprising that he classifies the Abhisamayiilaf!Lkara within the latter group, thus implying that the Prajfiiipiiramitiisutras that it comments on are provisional. This point is criticized by Blo gros mtshuJis med (rGyud bla sgron me, A 13,1-4, B 246-247).

278

KazuoKano

Indian Buddhist doctrines as constituting elements of a global Madhyamaka philosophy.90 . rNog has to face another tension as he, on the one hand, identifies Buddha-nature with emptiness and, on the other hand, presents Buddha-nature as the cause for becoming a Buddha. His rationale for emptiness functioning as a "cause" is repeatedly quoted in the later literature, although it is variously evalutated.91 Although Phywa pa mostly follows rNog's position, he judges the Buddha-nature doctrine to be a non-literal teaching (ayathiiruta), and his judgement might have had an influence on Sakya Pa:r;lC;lita who, referring to RGV 1.156-157, judges the teaching of the RGV to be provisional (neyiirtha).92 Sa skya PaJ).c;lita's position was criticized by BIo gros mtshuIis med, who, like Dol po pa, took the teaching of the RGV to be definitive. 93 Thus, the Buddha-nature doctrine 90 On the question of "what kind of Madhyamaka" rNog subscribed to, the later Tibetan tradition and modern scholars usually classify him as a "Svatantrika Madhyamaka." The division between Svatantrika and PrasaIigika itself is however not attested in such terms in rNog's own works. rNog prescribes a reading of Nagarjuna that integrates the agenda of the epistemological school of Dharmaklrti. In his sPrins yig bdud rtsi'i thig Ie verse 10, rNog says that one should understand the emptiness doctrine taught in Nagarjuna's works on the basis of Dharmaklrti's "beautiful philosophy" (chos rnams ran biin ston pa'i tshulla 'jug pa'i sgo II yan dag rigs tshogs klu sgrub ial nas gsuns pa de II rigs pa'i dban phyug rnam 'grel mdzad pa'i giun mdzes las II gsal bar rtogs nas lugs nan gian kun rtswa biin dor /I). See Kano 2007 and 2009. 91 See Kano 2006, chapters 5 and 6.
92 See Sa skya PaI).<;lita, sDom gsum rab dbye, I.139-140: sems can rnam la sans rgyas kyi II sfiin po yod par gsuns pa ni II dgons pa yin par sef par bya II de yi dgons gii ston fiid yin II Sa skya PaI).<;iita thinks that if a teaching involves an intention (dgons pa) which is not literally expressed, then this teaching is a provisional one (drans don, neylirtha) since it needs to be interpreted non-literally. 93. BIo gros mtshuIis med took both the Prajfilipliramitli and the Buddha-nature doctrine as definitive teaching (nftlirtha), and considered that the statement of a purpose (prayojana) does not make a teaching non-literal (see rGyud bla sgron me, A 379,6-380,1, B 426-427: dgos pa

rNog Blo ldan ses nib on the Buddha-nature doctrine

279

was accepted differently from the viewpoint of each school, and interpreted in ways that fitted their respective doctrinal systems. The interpretation of the RGV is symptomatic of the kind of issues Tibetan Buddhists had to confront as they imported" almost simultaneously, Indian doctrines that, due to their gradual development in various periods and areas in India, sometimes contradict each other. Tibetan scholars such as rNog strove to construct a consistent philosophical system out of this disparate material. In this regard, rNog played a pioneering role in the introduction of Indian Buddhist philosophical traditions to Tibet in the "Later Dissemination" (phyi dar) period, as he aimed at a systematization of Indian Buddhist philosophy centered on Madhyamaka. Bibliography
Sanskrit and Tibetan works
bKa' gdams chos 'byun gsal ba'i sgron me. Las chen Kun dga' rgyal mtshan. Lhasa: Bod ljons mi dmans dpe skrun khaIi, 2003. bKa' gdams gsun 'bum = bKa' gdams gsun 'bum phyogs bsgrigs. Ed. dPal brtsegs bod yig dpe rfiln fib 'jug khaIi. Chengdu: Si khron mi rigs dpe skrun khaIi, 2006 (vols. 1-30) and 2007 (vols. 31-60). 'Bras spuns dkar chag = 'Bras spuns dgon du biugs su gsol ba'i dpe min dkar chag; Ed. dPal brtsegs bod yig dpe rfiiIi fib 'jug khan. Beijing: Mi rigs dpe skrun khaIi, 2004. bsTan rim =bDe bar gsegs pa'i bstan pa rin po che la 'jug pa'i lam gyi rim pa rnam par bSad pa. Gro lun pa Blo gros 'byuIi gnas. A: xylograph, 548 folios, preserved in the Bihar Research Society, Patna (cf. Jackson, D. "The "Miscellaneous Series" of Tibetan Texts in the Bihar Research Society, Patna: A Handlist," pp. 164-165. Tibetan and Indo-Tibetan Studies, vol. 2. Stuttgart: Franz Steiner Verlag, 1989); B: In bKa' gdams gsun 'bum, vol. 4, 35-735 and vol. 5, pp. 3-242. D Derge Kanjur and Tanjur. Deb ther snon po. 'Gos 10 tsii ba gZon nu dpal. Chengdu: Si khron mi rigs dpe skrun khaIi, 1984.

brjod pas sgra ji btin pa ma yin par 'gyur na bde bar gsegs pa'i gsun rabs la dgos pa med pa gan zig yod).

280

KazuoKano

Lanko.vato.rasatra. (a) The Lanko.vato.ra satra. Ed. B. Nanjio. Kyoto: The Otani University Press, 1923; (b) A Revised Edition of the Lanko.vato.rasatra: K~a1}ika-parivarta. Ed. J. Takasaki. Tokyo, 1981. Madhyamakahrdaya. Bhaviveka. Ed. Ch. Lindtner. Madhyamakahrdayam of Bhavya. Madras: Adyar Library and Research Center, 2003 .. Madhyamako.loka. KamalasIla. D 3887; P 5287. Madhyamako.vato.ratfko.. Jayananda. D 3870; P 5271. Madhyo.ntavibhaga/bha~ya (attrib. to Vasubandhu). Ed. G. Nagao. Tokyo: Suzuki Research Foundation, 1964. Maho.yo.nasa'!lgraha. Tr. and ed. G. Nagao. ShiJdaijiJron. 2 vols. Tokyo: K6dansha, 1982/87. Maho.yo.nottaratantraso.stropadesa. S ajj ana. Crit. ed. in Kano 2006, appendixB. Maho.yo.nottaratantratippal}-f. Vairocanarak~ita. Crit. ed. in Kano 2006, appendix C. mNon rtogs sgron me =mNon par rtogs pa'i rgyan 'grel rin po che'i sgron me. rNog BIo Idan ses rab. A: Lotsaba'i bsdus don by Rnog Lotsaba Blo ldan ses rab. Dharamsala: Library of Tibetan Works and Archives, 1993; B: In bKa' gdams gsun 'bum, vol. 1, pp. 125-201. Malamadhyamakako.riko.. Nagarjuna. Ed. L. de La Vallee Poussin. Malamadhyamakako.riko.s (Madhyamikasatras) de No.go.rjuna avec la Prasannapado., Commentaire de Candrakfrti. Bibliotheca Buddhica IV. St. Petersburg: Russian Academy of Sciences, 1903-13 (repr. Osnabrock, 1970). P Peking Kanjur and Tanjur. RGV/V = Ratnagotravibhaga/-vyo.khyo.. Ed. E. Johnston. Patna: The Bihar Research Society, 1950. rGyud bla don bsdus = Theg pa chen po rgyud bla ma'i bstan bcos kyi don bsdus pa. rNog BIo Idan ses rab. A: Theg chen rgyud bla ma'i don bsdus pa, Commentary on the Ratnagotravibhaga by Rnog Lotsaba Blo ldan ses rab. Dharamsala: Library of Tibetan Works and Archives, 1993; B: In bKa' gdams gsun 'bum, vol. 1, pp. 289-369. rGyud bla don gsal = Theg pa chen po rgyud bla ma'i bstan bcos kyi tshig dan don gyi cha rgya cher bsfiad pa phra ba'i don gsal ba. Phywa pa Chos kyi sen ge. In bKa' gdams gsun 'bum, vol. 7, pp. 163-345. rGyud bla legs Mad = Theg pa chen po rgyud bla ma'i bstan bcos legs par Mad pa. Ron ston Sakya rgyal mtshan. Chengdu: Si khron mi rigs dpe skrun khan, 1997. rGyud bla me Ion = Theg pa chen po rgyud bla ma'i bstan bcos kyi 'grel Mad de kho na fiid rab tu gsal ba'i me Ion. 'Gos 10 tsa ba gion nu dpal. Ed. K.-

 rNog BIo ldan ses rah on the Buddha-nature doctrine

281

D. Mathes. Nepal Research Centre Publications, no. 24. Stuttgart: Franz Steiner Verlag, 2003. rGyud bla sgron me = Theg pa chen po rgyud bla ma'i bstan bcos kyi nes don gsal bar byed pa'i rin po che'i sgron me. gSaD. phu pa BIo gros mtshuns med. A: Tezu: Tibetan Nyingmapa Monastery, 1974 (repr. bKa' .gdams gsun 'bum, vol. 17); B: In The Collection Works of the Ancient Sa-skya-pa Scholars, vol. 3, 239-565. Kathmandu: Sakya College, 1999. rGyud bla !fka = Theg pa chen po rgyud bla ma'i {fka (or Theg pa chen po'i nes don gyi mdo sde'i dgons pa 'grel par byed pa). rGyal tshab Dar ma rin chen. Collected works (2:01 edition). Vol. ga, 1a-230a3. New Delhi: Mongolian lama guru deva, 1982. sDom gsum rab dbye. Sa skya PaI).gita Kun dga' rgyal mtshan. Ed. J. Rhoton. A Clear Differentiation of the Three Codes: Essential Differentiations among the Individual Liberation, Great Vehicle, and Tantric Systems. New York: State University of New York Press, 2002. T TaishO Sinshii Daizokyo. Tarkajviilii Madhyamakahrdayavrtti. Bhiiviveka (?). D 3856; P 5256. Tho yig = dPe rgyun dkon pa 'ga' zig gi tho yig don giier yid kyi kunda bzad pa'i zla 'od 'bu gyi siie mao A khu Chili Ses rab rgya mtsho. Ed. L. Chandra. Materials for a History of Tibetan Literature. Part 3. SataPitaka Series, vol. 30, pp. 503-601. New Delhi: International Academy of Indian Culture, 1963 (repr. Kyoto: Rinsen Books, 1981). Trif!!sikiibhii-rJa. Sthiramati. Ed. H. Buescher. Sthiramati's Trif!!sikiivijiiaptibhii-rJa: Critical Editions of the Sanskrit Text and its Tibetan Translation. Vienna: Osterreichische Akademie der Wissenschaften, 2007. Zu chen thob yig = dPal ldan bla ma dam pa rnams las dam pa'i chos thos pa'i yi ge don giier gdens can rol pa'i chu gter. sDe dge Zu chen Tshul khrims rin chen. Record of Teachings Received. 2 vols. Dehradun: D. Gyaltsan, 1970.

Secondary sources
Cabez6n, J. Forthcoming. "The Madhyamaka in Gro lun pa's Bstan Rim chen mo." In Proceedings of the Eleventh Seminar of the International Association of Tibetan Studies. KiJnigswinter 2006. Beitriige zur Zentralasienforschung. Halle: International Institute for Tibetan and Buddhist Studies. Eckel, M. 2008. Bhiiviveka and His Buddhist Opponents. Cambridge: Harvard University Press. Hookham, S. 1991. The Buddha Within: TathagatagarbhaDoctrineAccording to the Shentong Interpretation of the Ratnagotravibhaga. New York: State University of New York Press.

282

Kazuo Kano

Jackson, D. 1993. "rNog Lo-tsa-ba's Commentary on the Ratnagotravibhaga." Forword to: Theg chen rgyud bla ma'i don bsdus pa. Commentary on the Ratnagotravibhaga by rNog Lotsaba Blo ldan ses rab. Dharamsala: Library of Tibetan Works and Archives. - 1996. "The bsTan rim (,Stages of the Doctrine') and Simil;u Graded Expositions of the Bodhisattva's Path." In Tibetan Literature: Studies in Genre. Essays in Honor of Geshe Lhundup Sopa. Ed. 1. Cabezon and R. Jackson. New York: Snow Lion, pp. 229-243. Kano, K. 2001. Tibetto ni okeru HoshOron no juyo to tenkai. M.A. Thesis. Kyoto University. - 2003. "Hoshoron kenkyu (1): Phywa pa ni yoru Hoshoron 1.26 kaishaku." Indogaku bukkyo gaku kenkyii 51(2), pp. 109-111. - 2006. rNog Blo ldan ses rab's Summary of the Ratnagotravibhaga: The first Tibetan Commentary on a Crucial Source for the Buddha-nature Doctrine. PhD thesis, Hamburg University. - 2007. "rNog Blo ldan ses rab cho Shokan kamo no shizuku: Kotei text to naiyou gaikan." Koyasan daigaku mikkyobunka kenkyiijo kiyo 20, pp.I-58. - 2008. "rNog Blo ldan ses rab's Topical Outline of the Ratnagotravibhaga Discovered at Khara Khoto." In Contributions to Tibetan Literature. Proceedings of the Eleventh Seminar of the International Association for Tibetan Studies, Konigswinter 2006. Beitriige zur Zentralasienforschung. Ed. O. Almogi. Halle: International Institute for Tibetan and Buddhist Studies, pp. 127-194. - 2009. "rNog Blo ldan ses rab cho Shokan kamo no shizuku: Yakuchuhen." Koyasan daigaku mikkyobunka kenkyiijo kiyo 22, pp. 121-178. Kramer, R. 2007. The Great Tibetan Translator: Life and Works of rNog Blo ldan ses rab (1059-1109). Collectanea himalayica 1. Miinchen: Indus Verlag. Lobsang Dorjee, R. 1999. Five Treatises of Acarya Dipamkarasrijnana. Bibliotheca Indo-Tibetica Series, no. 41. Sarnath: College for Higher Tibetan Studies. Mathes, K.-D. 2007. A Direct Path to the Buddha Within: Go Lotsawa's Mahamudra Interpretation of the Ratnagotravibhaga. Studies in Indian ana Tibetan Buddhism. Boston: Wisdom Publications. Ogawa, 1. 1969. Indo daijobukkyo ni okeru nyoraizo, busshO no kenkyii: Darumarinchen zo Hoshoron shakuso no kaidoku. Kyoto: Buneido Shoten. . Schmithausen, 1.. 1971. "Philologische Bemerkungen zum Ratnagotravibhaga." Wiener Zeitschriftfur die Kunde Sudasiens 15, pp. 123-177.

"rNog Blo ldan ses rab on the Buddha-nature doctrine

-

283

1973. "Zu D. Seyfort Rueggs Buch 'La Theorie du Tathiigatagarbha et du Gotra'." Wiener Zeitschrift fur die Kunde Sudasiens 17, pp. 123-160. Seyfort Ruegg, D. 1969. La Theorie du Tathiigatagarbha et du Gotra: Etudes sur Soteriologie et la Gnoseologie du Bouddhisme. Publications de l'Ecole Fran9aise d'Extreme-Orient 70. Paris: Ecole Fran9aise d'Extreme-Orient. - 1973. Le Traite du Tathiigatagarbha du Bu ston Rin chen grub: Traduction du De biin gsegs pa'i sfiin po gsal zin mdzes par byed pa'i rgyan. Publications de l'Ecole Fran9aise d'Extreme-Orient 88. Paris: Ecole Fran9aise d'Extreme-Orient. - 1977. "The gotra, ekayiina and tathiigatagarbha theories of the Prajfiaparamita according to Dharmamitra and Abhayakaragupta." In: Prajfiiipiiramitii and Related Systems: Studies in Honor of E. Conze. Berkeley: Berkeley Buddhist Studies, pp. 283-312. 1981. The Literature of the Madhyamaka School of Philosophy in India. A History of Indian Literature 7. Ed. J. Gonda. Wiesbaden: Otto Harrassowitz. Takasaki, J. 1980. "Nyoraizo wo meguru ronso: Shaben zo Chiiganshinron shomon sinzitsu kecchaku sho wo sozai to shite." In Bukkyo shisoshi 3: Bukkyo naibu ni okeru taiTOn - indo. Kyoto: Heirakuji shoten (repr. Nyoraizoshiso. Tokyo: Hozokan, 1988-1989), pp. 217-258. - 1982. "Nyoraizoshiso to alayashiki: Yuishikisetsu tono kosho." Ed. Hirakawa et aI., Koza Daijobukkyo 6. Tokyo: Shunjiisha, pp. 151-183. Wangchuk, D. 2005. "The rNin-ma Interpretations of the Tathagatagarbha Theory." Wiener Zeitschriftfur die Kunde Sudasiens 48, pp. 171-213. Zimmerman, M. 2002. A Buddha Within: The Tathagatagarbhasiitra: The Earliest Exposition of the Buddha-nature Teaching in India. Bibliotheca Philologica et Philosophica Buddhica VI. Tokyo: The International Research Institute for Advanced Buddhology, Soka University.

Making and remaking the ultimate in early Tibetan readings of Santideva

Kevin Vose

Santideva's Bodhicaryavatara has long been celebrated, alongside Candraklrti's Madhyamakavatara, for its explication of emptiness from the PrasaIigika-Madhyamaka viewpoint, set within a broader presentation of the Mahayana path structure. While the features that these two texts share - and, indeed, the features of PrasaIigika itself - have been oft-debated in the history of Tibetan Madhyamaka, we have good textual and doctrinal evidence for associating them. One compelling reason is Santideva's declaration that "the ultimate is not a referent of awareness,"l a claim that echoes CandrakIrti's statement that the ultimate "is ineffable and just not a referent of consciousness."2 Further, in explication of Santideva's assertion, Prajfiakaramati's (c. 950-1030) commentary on the Bodhicaryavatara quotes the Madhyamakavatara four times, linking these two texts' views on this central Madhyamaka doctrine. 3 The kind of ultimate suggested in these passages, an ultimate that transcends thought and language, would prove to be a stumbling block for those early Tibetan Madhyamikas with strong com1 Stanza IX.2c; La Vallee Poussin, Prajfziikaramati's Commentary (1905), 352: buddher agocaras tattvaf!!.
2 La Vallee Poussin, Madhyamakiivatiira, 109,2-3: don dam pa'i bden pa bstan par 'dod pas de ni brjod du med pa'i phyir dan ses pa'i yul ma yin pa fzid kyi phyir dnos su bstan par mi nus pas. 3 Prajfiakaramati cites Madhyamakiivatiira V1.23, 25, 28, and 29 in his comments to Bodhicaryiivatiira IX.2; La Vallee Poussin, Prajfziikaramati's Commentary (1905), 353, 361, and 366.

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 285-318

286

Kevin Vose

mitments to the Buddhist epistemological tradition. Logic-minded Madhyamikas, particularly those connected with gSan phu Ne'u thog Monastery, tended to reject Candraklrti's philosophy following its spread in Central Tibet around the year 1100, in large part because of its perceived difficulties in explaining how one realizes an ultimate that transcends human intellect. 4 In contrast, the Bodhicaryiivatiira, having long since been translated, retranslated, and commented upon, was universally acclaimed, making an accounting of its views incumbent upon any author. Rather than accept or reject Santideva's seemingly transcendent ultimate and the host of problems attendant on this view, early bKa' gdams pa scholars found a variety of ways to interpret it. How one interpreted Santideva's ultimate, whether aligning with Candraklrti's transcendent portrayal or with gSan phu's logic-based model, in turn became a dividing line for a series of categories of Madhyamaka, including the well-known Svatantrika-Prasangika divide. The authority of the Bodhicaryiivatiira was forged at gSan phu by the monastery's intellectual founder, rNog BIo ldan ses rab (1059-1109), who is credited with establishing the final version of the Tibetan translation of the text. The colophon of the Tibetan translation states that the text was first translated from a Kashmiri exemplar by Ka ba dpal brtsegs (prior to 840), then was revised in accordance with a central Indian exemplar and its (unnamed) commentary by Rin chen bzan po (958-1055) and Shiikya blo gros (eleventh century). Finally, rNog - apparently on the basis of no new manuscripts or commentaries, but under the guidance of his Kashmiri teacher Sumatiklrti - corrected and finalized the text. 5 Despite the credit given to rNog, early Tibetan commentaries reveal a variety of readings of the root text, suggesting that it circulated in many forms during this period and not just in rNog's "finalized" version. 6

4 Candraldrti's twelfth-century ascension and the debates it touched off in Tibet are treated in Vose, Resurrecting Candrakfrti. 5 This information is drawn from the colophon to the Tibetan translation found in the bsTan 'gyur; sDe dge edition, vol. ya, 40a5-7. 6 Just what constitutes rNog's finalized translation of the Bodhicarya-

Making and remaking the ultimate

287

rNog is known to have written both a commentary and a topical outline on the Bodhicaryavatara? The text likewise figured prominently at gSan phu in the generations following rNog, as a series of scholars in teacher-to-student relationship composed commentaries on it: rNog's student rGya dmar pa Byan chub grags; rGya: dmar pa's student Phya pa Chos kyi sen ge; and Phya pa's student gTsaIi nag pa brTson 'grus sen ge. 8 To these we may add the commentary of the second Sa skya pa hierarch, bSod nams rtse mo (1142-

va tara is difficult to pinpoint. Akira Saito has analyzed Bu ston's suspicion that the version of the Bodhicaryavatara available to him (which he included in his bsTan 'gyur collection) contains unwarranted "emendations" made by gTsan nag pa; see Saito, "Bu ston on the sPyod 'jug," 79-85. In one example, which he takes to be representative, Saito (p. 84) suggests that "the alteration [of rNog's translation] appears to have been made with rather careless consultation of the old translation(s)." A thorough evaluation of this textual conundrum will require an examination of the Bodhicaryavatara stanzas embedded in the various Indian and Tibetan commentaries, as compared to the stanzas preserved in the bsTan 'gyur editions and in the Dunhuang manuscript (Stein 628) edited in Saito, A Study of the Dun-huang Recension. My initial investigation shows that gTsan nag pa's commentary offers readings of the Bodhicaryavatara stanzas that accord with the stanzas preserved in the bsTan 'gyur but not with those found in the Dunhuang version (in cases where the bsTan 'gyur edition and Dunhuang version disagree). If gTsan nag pa was indeed Bu ston's culprit, he does not seem to have been utilizing a translation of the Bodhicaryavatara related to the Dunhuang version. 7 For an overview of rNog's compositions, see Kano, rNgog Bloldan-shes-rab's Summary, 125-128 and Kramer, The Great Translator. Fragments of rNog's commentary are cited in bSod nams rtse mo's commentary (discussed below); the whereabouts of the complete commentary remain unknown. I thank Kazuo Kano for alerting me to his discovery in Lhasa of rNog's sPyod 'jug gi bsdus don, a "topical outline" of Santideva's text, which he is now preparing for publication. 8 The available texts are rGya dmar pa Byan chub grags's Bymi chub sems dpa'i spyod pa la 'jug pa'i tshig don gsal bar Mad pa; Phya pa's sPyod 'jug bsdus don, a topical outline (Phya pa's full commentary is not presently known); and gTsan nag pa brTson 'grus sen ge's sPyod 'jug gi rnambsad.

288

Kevin Vose

1182), which according to its colophon relates the comments of his teacher Phya pa and also occasionally cites rNog's commentary.9 These four generations of commentaries allow us to see the evolution of gSan phu exegesis, revealing a variety of ways to accommodate and even champion the transcendent ultimate found in the Bodhicaryavatara, while at the same time holding to gSan phu's emphasis on inferential logic. These early commentators' solutions to the problems attendant upon Santideva's ultimate took the shape of two interrelated discussions of the nature of ultimate truth and of the cognitive processes involved with realizing it. These in turn gave rise to distinct Madhyamaka categories to classify views on each. The ultimate: "like an illusion" or "not abiding at all?" bKa' gdams pa discussions of ultimate truth, and the Madhyamaka classifications these discussions engendered, center round Santideva's famous proclamation, "The ultimate is not a referent of awareness; awareness is said to be conventional,"lo as well as the
9 bSod nams rtse mo, sPyod pa la 'jug pa'i 'grel pa. In the colophon (515.2,4-6), we read: "Indeed, the Commentary composed by the lord (btsun pa) is exceedingly clear. However, ... I write this for the ease of realization of myself and those like me. For the sake of easy realization even of the wisdom chapter, the spiritual friend bSod nams rtse mo clearly arranged (fie bar sbyar) this from the concise (tshig bsdus, "a summary") and difficult to understand (go dka' ba) Explanation of Engaging in the Bodhisattva's Practices composed by the monk ehos kyi Sen ge." In including bSod nams rtse mo's commentary in this discussion of early bKa' gdams pa commentaries, I do not intend to portray him as a bKa' gdams pa, but rather take his attribution of "arranging" his teacher's comments on the' Bodhicaryavatara as rendering his comments germane to this investigation. As will be seen, his comments would prove to be more faithful to Phya pa's views than those of another of Phya pa's students, gTsan nag pa. Even still, the subtle criticism implicit in calling Phya pa's summary "difficult to understand" may suggest a certain distance between teacher and student that we would be unlikely to see in later master-todisciple relationships within the established orders of Tibet. 10 La Vallee Poussin, Prajfiakaramati's Commentary (1905), 352:

Making and remaking the ultimate

289

stanzas that follow this, which discuss how more advanced yogis' views "harm," or invalidate, those of less advanced yogis.u This latter consideration allows commentators to flesh out just who harms whom, a doxographical project that in some treatments includes higher and lower types of Madhyamikas,12 Unfortunately, the continued absence of rNog Blo ldan ses rab's Bodhicaryiivatiira commentary prevents us from exploring his views on these stanzas. However, in his extant works, rNog touches upon Santideva's proclamation on the ultimate, albeit in a quite different context, his commentary on the Ratnagotravibhiiga. There, rNog relates "The ultimate is not a referent of speech and ... is not a referent of conceptuality, since conceptuality is conventional."13 This clearly
buddher agocaras tattvaf!2 buddhil;. saf!2vrtir ucyate II The Tibetan (vol. ya, 31al) reads: don dam blo yi spyod yul min I blo ni kun rdzob yin par brjod II Saito ("Santideva in the History of Madhyamika," 261, n. 3) reports that the Dunhuang version for pada dreads: blo dan sgra ni kun rdzob yin II = buddhiJ:! sabdas ca saf!2vrtiJ:! ("awareness and speech are conventional"), which would seem to associate "awareness" and "speech" in the way rNog (just below) associates "conceptuality" and "speech." 11 Stanza4ab states this concisely (La Vallee Poussin, Prajiiakaramati's Commentary [1905], 370): biidhyante dhfvise!felJa yogino 'py uttarottaraiJ:! I "Through refinements of yogis' awareness, the progressively higher harm [the lower]." 12 Stanza 4c, which notes how higher yogis "harm" lower yogis, would seem to be germane to this discussion of Madhyamaka classification. It reads (La Vallee Poussin, Prajfiiikaramati's Commentary [1905], 371): dr!ftiintenobhaye!ftena; "through examples accepted by both [parties]." This could well be an allusion to Candraklrti's Prasannapadii critique of Bhaviveka's logical procedures, a discussion that led Tson kha pa to distinguish between CandrakIrti's "Prasangika" Madhyamaka and Bhaviveka's "Svatantrika" view (on Candraklrti's critique and Tson kha pa's distinction, see Ruegg, Two Prolegomena to Madhyamaka Philosophy). However, none of the bKa' gdams pa or Sa skya pa scholars whose commentaries I examine here make this association. 13 Kano's edition of the Tibetan reads (Kano, rNgog Blo-ldan-shesrab's Summary, 286): rdo rje gnas 'di bdun dmigs pa med pa'i ye ses kyis rtogs par bya ba tsam yin gyi I brjod du med pa yan don dam pa'i ran biin yin pa'i phyir ro II don dam pa ni nag gi yul ma yin pa'i phyir te I rnam par

290

Kevin Vose

parallels Santideva's statement, substituting "conceptuality" for "awareness," and suggests that, in rNog's estimation, the ultimate is only beyond the ken of conceptuality but can be accessed by non-conceptual, yogic awareness. This would delimit the ultimate's transcendence, returning it to the realm of (highly refined) human intellect, an approach characteristic of later gSaIi phu authors, such as Phya pa. In the generation following rNog, available sources reveal further discomfort with a literal reading of Santideva's proclamation. Discussing how the two truths are divided, rGya dmar pa first notes that "the basis of division" (dbye ba'i gii) into two truths is "mere object of knowledge, taken as a referent by awareness."14 This would entail that the ultimate, like the conventional, can be brought within the scope of human awareness: it is knowable. Awareness itself is of two types: "All awarenesses that engage objects of knowledge are counted as only two: mistaken consciousnesses that engage erroneously and reasoning consciousnesses that engage non-erroneously."ls As rGya dmar pa explains, a "reasoning consciousness" (rigs pa'i ses pa) is the type of consciousness (which is here used synonymously with "awareness") that knows ultimate truth, through a logical process of investigating the final nature of phenomena. 16 The ultimate would seem then to be a referent of awareness, directly contradicting Santideva.
rtog pa ni kun rdzob yin pas don dam pa rtog pa'i yul ma yin pa'i phyir ro II "The seven vajra topics are only to be realized by the wisdom that lacks an intentional object; they are ineffable due to being the nature of the ultimate. This is because the ultimate is not a referent of speech and because the ultimate is not a referent of conceptuality, since conceptuality is conventional." Kano's translation is on p. 383. 14 rGya dmar pa, Tshig don gsal bar Mad pa, 125,3 (59a3): bios yul du bya pa'i ses bya' tsam. An interlinear note on ses bya' tsam explains the phrase as ses bya ma yin pa las log pa. lS rGya dmar pa, Tshig don gsal bar Mad pa, 127,1 (60a1): ses bya la 'jug pa'i blo' mtha' dag ni nor bar 'jug pa 'khrul ses dan I ma nor bar 'jug pa rigs pa'i ses pa gfiis kho nar grans nes pas na II 16 Ultimate truth is "true in the perspective of a reasoning [consciousness]." rGya dmar pa, Tshig don gsal bar Mad pa, 127,1 (60a1): rigs pa'i

Making and remaking the ultimate

291

However, when commenting on Santideva's proclamation, rOya dmar pa seems to contradict himself:
[The ultimate] is passed beyond all referents of awareness, conceptual and non-conceptual. It also has no proliferations (spros pa) because it is not a determined referent of speech or conceptuality.... The ultimate is not to be characterized by any quality and has no characteristic; thus [Santideva] said, "not a referent of awareness." ... Since all awarenesses are mistaken and conventional, all referents of awarenesses are conventional truths. Thus it is established that the opposite of conventional truth, ultimate truth, is not a referent of awarenessY

Whereas rOya dmar pa's previous discussion divided awareness into "mistake" and "reason," here awareness is uniformly mistaken, functioning only in the conventional, dualistic world of characteristics and the characterized. How are we to reconcile rOya dmar pa's conflicting positions? We might read "not a determined referent of speech or conceptuality" as delimiting the scope of awareness, the same move that rNog made, thereby allowing for a nonconceptual form of awareness that accesses the ultimate. However, the first sentence of the quote, implicating both conceptual and non-conceptual awareness, would seem to preclude this interpretation. To make sense of rOya dmar pa's earlier statement, that the two truths divide "objects of knowledge" (ses bya), we would have to tease out some kind of consciousness (ses pa) that is not included in "all awarenesses." The ultimate would be accessible to (some
nor bden pa. 17 We can note in this passage that rGya dmar pa uses spyod yul (found in Santideva's text) and yul (in rGya dmar pa's comments) interchangeably. Both are used to translate the Sanskrit vi~aya, but typically only the former is used to translate gocara, which is the reading found in the Sanskrit of Santideva's text. rGya dmar pa, Tshig don gsal bar bSad pa, 127,2-4 (60a2-4): rtog pa dan mi rtog pa'i blo'i yul thams cad las 'das pa'o II spros pa med pa yan yin te I sgra dan rnam par rtog pa'i zen pa'i yul ma yin pa'i phyir ro II ... chos gan gis kyan mtshon par byar myed pa mtshan fiid med pa fiid don dam pa yin pas I blo'i spyod yul min zes bya' ba'o I ... blo thams cad 'khrul pa kun rdzob pa yin pa'i phyir blo'i yul mtha' dag kun rdzob kyi bden pa yin pas I kun rdzob kyi bden pa las bzlog pa don dam pa'i bden pa blo'i yul ma yin par grub po I

292

Kevin Vase

kind of) consciousness but would be beyond the scope of conceptual and non-conceptual awareness. rGya dmar pa, unfortunately, leaves the discussion unresolved. For his part, rGya dmar pa introduces a further distinction when explaining the relationship between the two truths (in technical terminology, the "meaning of the division into two truths" [bden pa gfiis kyi dbye ba'i don]). He relates, "[The two truths] are inexpressible as the same or other due to not being one and not being different," a position that he identifies as the view of the "utterly non-abiding system (rab du mi gnas pa'i lugS)."18 He provides an alternative view, that of the "illusionists" (sgyu ma lta bu) who hold that "the two truths are two qualitative divisions in a single entity, just like product and impermanence;" however, he notes that Santideva's position is the former. 19 This discussion does not answer the previous conundrum of how the ultimate is an object of knowledge but not a referent of awareness. However, we can detect some amount of harmony between these two stances and the two Madhyamaka viewpoints rGya dmar pa here mentions: The "illusionist" position holds that the two truths are "qualitative divisions" (chos kyi dbye' ba) of an entity (dflOS po), with "entity" being (in some presentations) equated with "object of knowledge." This suggests a stronger status for ultimate truth, making it a knowable phenomenon, accessible to analysis. In contrast, the "non-abiding" position states that the relationship between the two truths is simply ineffable, transcending speech in the same way that Santideva declared that ultimate truth transcends awareness. 20

18 rGya dmar pa, Tshig don gsal bar Mad pa, 126,3-4 (59b3-4): gcig pa yan ma yin tha dad pa yan ma yin pas de iiid dan gian du brjod du med pa .. , dbye' ba'i don 'di ni rab du mi gnas pa'i lugs la ltos ste rnam par biag pa'o / 19 rGya dmar pa, Tshig don gsal bar Mad pa, 126,4-5 (59b4-5): sgyu' ma Ita bu' ni byas pa dan mi rtag pa biin du bden pa gfiis dnos po cig la chos kyi dbye' ba gfiis su 'dod mod kyi / slob dpon 'di'i lugs ma yin pas Ina {read: sna} ma ltar yin no / 20 Unless, ala rNgog, we pair "ineffable" with "beyond conceptuality," rather than "beyond awareness."

Making and remaking the ultimate

293

Looking beyond rGya dmar pa's work, we see that this division of Madhyamaka is widely referred to - and widely rejected - in early bKa' gdams pa literature. 21 rNog, his student Gro luIi pa Blo gros 'byun gnas (c. 1040-1120),22 and Phya pa all refer to this as a potential bifurcation of Madhyamaka according to positions on ultimate truth. They all argue against it, but for different reasons. Jose Cabez6n points out that rNog and Gro lun pa see both of these potential divisions of Madhyamaka as, instead, deviations from the middle way.23 "Illusionists" assert (in this portrayal) that the illusory appearance of phenomena is ultimate truth, which in rNog's and Gro lun pa's estimation amounts to an extreme of realism, as this illusory nature would stand through reasoned analysis. Additionally, Gro lun pa is keen to point out that the "illusionist" view is not that of Santarak~ita, although he does not say who held it. 24 The "non-abiding" position carries several possible valences in rNog's and Gro lun pa's work: rNog seems to equate the position both with the meditational practice of "no [thing] abiding" in the mind (yid la mi gnas pa) and with the ontological claim that no phenomenon abides (chos kun mi gnas); these would be extremes of quietism (from a practice standpoint) and nihilism (from an on-

21 This bifurcation of Madhyamaka is also rejected by Tson kha pa and his dGe lugs pa followers; see Napper, Dependent-Arising and Emptiness, 403-440 (Appendix I: "The Division of Madhyamikas Into ReasonEstablished lllusionists and Proponents of Thorough Non-Abiding") and Ruegg, Three Studies, 96-1Ol. 22 Jose Cabez6n discusses the problems attendant upon Gro lun pa's dates, as well as this provisional estimate, in "The Madhyamaka in Gro lung pa's Bstan Rim chen mo." 23 My discussion here is based on that in Cabez6n, "The Madhyamaka in Gro lung pa's Bstan Rim chen mo," which relies on rNog Blo ldan ses rab, sPrift yig bdud rtsi thig le and Gro lun pa Blo gros 'byun gnas, bsTan rim chen mo. 24 Gro lun pa criticizes the view that "Siintarak~ita and others posit illusion as the ultimate;" Gro lun pa, bsTan rim chen mo, 805: slob dpon ii ba 'tsho la sogs pa ni sgyu ma don dam pa bied pa ste / While he does not tell us who the "others" are, they likely include Kamalaslla.

294

Kevin Vase

tological perspective).25 Gro lun pa, instead, states that the "nonabiding" view holds that the non-abiding of phenomena is itself established by a "reasoning consciousness," which, similar to allowing that an illusory nature withstands reasoning, suggests that something stands out of emptiness - albeit, in this case, that something would be non-existence. 26 This perspective would then represent an overly reified ultimate. In all interpretations, rNog and Gro lun pa reject both the "illusionist" and "non-abiding" positions and so discount their use for dividing Madhyamaka. In his recently published doxography, Phya pa takes a quite different approach to this division, claiming that since all Madhyamikas assert both the utter non-abiding and the illusory appearance of all phenomena (not neither), there can be no division of Madhyamaka along these lines. The important issue, instead, is just what "withstands analysis" (dpyad bzod pa). If one holds that appearances withstand analysis - that they are not utterly non-abiding - one would not claim that they are like illusions; this, then, would make one a realist. If one does not hold that appearances are "mere affirming negatives" (ma yin dgag pa tsam), one deprecates the conventional and becomes a nihilistY For Phya pa, then, to say that phenomena are "utterly non-abiding" and to say that they are "like illusions" come back to the same point: appearances do not withstand analysis and so exist only illusorily. Further, Phya pa states, ''All Madhyamikas, due to not asserting that illusory [phenomena] withstand analysis, assert that only utter non-establishment withstands analysis; there is [thus] no difference

25 rNog's summary criticism reads (sPrin yig bdud rtsi thig Ie, 708): sgyu ma gfiis med chos kun mi gnas dbu ma yi lugs gfiis rnam 'byed de yan rmons pa mtshar bskyed yin; "Distinguishing two Madhyamaka systems [according to those who hold that] illusion is non-duality and [those who hold that] all phenomena do not abide amazes [only] fools." 26 This position would also make emptiness an "affirming negative" (ma yin dgag), as the existence of "non-abiding" would be affirmed. This view is unacceptable to Gro lun pa and to all the bKa' gdams pa authors treated here. 27 Phya pa, bDe bar gsegs pa, 65,6ff.

Making and remaking the ultimate

295

[among Madhyamikas] at all."2B "Utter non-establishment" (cir ywi rna grub pa) appears to be synonymous with "utterly non-abiding" in Phya pa's usage, an equation not reflected in rNog's orGro lun pa's discussions. Holding that "utter non-establishment withstands analysis" saves this portrayal of the ultimate from the extremes of quietism and nihilism that rNog argued against, as Phya pa's ultimate would be reached by a "reasoning consciousness" - it would "withstand" the analysis of that consciousness - and would be counted as an existent "object of knowledge." Where Gro lun pa criticized the "non-abiding" ultimate for affirming the existence of "non-abiding," Phya pa draws a fine line: the ultimate withstands analysis but is not established by reasoning; it is still "utterly not established."29 While Phya pa's portrayal might seem to solve some of the problems in the "illusionist" and "non-abiding" positions adduced by rNog and Gro lun pa (by redefining those positions) and, further, might seem to ameliorate divisions within Madhyamaka, two points bear consideration. First, he introduced this discussion by noting "someone claims there are two systems of assertions on ultimate truth."30 However, Phya pa's discussion of the "illusory" status of phenomena centers on the conventional world. The real problem would not arise from claiming that conventional phenomena are illusory but would arise from the claim that an "illusory" nature is ultimate truth, which is the position rNog and Gro lun pa both rejected. Elsewhere, Phya pa too argues against this position at length, which he characterizes as the view that the union of appearances and emptiness (snan stan giiis tshags) is the ultimate; to this claim, he points out that each of the components would per force be

2B Phya pa, bDe bar gsegs pa, 67,2: dbu ma ba thams cad kyis sgyu ma lta bu iiid ni spyad par [dpyad bzod par?} mi 'dod pas I cir ymi ma grub pa kho na dpyad bzod par 'dod pa la tha dad gtan med pa yin no II 29 For a more complete discussion of Phya pa's views on the ultimate, see Vose, Resurrecting Candrakfrti, 92-99. 30 Phya pa, bDe bar gsegs pa, 65,6-7: don dam pa'i bden pa la yan kha cig bden pas stan pa'i snan pa sgyu ma lta bur smra ba dan I ma yin dgag du bden pa gan du yan rab tu mi gnas par smra ba'i lugs giiis yod zer ba.

296

Kevin Vose

the ultimate, making ordinary appearances ultimately true. 31 Phya pa, then, follows his gSan phu forebears in rejecting the ultimate validity of the "illusionist" view. 32 The frequency and vehemence of this refutation in early bKa' gdams pa literature suggests that it was a powerful current outside ofbKa' gdams pa circles in twelfthcentury Tibet. 33 The second point to consider is that, as Phya pa was well aware, not all those who call themselves Madhyamikas assert that the ultimate bears analysis. Such a position would be labeled "realist" by other Tibetan Madhyamikas, and constituted one of the chief points of contention between Phya pa and twelfth-century support-

31 Phya pa emphasizes the distinction between the absolute negative, emptiness, that a reasoning consciousness realizes and the affirming negative, the collection of appearance and emptiness (the illusion-like nature), that inference comprehends. The inference that proves "entities are empty of a true nature" must realize an affirming negative, because it associates the inferential subject, "entity," with the predicate, "emptiness." However, a reasoning consciousness realizes only the absolute negative, "empty of a true nature," based on this inference. Phya pa, dBu ma sar gsum, 93,14-97,14; especially 94,14-18. 32 Phya pa adopts a position very close to the "illusionist" view on "the meaning of the division into two truths." Above, we saw rGya dmar pa report the "illusionist" position to hold that "the two truths are two qualitative divisions (chos kyi dbye' ba) in a single entity." Phya pa held that "the two truths are only different isolates in a single entity;" Phya pa, dBu ma sar gsum, 10,12: bden pa gfiis no bog cig la ldog pa tha dad pa kho na'o II 33 The distinction between "illusionist" Madhyamaka and "utterly non-abiding" Madhyamaka may have arisen in proto-bKa' brgyud pa sources. The two terms are used to denote types of Madhyamaka in Maitripada/Advayavajra's Tattvaratniivalf, where the terms are given as miiyopamiidvayaviidin and sarvadharmiiprati~thiinaviidin. On MaitrIpada's views, see Mathes, "Blending the Slitras with the Tantras," 201227. Ruegg reports that the distinction is also found in sGam po pa bSod nams rin chen's work, Tshogs chos legs mdzes ma, in volume one of his gSun 'bum (Delhi: Khasdub Gyatsho Shashin, 1975), ca, folio 85a; see Ruegg, Three Studies, 32-35.

Making and remaking the ultimate

297

ers of CandrakIrti. 34 His claim, then, that there is no division within Madhyamaka is not a claim that there is no dispute within Madhyamaka. Instead, "his is the highly polemical claim that those who disagree with him are not even Miidhyamikas. Phya pa's reconfiguration of the "illusionist" and "non-abiding" positions allows him to accept his teacher rGya dmar pa's claim that Siintideva held the "non-abiding" view, without subscribing to the transcendent ultimate that such a position may have entailed. Additionally, like his gSaIi phu predecessors, Phya pa can reject the notion that Siintarak~ita and KamalaSUa hold an "illusionist" ultimate, while at the same time accepting the conventional validity of "illusion-like" phenomena. 35 These steps are important for preserving the gSail phu interpretation of Madhyamaka, which championed the works of all of these authors: Siintideva, Siintarak~ita, and KamalaSUa. With the rise of CandrakIrti's importance in Tibet during Phya pa's lifetime, preserving Siintideva's affiliation with views that now came to be labeled "Sviitantrika" - and thereby dissociating Siintideva from the new Priisailgika - required interpretive finesse. 36 Furthermore, in Phya pa's treatment, the "nonabiding" view comes to represent a well-acceptable Madhyamaka stance; in Phya pa's view, all Miidhyamikas espouse it. This interpretation would make it possible for Phya pa's students to stand behind the "non-abiding" label and, as we will see in the next section, to subdivide the position further.
34 Twelfth-century Prasaitgika views on ultimate truth are treated in Vose, Resurrecting Candrakfrti, 88-92. 35 As noted above (n. 32), Phya pa's position on "the meaning of the division into two truths" is nearly identical with the "illusionist" position that rGya dmar pa reports. 36 The Indian commentaries on the Bodhicaryavatara that treated the text from a Yogacara-Madhyamaka standpoint, discussed by Saito ("Siintideva in the History of Madhyamika," 259), are important precedents for the early gSait phu interpretation. Santarak~ita's citation of the Bodhicaryavatara in the Tattvasiddhi may also be taken as indicating a compatibility between Santarak~ita's and Siintideva's views, although it seems doubtful that the Tattvasiddhi's author is the same Siintarak~ita who wrote the Madhyamakiilarp,kara.

298

Kevin Vose

One of those students who did not take up the "non-abiding" label and who would reject his teacher's Madhyamaka views in favor of the new Prasangika was gTsan nag pa brTson 'grus sen ge. gTsan nag pa'sdiscussion of Santideva's ultimate evinces some of the same tensions we saw in rGya dmar pa's explication. gTsan nag pa holds that ultimate truth is an "object of knowledge" but, "an instance of the ultimate is passed beyond referents of awareness."37 A significant difference from rGya dmar pa's presentation is that gTsan nag pa credits his position to Candraklrti, noting that ultimate truth is "found in the perspective of a final reasoning consciousness, ... just as the Master Candraklrti said."38 Unlike his gSan phu forebears, gTsan nag pa takes Prajfiakaramati's lead in drawing Candraklrti's exposition of the two truths into his own explanation of Santideva's stanza. While gTsan nag pa's position on the ultimate may appear identical to rGya dmar pa's stance two generations prior, Candraklrti's presence is crucial, as an examination of gTsan nag pa's division of Madhyamaka reveals. gTsan nag pa interprets Santideva's statement that higher yogis harm lower yogis as implying three successively higher levels of Madhyamaka. 39 The first type holds that "an affirming negation, a composite of appearance and emptiness" is the ultimate. This, 37 gTsan nag pa, sPyod 'jug gi rnam Mad, 557,2-3 (36a2-3): ses bya'i lus bden gfiis su rnam par giag pa and 557,7-8 (36a7-8): don dam pa'i mtshan gii' ni blo'i yullas 'das pa. 38 gTsan nag pa, sPyod 'jug gi rnam Mad, 557,5 (36a5): mthar thug rigs pa'i ses nor rfied pa. 39 gTsan nag pa initially speaks of two types of Madhyamikas, but then points out a third Madhyamaka view that invalidates these two (sPyod 'jug gi rnam Mad, 561,2-3 [38a2-3]): dbu' ma'i rnal 'byor fiid la ston pa mtshan mar lta ba gfiis yod de I bden pa'i dnos pos ston pa'i ses bya ma yin dgag don dam par smra ba dan I ston fiid med dgag don dam par smra ba'o II de dag la yan dbu' ma chen po'i rigs pas gnod te I "Madhyamika yogis have two views on the mark of emptiness: (1) those who hold that an affirming negative, an object of knowledge that is empty of true entity, is the ultimate; and (2) those who hold that an absolute negative, emptiness, is the ultimate. (3) The reasoning of Great Madhyamaka harms those also."

Making and remaking the ultimate

299

of course, is the "illusionist" position already critiqued by rNog, Gro lun pa, and Phya pa. This position is invalidated (or "harmed") by the second type of Madhyamika, "those who hold that an absolute negative (rned dgag), emptiness, is the ultimate." As noted, this view appears to have been a very common position among early bKa' gdams pa authors, as it continued to be in the development of Tibetan Madhyamaka. This position, then, would seem to be unassailable; yet gTsan nag pa states, "The reasoning of Great Madhyamaka (dbu rna chen po) harms those, also." What is wrong with holding emptiness to be the ultimate? And, what does "Great Madhyamaka" signify? gTsan nag pa tells us
Since the entity to be negated is not established, the negation too is not established, like the death of a barren woman's child. In that way, it is realized that ultimately phenomena - entites, non-entities, and so forth - do not exist at all. ... [This] also dispels the idea that the ultimate is established as an analytical referent of awareness. 40

These points suggest that "Great Madhyamaka" upholds a strong interpretation of Santideva's transcendent ultimate: the reasoning process that negates "true entity" (bden pa'i dftos po) itself has no standing and "ultimately phenomena do not exist at all." In this interpretation, the seemingly unproblematic position that "emptiness is the ultimate," while invalidating the "illusionist" position, still assigns too strong a status to negation. 41 Something still stands out of emptiness. The position is further criticized for a perceived waffling from the ultimate's transcendence; only "Great Madhyamaka" understands that the ultimate passes beyond the scope of human intellect.
40 gTsaIi nag pa, sPyod jug gi rnam biad, 561,3-4 (38a3-4): dgag bya'i dflOS po ma grub pas de bkag pa ymi mi 'grub ste I mo sam gyi bu si ba biin ies bya'o I de ltar na don dam par dnos po dan dnos med lasogs pa'i chos 'ga' yan med par rtogs par 'gyur ro II ... don dam blo'i spyod yul du grub par 'gyur sfiam pa'i rtog pa 'an bsal ba yin no II 41 Pa tshab, too, points out that since the object of negation is not established, the negation of it lacks status (dBu' ma rtsa ba'i ses rab kyi ti ka, 49: dgag bya'i ran biin ma grub pas na bkag pa yan mi 'thad de); see Georges Dreyfus and Drongbu Tsering, "Pa tshab and the origin of PriisaIigika," in this volume.

300

Kevin Vose

gTsaIi nag pa's invocation of CandrakIrti (to make essentially the same point that rGya dmar pa made) can be understood in the context of his Madhyamaka rankings. "Great Madhyamaka" aligns Santideva's ultimate with CandrakIrti's and gives both a transcendent reading. In showing how Great Madhyamaka trumps those who hold emptiness to be the ultimate, gTsaIi nag pa attacks a position very much like the one Phya pa argued for and attempts to answer Phya pa's critique of CandrakIrti and his newfound followers. Phya pa holds that an absolute negation, emptiness, is the ultimate and that emptiness withstands analysis, a position that gTsaIi nag pa understands as an overly reified view. Where Phya pa held that the ultimate is an object of knowledge, gTsaIi nag pa criticizes those who "think that the ultimate is established as an analytical referent of awareness." We can recall that Phya pa emphasizes that the ultimate is "utterly non-established." gTsaIi nag pa may here misrepresent his former teacher's position in order to draw a clear separation from his own view, which itself offers a significant softening from both CandrakIrti's and Santideva's proclamations. Jayananda, for one, took CandrakIrti very literally and claimed that the ultimate was not even an object of knowledge. Phya pa pointed out a number of problems with this position, the most significant being that the ultimate - realization of which is soteriologically necessary - could not be known.42 Here, we see gTsaIi nag pa arguing that the ultimate is an "object of knowledge" but is not "established as an analytical referent of awareness," a position he arrives at in response to Phya pa's critique. Among gTsaIi nag pa's chief concerns are defending and championing CandrakIrti's interpretation of Madhyamaka, here reading that interpretation (rightly or wrongly) into Santideva's text. "Great Madhayamaka," then, is gTsaIi nag pa's term for what others in this time period began to call "PrasaIigika.''43

42 For the debates between Jayiinanda and Phya pa over the status of the ultimate, see Vose, Resurrecting Candrakfrti, 88-99. 43 In his contribution to the present volume, Thomas Doctor shows that rMa bya, likewise, calls Candraklrti's views "Great Madhyamaka."

Making and remaking the ultimate

301

Buddha perception: PrasaIigika v. Svatantrika A second set of divisions of Madhyamaka found in these early commentaries takes as its chief criterion the status of those transformed by their realization of the ultimate: Buddhas. These divisions are found in exegeses of Santideva's answer to the objection that the Madhyamaka understanding of emptiness renders the passage from saJ?1siira to nirvaI).a pointless. Paul Williams noted that where the received Sanskrit of Santideva's text here offers precision, the Tibetan translation of it offers ample room for interpretation.44 Translating from the Sanskrit, we can read the objection as "If what is ceased ultimately is cycling in saJ?1siira conventionally, then the Buddha, too, would be cycling in saJ?1siira; thus, what use would the practices of enlightenment be?,'45 In the Tibetan translation, "what is ceased" (nirvrtal.z) is rendered mya nan 'das, allowing the possibility that "nirvaI).a" is at issue. Additionally, the verbal sense of saJ?1siira is lost, as are the adverbial usages of "ultimate" and "conventional." The Tibetan, then, offers several possibilities, including the very straightforward rendering, "If the ultimate is nirvaI).a and saJ?1siira is the conventional, the Buddha too would be saJ?1siira; thus, what use would the practices of enlightenment be?''46 The ambiguity of the Tibetan translation allows for a number of interpretations, not all of which would seem coherent when reading the received Sanskrit. Williams pointed out that many Indian and 44 Williams, "On PrakrtinirviiIJ-alPrakrtinirvrta in the Bodhicaryiivatiira," 522-523. 45 The Sanskrit of stanzas IX.l3cd and 14ab reads (La Vallee Poussin, Prajfiiikaramati's Commentary [1907], 385): nirvrtaJ:t paramiirthena sa-/!lvrtyii yadi saf!!saret II buddho 'pi saf!!sared evaf!! tataJ:t kif!! bodhicaryayii I While redundant, my "cycling in saf!!siira" conveys the Sanskrit verbal use of the word. 46 The Tibetan, edited from the canonical versions, is found in Oldmeadow, A Study of the Wisdom Chapter, vol. 2, 34: gal te don dam mya nan 'das II 'khor ba kun rdzob de ita na II sans rgyas kyan ni 'khor 'gyur bas II byan chub spyod pas ci zig bya II The Dunhuang version does not differ substantially (Saito, A Study of the Dun-huang Recension, 50).

302

Kevin Vose

Tibetan commentators read this objection as embodying a misunderstanding of the difference between "natural nirvaI).a" (or "natural cessation;" prakrtinirvtilJa/prakrtinirvrta) and the attainment of nirvaI).aY Prajiiakaramati reads the passage in just this way, having the objector state that "a Buddha, who has cessation due to abandoning all defilements, also would be cycling in saf[lstira."48 The identity between natural nirvaI).a and saf[lstira that Madhyamikas accept leads the objector to contend (mistakenly) that the attainment of nirvaI).a would leave one still in saf[lstira. Prajiiakaramati's reading of the objection makes for an easy Madhyamaka answer: despite the "natural cessation" of all phenomena, the attainment of nirvaI).a depends on attaining the cessation of ignorance (avidytinirodha), at the root of the twelve links of dependent arising, as explained in a lengthy quote from the Stilistambasiitra. 49 Upon the cessation of ignorance and the remaining links, "there would be no cycling in saf[lstira."50 rNog BIo ldan ses rab, bSod nams rtse mo tells us, took a similar approach to explicating Santideva's question and answer. On rNog's reading, the objector claims that since all things are naturally nirvaI).a, which is in the end no different from saf[lstira, those
47 Williams, "On PrakrtinirviiT}a/Prakrtinirvrta in the Bodhicaryiivatiira," 522ff., especially 525-526 where Williams notes, "The distinction between innate 'enlightenment' and that attained through following the path means that the prakrtinirviiT}a is almost universally employed in Tibet to explain the opponent's objection and its solution." As will be seen, the early bKa' gdams pa materials present an important counter to this claim. 48 La Vallee Poussin, Prajfiiikaramati's Commentary (1907), 385,1718: buddho 'pi sarviivaraT}aprahiiT}ato nirvrto 'pi sal!lsiiret / The Tibetan translation here renders nirvrta as mya nan las 'das pa, consistent with the handling of Santideva's stanza. However, in Prajfiakaramati's introduction to this passage, his usage of prakrtinirvrta (385,7) is rendered ran biin gyis ldog pa. For the edited Tibetan of the complete passage, see Oldmeadow, A Study o/the Wisdom Chapter, vol. 2, 34-35. 49 La Vallee Poussin, Prajfiiikaramati's Commentary (1907), 386,12ff. 50 La Vallee Poussin, Prajfiiikaramati's Commentary (1907), 389,12: sal!lsiiraT}al!l na syiit /

Making and remaking the ultimate

303

who attain nirvax;ta are in actuality still in sa1flSara; so, the objector wonders, what is the point of the practices of enlightenment? rNog explains (according to bSod nams rtse mo) that while all things are ultimately naturally nirvax;ta, conventionally there is a difference between the attainment of nirvax;ta and sa1flsara according to "the extinction or non-extinction of the adventitious causes that generate affiictions."51 rNog's explanation, then, may have relied on that of Prajfiakaramati, or at minimum was drawn from consultation of a Sanskrit version of Santideva's text related to that preserved in Prajfiakaramati's commentary, a version that facilitates a "natural nirva1'}a versus attained nirva1'}a" reading of the objection and answer. Surprisingly, Prajfiakaramati's and rNog's explication of this passage was not adopted by other early bKa' gdams pa authors. One reason for this departure is the above-noted ambiguity of the Tibetan translation of Santideva's question. A second reason is Santideva's somewhat cryptic answer: "If the causes do not have their continuum cut, illusion also is not stopped; upon the causes having their continuum cut, it does not arise even conventionally."52 Following Prajfiakaramati, "the causes" would be ignorance, desire, and hatred - the causes of sa1flSara; the cutting of their continuum would refer to the reversal of the twelve links of dependent arising. The parallel construction of this passage implies that what "does not arise even conventionally" is "illusion." However, the variety of interpretations left open by the objector's question allows Tibetan authors to develop various readings of just what has its "continuum cut" and just what Santideva claims "does not arise," readings that give rise to distinct classifications of Madhyamaka. 51 bSod nams rtse mo, sPyod pa la 'jug pa'i 'grel pa, 495.3,4 (295b4): non mons pa skyed pa'i rgyu glo bur ba zad ma zad kyis 'byed do. 52 La Vallee Poussin, Prajniikaramati's Commentary (1907), 386,67: pratyayiiniim anucchede miiyiipy ucchidyate na hi I pratyayiiniirrz tu vicchediit sarrzvrtyiipi na sarrzbhavaJ:z II Oldmeadow, A Study of the Wisdom Chapter, vol. 2, 35: rkyen nams rgyun ni ma chad na II sgyu ma'ang ldog par mi 'gyur gyi II rkyen nams rgyun ni chad pas na II kun rdzob tu yan mi 'byun no II While the original has no pronoun in the final piida, I insert "it" as a placeholder for the interpretations discussed below.

304

Kevin Vase

The clearest statement of this alternate interpretation is found in bSod nams rtse mo's commentary, our source for the previous presentation of rNog's reading. Having given rNog's position, bSod nams rtse mo notes, "That explanation is not very good. Regarding this, the great commentator explains in this way."53 Following this unnamed "great commentator" (likely to be Phya pa), he claims that rNog has misunderstood the objection, for in Sautrantika, "the intentional objects of nirvii1}a are ultimate and thoroughly afflicted intentional objects are conventionaL" Given this, the Sautrantika opponent here asks
Would the Buddha also have awareness of appearances or would he not? If not, it follows that the Buddha is a non-entity or is matter. If Buddha has awareness of appearances, is that conventional or ultimate? If ultimate, [your] thesis stating that awareness is conventional deteriorates. If conventional, it follows that [the Buddha] is sal'(tsiira due to having conventional, thoroughly afflicted proliferations. If you accept that, what is the point of the bodhisattva practices?54

Rather than charging Madhyamaka with a suffering Buddha through the confiation of natural nirvii1}a and the attainment of nirvii1}a - in this interpretation the objector is concerned about what a Buddha can perceive. The equations that the Tibetan translation allows between nirvii1}a and the ultimate, on one hand, and
53 bSod nams rtse mo, sPyod pa la 'jug pa'i 'grel pa, 495.3,5 (295b5): Mad pa de ha cali legs pa ma yin no II 'di la 'grel chen byed pas 'di ltar 'chad de I 54 bSod nams rtse mo, sPyod pa la 'jug pa'i 'grel pa, 495.3,5-495.4,2 (295b5-296a2): spyir mdo sde pa rali gi grub mtha' mya lian las 'das pa'i dmigs pa ni don dam pa yin la I kun nas non mons pa'i dmigs pa ni kun rdzob ces bsams nas rgol ba ni I gal te don dam mya nan 'das II 'khor ba kun rdzob de Ita na II sans rgyas kyan ni 'khor 'gyur bas II byali chub spyod pas ci zig bya II zes smos te I gal te don dam pa ni mya lian las 'das pa yin na 'khor ba ni kun rdzob kyi spros pa yin na sans rgyas la'ali snan bcas kyi blo yod dam med I med na salis rgyas dlios med dam bems por thalIa I yod na de kun rdzob yin nam don dam yin I don dam yin na blo ni kun rdzob yin par brjod kyi dam bca' nams la I kun rdzob yin na kun rdzob kun nas non molis spros pa yod pas 'khor bar thallo I de 'dod na byali chub spyod pas ci Zig bya zes rgol ba na I

Making and remaking the ultimate

305

sa1!lsiira and the conventional, on the other, leaves the Buddha perceiving only an ultimate emptiness, without the ability to perceive conventional appearances. If a Buddha perceives conventional appearances, the objector reasons, that Buddha must have "conventional, thoroughly afflicted proliferations," which the bodhisattva path was to have removed. bSod nams rtse mo's interpretation of the issue is not unique to him but is first found in rGya dmar pa's commentary. rGya dmar pa further reads the objector to equate "mistake" (,khrul pa, bhriinta) and awareness (bla, buddhi); a Buddha's perception of appearances would entail a mistaken Buddha. 55 rGya dmar pa and bSod nams rtse mo thus see a debate on just what connection a Buddha has with the conventional world. Specifically, if nirviilJa is the ultimate, do those who reach nirviilJa perceive the conventional, which has (in this interpretation) been declared to be sa1!lsiira? And if they do, do these perceptions entail on the part of Buddhas ignorance, the root of sa1!lsiira? rGya dmar pa notes that Madhyamikas have two possible answers to this objection, which divides them into two camps: "those who assert that wisdom has its continuum cut" and "those [who assert] that wisdom does not have its continuum cut." rGya dmar pa clearly reads Santideva's answer ("upon the causes having their continuum cut, it does not arise even conventionally") in a new way: here, the issue is whether awareness continues through the transformation to buddhahood or, alternatively, whether it is a Buddha's wisdom that "does not arise." On the first group's answer, rGya dmar pa tells us
Some Madhyamikas assert that since all awareness is mistaken, when mistake is extinguished awareness itself does not exist and thus wisdom has its continuum cut; "even conventionally" wisdom does not exist. These assertions are not reasonable. . .. Even though mistake is extinguished, wisdom is not stopped .... Since [Buddhas] see (gzigs pa) illusory dependent arising as just illusion without the capacity to apprehend it as true, they are not mistaken. 56 rGya dmar pa, Tshig don gsal bar Mad pa, 133,3-4 (63a3-4). rGya dmar pa, Tshig don gsal bar bsad pa, 134,6-135,1 (63b6-64a1):

55 56

306

Kevin Vose

Accepting a strong correlation between nirvii1Ja and perceiving the ultimate, emptiness, the first camp holds that the transformation to a Buddha eradicates all awareness, which is necessarily mistaken. Without awareneSs - the instrument that perceives conventional appearances - wisdom (which is here understood to be a type of awareness) cannot arise. With mistaken awareness extinguished, Buddhas are left in emptiness. This notion is unacceptable to rGya dmar pa, a partisan of the second camp, who instead holds that mistake and awareness are separable. Rather than read Santideva's answer as entailing a wisdom-less Buddha, rGya dmar pa nuances his response, writing "When, at buddhahood, 'the causes' that are collected in sa1'(tsiira 'have their continuum cut,' that [awareness] collected in sa1'(tsiira 'even conventionally does not arise."'57 Whereas the first group understands Santideva to claim that "even conventionally, wisdom does not arise" due to all awareness being extirpated along with ignorance, this interpretation limits the negation to those states of mind "collected in sa1'(tsiira." While "mistake" characterizes awareness in sa1'(tsiira, it is not a necessary quality of awareness. Wisdom is a Buddha's awareness, stripped of mistake by way of a Buddha's overcoming the conception that ordinary appearances are true. rGya dmar pa notes a further significant difficulty in the first interpretation: "If wisdom did not exist, the Buddha would not exist; thus, the Buddha's teachings themselves would not exist and the sangha who realize them would not exist."58

'on kyan blo thams cad 'khrul pa yin pas 'khrul pa zad pa'i tshe blo iiid med pas ye ses rgyun chad do ies kun rdzob du yan ye ses med do ies dbu' ma pa kha cig 'dod pa ni mi rigs ste / ... 'khrul pa zad kyan ye ses de mi ldog go / ... rten 'brei sgyu' ma la bden par 'dzin pa mi mna' bar sgyu' ma iiid du gzigs pas ma 'khrul ies bya'o / 57 rGya dmar pa, Tshig don gsal bar Mad pa, 133,7-8 (63a7-8): sans rgyas pa'i tshe' 'khor bas bsdus pa'i rkyen rnams rgyun ni chad pas na / 'khor bas bsdus pa kun rdzob du yan mi 'byun no / 58 rGya dmar pa, Tshig don gsal bar Mad pa, 134,3 (63b3): ye ses myed na sans rgyas iiid med pas / de'i bstan pa ilid med cin de la bsgrub pa'i dge 'dun med pas /

Making and remaking the ultimate

307

rGya dmar pa's portrayal of these two Madhyamika explanations appears in similar form in bSod nams rtse mo's explication. Having already declared that the "natural nirviilJa v. attained nirviilJa" interpretation of rNog misses the point and reoriented the problem around the issue of what a Buddha perceives, bSod nams rtse mo gives two possible answers to this problem, which he further identifies as the PrasaIigika and Svatantrika positions. He begins with the PrasaIigika answer, which reads Santideva's "causes" as "the causes of awareness of appearances:"
'Upon the causes having their continuum cut:' The diamond-like meditative equipoise cuts the continuum of all signs and conceptuality, thereby cutting the continuum of afflictions. Cutting that [further] cuts the continuum of actions. Cutting the continuum of that [further] cuts the continuum of awareness of appearances, whereby sarrzsiira, 'even conventionally, does not arise.'59

This equation of samsiira and perceiving conventionalities works well with the Tibet~n translation of Santideva's stanza. However, bSod nams rtse mo cannot accept this equation, as it either leaves Buddhas in the dark or, as the objector has claimed, leaves all manner of realized beings in saY[tsiira. Instead, bSod nams rtse mo states "We do not assert that all conventionalities are saY[tsiira, nor do we assert that all nirviilJas are ultimate; we posit [them] as ultimate or conventional through bearing or not bearing analysis."60 He endorses the Svatantrika reading of Santideva's answer, that "'Upon causes,' that is, actions and afflictions, 'having their continua cut, even conventionally' that is,

59 bSod nams rtse mo, sPyod pa la 'jug pa'i 'grel pa, 495.4,4-495.4,5 (296a4-296a5): rkyen rnams rgyun ni chad pa na* ste rdo rje ita bu'i tin ne 'dzin gyis mtshan ma dan rnam par rtog pa thams cad rgyun bead pas non mons pa rgyun bead / de bead pas las rgyun bead / de rgyun bead pas snan bcas kyi blo rgyun bead pas 'khor ba kun rdzob tu'an no i.es bya ba'o / * The text incorrectly reads ma chad pa for chad pa na. 60 bSod nams rtse mo, sPyod pa la 'jug pa'i 'grel pa, 496.1,1 (296b1): kho bo cag kun rdzob thams cad 'khor bar khas mi len / mya nan 'das pa thams cad don dam du khas mi len te / dpyad bzod mi bzod kyis don dam kun rdzob tu 'jog go /

308

Kevin Vase

even to Buddhas' conventional wisdom, [illusion] 'does not arise' as saY(lsara."61 Appearances themselves are not cut off by the transformation to buddhahood, afflictions are. As with rGya dmar pa, appearances have a somewhat neutral status: illusory appearances can "arise as saY(lsara" to the afflicted or they can be perceived by a Buddha's conventional wisdom without these appearances serving as a cause of suffering. As for rGya dmar pa, afflictions and awareness are separable. There is a point to the practices of enlightenment, after all. Before considering what we can glean from these divisions of Madhyamaka, we must ask a more basic question: was this alternate interpretation of Santideva's objection and answer simply a misunderstanding of the original passage? The issue of a Buddha's ability to perceive ordinary appearances would be nearly impossible to derive from the Sanskrit versions that we have of Santideva's text. Prajfiakaramati's lengthy commentary takes a straightforward approach to this passage, explaining the distinction between saY(lsara and nirvalJa with recourse to the twelve links of dependent arising. While my investigation of Indian commentaries is far from complete, we can note that Vibhuticandra's interpretation of this passage suggests a concern with the relation between "mistake" and "appearances," and whether a Buddha's ability to perceive the latter entails possession of the former. He writes, "Ultimately, saY(lsara just does not exist; the appearances of saY(lsara due to mistake [lasts] for as long as ignorance exists. Having dispelled ignorance through the arya path, those [appearances of saY(lsara] also do not exist."62 Vibhuticandra's reading echoes the Prasailgika interpreta61 bSad nams rtse mo, sPyod pa la 'jug pa'i 'grel pa, 496.1,2-3 (296b23): rkyen las dan non mons pa rgyun chad na kun rdzob tu'an ste sans rgyas la ye ses kun rdzob yod kyan 'khor bar mi 'byun ies bya ba dbu ma ran rgyud pa rnams Mad pa byed do / In reading "even conventionally" as "even to Buddhas' conventional wisdom," bSod nams rtse mo must read Santideva as assuming that Buddhas' meditative absorption (mnam biag ye ses) contains no appearances at all; the point here, as bSod nams rtse mo sees it, is that appearances, while occurring in Buddhas' conventional wisdom, are not the affiicted appearances of sal!lsara. 62 Bodhicaryavataratatparyapanjika Vise,yadyotanf, 260b7-261al: don

Making and remaking the ultimate

309

tion that bSod nams rtse mo sketched and rejected. Vibhuticandra's commentary, however, was composed later than all of the Tibetan commentaries discussed herein and so could not serve as bSod nams rtse mo's referent. 63 Vibhuticandra may be repeating an interpretation from an earlier Indian commentary on Santideva's text (which would likely require an alternate version of the Sanskrit that would allow for this interpretation), suggesting that Tibetan commentators may have had some Indian precedent for reading this passage as concerning a Buddha's perceptive abilities. At the very least, we cannot dismiss this Tibetan interpretation on the grounds of poor philology. Returning to the issue of Madhyamaka classifications drawn from this passage, it is clear that rGya dmar pa and bSod nams rtse mo refer to the same debate, portraying the two Madhyamaka positions in very similar terms and themselves taking the same side in the dispute. What are we to make of bSod nams rtse mo's addition of labeling the two positions "Prasangika" and "Svatantrika?" In the generation between rGya dmar pa and bSod nams rtse mo, Candraklrti's philosophy began to attract a strong following. Among the issues that Phya pa, who was bSod nams rtse mo's teacher of the Bodhicaryiivatiira, found the most distressing about this development were the "Prasangika" (although he did not use the term) portrayals of the ultimate and of buddhahood. We see Jayananda, the twelfth-century commentator on the Madhyamakiivatiira, interpret CandrakIrti's claims that the ultimate "is ineffable and just not a
dam par 'khor ba med pa fiid 'khrul pas 'khor ba'i man ba ni gti mug yod pa ji srid du'o II 'phags pa'i lam gyis gti mug spans nas de 'an med par 'gyur ro / 63 In his colophon (Bodhicaryiivatiiratiitparyapafijikii, 28Sa7), VibhUticandra states that he translated his text into Tibetan himself at 'BriIi mtshams, making it highly unlikely that it would have been known to bSod nams rtse mo, as Vibhuticandra made the first of his three trips to Tibet in 1204, in the company of SiikyasrThhadra, twenty-two years after bSod nams rtse mo died. On Vibhuticandra's travels, see Stearns, "The Life and Tibetan Legacy," 127-146. It remains possible that VibhUticandra wrote his commentary in Tibet and in this passage offered something of a response to bSod nams rtse mo.

310

Kevin Vose

referent of consciousness" and that Buddhas have no mind very literally.64 Since the ultimate is beyond awareness, Jayananda tells us, realization of the ultimate must entail the cessation of awareness and with it the ability to perceive appearances. 65 Phya pa argued at length against these positions, for reasons quite like bSod nams rtse mo's: Buddhas must have wisdom, wisdom is a type of awareness, and so awareness must continue in a purified form through the transformation to buddhahood; to hold otherwise, Phya pa says, one's position would resemble the nihilism of the Carvakas. 66 Phya pa, in turn, was criticized for his "reified ultimate," an ultimate that bears analysis and is accessible to awareness (although, as noted

64 La Vallee Poussin, Madhyamakiivatiira, 108,9-11: sans rgyas rnams la ni chos thams cad rnam pa thams cad du mnon par rdzogs par byan chub pa'i phyir I sems dan sems las byun ba'i rgyu ba gtan log par 'dod pa yin no I "We assert that for Buddhas, due to being manifestly and completely enlightened to all phenomena in all aspects, the movement of mind and mental factors has entirely ceased." Candraklrti makes similar statements in his discussion of buddhahood, particularly in his autocommentary to stanzas XII.8-9. 65 Jayananda, Madhyamakiivatiiratfkii, 146a7-146b1: ci yan thugs su chud pa med pa'i sgo nas byan chub pa'i phyir sems dan sems las byun ba'i rgyu ba gtan log par 'dod pa yin te Iiams su myon ba'i mtshan liid can gyi sems dan tshor ba la sogs pa sems las byun ba rnams kyi kun du spyod pa ste 'jug pa log par 'dod pa yin te I sems dan sems las byun ba rnams Jug pa ma yin no ies pa'o II des na ci yan snan ba med pa yin no ies pa'i tha tshig ste I rnam par rtog pa thams cad 'gag pa'i phyir ro I "Since enlightenment is by way of not knowing at all, we assert that the activities of mind and mental factors - feeling and so forth - having the character of experiencing, have ceased their engagement; there is no engagement of mind and mental factors. Therefore, there is no appearance at all, because all conceptuality has been blocked." Jayananda's views on ultimate truth and buddhahood are discussed in detail in Vose, Resurrecting Candrakfrti, 88-92 and 112-122. 66 Phya pa chos kyi sen ge, dBu ma sar gsum, 72,16-76,5. This passage is translated in Vose, Resurrecting Candrakfrti, 159-164 and analyzed at 122-131.

Making and remaking the ultimate

311

above, Phya pa holds the ultimate to be "utterly not established"), a position that Jayananda rejects as "Svatantrika."67 bSod nams rtse mo remains the earliest known author to pair the terms "Prasangika" and "Svatantrika," here using them to distinguish alternative visions of ultimate truth and buddhahood. His usage - and the absence of these terms in rGya dmar pa's discussionholds two significant consequences for our understanding of how Tibetans read Santideva. Prior to the widespread dissemination of Candraklrti's texts in the early twelfth century, Santideva was not a Prasangika. While rGya dmar pa refers to variant interpretations of Santideva (without calling these interpretations "Prasangika" or "Svatantrika"), he reads Santideva's views as consistent with those of Srlgupta and Jiianagarbha; and he seems to have had Indian precedent for doing SO.68 Further, even as Candraklrti's fame grew, Prasangika became only one option for reading Santideva; as bSod nams rtse mo's work evinces, Santideva could be read as a Svatantrika. Santideva became a Prasangika only after generations of Tibetan debate. We can recall that gTsan nag pa, writing in the same generation as bSod nams rtse mo, invoked Candraklrti's transcendent ultimate to explain Santideva's presentation; his "Great Madhyamaka" unites Candraklrti and Santideva around a transcendent reading of ultimate truth. Intriguingly, gTsan nag pa says nothing about the Prasangika-Svatantrika divide over a Buddha's ability to perceive appearances. Rather, his explanation of this passage in Santideva returns to a portrayal similar to Prajiiakaramati's: at issue is a misunderstanding on the part of the objector between natural nirviilJa and the attainment of nirviilJa. He has the objector ask,

67 Jayananda, Madhyamakilvatiiratfkii, 120a. 68 rGya dmar pa reads Santideva's portrayal of the two truths as consistent with that of Jiianagarbha (Tshig don gsal bar Mad pa, 127 [60aJ) and refers to SrIgupta to explain a Buddha's perception of ordinary appearances (Tshig don gsal bar Mad pa, 135 [64aJ). Saito notes the YogacaraMadhyamaka interpretation of the Bodhicaryiivatiira in two early Indian commentaries ("Santideva in the History of Madhyamika," 259).

312

Kevin Vose

If it is suitable that even natural nirviilJa is sa,!!siira conventionally, it is not contradictory that the attainment of buddhahood through making effort on the path also is sa'!!siira conventionally; thus, what would be the point of attaining a buddhahood that does not cast off the sufferings of sa'!!siira?69

gTsaJi nag pa quickly answers that Santideva's discussion teaches that upon the attainment of buddhahood, the causes of sarrtsiira are cut off.7 This brief answer and the failure to discuss the issue of a Buddha's perception - following three generations (if we include Phya pa) of commentaries that discuss the issue - suggests that gTsaJi nag pa's portrayal of a transcendent ultimate does not provide him with a good answer to rGya dmar pa's and bSod nams rtse mo's criticisms. An ultimate beyond the ken of the intellect, realization of which entails the cessation of awareness, provides no nuanced way of presenting how a Buddha's "conventional wisdom" arises. Rather than answering bSod nams rtse mo's critique of the PrasaJigika "blind Buddha," gTsaJi nag pa changes the subject. One senses that the early Tibetan tradition following Candrak'irti had to set aside the thorny issue of where an utterly transcendent ultimate left a Buddha's perception of ordinary appearances. The early evolution of Madhyamaka categories We have seen how Santideva's statements on ultimate truth and the distinction between sarrtsiira and nirviilJa gave rise to competing interpretations, which in turn served as criteria for classifying Madhyamikas. Santideva's ultimate could be "like an illusion" or "utterly non-abiding;" his buddhahood could be PrasaJigika or Svatantrika. One further discussion suggests how we might trace an evolution among these twelfth-century Madhyamaka categories. Rather than a division between "illusionists" and "utterly non-abiding-ists," Grags pa rgyal mtshan, bSod nams rtse mo's
69 gTsan nag pa, sPyod jug gi rnam Mad, 562,7-8 (38b7-8): ran biin gyis mya nan las 'das !ryan kun rdzob du 'khor bar run na lam la 'bad pas sans rgyas thob !ryan kun rdzob du 'khor ba mi 'gal bas 'khor ba'i sdug bsnal mi 'dar ba'i sans rgyas 'thob pas ci bya ies pa la I 70 gTsan nag pa, sPyod jug gi rnam Mad, 562,8 (38b8).

Making and remaking the ultimate

313

younger brother, writes of a division within the "utterly non-abiding" Madhyamaka position between "Continuum Cutting Utterly Non-Abiding [Madhyamikas]" and "Union Utterly Non-Abiding [Madhyamikas]."71 In his discussion of these two groups, instead of these unwieldy names Grags pa rgyal mtshan calls the former "PrasaiLgika" and the latter "Svatantrika." His explanation of the two categories maps very closely onto the divisions we have seen rGya dmar pa and bSod nams rtse mo make: Prasangikas hold that "Since Buddhas are always in meditative absorption, they only [have] ultimate [awareness];"72 ordinary awareness perceiving appearances has had its continuum cut. The Svatantrika position, which Grags pa rgyal mtshan endorses, instead holds that "Buddhas' non-conceptual minds are ultimate; as [their] pure worldly wisdom is supported by [non-conceptual] wisdom, it accords with the conventional."73 Grags pa rgyal mtshan's notion of 71 Grags pa rgyal mtshan, Rin po che'i !jon sin, 21.4,1 (42b1): rgyun chad rab tu mi gnas pa'i 'dod tshul dan / zun 'jug rab tu mi gnas pa'i 'dod tshul. 72 Grags pa rgyal mtshan, Rin po che'i !jon shin, 21.4,5-6 (42b5-6): dbu ma thaI 'gyur pa dag ni ... sans rgyas ni dus rtag tu mfiam par gtag pa yin pas don dam pa 'ba' zig go tes zer ro / This passage is set within a discussion of what kinds of awareness ordinary and enlightened beings possess (21.3,6-21.4,1 [42a6-42b1]: blo thams cad bsdu na blo kun rdzob pa dan / blo don dam pa tes bya ba gfiis yin pas na / blo gfiis po de gan zag gan gi rgyud la ldan ze na / "When encompassing all awareness, there are two: conventional awareness and ultimate awareness. Thus, one might ask, 'Which persons have these two awarenesses in their continua?"'). Grags pa rgyal mtshan goes on to list four positions (the "Hearer" position, the Yogiiciira position, the PriisaIigika position, and the Sviitantrika position), detailing where each school stands on the mental states of "ordinary beings" (so so skye bo), iiryas, and Buddhas. 73 Grags pa rgyal mtshan, Rin po che'i !jon sin, 22.1,1-3 (43a1-3): dbu ma ran rgyud pa dag ... sans rgyas kyi thugs rnam par mi rtog pa ni don dam pa yin la / dag pa 'jig rten pa'i ye ses ni ye ses la dmigs nas rnam grans kyi kun rdzob yin no / For a more substantial discussion of this passage that discusses Grags pa rgyal mtshan's views on "concordant ultimates" and "concordant conventionals," see Vose, Resurrecting Candrakfrti, 104-107.

314

Kevin Vase

"pure worldly wisdom" mirrors his brother's "conventional wisdom;" both show how Buddhas can continue to perceive conventional appearances along with their realization of emptiness. Grags pa rgyal mtshan further notes that both kinds of wisdom are necessary to explain non-abiding nirviilJa (aprati~rhitanirviilJa, mi gnas pa'i mya nan las 'das pa), in which Buddhas are both fully realized and fully able to aid sentient beings. 74 Grags pa rgyal mtshan's discussion suggests that, at least in one reckoning, Prasailgika and Svatantrika evolved out of the "utterly non-abiding position" and that neither could be mapped onto the "illusionist" position, which was widely rejected in bKa' gdams pa circles. Having cast aside the "illusionist" view, early bKa' gdams pa authors (and their Sa skya pa students) found significant enough philosophical differences to split further the Madhyamaka view. Despite a mutual adherence to the "non-abiding" of all phenomena, disputes over buddhahood produced, at least in part, the PrasailgikaSvatantrika divide.75 The only outlier category is gTsail nag pa's "Great Madhyamaka," interestingly posited by the one Candraklrti supporter in this group of commentators. Above, we saw that gTsail nag pa's "Great Madhyamaka" refers to Candraklrti's revivalists, whom others in this time period would call "Prasailgikas." The similarities between gTsail nag pa's views and those of Jayananda, as well as the positions that bSod nams rtse mo and Grags pa rgyal mtshan criticize (and label "Prasailgika"), strengthens this identi-

74 After stating the Prasailgika position, Grags pa rgyal mtshan notes, "That is not correct; it incurs the fault that it would absurdly follow that Buddhas would not enter non-abiding nirva;(ta" (Rin po che'i ljon sin, 21.4,6 [42b6]: de'an yan dag pa ma yin te / sans rgyas rnams ni mi gnas pa'i mya nan las 'das pa labiugs pa ma yin par thai 'gyur ba'i skyon yod do I). Following his statement of the Svatantrika position, he notes, "Thus, there is non-abiding nirvii(ta" (Rin po che'i ljon sin, 22.1,3-4 [43a3-4]: des na mi gnas pa'i mya nan las 'das pa'an yin la I). 75 This, of course, discounts those like rNog and Gro lUIi pa who rejected the "non-abiding" position out of hand. However, as noted above, both understood the "non-abiding" view quite differently from how it is portrayed by rGya dmar pa and Phya pa.

Making andremaking the ultimate

315

fication. Why then doesn't gTsaIi nag pa take up the "PrasaIigika" moniker? To sketch a tentative answer, we can note that bSod nams rtse mo's discussion of PrasaIigika and Svatantrika, when explicating Santideva's distinction between saJ?'lsara and nirva1'}a, skews the playing field against the former, equating it with the clearly problematic stance that awarenesS is cut off upon buddhahood. Rather than take up the "PrasaIigika" side of the debate on this issue and thereby attempt to defend a blind Buddha - gTsali nag pa ignored the dispute altogether. At least on this issue, the Svatantrika position had a clear advantage over Prasaligika in twelfth century Tibet?6 In contrast, "Great Madhyamaka" would represent gTsali nag pa's attempt to divide up Madhyamaka with Candraklrti on top. Rather than make a case for CandrakIrti's superiority on the issue of a Buddha's perception, gTsali nag pa uses CandrakIrti's portrayal of a transcendent ultimate, utterly distinct from mundane objects of knowledge, to valorize the "Great Madhyamaka" view. In the context of early Tibetan Bodhicaryavatara commentaries, these evolving Madhyamaka categories attempt to give the author's system top billing and to claim Santideva as their own. gTsali nag pa aligns Santideva's ultimate with CandrakIrti's, adopting him into Great Madhyamaka. The host of Svatantrika-Ieaning commentators instead nuanced Santideva's stanzas on ultimate truth and the saJ?'lsara-nirva1'}a distinction in order to forge a more coherent model of ultimate truth and those transformed by realization of it. rGya dmar pa and bSod nams rtse mo would have us believe that Santideva's "true thought" aligns with what would come to be called "Svatantrika." Santideva, then, was called many things by early Tibetan commentators. Fidelity to Santideva is difficult to judge given the poetic nature of his work. It may well be easier to square Santideva's statement that "the ultimate is not a referent of awareness" with a tran-

76 This is not to say that no Candraldrti partisans adopted the label "Priisaligika" in this period. Georges Dreyfus and Drongbu Tsering's article in this present volume shows that Pa tshab Ni ma grags indeed used the term "PriisaIigika" for his CandrakIrti-inspired Madhyamaka.

316

Kevin Vose

scendent reading, consistent with how Candraklrti's twelfth century revivalists portrayed his views. However, reading Santideva as advancing views similar to those of Srlgupta, Jfianagarbha, Santarak~ita, or KamalaSIla held historical credibility and remained a viable option for early Tibetan commentators. We see in these commentaries a variety of ways to reconcile Santideva's seemingly transcendent ultimate with his broader project of mapping the practices of enlightenment. Harmonizing the ultimate with the path leading to its realization is an issue at the very core of Mahayana Buddhism; harmonizing realization of the ultimate with the kind of epistemological program that gSari. phu Ne'u thog was known for would remain a defining feature of Tibetan Buddhist scholasticism. Early Tibetan Bodhicaryiivatiira commentaries provide a glimpse of how these processes played out in the making and remaking of one of the preeminent figures of Buddhist India.

References
Sanskrit and Tibetan works cited
Grags pa rgyal mtshan. rGyud kyi mnon par rtogs pa rin po che'i ljon sin. In Sa skya pa'i bka' 'bum, volume 3. Tokyo: Toyo Bunko, 1968. Gro lun pa BIo gros 'byun gnas. bsTan rim chen mo. Mundgod: Library of His Eminence Trijang Rinpoche, 2001. Also, bKa' gdams gsun 'bum, volumes 4 and 5. rGya dmar pa Byan chub grags. Byan chub sems dpa'i spyod pa la 'jug pa'i tshig don gsal bar Mad pa. In bKa' gdams gsun 'bum, volume 6,11-175. Jayananda. Madhyamakilvatilratfkil (dBu ma la 'jug pa'i 'grel Mad). bsTan 'gyur, sDe dge edition, Tohoku 3870; dbu ma, volume ya. bKa' gdams gsun 'bum phyogs bsgrigs. Chengdu: dPal brtsegs bod yig dpe riii:f1 fib 'jug khan, 2006. rNog Blo ldan ses rab. sPrin yig bdud rtsi thig Ie. In bKa' gdams gsun 'bum, volume 1,707-710. Phya pa Chos kyi sen ge. dBu ma sar gsum gyi ston thun, edited by Helmut Tauscher. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien Universitat Wien, 1999. - bDe bar gsegs pa dan phyi rol pa'i giung mam par 'byed pa. In bKa' gdams gsun 'bum, volume 9, 7-73.

Making and remaking the ultimate

317

- sPyod jug bsdus don. In bKa' gdams gsun 'bum, volume 7, 131-145. Prajfiakaramati. Bodhicaryavatarapafijika. See La Vallee Poussin. Santideva. Bodhicaryavatara (Byan chub sems dpa'i spyod pa la jug pa). bsTan 'gyur, sDe dge edition, Tohoku 3871; dbu ma, volume ya. bSod nams rtse mo. sPyod pa la jug pa'i 'grel pa. In Sa skya pa'i bka' 'bum, volume 2. Tokyo: Toyo Bunko, 1968. gTsan nag pa brTson 'grus sen ge. sPyod jug gi rnam Mad. In bKa' gdams gsun 'bum, volume 13, 487-581. VibhUticandra. Bodhicaryavataratatparyapafijika Vise~adyotanf(Byan chub kyi spyod pa la jug pa'i dgons pa'i 'grel pa khyad par gsal byed). bsTan 'gyur, sDe dge edition, Tohoku 3880; dbu ma, volume sa.

Western language works cited

Cabez6n, Jose Ignacio. "The Madhyamaka in Gro lung pa's Bstan Rim chen mo." Proceedings ofthe Eleventh Seminar ofthe International Association of Tibetan Studies. Forthcoming. Kano, Kazuo. rNgog BIo-ldan-shes-rab's Summary ofthe Ratnagotravibhiiga: The First Tibetan Commentary on a Crucial Source for the Buddhanature Doctrine. Ph.D. Dissertation, Hamburg University, 2006. Kramer, Ralf. The Great Translator: Life and Works of rNgog BIo ldan shes rab (1059-1109). Munich: Indus Verlag, 2007. La Vallee Poussin, Louis de. Prajfiakaramati's Commentary to the Bodhicaryavatara of 9antideva, Bibliotheca Indica. Calcutta: Asiatic Society of Bengal, 1905-1914. - Madhyamakavatara par Candrakfrti, Bibliotheca Buddhica 9. Osnabriick: Biblio Verlag, 1970. Mathes, Klaus-Dieter. "Blending the SUtras with the Tantras: The Influence of Maitrlpa and his Circle on the Formation of Sutra Mahamudra in the Kagyu Schools." In Tibetan Buddhist Literature and Praxis: Studies in its Formative Period 900-1400, Proceedings of the Tenth Seminar of the International Association of Tibetan Studies, Oxford 2003, volume 4, edited by Ronald M. Davidson and Christian K. Wedemeyer, 201-227. Leiden: Brill, 2006. Napper, Elizabeth. Dependent-Arising and Emptiness. Boston: Wisdom Publications, 1989. Oldmeadow, Peter R. A Study of the Wisdom Chapter (Prajfiapiiramita Pariccheda) of the Bodhicaryavatarapafijika of Prajiiakaramati. Ph.D. Dissertation, Australian National University, 1994. Ruegg, David Seyfort. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy, Studies in Indian and Tibetan Madhyamaka

318

Kevin Vose

Thought, Part One. Vienna: Arbeitskreis flir Tibetische und Buddhistische Studien Universitat Wien, 2000. - Two Prolegomena to Madhyamaka Philosophy, Candrakfrti's Prasannapadil Madhyamakavrttif:t on Madhyamakakilrikil 1.1 and TSOli kha pa Blo bzan grags pa / rGyaltshab Dar ma rin chen's dKa' gnad/gnas brgyad kyi zin bris, Annotated Translations, Studies in Indian and Tibetan Madhyamaka Thought, Part 2. Vienna: Arbeitskreis flir Tibetische und Buddhistische Studien Universitat Wien, 2002. Saito, Akira. "Santideva in the History of Madhyamika." In Buddhism in India and Abroad, edited by Kalpakam Sankarnarayan, Motohiro Yoritomi, and Shubhada A. Joshi, 257-263. Mumbai and New Delhi: Somaiya Publications, 1996. - "Bu ston on the sPyod 'jug (Bodhisattvacaryilvatilra)." In Tibetan Studies: Proceedings of the 7th Seminar of the International Association for Tibetan Studies, edited by Helmut Krasser, Michael Torsten Much, Ernst Steinke1lner, and Helmut Tauscher, 79-85. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften, 1997. -A Study of the Dun-huang Recension of the Bodhisattvacaryiivatiira. Mie: Mie University, 2000. Stearns, Cyrus. "The Life and Tibetan Legacy of the Indian Mahiipar:zita Vibhuticandra." Journal of the International Association of Buddhist Studies 19, no. 1 (1996): 127-161. Vose, Kevin. Resurrecting Candrakfrti: Disputes in the Tibetan Creation of Prilsangika. Boston: Wisdom Publications, 2009. Williams, Paul. "On Prakrtinirvilr:za/Prakrtinirvrta in the Bodhicaryiivatiira: A Study in the Indo-Tibetan Commentarial Tradition." Etudes Asiatiques 46, no. 1 (1992): 516-550.

The origin of the theory of definition and its place in Phya pa Chos kyi sen ge's philosophical system* Pascale Hugon

Introduction Epistemology had already been considered a significant element of the Buddhist corpus at the time of the Early Diffusion (sJia dar) of Buddhism in Tibet, as the first and partial translations of the works of the Indian forefathers - Dignaga (ca. 480-540) and Dharmaklrti (ca. 600-660) - and of their commentators testify.l Though we do not know much about the "old epistemology" (tshad rna rfiiJi rna) of the early years of the Later Diffusion (phyi dar),2 the relevance of this field for Tibetan scholars was confirmed with the

* Work on this paper has been generously supported by the Austrian Science Fund (FWF) in the context of the FWF-Projekt P19862 "Philosophische und religiOse Literatur des Buddhismus." I wish to thank Cynthia Peck-Kubaczek for correcting my English, and Kevin Vose who provided a number of relevant suggestions and references. 1 This is demonstrated notably by the existing translations and translations' in progress of epistemological works recorded in the IHan dkar catalogue; cf. Lalou 1953. 2 The "old epistemology" is associated in the Blue Annals (Deb ther sflOn po 64,1-2, 65,1 and 83,2) with the name of Khyun po Grags se who, according to 'Gos Lo tsa ba gZon nu dpal, composed several works on the subject, and with the translation of Dharmaklrti's PramalJ-avarttika and in particular that of Devendrabuddhi's PramalJ-avarttikapafijikti by rMa Lo tsa ba dGe ba'i blo gros. See also van der Kuijp 1983: 1-3.

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 319-368

320

Pascale Hugon

rise of the "new epistemology" (tshad ma gsar ma) associated with the endeavors of the translator and scholar rNog BIo ldan ses rab (1059-1109), alias rNog Lo tsa ba (hereafter: rNog Lo). Along with his contribution to the .translation (or re-translation) of the founding treatises together with their commentaries, rNog Lo initiated through his exegeses a line of interpretation that prevailed until the 13th century and made the monastery of gSari. phu Ne'u thog the leading institution of the period for the study of epistemology. Recent findings, including in particular the material published in the bKa' gdams gSU1i 'bum, allow us to evaluate anew the contributions of rNog Lo and his successors, among them the famous Phywa (/phya) pa Chos kyi sen ge (1109-1169; hereafter: Phya pa), who stands as a key figure in the development of a genuine Tibetan epistemology. In the available works one can trace various factors as well as distinct phases that account for the transformation of Indian pramii;~ta into Tibetan tshad ma. The process through which the imported Indian corpus was then appropriated is characterized by a significant amount of "re-formatting" of the source material, both in form and contents, as well as by significant innovations. This paper deals with the early developments of one of these innovations, which came to occupy a significant place in Tibetan epistemology: the theory of definition. While DharmakIrti did not offer a methodical discussion on the topic of definition per se3 - as Indian epistemology in general did not - the theory of definition, which deals essentially with the nature of and relation between the elements of a definition and with the rules that warrant a correct definition, figures as a core element
3 Although DharmakIrti deals with the definition of numerous notions, he does not discuss the terms of definition themselves or their relations. The closest he gets to giving rules of definition is his discussion of the definition of the thesis (pak~a) in PV IV.85-86 (=PVin ill.23-24). There he analyses Dignaga's defining characteristics as ways to rule out nonpervasion (avyapti) and excessive extension (ativyiipti). He also mentions theses that are invalid due to the absence of the defining characteristic (alak~a1Javrtti). This passage is quoted, for instance, by Sa skya Pru;,:t~ita to support his threefold division of the faults pertaining to the definiens (Rigs gter 190,15-26).

The origin of the theory of definition

321

of Tibetan epistemological treatises since Phya pa at least.4 Even Sa skya PaI).Q.ita Kun dga' rgyal mtshan (1182-1251), an author prompt to denounce "self-fabricated" (rali bzo) Tibetan theories, adopts it as a bona fide part of his epistemological system. The popularity of the theory of definition is noted by gSer mdog pal). chen Sakya mchog ldan (1428-1507) who reckons that by the 15th century it had become "like milk at the market place."s However widespread the theory became, not all Tibetan thinkers shared the same enthusiasm for this innovation, which some saw as a mere sterile proliferation. 6 Tibetan presentations of the theory of definition are indeed of substantial size - such presentation already occupies one fifth of Phya pa's Resume of Epistemology - and tend to expand over time insofar as subsequent authors add their own contribution after presenting - and if need be, refuting - preceding positions on the subject. While this leads to massive and often complex expositions, this also allows us to trace the evolution of the views on the subject and to fill in some blanks (with due caution) regarding the position of thinkers whose works are not available. My intention here is not to go through a detailed account of the Tibetan theory of definition and all the subtleties it involves?
4 The earliest occurrence that is currently available of this theory in a fully developed and systematic form is found in the epistemological works ofPhya pa (see below). Sakya mchog ldan ascribes to him the very origin of this theory (see dGa' byed 18b6-7, cited and translated in van der Kuijp 1983: 77-80) and affirms that it was unknown from rNog Lo up to rGya dmar pa (see Pham byed 135b7, translated in van der Kuijp 1983: 66). There is however, as I will discuss below, evidence of Tibetan contributions prior to Phya pa. S dGa' byed 18b7: phyi ma'i dus 'dir rnam giag de 'tshari 'dus kyi 'a ma Ita bur gyur. Cited and translated in van der Kuijp 1983: 79-80. 6 Go rams pa bSod nams sen ge (1429-1489) mentions for instance complaints in this regard by dPaIi Lo tsa ba (probably dPaIi Lo tsa ba Blo gros brtan pa [1276-1342]). Cf. Rigs gter gSal byed 43b3-4, translated in van der Kuijp 1983: 67. 7 A basic study of the nature of the elements of definition and their relationships based on gTsaIi nag pa's bsDus pa and on the De kha na

322

Pascale Hugon

My goal in this paper is rather to examine possible sources for the elaboration of the specific model involved in the Tibetan theory. Indeed, while a definition commonly involves two elements, what is defined and what defines it - as in "man is a rational animal" or "water is Hp" - the particularity of the Tibetan model is that it is conceived as involving not two, but three elements: alongside the definiens (Tib. mtshan fiid) and the definiendum (Tib. mtshon bya), one posits a "definitional basis" (mtshan gii), whose function will be discussed below. This triad - often given in the abbreviated form: mtshan/mtshon/gii - can be illustrated by the stock-example of the definition of cow, which involves: "cow" (the definiendum), "having a hump, a dewlap, etc." (the definiens), and "(a) white calf" as the definitional basis. What motivated Tibetan thinkers to go from a two-term model - well established in Indian texts with the pair lak~afLa/lak~ya - to a three-term definition? My point of departure in an attempt to answer this question will be the hints given in the tradition itself that were pointed out by van der Kuijp in his introductory assessment of the emergence of the theory of definition, 8 hints which we are now in a position to better evaluate thanks to newly available material. In particular, I will consider remarks hinting at the role of the translation of the relevant terms and discussions of "lak~afLa" in the context of the Abhisamayiilankara-related literature. I will then argue that, although it is not explicitly acknowledged by the Tibetans, one can trace a possible influence coming from Indian epistemological works and specifically from Dharmottara's writings. Further, I will reflect on the place and role of the theory of definition in Phya pa's works. Although in his epistemological writings Phya pa ascribes it the role of a prolegomenon to the definition of valid cognition (Tib. tshad ma), I will suggest that Phya pa's con~ern with the theory of definition is not so much linked with

fiid bsdus pa can be found in Fukuda 2003. As Phya pa's epistemological works were not available at the time, Fukuda only refers to Phya pa's sTan thun. Onoda 1984 deals with the question on the basis of later bsdus grwa literature. 8 Van der Kuijp 1983: 66-67.

The origin of the theory of definition

323

defining valid cognition - or for that matter, with defining cows or trees - but rather addresses questions related to specific issues in Madhyamaka philosophy. A summary of Phya pa's position Before proceeding to trace the origin of the specifics of the Tibetan model, let me provide a brief overview of Phya pa's position. Phya pa gives a full-fledged presentation of the theory of definition in both his Epistemology, the Dispeller of the Darkness ofMind (Tshad ma yid kyi mun sel, hereafter: Mun sel) - a work also known as his Resume of Epistemology (Tshad ma'i bsdus pa) - and in his extensive commentary on Dharmaklrti's Pramiir;aviniicaya (hereafter: 'Od zer).9 Phya pa's presentation comprises three main sections:
1. An analysis of the type of distinction existing between the elements of a definition (Mun selllb6-16b7, 'Od zer 6bl-lOa3)

II. Their respective definitions (Mun sel 16b7-27aS, 'Od zer lOa4ISb7)

m. The faults related to each element when stating a definition (Mun

seI27aS-32a2, 'Od zer ISb7-20b3)

What I call "theory of definition" is introduced in these texts in terms of "mtshan Rid kyis mtshon bya mtshon pa'i tshul" ("how to apply [lit. to define] a definiendum by means of/due to a definiens"). This formulation veils a significant particularity of the Tibetan model, which is, as I noted in the introduction, the presence of a third element: the definitional basis (mtshan gii). This notion is defined minimally as being the basis for the other two elements. Its role can be better understood in the light of the following contexts: One can notice first that when discussing specific definitions (the definition of valid cognition, of a correct logical reason, of
9 See respectively Mun sel lla6-32a2 and 'Od zer 6aS-20b3. Phya pa's presentation of definition in these two texts follows roughly the same outline, but surprisingly one finds few identical passages; although the ideas correspond, examples and arguments are formulated for the most part with different wording.

324

Pascale Hugon

ultimate truth, etc.), the author strives not only to establish a proper definiens for a given definiendum, but also to identify "what kind of things" are likely to have these definiens and definiendum. For example, when defining "ultimate truth," Phya pa not only aims at identifying the criterion for being "ultimately true," but also at identifying what can be grasped as ultimate truth (is it a simple negation, an implicative negation, both, neither?). In unproblematic cases, this identification amounts to giving an example or an illustration: a pot when defining causal efficiency, a si7Jlsapii when defining tree, etc. Next, in the context of the presentation of the theory of definition, the definitional basis appears as a specific case to which a given definiendum applies, or must be shown to apply, owing to the relevant definiens. "Defining" does not refer here to the association of a definiens and a definiendum (although this must be presupposed), but rather to the application of a definiendum to a definitional basis.10 The definiens is presented as a justification for its application (or applicability). The statement of a definition can thus be seen as an argument that establishes that the definiendum . applies to the definitional basis, such as for the definition of "cow:" "[this] white calf is a cow, owing to the hump, dewlap, etc." The definition of "man" according to this model would amount to stating or establishing, for instance, that "John is a man" owing to John's having the definiens of man, "being a rational animal." The relation between this type of formulation and inference will be discussed below. The definiendum (mtshon bya), for its part, is described as a "convention" (tha sfiad, Skt. vyavahiira). What constitutes a "convention" is identified by Phya pa using terms borrowed from the theory of exclusion (apoha), which, in DharmakIrti's system, gov-

10 Phya pa presents a more detailed three-step procedure leading to the application of the definiendum to the definitional basis when he deals with the definition of a logical reason that is a property of the subject (pak~adharma). See Mun sel 43b2-6 and 'Od zer 82a5-7, that repeat rNog Lo's presentation in dKa' gnas 214,18-216,3.

The origin of the theory of definition

325

ems the formation and application of concepts and words. He thus presents the definiendum as:
Fit to be determined by a word as unmixed with all that has a property directly contradictory to the definiens, or fit to be determined by a word that is a non-erroneous exclusion of what is other that eliminates all that is mutually contrary with the definiens.11

The application of a definiendum hence presupposes a process of exclusion involving two complementary classes generated in function of the definiens. The convention applies to the elements of one class by way of excluding members of the complementary class. It can, but need not, be associated with a word. However, it does not itself have the nature of a "name" (min)P By defining, i.e., applying a convention to a definitional basis, one situates the element under consideration within a framework of worldly customs that include - but are not limited to - transactional usage and linguistic denominations.13 The definiens (mtshan i'lid) is presented as the "motive for the application of the convention" (tha sfiad kyi rnam 'jog gi rgyu mtshan); it is also the "motive for eliminating the heterogeneous, i.e., the opposite" ('gal zla rigs mi mthun gcod pa'i rgyu mtshan),

11 Mun seI25b7-8: des na mtshan fiid dan dnos 'gal gyi chos dan ldan pa mtha' dag dan ma 'dres par sgras zen du run pa 'am I mtshan fiid dan phan tshun spans 'gal mtha' dag sel pa gzan sel phyin ci log (read: ma log) pa'i sgras zen du run pa'o II 12 On this point, see the discussion in Mun seI25b6-7. 13 Knowing the convention "cow" may involve knowing the word that is to be applied to animals which have a hump and a dewlap ("go" in India, "ba Ian" in Tibet, etc.), and conversely, knowing what to look for when being asked to bring a "cow." It also involves knowing "what to do" with such things in a specific situation. Note however that defining is not limited to entities (dnos po) endowed with causal capacities; definitions are also applicable in the realm of non-entities (dnos med) (see n. 18) and for second-order properties: conventions themselves can become definitional bases in a definition. For instance, one can apply to "the convention 'cow'" the definiendum "being a definiendum."

326

Pascale Hugon

i.e., the motive for the above-mentioned division into two c1asses.14 One can say that it plays the role of a sufficient and necessary condition for the applicability of the convention. is Note that while the theory of definition provides the criteria that allow the "testing" of a putative definiens, it does not give a procedure to identify the definiens involved in an established convention.16 The first section of Phya pa's theory of definition is dedicated to an analysis as to how these three elements are distinct (tha dad)17
14 In 'Od zer 6a9, Phya pa describes the definiens as "the motive for eliminating the heterogeneous, i.e., the opposite" ('gal zla rigs myi mthun gcod pa'i rgyu mtshan) and the definiendum, reciprocally, as "that which is to be determined as specific after having distinguished the heterogeneous" (rigs myi mthun las phye nas thun man ma yin par zen par bya ba). 1S

See Mun seflla7: yod na ... tha sfiad jug fa med na mi jug pa ...

16 That the identification of the definiens for someone learning a convention pertaining to entities such as cows or trees should always be spontaneous is contradicted by the case, which will be discussed below, of the "dimwit" who has not been able to grasp the definiens of tree (or cow) properly. 17 Phya pa considers four options as to how two elements (x and y) can be distinct (tha dad):

1. dflOS po tha dad: x and yare causally efficient entities and mutually empty, such as a pot and a pole. 2. gcig pa bkag pa'i tha dad: x and yare distinct insofar as they are not one. One rejects the identity of the two, but as they do not exist conjointly, they are not, properly speaking, "distinct entities," as for example an actual pot and the pot in the preceding moment. 3. tha dad med pa la rnam gralis kyis brjod pa: x and yare distinct only in terms of expression, but there is only one concept (ldog pa). In such a case, we are dealing with synonyms, as for example sili and ljon pa (both meanVIg "tree"). 4. lio bo cig gi ldog pa tha dad: x and yare distinct concepts for a unique, undifferentiated nature. Their distinctiveness is expressed in terms of being separated (phye ba) from distinct eliminanda. This echoes the formulation "excluded from different bases of exclusion" that one finds in the apoha theory. The classical example is that of "produced" and "impermanent," two properties of the pot, which correspond to two ways of conceiving the pot, namely, as something that is excluded from what is non-produced, and as something that is excluded from what is permanent.

The origin of the theory of definition

327

with the conclusion that the definitional basis, definiens and definiendum have an undifferentiated nature (no bo dbyer med),18 but are distinct concepts, or "distinguishers" (ldog pa tha dad), as exemplified by the properties "produced" and "impermanent" in the case of a pot. This distinction, in the case of the definiens and the definiendum, is paired with their characterization as beiIlg, for the first, "substantially existent" (rdzas yod, dravyasat) and for the second "nominally existent" (btags yod, prajfiaptisat),19 and with the characterization of their instances as being, respectively, a state of affairs (don, artha) and a convention (tha sfiad, vyavahara). In the second section, the theory of definition is applied autoreferentially, that is, the definiens and the suitable definitional bases are identified for each of the three elements. This leads to complicated discussions, particularly in the case of "definiens," focusing on the questions: "Does a definiens require a definiens?" and "if it does, would this lead to infinite regress?" Against some of his predecessors, who contend that a definiens does not require a definiens, Phya pa gives as the definiens of a definiens a set of three properties (chos gsum) that address its nature and its relation with the two other elements: the definiens must be substantially existent, it must not be "different in meaning" (don ldog gian) from the

Cf. Mun selllb7-8 and 12b9-13al. 18 This undifferentiated nature can be either real or unreal, cf. Mun sel 12b4-5: des na gii dfws po yin na mtshan fiid dan mtshon bya dnos po cig yin la gii dnos med yin na'an de gfiis tha dad med pa yin pas mtshan fiid dan mtshon bya no bo dbyer med pa la ldog pa tha dad du gnas pa yin no II Parallel in 'Od zer 7b1-2. 19 These notions are defined in terms of the comprehension depending or not depending on the apprehension of another property as being the cause for its establishment. Cf. Mun sel12b5-6: ldog pa'an gfiis te rdzas su yod pa dan rtags [=btags] su yod pa'o I rdzas yod ni chos gian rnam 'jog gi rgyur bzun pa la ma ltos par rtogs par bya ba ste bden pa rtogs pa lasogs pa don gyi chos rnams so II btags yod ni chos gian rnam 'jog gi rgyur bzun pa la ltos nas rtogs par bya ba ste tshad ma'i tha sfiad lasogs pa tha sfiad kyi chos rnams sam tha sfiad des khyad par du byas pa'i chos rnams so I Parallel in 'Od zer IOb3.

328

Pascale Rugon

definiendum,20 and it must occur for the definitional basis. 21 This "triple property" amounts to the criterion that warrants a correct definition and serves to delineate the faults of the definition described in the third and last section. Possible sources for the early developments of the theory of definition Although the Tibetan theory of definition does not replicate an existing Indian model, one can reflect on the possible infLuence(s) that led Tibetan thinkers to develop a triadic model of definition involving the particularities mentioned above. As pointed out by van der Kuijp, the Tibetan tradition provides some hints in this regard: dPaIi Lo tsa ba, whom Go rams pa mentions as a detractor of the theory of definition, claims that all these discussions on definition arose due to inconsistencies in the translation of the Sanskrit terms involved; what were originally different translations of the same term became distinct elements in the definition.22 According to 20 This criterion has the goal (in addition to the elimination of putative definiens that would be contrary to the definiendum, too broad, or too restricted) of eliminating cases of properties that, although co-extensive with the definiendum, are "different in meaning," a notion that Phya pa clarifies by explaining that they are "not directly contradictory to the heterogeneous class" (Mun sel 28('og)a3: rigs mi mthun dan dnos 'gal ma yin). Hence, for instance, "causally efficient" is held not to be the definiens of "impermanent" because it is not directly contradictory to "permanent." The proper definiens of "impermanent" is "not remaining for a second moment" (dus gfiis par mi sdod pa). 21 Mun se117a4: des na mtshan fiid kyi mtshan fiid ni I chos gsum dan ldan pa ste I ran ldog nes pa rnam 'jog gi rgyu mtshan la mi ltos pas rdzas yod yin pa dan I mtshon bya las don gyi ldog pa gian du mi gnas pa dan mtshan gii la ldan pa'o II 22 See n. 6 for the reference. As noted by van der Kuijp (1983: 67 and n. 249), the passage quoted by Go rams pa mentions that dPang Lo tsa ba points to the inconsistency in the translation of the term "lak~a-t:ta," whereas, as will be discussed below, it is rather the term "lak~ya" that is inconsistently translated as either mtshan gii or mtshon bya. Sakya mchog Idan points out that the expression "mtshon par bya ba" (i.e., lak~ya) has

The origin ofthe theory of definition

329

Sakya mchog ldan, the theory of definition was well known thanks to commentaries on two verses of the Abhisamayiilankiira (AA), namely AA IV.13 and IV.3l. 23 Following Sakya mchog ldan's hint to the existence of discussions of definition outside the scope of epistemological treatises, I consider below the relevant passages of the AA and a number of Tibetan commentaries thereupon, which additionally will provide a starting point to evaluate dPan Lo tsa ba's claim.

Abhisamayiilankiira literature
The two verses mentioned by Sakya mchog ldan occur in the fourth section of the AA, which deals with the lak~aT.ta (meaning here a specific mark or characteristic feature) of the training of a Bodhisattva. 24 Verse AA IV.l3 states:
two uses - it can indicate either the mtshon bya (for instance the verbal convention "cow") or the mtshan gii (e.g., the white calf) - but also mentions the double use of the word "mtshan fiid" (lak~a1}a), not however corresponding to mtshan gii and mtshon bya, but to "that by which one defines" (gail gis mtshon par byed pa) and "the basis on which it is to be defined" (gail la mtshon par bya ba'i gii); that is, it would corespond to the mtshon byed and to the mtshon gii. See Pham byed 122a7-b6. 23 Sakya mchog ldan affirms: ''Although there is no source for formal applications of definition in the texts of the author of the Seven Treatises, relative to the instances in the commentarial text of the Noble Maitreya in the verses [... (see below)] and to the elaboration in the commentaries on this it was clearly known to all logicians of the Snow Land from the Lord of Logic Phya pa hitherward (tshun chad)*." (Pham byed 135b5-7; cited in van der Kuijp 1983: 66) * I take "tshun chad" to express a timeframe from the point of view of Sakya mchog ldan. While Sakya mchog ldan claims in the passage that follows that the theory of definition was not known between the time of rNog Lo and that of rGya dmar pa (who was one of Phya pa's teachers), he mentions elsewhere some views on definition by Phya pa's predecessors. Van der Kuijp (1983: 66-67) suggested that the latter could have dealt with definition in the context of Abhisamayiilailkiira commentaries whereas Phya pa was the first to deal with it in the context of epistemology. 24 This fourth section deals with the Perfection of Wisdom consisting

330

Pascale Hugon

A mark (lak~WLa) is to be known as that by which [something] is characterized (lak~yate). And it is threefold: i) the mark of cognition, ii) the distinctive mark, and iii) the mark of the activity. What is characterized is the nature (svabhava) [of the training of a Bodhisattva].2s

In verse IY.31b-d, the "nature" that is being characterized is presented as constituting itself a fourth kind of characteristic:
The nature has a sixteen-fold disposition; since [it] is characterized as if it had [these sixteen as] a mark, [it] is admitted as a fourth mark. 26 in "enlightenment concerning all the aspects" (sarviikiiriibhisambodha), the so-called aspects being topics in the understanding of entities. After presenting the aspects - divided into three according to the three kinds of omniscience (omniscience simpliciter [sarvajiiata], knowledge of the path [margajiiata], and knowledge of all aspects [sarvakarajiiata]) - the author deals with the "endeavors" (prayoga) or training aimed at obtaining this particular enlightenment. He describes the suitable endeavors, the method of training (listing twenty endeavors), followed by the virtues, faults and features relative to the training. The lak~WLa are here marks, or characteristic features, of the training, rather than a definiens strictly speaking. "Characterization" differs from "definition" in that the mark that is stated is not specific to the given lak~ya. Sakya mchog Idan states this difference in Pham byed 122b2-3 in the following terms: the characterization (mtshon par byed) can be either specific (thun mon min par) or non-specific (thun mon du); for instance, "heat" characterizes "fire" specifically (i.e., it is its definiens), but "impermanence" characterizes the Truth of suffering non-specifically (i.e., although it is a characteristic feature of this Truth, it is not its definiens). 2S AA IV.13, p. 19: lak~yate yena taj jiieyaY(! lak~a1JaY(! trividhaY(! ca tat I jiiiinaY(! vise~aJ:z karitraY(! svabhavo yas ca lak~yate II (Tib. p. 35: gan gis* mtshon de mtshan fiid du II ses bya de yan rnam pa gsum II ses pa khyad par byed pa ste II no bo fiid gan mtshon bya yan II). *The text edited in the sDe dge edition (7a6) has the faulty reading "gi." 26 AA IY.31, p. 22: [tathatanupalambhas cal svabhavaJ:z ~o4aatmakaJ:z I lak~mevallak~fva* lak~yate ceti caturthaY(!** lak~a1JaY(! matam II (Tib. p. 40: [de biin iiid ni mi dmigs daTi II] no bo fiid bcu drug bdag fiid II mtshon bya Ita bur* mtshon pas na II mtshan iiid bii par bied pa yin II) *Obermiller and Stcherbatsky emend to "lak~meva" the reading "lak~mfva" found in the two mansucripts of AA. Wogihara likewise adopts this emendation in his edition of Haribhadra's AbhisamayalaTikaraloka (AAA)

The origin of the theory of definition

331

This passage of the AA clearly evokes a two-term model of characterization involving something that characterizes - the lak~m:La - and something that is characterized by it (tena lak~yate).

against the readings "lak:jmlva" and "lak:jlva" found in the manuscripts. The manuscript of the Abhisamayiilmikiirakarikiisiistravivrti (AAKSV) edited by Amana reads "lak:jlva." My translation follows Haribhadra's gloss: "The endeavor regarding the three kinds of omniscience is characterized by these sixteen natures like if it had [them as] a mark; therefore one admits [as] a fourth [mark] the mark of the nature" (AAKSV 67,12-13: ity eval!l :jorf.asasvabhiivair yathiivat trisarvajfiatiiprayogo lak:jalJaviin iva lak:jyata ity eval!l caturthal!l svabhiivalak:jalJaf!l matam iti). This gloss in terms of "lak:jalJaviin iva" (Tib. 113a5: mtshan fiid dan [dan pa biin du) would support the reading "lak:jlva." Regarding the Tibetan translation of the AA, the Tibetan block-prints of Aga monastery consulted by Obermiller and Stcherbatsky read "mtshon bya Zta bur." So does the Tibetan translation of the AA in the sDe dge edition (Sa2) and that of the AAKSV in the Peking and Narthang editions (Amana 2000: 66) whereas the sDe dge edition of the latter reads "mtshon cha Ita bur" (112b6), a reading also found in the Tibetan edition of the AAA used by Wogihara. The Tibetan "mtshon bya" is attested as a translation of "lak:jman," but the latter has the sense of a "mark" rather than of what is being marked. In view of the fact that rNog La is responsible for the translation of AA as well as for the revision of both AAA and AAKSV; it is possible that his choice of "mtshon bya" reflects the identification of "mtshon bya" and "mtshan fiid dan ldan pa." As an alternative to "mtshon bya lta bur," one finds, reflecting explicitly Haribhadra's understanding, "mtshan fiid can ltar" in the Tibetan translation of this same verse quoted in a commentary on the A:jtasiihasrikii by Abhayakaragupta (late 11th-early 12th c.) and translated by himself (Aryii:jtasiihasrikiiprajfiiipiiramitiivrttimarmakaumudf, Tib. 146b4). As further alternative, AA IV.31 is cited with the reading "mtshan gii lta bur" in Ratnakarasanti's Aryii:jtasiihasrikiiprajfiiipiiramitiipafijikiisiirottamii (Tib. 142a6), a work translated by SubhTIti(srI)anti and Sakya blo gros around the middle of the 11th century. This translation is reflected also for instance in the commentary by sGros ruin pa (13th c.), who glosses the expression by "mtshon par bya ba'i mtshan gii ita bu" (mNon par rtogs pa'i rgyan gyi rnam par 'byed pa'i 'grel ba 262,3) . * I follow the reading of this verse in AAA and AAKSV AA's edition alone reads "caturthalak:jalJaf!l."

332

Pascale Rugon

As no less than ten Tibetan commentaries on the AA have recently become available in the first thirty volumes of the bKa' gdams gsuri 'bum, we have ample opportunity to evaluate what Tibetan thinkers made of this discussion. 27 Commentaries that postdate Phya pa clearly demonstrate the popularity of the triadic model of definition. For instance, sGros rfiiil pa 'Od zer mgon po and gZon nu rin chen (both writing around the 13th century or later) analyze this passage of the AA using a model involving mtshan fiid, mtshon bya and mtshan gii,28 If these authors make use of this model, they do not however provide at this occasion an extensive theoretical presentation of its specifics. 29
27 I mention here by way of reference the name of the author, a date when available, and the volume and text number as given in the dkar chag. rNog Lo (1059-1109) I.13; Ar Byan chub ye ses (11th c.) II.9; Khu Ses rab brtson 'grus (1075-1143) X.4; 'Jam gsar ba Ses rab 'od zer XIY.5; sGros ruin pa 'Od zer mgon po (13th c.) XV.4; 'Chus Dar rna brtson 'grus (1117-1192) XIX.3; gZon nu rin chen (13-14th c.) XXI.4; Sakya'i dge slon Chos grags XXIII.7; dGe slon Ses rab grags (early 11th c.) XXVI.4; bSod suom pa Rin chen rgyal mtshan XXIX.1l. 28 sGros rUin pa (mNon par rtogs pa'i rgyan gyi rnam par 'byed pa'i 'grel ba 256ff.) explains the process of defining in terms of: mtshan gii la mtshan fiid kyis mtshon bya mtshon tshul. The mtshan gii is identified as the endeavors of the Bodhisattva, and the mtshon bya as the cause or specificity of the three types of omniscience (ses pa de dag gi rgyu'i tha sfiad). gZon nu rin chen (mNon par rtogs pa'i rgyan gyi 'grel ba bcud kyi thig 226ff.) makes roughly the same identification of the three elements as sGros rUin pa. The fourth mark, the nature, is presented in terms of mtshon bya Ita bu'i mtshan fiid, in opposition to the other three, which are qualified as mtshon bya dlios kyi mtshan fiid. The three-term definition is illustrated by the example of smoke being defined as "logical reason" owing to the definiens "[being a] property of the subject, [having] positive and negative concomitance [with the property to be proved]," in other words, to its satisfying the triple characteristic (trairupya) for being a genuine logical reason. 29 The titles of the subdivisions recall theoretical discussions, for instance "mtshan gii la mtshan fiid kyis mtshon bya ji !tar mtshon pa'i tshul," or "mtshan mtshon gii gsum lios gzuli pa." But the contents of

The origin of the theory of definition

333

To find out whether it is in this context that the theory of definition originaUydeveloped we must turn to earlier commentaries. Of special interestto us is the one composed by rNog Lo, who also translated the AA together with the Pandit Go mi 'chi med and revised earlier translations of Haribhadra's commentaries.
rNog La's commentary on the Abhisamayiilalikiira

In rNog Lo's commentary on these verses (27a2ff.), one does not see the application of a three-term model of definition. One notices however that what is being characterized, the "nature of the training," is presented in two ways so as to constitute the following pairs with lak~a1Ja:
i. mtshan gii - mtshan fiid (kyi chos) ii. mtshon bya'i chos -mtshon byed kyi chos

Thus, while "lak~a1Ja" is expressed either by its Tibetan equivalent "mtshan Hid" or through its function as "the property that characterizes" (mtshon byed kyi chos), one sees the use of two different Tibetan expressions for the counterpart of "lak~a1Ja:" one corresponds to the idea of a "basis of characterization," to be understood as "a basis for the defining characteristic;" the other to that of a "property that is to be characterized." These alternatives appear to be a consequence of the way this element is apprehended in relation to the "lak~a1Ja." There does not, however, seem to be an essential difference between the two at this stage, as the "property that is to be characterized" is also said to be "the foundation of what characterizes" (mtshon byed kyi Itos sa), thus echoing the notion of "basis" (gii) occurring in the first pair.

these subdivisions only address the application in the case under consideration. Note however that gZon nu rin chen, in his "extensive explanation" (mNon par rtogs pa'i rgyan gyi 'grel ba bcud kyi thig 227:ff.), explains the distinction between mtshan fiid and mtshon bya in terms of the oppositions don ldog/tha sfiad ldog and rdzas yod/btags yod.

334

Pascale Hugon

Ar Byan chub ye ses's commentary

Another early commentary of interest is that of Ar Byan chub ye ses (11th c.). In his discussion of AA IV.13-31, this author uses at some points the terms "mtshan gii" and "mtshon bya" together with "mtshan fiid" as part of a triad,30 but he does not provide a theoretical presentation of these notions or an analysis of their nature and relationship. His general outline follows, however, the two-term model suggested in the Sanskrit text, distinguishing between "what characterizes" (mtshon byed) and "what is to be characterized" (mtshon bya). A reader of Ar Byan chub ye ses's commentary completed this outline with a marginal note providing an interpretation of the verses of the AA according to a three-term model of definition.31 This note identifies the three elements of the definition - mtshan gii, mtshon bya and mtshan fiid - and defines mtshan gii as "a basis where the mtshan fiid resides" (which is also Phya pa's definition). Moreover, our "scribbler" supplies a citation in support of the triad

30 Cf. mNon par rtogs rgyan gyi 'grel ba rnam par 'byed pa 168('og) a6-7: sbyor pa fii su 'am Ina po mtshan giir biag Ia I 'di rnams sans rgyas skyed par byed pa'am I sans rgyas Ia 'jig rten ston par byed pa yin te ies mtshon byar bya'o I de'i mkhyen pa gsum gyi rtogs pa skyed pa yin pa ni mtshan fiid du bya'o II See sGros rmIi pa's mention and criticism of this position in mNon par rtogs pa'i rgyan gyi rnam par 'byed pa'i 'grel ba 256. 31 This note can be only partly deciphered due to blurring. See ibid. 168('og)a2: ni rje btsun gyis II sans rgyas rnaf!!s kymi sef!!n [=sems can] Ia II phan gdags phyir na mtshan gii dan II mtshan fiid dan ni mtshan pa ste II ies bSad pas mtshan fiid gnas pa'i rten mtshan gii dan I mtsh[o]n bya dan I mtshan fiid do I dan [po] ni gii dan d [rnaf!!?] pas [bsfiam?] pa'i chos [thaf!!d?] skye pa med par [bsfiof!!s?] pa'i gnas pa bkag pa Zasogs pa'i sbyar ba fii shu'o 112 pa ni sbyor pa de Ia yod pa'i sbyor ba'i khyad par de biin gsegs pa skyed bar byed pa'af!! yan na de biin gsegs pa Za 'jig rten ston par byed '0 113 pa mtshan fiid. (1 transcribe the bindu as f!!. Square brackets indicate unclear readings, square brackets with a question mark uncertain readings, and .. stands for an illegible character.)

The origin of the theory of definition

335

that can be identified as Mahiiyanasutralalikara (MSA) XI.36. 32 This verse mentions three terms: lak~ya, lak~atJa and lak~a1Ja, translated in Tibetan respectively as mtshan gii, mtshan iiid and mtshon pa. The author of the interlinear note seems to want to identify "mtshon pa" with "mtshon bya" (an alternative occasionally found in Phya pa's works). Are we dealing here with a Sanskrit precedent for the Tibetan triadic model of definition? Such a possibility is not supported by the sense of this passage, as the third term, "lak~a1Ja," does not refer to a third element, but to the process through which the indicand (lak~ya) and the indice (lak~a1Ja) are linked. 33 But although there would thus be no parallelism between the three terms found in MSA XI.36 and the three elements of the Tibetan model of definition, the identification performed in this marginal note is found also in Phya pa's commentary on the MSA, where Phya pa illustrates MSA XI.36 with the definition of cow, involving the three elements "white calf," "hump and dewlap, etc." and "COW."34
32 MSA XI.36 (Tib. XII.36): lak~yaT(l ca lak~a1JaT(l caiva lak~a1Jii ca prabhedataJ:t / anugrahiirthaT(l sattviiniiT(l sambuddhaiJ:t samprakiiSitii. Levi translates: "L'Indicand, l'Indice, l'Indication ant ete expliques dans leurs sections par les Bouddhas pour rendre service aux creatures." In a more recent publication dealing with Buddhist semiotics, D'Amato formulates the triad in terms of "signified, signifier, and signification" (2003: 91). In the given passage of the MSA, lak~ya is mind, lak~a1Ja the three natures, and lak~a1Jii the stages of spiritual discipline. 33 I follow here Levi's understanding of the passage (1907: 116 n.36): "Toute expression est directe au indirecte. Si elle est indirecte, on se sert d'un indice (lak~a1Ja) pour faire entendre l' indicand (lak~ya, indicandum), Ie sens qu'on veut faire entendre; l' indication (lak~a1Jii) est la fonction qui met en rapport les deux autres termes." In the classical example "gangiiyiiT(l gho~aJ:t" ("a stable on the Ganges"), the lak~a1Jii is the process of association that allows the understanding of the indicand as "a stable on the bank of the Ganges" through the indice (i.e., the statement "a stable on the Ganges"). 34 Theg chen mdo sde rgyan gyi legs Mad rgyan fii 'od gsal ba 473-474. None of the other five commentaries on the MSA in the bKa' gdams gsun 'bum - including the one composed by rNog Lo (mDo sde rgyan bsdus don, see 16b8ff. on this verse) - hint at a parallel with the theory of defini-

336

Pascale Hugon

Thus, although the Abhisamayiilafzkiira and, as we have seen, the Mahiiyiinasutriilafzkiira as well, contain a discussion involving the key terms "lak~m:La" and "lak~ya" (or "yad lak~yate" for the first), this preliminary look into the early Tibetan commentaries did not reveal any theoretical discussions on the nature of the elements of definition (or characterization). It showed nevertheless that the triadic model was known from early on and was applied in this context. Although it is not used by rNog Lo in his commentary on the AA, we find there the suggestion of an early differentiation concerning the counterpart of lak~a1'.za: the counterpart is alternately expressed with the terms "mtshan gii" and "mtshon bya." Further evidence concerning the use of these two terms can be gathered from rNog Lo's epistemological writings.
rNog Lo's epistemological works

Looking at the epistemological works of rNog Lo that are actually available - a synopsis of Dharmottara's Nyiiyabindutfkii35 and a commentary on the difficult points of the Pramiil,laviniscaya (dKa' gnas) - one can see in at least one case a clear echo of the indeterminacy with regard to the formulation, in Tibetan, of the counterpart of lak~al,la in a two-term definition. Namely, when discussing DharmakIrti's definition of perception (pratyak~a, Tib. mfzon sum), rNog Lo introduces the views of several Indian commentators (Dharmottara, KamalaSIla, Santabhadra) on the function of the two elements "direct perception" and "devoid of conceptualization and non-erroneous." The question here is how these two, conceived respectively as definiendum and definiens (lak~ya/lak~al,la), are to be identified with the elements of the pair "what has been posited" (anuviida, ... anudya) / "what is to be affirmed" (vidheya) (Tib. rjes su brjod pa/sgrub pa), which one can assimilate to the semantic

tion in this context. 35 This work, whose first page bears the title mTshan fiid kyi chos, is erroneously identified as a synopsis of the Pramill}-avini.scaya (rNam lies bsdus don) by the editors of the bKa' gdams gsuli 'bum.

The origin of the theory of definition

337

notions of "theme" and "rheme."36 What interests us here is how rNog Lo formulates the notions of Zakya and Zakaf}a. 37 In both works, rNog Lo uses the pair of Tibetan terms "mtshon bya" and "mtshan fiid" when paraphrasing the positions of Indian commentators as well as when giving his own opinion. If we turn now to his translation of Dharmottara's Pramiif}aviniscayatfkii (PVin'D, one notices that in the corresponding context the pair Zakya/Zakaf}a is systematically translated with mtshan gii/mtshan fiid. 38 Thus, in this discussion, rNog Lo uses "mtshon bya" for Zakya in his own works, whereas he uses "mtshan gii" to translate this term in the
PVinT

Although one finds in the preceding case "mtshon bya" and "mtshan gii" as two alternative renderings of "Zakya," rNog Lo clearly differentiates between the two when he applies a triadic model of definition. He also states explicitly in dKa' gnas that mtshan gii and mtshon bya are distinct, as exemplified in the case of "smoke" and "logical reason" ("smoke" being the logical

36 See dKa' gnas 86,18-90,21. In brief, the "theme" is the topic, what is being talked about, while the "rheme" is what is being said about it, namely, thenew information provided in the given statement. In the discussions mentioned here, the question is whether "devoid of conceptualization, etc." is affirmed of a posited "perception," or if "perception" is affirmed of a posited "cognition devoid of conceptualization, etc." 37 The Sanskrit pair is attested in this context in KamalaSIla's TSP 449,11-12 ad TS 1213 and in a parallel discussion in Jinendrabuddhi's PST I 37,5. Although absent from NBT ad NB lA, these expressions can be found in Durvekamisra's DhPr 41,6-15. 38 This portion of the PVinT (namely PVinT I 37a-38a) is not available in Sanskrit. However, a citation in the DhPr confirms that the Sanskrit original contained the pair lak~yallak~al}a. Cf. PVinT I 37b2-3: mtshan gii dan mtshan fiid grub pa la ni mtshan fiid rjes su brjod nas mtshan gii fiid ses par byed la I mtshan gii dan mtshan fiid kyi dnos po ma grub pa'i nag la ni mtshan fiid sgrub par byed do II and DhPr 41,10-11: yataJ:z prasiddhe lak~yalak~al}abhiive lak~al}anuvadena lak~yar!2 vidheyam, aprasiddhe tu lak~al}avakyena lak~al}am eva vidheyam, na lak~yam iti nyaya iti I

338

Pascale Hugon

reason invoked to prove the presence of "fire").39 It is interesting to note that this remark occurs precisely when rNog Lo comments on Dharmottara's arguments against KamalasITa's position on the theme/rheme controversy. KamalasITa's position is that the definiendum "direct perception" is affirmed once "devoid -of conceptualization and non-erroneous" has been posited as a topic. 40 Dharmottara contends that in this case, the definiens would have to be established; he offers two alternatives: it is established without a basis (gii) - which is impossible - or with a basis, namely, an instance of perception (miwn sum fiid kyi gii). In this last case, saysDharmottara, the lak.~ya would be established; being already established, it would not be what is to be affirmed once the definiens is posited.41 What Dharmottara implies here is that "perception" is established because a cognition which is a perception is established. This, according to rNog Lo amounts to confusing mtshan gii and mtshon bya. rNog Lo is here implying that behind the unique Sanskrit term "lak~ya," here referring to "perception," two notions have to be distinguished: an instance of perception which is established insofar as it is experienced, and the definiendum "perception," a characterization which is to be applied to such an instance. Note that rNog Lo's remark would have made little sense if he had in this text, as in his translation of the PVinr, previously used the term mtshan gii rather than mtshon bya for the counterpart of mtshan fiid. I will come back to the issue of translation below.

39 dKa' gnas 88,18: mtshon bya dan mtshan gii yan tha dad de / du ba dan gtan tshigs fiid biin no II 40 TSP 449,11 ad TS 1213: tatrajiiiinasya kalpaniiporf-hatvam abhriintatvaTfl ciiniidya pratyak~atvaTfl vidhfyate. Hence, according to him, the definiens is the theme -and the definiendum the rheme. Dharmottara, _ on the other hand, contends that the definiendum "perception" is the theme and the definiens the rheme. According to him, "perception" is something which everyone can acknowledge, and the character of being "devoid of conceptualization and non-erroneous" is what is affirmed of this established topic. 41 See dKa' gnas 88,1-8 for a presentation of Dharmottara's refutation (corresponding to PVinTr 38a3ff.), and 88,9-19 for rNog La's criticism.

The origin of the theory of definition

339

While rNog Lo uses the triadic model in his epistemological works, he does not provide a systematic presentation of the nature of the three elements or of their relationship - to some extent this can be expected in view of the nature of the available works. Still, a number of "technical remarks" prefiguring what is found ill Phya pa's extensive presentation let us guess that some sort of theoretical framework had already been worked out at that time,42 countering Sakya mchog Idan's claim that the theory of definition was not known until Phya pa. 43 That rNog Lo may have provided a more systematic account in a work that is as yet unavailable remains a possibility. The same remark holds true for rNog Lo's successors, whose contributions to the subject can be gleaned for instance from the numerous citations in the De kho na Hid bsdus pa, a work probably written in the 12th century.44 This work, in which the conflicting views concerning definition are nominally identified, testifies to a significant amount of discussion focused on theoretical issues by students of rNog Lo and contemporaries of Phya pa. 4S

42 These remarks mainly address the nature and relationship between mtshan fiid and mtshon bya: i) they represent, respectively, the reason of imputation and the property imputed (btags pa'i chos) (dKa' gnas 166); ii) the first is substantially existent, the second nominally existent (p. 166); iii) they do not deviate from one another, but are distinct properties as objects of thought (p. 303); and iv) the relationship between mtshan fiid and mtshon bya can be paired with the role of liriga and lirigin (p. 487). In addition, in the context of the definition of a logical reason that is a property of the subject (pak~adharma) (p. 214), rNog Lo describes the three-step process of "establishing a convention" (see also n. 10 above). 43 See notes 4 and 23 above. 44 This work is attributed to KlOIi. chen Rab 'byams pa. However, insofar as its author mentions the positions of thinkers up to Phya pa and those of Phya pa's contemporaries, but does not seem to be aware of any of the innovations brought by Phya pa's student gTsan nag pa, it is more likely to have been written in the 12th century, at a time when gTsan nag pa was not yet active as a teacher. On this text, see van der Kuijp 2003. 45 The authors identified in the De kho na fiid bsdus pa are (with, in parentheses, the number of times their name occurs with regard to the theory of definition): Jo btsun (3), Gans pa (3), rGya (8), Me dig pa (2),

340

Pascale Hugon

Translation issues
Let us consider the question of translation. As I mentioned above, dPaIi. Lo tsa ba attributes the emergence of the three-term model of definition to confusion in the translation. We have seen that rNog Lo uses "mtshan gii" and "mtshon bya" as alternative Tibetan equivalents for "lakrJa" in his epistemological works. One can wonder if this fluctuation echoes a generalized phenomenon in translations of the Buddhist corpus. Are the two renderings equivalent? Is one of them the preferred translation? A preliminary search through approximately three hundred works of the bKa' 'gyur and bsTan 'gyur that have been entered by the Asian Classics Input Project,46 although far from exhaustive, yielded results that I deem significant: in this selection, the term "mtshan gii" occurs 297 times (in 36 texts), in the majority of cases in conjunction with "mtshan Hid" (only five times in conjunction with "mtshan"); one finds also 413 occurrences (but in only seven of the three hundred texts) of the expression "mtshan Hid kyi gii," all of them in conjunction with "mtshan flid." In comparison, "mtshon bya" comes up only 38 times (in 19 texts).
Byan chub skyabs (4), Phya pa (10). Regarding their identification, see . van der Kuijp 2003: 415-417. According to Chos kyi rgyal po's Zhib mo rdo rje (translated in Stearns 2001), Me dig pa had been the "assistant teacher" (zur chos) of KhYUIi Rin chen grags, one of the most important disciples of rNog Lo (see Stearns 2001: 134 and 137). The De kho na iiid bsdus pa also refers twice to "Lo tsa ba" (i.e., rNog Lo), and once to "some old Tibetans" (bod bgres po kha cig). The discussions in which their names and positions are mentioned are: the nature and relation of mtshon bya and mtshan fiid (Jo bstun, Gans pa, rNog); mtshan gti (Phya pa, Byan chub skyabs); mtshan fiid kyi mtshan fiid (bod bgres, rNog Lo, Phya pa, Byan chub skyabs); mtshon bya'i mtshan iiid (rGya); mtshan gti'i mtshan fiid (Gans pa, rGya); the faults of a definition (Me dig pa refuted by rGya, Gans pa refuted by rGya, Phya pa, Byan chub skyabs); the faults linked to mtshan gti (Me dig pa refuted by rGya and by Jo btsun, Jo bstun refuted by rGya). 46 The e-texts consulted can be accessed on http://www.asianc1assics. org. This search was greatly facilitated by a program developed by my colleague Horst Lasic, to whom I am most grateful.

The origin of the theory of definition

341

To give some examples, one sees the pair mtshan gii/mtshan fiid for lak~yallak~a1}a figuring in the translation of certain major works, such as Nagarjuna's Mulamadhyamakakiirikii (MMK; more precisely in the 5th chapter), translated in the early 8th century, as well as in Vasubandhu's Abhidharmakosa and bhii~ya translated by Jinamitra (around 800).47 "mTshan gii" also appears inthe translation of Candraldrti's works,48 and in Jayananda's translation of his own Madhyamakiivatiiratfkii.49 Extensive discussions involving the pair mtshan fiid kyi gii/ mtshan fiid are located in Buddhapalita's Mulamadhyamakavrtti and in Bhaviveka's Prajiiiipradfpa, as well as in Avalokitavrata's Prajfiiipradfpatfkii, which were all translated by Jfianagarbha and Cog ro Klu'i rgyal mtshan around 800. As these discussions take place when the authors comment on MMK V, where the discussion of lak~ya and lak~a1}a occurs, one can surmise, although the Sanskrit is not extant, that in these commentaries "mtshan fiid kyi gii" translates "lak~ya," whereas the term is translated as "mtshan gii" when it occurs in the verses of MMK. The occurrences of "mtshon bya," much fewer in number in the sample considered, can be roughly grouped into four categories: i) the AA verse discussed above and its citation in vari47 See AK II.50cd (Tib. 6al) and AKBh 83,19-21 (Tib. 86b7). AKBh 78,4 ad II.46ab (Tib. 82bl) translates lak~aT}allak~a with mtshan nidi mtshan nid kyi gii. In this passage, Vasubandhu provides interesting remarks concerning the fact that the definiens of X is not different from X, taking as an example the definiens of cow:" "The dewlap, the tail, the hump, the hoof, the horns, which are the defining characteristics of cow, do not differ from the cow." (AKBh 78,6: silsnillilngalakakudasaphavi~il T}ildfni ca gotvalak~aT}ilni gor nilnyilni; Tib. 82b2-3) "Xuanzang has, instead, according to de La Vallee Poussin's translation (1971: 230), a zebra. 48 See note 53. 49 See Madhyamakilvatilratfkil 131b5, 151a2, or the statement on 274b2: mtshan gii ni mtshan nid kyi sgo nas mtshon par bya ba'i dnos po'o I me ni sreg par byed pa'o I This statement is cited as an example of the identification of mtshon bya and mtshan gii by Glo bo mkhan chen in Rigs gter Ni ma 159,21-22.

342

Pascale Rugon

ous commentaries,50 ii) mtshon bya occurring in conjunction with mtshan fiid, iii) mtshon bya occurring in conjunction with mtshon byed, and iv) rare cases of mtshon bya occurring alone. Note that "mtshon bya" does not always stand for "lak~ya." It can also be a form of the verb "lak~" (lak~yate, etc.), although "mtshon par bya ba" is more frequent for the latter. With regard to ii), mtshon bya occurs in conjunction with mtshan fiid in both older translations, like the Mahayanasutralalikaratrka of Asvabhava translated by SakyasiIp.ha and dPal brtsegs around 800, and later ones such as the Pramar:tasamuccayatrka (PST) of Jinendrabuddhi translated by dPail Lo tsa ba, hence sometiIne at the beginning of the 14th century. 51 As for iii), "mtshon byed" appears as an alternative to "mtshan fiid" to translate the term "lak~ar:ta." Thus for instance, the compound "mtshon bya mtshon byed kyi dlios po" stands for "lak~yalak~ar:tabhava" in the translation of PrajfUikiiragupta's Pramar:tavarttikalalikara.52 This is maybe not incidental, as this work was translated by rNog Lo and Bhavyaraja at the end of the 11th c., and we have met already with the pair mtshon bya'i chosl mtshon byed kyi chos in rNog Lo's commentary on the AA. If it is uncommon for both translations of "lak~ya" to be found in the same text, the case does however arise, as for instance in the translation of the Prasannapada by Pa tshab Ni rna grags; there, the pair lak~yallak~ar:ta is usually translated as "mtshan giilmtshan fiid," but on one occasion "mtshon bya" is found for "lak~ya" instead. 53
so. Seen. 26. 51 See for instance the discussion referred to in note 37 and the passage cited i1;l note 58. 52 Pramiir:taviirttikiilalikiira 211,27; Tib. 196a6. 53 The Tibetan "mtshan fiid phyin ci log brjod pas byas pa'i mtshon bya phyin ci log yod par 'gyur na ni" (20a6) corresponds to the Sanskrit "yadi ... viparftalak~a1Japra1JayanaYfl krtaYfl lak~yavaiparftyaYfl ... syiit" (59,1-2). Another occurrence of the term "mtshon bya" paired with "mtshan fiid" in this translation turns out to have no equivalent in the Sanskrit. Cf. 4b3-4: brjod bya dan rjod byed dan / mtshan fiid dan mtshon bya la sogs

The origin of the theory of definition

343

It would thus appear that the translations "mtshan gii" and "mtshan fiid kyi gii" were preferred for "lak~ya." And although the Mahiivyutpatti (no. 6975) prescribed "mtshan gyi gii" as the translation for "lak~yam" - a translation actually seldom used 54 - which would rather give support to these renderings, this did not prevent translators from adopting the alternative translation "mtshon bya" from early on. The discussions on the two members of Dharmaklrti's definition of perception already mentioned above in connexion with rNog Lo provide a pertinent example of how these alternative renderings were used equivalently: Jinamitra's translation of VinItadeva's commentary on the Nyiiyabindu, which dates to the 9th c., has mtshan fiid kyi gii associated with mtshan fiid in this context. 55 This is also found in SubhiltisrI and dGe ba'i blo gros's (active in the 11th century) translation of Devendrabuddhi's PramiilJaviirttikapafijikii, while the shorter form mtshan gii is substituted for mtshan fiid kyi gii in their translation of Sakyabuddhi's sub commentary thereon, the PramiilJaviirttika(rkii.56 In turn, "mtshon par bya ba"

pa'i spros pa rnams, whereas the Sanskrit is: abhidheyiidilak~m:!asya prapaficasya (11,6). 54 In the selection of texts considered, I only found the expression "mtshan gyi gii" in the translation of the Lankiivatiirasutra in conjunction with "mtshan" and with "mtshan fiid" - both pairs corresponding here to the Sanskrit lak~yallak~W!a. See for example "mtshan dan mtshan gyi gii dan II brjod par bya ba dan bral ba" (111b7-112a1) for the Skt. "lak~yalak~m:liibhidheyarahitii" (19,23), and 115a4 "mtshan gyi gii dan mtshan fiid med" for the Skt. "na lak~yii na ca lak~alJam" (verse 2.104). "mTshan gyi gii" also occurs on a single occasion in conjunction with "mtshan fiid" in a commentary on the Prajfiiipiiramitii by Daql~trasena, the Aryaatasiihasrikiipaficavil'(latisiihasrikii~tiidaasiihasrikiiprajfiii piiramitabrha(tfkii, translated by Surendrabodhi and Ye ses sde. 55 NBTv 4a7-b1: mnon sum ies bya bas ni mtshan fiid kyi gii bstan to II rtog pa dan bral iin ma 'khrul pa'o ies bya bas ni mtshan fiid bstan te ... 56 PramiilJaviirttikapafijikii 168b2: mnon sum ies bya ba 'dis ni mtshan fiid kyi gii bstan to II rtog dan bral ba can ies bya ba ni mtshan fiid yin te; PramiilJaviirttikatfkii Ne 180b1-2: mtshan gii bstan to ies bya ba ni mnon sum ni mtshon par bya ba'i gii fiid yin pa'i phyir ro II rtog pa dan

344

Pascale Hugon

is the translation adopted for "lak:jya" in the Tibetan version of Kamalaslla's Tattvaswigrahapafijika (TSP) preserved in the bsTan 'gyur, dating from around 1100,S? and "mtshon bya" is used, in a parallel passage, in the translation of the P by dPan Lo tsi! ba. 58 As for rNog Lo, we saw that he adopts the term "mtshan gii" for the translation of PVinr, while using the term "mtshon bya" in his paraphrasing of the same passage in the dKa' gnas. The fact that rNog Lo does not use exclusively "mtshan gii" in his translation of the PVin'T'9 suggests, I believe, that while translating this work, he was not (yet?) concerned with a triadic model of definition and used the two terms either equivalently, or in view of the nuances seen in his AA commentary, discussed above. It is possible that he used "mtshan gii" when following an earlier usage such as found in Jinamitra's translation of Vinltadeva's Nyayabindurfka. But in the dKa' gnas, as rNog Lo started to analyze definitions according to a triadic model, discussions originally involving twoelement definitions had to be reformulated in a way compatible

Sr

bral ba can ies bya ba'i mtshan fiid kyis mnon sum mtshon par bya ba yin pa'i phyir ro I The same pair mtshan giilmtshan fiid occurs in the context of the definition of pramil"(la eNe 72a3). My thanks to Chizuko Yoshimizu for this last reference. 5? See TSP Ye 25a6-7: 'dir mnon sum ni tshon (read: mtshon) par bya ba yin te I de'i mtshan fiid bstan pa'i skabs yin pa'i phyir ro I Skt. 449,1112: lak~yam atra pratyak~am, tallak~a"(lasyaiva prastutatvilt. An earlier "translation in progress" of this text is already mentioned in the IHan dkar catalogue; cf. Lalou 1953: 337. 58 PSTI 18a3: 'di ni {i.e., mnon sum} mtshon bya'o II rtog pa dan bral ba ies pa mtshan fiyid do I Skt. 37,5: etal lak~yam I kalpanilpoham iti
lak~a"(lam

59 One finds another occurrence of the pair mtshan fiidlmtshan gii in a passage in the third chapter (PVinT III 48b6). The term "mtshon bya" occurs elsewhere in rNog Lo's translation of the PVinT, but unfortunately we cannot establish the Sanskrit reading of these passages. See notably the opposition between mtshon byed kyi tshul and mtshon bya'i tshul in PVinT II 237b7, and the statement "mtshan fiid la gnod par byas na mtshon bya mtshon par mi nus te" in PVinTII 277bl.

The origin of the theory of definition

345

with it. The distinction between mtsha.n gii and mtshon bya being established at that point, it possibly influenced later translations. There was thus indeed, as dPaIi. Lo tsa ba claimed, some kind of "inconsistency" at work in Tibetan translations. It remains to be substantiated whether the adoption of the triadic model of definition was influenced by these distinct translations. One could suggest that the triadic model was inspired, more than by the translations themselves, by the nuances in expressing the counterpart of lak~a1}a attested in Tibetan commentaries (the choice of translation itself can naturally also be seen as expressing such nuance). In the corpus considered above, the pair lak~a1}a/lak~ya is mostly treated according to the subject-property model, a model made explicit by the translators and commentators who adopt the rendering "mtshan gii" for lak~a. One can see the definitional basis in the Tibetan model as an echo of the presentation of the lak~a as "the basis of the lak~a1}a" (mtshan iiid kyi gil), or of expressions such as used by Sakyabuddhi, who glosses "mtshan gii" in terms of "the basis which is to be defined [owing to the definiens]" (mtshon par bya ba'i gii).60 In the Tibetan model, the definitional basis is a basis of the lak~a1}a. If the Tibetan model goes beyond the dual structure with the introduction of the definitional basis as a distinct member, its specificity also resides in the nature ascribed to the definiendum, which comes to be understood as a "convention." Considering this, one can, I would suggest, identify a stronger source of influence accounting for the adoption of the three-term model, which can be located in Dharmottara's works.

Dharmottara on definition
While Dharmaklrti does not deal with definition in any remarkable way, Dharmottara for his part devotes a few passages to the question in his commentary on the PVin as well as in some parallel passages in the NBT, in particular in discussions related to the
60 See the passage cited in note 56. Unfortunately, no Sanskrit is available for this portion of the commentary.

346

Pascale Hugon

topic of "non-perception" (anupalabdhi). Commenting on the first reference to non-perception in PVin I (3,9ff.), Dharmottara devotes a long excursus to the demonstration that proofs relying on the nonperception of an object establish a "convention" (vyavahara, Tib. tha sfiad), namely, the convention of the absence of the non-perceived object. At the beginning of his argumentation, Dharmottara distinguishes between two kinds of logical reasons:
- "definitional" logical reasons (mtshan fiid kyi gtan tshigs I mtshan fiid pa'i gtan tshigs) - "factual" logical reasons (dnos po'i [gtan tshigsj).61

The first has as its object a convention (tha sfiad) - or the fitness for the [application of a] convention (tha sfiad du ruft) - whereas the second proves a real property. This distinction is a quite likely source for the division, common in Tibetan works, between the "proof of a convention" (tha sfiad sgrub) and the "proof of a state of affairs" (don sgrub).62 In the first case, Dharmottara describes the logical reason as a definiens (mtshan fiid), which is the "motive for the convention" (tha sfiad kyi rgyu mtshan, Skt. vyavahiirasya nimittam). Definiens and convention (Dharmottara does not describe the second as a definiendum) are linked by a relation that is purely a matter of custom: to ascertain this relation, it is enough to record a single application of the convention in the presence of the said definiens.

61 PVinT I 23b4-6: gtan tshigs ni rnam pa gfiis te I tha sfiad kyiyul can mtshan fiid kyi gtan tshigs de ni Ian cig tha sfiad zugs pa fiid las tha sfiad dan 'brei par ses par 'gyur ba yin te I dper na ba Ian gi tha sfiad kyi rgyu mtshan Ikog sal la sogs pa dan Idan pa'i mtshan fiid kyi yul la Ian cig tha .sfiad zugs pa las kyan 'brei bas ses par 'gyur ba Ita bu ste I 'di ni mtshan fiid pa'i gtan tshigs so II gzan ni dnos po'i ste I gan zig dnos po'i chos sgrub par byed pa'o II (My thanks to Birgit Kellner for pointing out this passage to me.) Regarding the first kind of reason, Dharmottara also speaks of a "definitional inference" (mtshan fiid pa'i rjes su dpag pa [26bl]). 62 The distinction is made already in dKa' gnas 414,6-7; see also Mun sel72a5, Rigs gter 297.

The origin of the theory of definition

347

All proofs establishing conventions are to be understood according to this. model: proofs such as "the convention 'cow' can be applied to this white calf, because it has a hump and a dewlap;" proofs relying on the non-apprehension of X, which establish the convention "there is no X;" but also proofs such as "this is a tree because it is a simsapa," with which Dharmaldrti illustrates an inference based on an essential property (svabhavahetu).63 A question that arises in this regard is: why would one ever need to prove a convention? Dharmaklrti states: "the convention is established for the dimwit."64 Elaborating on this statement, Dharmottara presents various cases involving a dimwit (muha, jaa) who is in need of a proof concerning a convention. 65 The dimwit indeed turns out to be unable to re-apply a previously learned convention. His inability is attributed alternatively to forgetfulness, or to the fact that he did not grasp the motive (nimitta) of application properly in the first place: he would thus imagine, for instance, that the size of the cow present at the time of learning the convention matters, whereas it is just the presence of a hump and a dewlap that is the motive; hence he subsequently fails to apply the convention "cow" to a cow of a different size. 66 In such cases, an inference is necessary for the dimwit to grasp the convention. This discussion sheds light on Dharmottara's understanding of "convention." Dharmaklrti presents the application of a convention as an event subsequent to the establishment of an agreement (brda, Skt. sanketa).67 In Dharmottara's writings, this agreement is specifically associated with the determination of a word (sgra).68
63 64

See PVinT IT 198aff. ad PVin IT 50,8. PVin I 3,14: vyavahiiral} siidhyate milrjha1'[l prati.

65 The case of the "dimwit" has been discussed in Kellner 2004: 32-35.
Examples involving a cow or a tree are found in PVinTI 26al, PVinT IT 198a, 202a, 202b, NBT 106,11-107,2 adNB IT.16, PVSVT33,17-21 (ref. Kellner 2004: 32-33), as well as in PVinTm ad IT1.86b.
66
67 See for instance PVSV 59,20-60,1: saTiketakiile tathiidr~tam eviiyam artha1'[l tatsambandhina1'[l vii vyavahiire 'pi pratipadyata...

68 PVinTI 55a5: brda ni sgra Ties par sbyor ba ste I skabs kyis 'dir khyad par daTi khyad par can brjod pa'i yul can yin par gzuTi Tio II This seems

348

Pascale Rugon

However, the discussion involving the dimwit's inability to apply the convention hints to the fact that proving a convention is not so much a matter of reminding someone of the word that is to be applied, but rather, of the criterion for the application of that word. Knowing a convention consists in knowing which distinctive features are relevant for the circumscription of a notion; it implies that one can thereby identify an individual as part of a class, to which a word mayor may not have been associated yet at the time of an agreement. Dharmottara's excursus, and in particular the notion of "definitionallogical reason," provides a plausible source for several integral constituents of the Tibetan model of definition: First, Dharmottara relates "definiens" with "convention:" the definiens is the logical reason that establishes the convention; it is also the motive for the convention (vyavahiirasya nimittam) - a function that finds a clear echo in the Tibetan model (tha sfiad kyi rnam 'jog gi rgyu mtshan). Although Dharmottara does not speak of definiendum, or identify explicitly the convention as being the definiendum corresponding to the definiens presented as logical reason, the identification would make sense in view of the Indian dual model of definition. Further, as Dharmottara posits the definiens as a logical reason, he integrates it into the structure of an inference. And an inference, in Indian logic, involves three terms: a subject, a logical reason and a property to be proved. It is very possible that the "definitional basis" found its place in the definitional model in analogy with the inferential model in general, and in particular with the model of "definitional inference" involving the triad subject,definiens and convention. Thus, although the definitional basis can have the function of an illustration, in view of the parallel with the inferential model, the definitional basis has a role analogical to the subject of inference rather than to the example.

also to be rNog Lo's position; he describes the establishment of a convention as the understanding that "this kind of thing must be designated by this kind of word." See dKa' gnas 70,15-18.

The origin ofthe theory of definition

349

One should note here that the notion of a "definitional logical reason" (Skt. lak~ar.tahetu) already figures in Dignaga's Pramar,tasamuccayavrtti .(PSV), where it qualifies the logical reason of the inference "substance is substance because it has substantiality" (dravya7[L dravyatvavattvad dravyam).69 According to Jinendrabuddhi's commentary on the given passage, such a lak~ar,tahetu is a definiens posited as a logical reason, which brings forth (hinoti) a definiendum via the exclusion of other meanings. As an example, Jinendrabuddhi gives the case of rigidity (kathinya), the definiens of earth (Prthivf). The logical reason under discussion, namely "possessing substantiality" (dravyatvayoga), is to be understood similarly as being the definiens of substance. The definitional logical reason is here opposed to a "probative logical reason" (lit. "a logical reason that makes one know" Ufiapakahetu]), left unexplained by Jinendrabuddhi, which, in view of the context, can presumably be identified with Dharmottara's "factual logical reason."70 It is probable, however, that the influence of the notion of "definitional inference" on the adoption of a triadic model of definition

69 PSVad PS II1.17, K 130a4-5 (mtshan fiid pa'i gtan tshigs), V 44b7 (mtshan fiid kyi gtan tshigs). My thanks to Sh6ryu Katsura, who gave me access to his Sanskrit reconstruction of this passage. 70 pSTm D161b2-4, P185a6-8 (B124b3-5). Katsura's provisional edition of this passage reads: lak~a1J.ahetau hfti / arthiintaravyiivrttena rupe1J.a lak~ya7!l (ms: lak~a7!l, Tib. mtshon cha) hinotrti pratyiiyatfti lak~a1J.am .eva hetur ihabhipretaJ:z, prthivyiiJ:z kiithinyavat / dravyatvayogena ca dravya7!l lak~yata ity asau lak~a1J.ahetuJ:z / sa ca dravyiid anyatriipi vartata iti gU1J.e~u prasanga uktaJ:z, na tu jfiiipakahetoJ:z, anaikiintikatviit / Jinendrabuddhi also mentions this type of reason when commenting on the logical reason "jagacchasaniid" in PSVad PS 1.1, cf. PSTI 10,1, also with the example "kiithinyiit prthivf." On his definition of lak~a1J.a, see the parallel passage in PST I 104,6: lak~a1J.a7!l hi niima yad arthantaravyiivrttena rupe1J.a lak~ya7!l gamayati / Dignaga himself does not oppose "lak~a1J.ahetu" to ''jfiiipakahetu.'' One only finds in his Nyiiyamukha a distinction between ''jfiiipakahetu'' (the cause for making something known) and "utpiidakao" or "kiira1J.ahetu" (the generative cause). See Katsura 1977: 122-124 (my thanks to Toshikazu Watanabe for this reference).

350

Pascale Hugon

would have come from Dharmottara, whose works rN-gog Lo, as translator and exegete, was most familiar with, rather than from Jinendrabuddhi, whose commentary was translated only at the beginning of the 14th century, or directly from Dignaga, given the laconism of the relevant passage in PSv. Such influence is not explicitly acknowledged by the Tibetans. Phya pa, in his systematic presentation of the theory of definition, does not even attempt to ground this theory in Indian sources. One can note, however, that Phya pa does make a connection between the topic of definition and that of logical reasons that consist in a definiens. Dharmottara's definitional inference finds its way into Phya pa's system within the opposition between the proof of a state of affairs and the proof of a convention (don sgrub/tha sfiad sgrub), but also reappears in the discussion concerning logical reasons and inferences regarding "what is commonly recognized" (prasiddha, Tib. grags pa). When dealing with the criterion "not opposed by what is commonly recognized" (aniriikrtaJ.z prasiddhena) occurring in Dignaga's definition of the thesis and Dharmaklrti's discussion thereof, Phya pa identifies "what is commonly recognized" as a type of logical reason and of inference where the logical reason makes understood a property based on linguistic usage (sgra la ltos pa'i ChOS).71 What is inferred here is the fitness to express or determine (brjod ruflJien run) by a word. Such inferences are further differentiated according to whether the word relies on a "correct exclusion of what is other" (gian sel phyin ci rna log pa) or not.72 The first type is illustrated by the inference of the convention
71 This type of reason is opposed to objectively grounded (dnos po stobs tugs, Skt. vastubalapravrtta) logical reasons and inferences. See Mun sel 79b9. 72 In,the first case, what is commonly recognized is a "convention" (tha sfiad kyi grags pa), in the second, an "agreement" (brda'i grags pa). The second type refers to the unlimited applicability of words (for instance, that of the word "moon" to any entity, or to any mental aspect), whereas in the first type, this applicability is limited in the framework of a definite convention (cf. 'Od zer 170bS-7). See a similar distinction in gTsail nag pa's bsDus pa (185b2-5) and mTshur ston's sGron rna (60b5-7). Sa skya Pal!9-ita criticizes this distinc-

The origin of the theory of definition

351

"cow" from the collection of hump, dewlap etc., i.e., the inference of a definiendum from a definiens, and by the inference of the convention "tree" from the logical reason "siY(tsapii" (here, one infers a definiendum on the basis of what is included in the definiens),73 Those are identical to two of the "definitional inferences" discussed by Dharmottara. It must be noted that the identification of the nature of the property to be proved in such a proof is precisely that which was given for the notion of convention in the context of the theory of definition, and Phya pa even refers to this previous context explicitly.74 Thus, in Phya pa's works, definition and inference proving a convention are connected, but the theory of definition is not presented within the context of the latter: it already has a place of its own. Keeping in mind this connection, and the possible influence of definitional inference on the building up of the model of definition, it is necessary to take a closer look at the parallels between the definitional model and the inferential model. Defining and inferring The definitional model offers a striking parallelism with inference. Additionally, as noted above, the formulation of a definition, with the definiens stated as the motive of application of the definiendum, hardly differs from that of a proof. Regarding the formal structure, in addition to the obvious parallelism between the respective triad of elements - subject, logical reason and property to be proved on
tion between brda and tha sfiad in Rigs gter 329. 73 See Mun sel 80a3-4: d-tios 'gal las tha dad du byed pa'i gtan sel phyin ci ma log pas ten nas brjod du ruti pa dpog pa ni tha sfiad kyi grags pa te / lkog sallasogs pa 'dus pa las ba lan du dpog pa ltar mtshan fiid kyis mtshon bya dpog pa'am / sa pas siti du dpog pa ltar mtshan fiid kyi khyab byas mtshon bya dpog pa lasogs pa yin no / and 80a4-5: des na gtan sel phyin ci ma log pa'i sgras zen run dpog pa la phyogs chos dan khyab pa tshan pa tha sfiad kyi grags pa yin la / 74 Mun sel 81a7-8: mtshan fiid dan dnos 'gal sel pa gtan sel pa phyin ci ma log pa'i sgras zen run zes snar mtshon bya'i ran btin du Mad pa de bsgrub bya'i chos yin no / See note 11.

352

Pascale Hugon

one side, and definitional basis, definiens and definiendum on the other - the definiens of the definiens given by Phya pa in terms of a "triple property" (chos gsum) appears to be a clear echo of the definiens of the logical reason in terms of a "triple characteristic" (tshul gsum). This correspondence however turns out to be limited; closer examination reveals that the elements of each structure do not share the same properties and that the parallelism of the "three properties" and the "three characteristics" is merely superficial. Thus for instance, the definiendum differs from the property to be proved insofar as it is not necessarily inferred: it is only in the case of "dimwits" that an inference is required; learned persons know it via perception. 75 Also, of the three properties of a correct definiens, if "being possessed by the definitional instance" corresponds to "being a property of the subject" in the case of the logical reason, the property of being substantially existent (rdzas yod) on the other hand does not find an equivalent in the triple characteristic. The criterion pertaining to the relation of the definiens to the definiendum (i.e., "not being of a different meaning") differs as well from the characteristics of positive and negative concomitance. Further mismatches ensue when comparing the faults relative to the logical reason and to the definiens. This is to say: although for any definition there can be a corresponding definitional inference (for instance for the benefit of a dimwit), definition and inference remain two different processes; the elements of definition, when used in a definition, do not have the same properties as the elements of an inference, and the criteria for a "correct definition" differ from those for a "correct inference." The theory of inference may have inspired the elaboration of the model of definition - plausible specific influences being the notions of "definitional logical reason" and "inference proving a convention" - but once the three-term model was adopted, definition got "a life of its own:" it developed independently of the theory of in75 Another difference pointed out by Phya pa is that the thesis requires the criterion of "not being opposed," whereas the definiendum does not. See Mun seI22a4-6.

The origin of the theory of definition

353

ference and dealt with its own web of issues. In later texts, such as gTsan nag pa's, one can notice a converse influence of the theory of definition on that of inference insofar as innovations made by the author in the context of definition generate parallel developments in the context of inference.76 The place of the theory of definition in Phya pa's works Finally, I would like to reflect on the place of the theory of definition in Phya pa's works. In both Mun sel and 'Od zer, the systematic presentation of the theory of definition is introduced as a theoretical framework presupposed for the definition of the fundamental notion of "valid cognition" (tshad rna). However, if one looks at the actual definition of valid cognition and of other epistemological notions, one finds that little practical use is made of the theory of definition. For the most part, it plays a role in evaluating putative definientia. In the theory of definition itself, the definition of valid cognition is hardly dealt with, and appears relatively unproblematic in comparison with the other complicated cases discussed in this context. The presentation of the theoretical framework of definition is indeed not conducted in a straightforward way, but, as is the case with most topics in Phya pa's writings, through a game of objections and responses involving problematic cases. A closer look at these discussions gives us an indication of what may have been more pressing concerns in Phya pa's development of the theory of definition. A particularly extensive discussion occurs with regard to the claim that the three elements of a definition share an undifferentiated nature; in some definitions this contention turns out to be in need of a proof, namely those definitions in which the definitional basis is a negation with a positive implication (rna yin dgag) and the definiens (or definiendum) a simple negation (rned dgag), or vice-versa.

76 One such innovation of gTsaIi. nag pa regarding the "modes of definition" (mtshon pa'i tshu!) and the "modes of proof" (sgrub pa'i tshu!) is studied in Rugon 2008a: 316ff.

354

Pascale Hugon

The illustration for the first configuration is:

a) applying the convention "empty" to the definitional basis "causally efficient" (alternatively, "entity") on account of the definiens "empty of causal efficiency that withstands analysis.'m

The reverse case is illustrated by two examples:

bi) applying the convention "[correct] logical reason" to the definitional basis ''neither one nor many" on account of the definiens "being a property of the subject and pervasion [by the property to be proved]." bii) applying the convention "knowable" to the definitional basis "emptiness" on account of the definiens "fit to be posited as an object of thought."78

Phya pa's proof of undifferentiability (dbyer med du bsgrub pa) is conducted solely on the basis of a series of reductio ad absurdum arguments, which can be summarized as follows: if A and B were not undifferentiated, then A could/would be non-B; the latter is either absurd, or leads to an absurdity. A number of objections are raised against the above cases.79 They concern on the one hand the notion of "undifferentiation" (dbye ba med) in comparison with that of identity (bdag gcig), and on the other hand, the conflict between properties of the definitional
77 Mun sel 13b2-3: mtshan gii ma yin dgag la mtshan fiid med dgag bkod pa ni don byed nus pa giir byas te dpyad bzod pa'i don byed nus pas ston pas ston pa fiid kyi tha sfiad du mtshon pa Ita bu'o / 78 Mun se116a5: med dgag mtshan giir byas pa la ma yin dgag mtshan fiid du bkod pa ni yan dag pa'i cig dan du bral la phyogs kyi chos dan khyab pa mtshan fiid du bkod nas rtags su mtshon pa 'am ston pa fiid blo'i yul du byar run pas ses byar mtshon pa Ita bu'o II These examples and the discussion that follows in Mun sel address the case of mismatching definitional basis and definiens (Mun sel13bl-16b7: mtshan gii dang mtshan fiid dbyer med du sgrub pa). According to Phya pa (Mun sel 16b7), the case of mismatching definitional basis and definiendum is to be treated similarly. A parallel discussion is found in 'Od zer 7b5-lOa3. 79 See respectively Mun seI14b4ff. and 16a7ff. Parallel in 'Od zer 8b3lOa3.

The origin of the theory of definition

355

basis and properties of the definiens, as well as between properties of the definitional basis and properties of the definiendum. Namely, if A and B are undifferentiated, is what is true of B also true of A? For instance concerning a), the "emptiness of causal efficiency that withstands analysis," being emptiness, withstands analysis. 8o If the definitional basis "entity" is undifferentiated from "empty of causal efficiency that withstands analysis," does it also withstand analysis? This would contradict the claim that an entity does not withstand analysis. Further, an entity is not ultimately true, but emptiness is. If both are undifferentiated, would an entity be ultimately true? Similar objections are raised for bii): the definiendum "knowable" has the property of not withstanding analysis (dpyad mi bzod); if emptiness is undifferentiated from "knowable," would it then also not withstand analysis? Further, "knowable" is a definitional basis of "relative truth," and "emptiness" is a definitional basis of "ultimate truth;" if both are undifferentiated, does it imply that "emptiness" is also a definitional basis of "relative truth;" isn't this contradictory with its being a definitional basis of "ultimate truth?" I will not enter here into the details of how Phya pa solves these questions - which he handles mainly through the technique of "argumentation by parallels."81 What is to be noted here is that these discussions have nothing to do with the definition of valid cognition, but are all related to issues in Madhyamaka philosophy. One can then wonder what is to be found pertaining to definition in Phya pa's works specifically dedicated to Madhyamaka. Phya pa's treatise on the ideas of the "three MMhyamikas from
80 The pervasion of "emptiness" (stan fiid), alternatively of "non-implicative negation" (med dgag), by "withstanding analysis" (dpyad bzad) which is postulated in the objection is affirmed by Phya pa, for instance in sTan thun 25,19-26,2 when he identifies the respective definitional bases of relative and ultimate truth. My thanks to Kevin Vose for this reference. 8! On this argumentative technique, see Rugon 2008b. One of Phya pa's favorite parallels in the present context is the case of the appearance of two moons, which, although erroneous with regard to the grasped object, is non-erroneous qua reflexive cognition.

356

Pascale Rugon

the East" (dBu ma sar gsum gyi ston thun; hereafter: sTon thun), sometimes identified as his Resume of Madhyamaka (dBu ma'i bsdus pa), provides some pertinent material. The first and longest section of this work (40 folios out of 46)82 dealing with "what is knowable" (ses bya) addresses two main issues: 1) the distinction between the two truths and 2) the definitions of the two truths. The latter involves the identification of the definiens and definitional basis of each truth, followed by the establishment that the given definiens indeed applies to the definitional basis. When defining the two truths, Phya pa does not engage in a theoretical presentation of the notions of definiens, definiendum and definitional basis, or of the triple property (chos gsum) of the definiens, which are all presupposed. The definitions of the two truths, however, provide interesting parallels with the above-mentioned discussions of Mun sel and 'Od zero First, a differentiation of the definitional basis in terms of implicative and non-implicative negation also occurs in the sTon thun: according to Phya pa, everything can be defined as ultimately true or as relatively true depending on the type of cognition that analyzes it (i.e., a cognition that takes it as the object of final analysis [mthar thug jalJ or not); however, at the level of conceptual identification (ldog pa) all non-implicative negations (med dgag) are instances to which the convention "absolute truth" can be applied, and all implicative negations (ma yin dgag) are definitional bases for the convention "relative truth."83

82 The work is divided into three main sections: i) what is knowable (ses bya), ii) the practice of the Great Vehicle (theg pa chen po'i spyod pa mfiams su len pa'i thabs), iii) the result: the buddhabhumi (spyod pa rjes su bsgrubs pa las 'bras bu rnam pa thams cad mkhyen pa'i sans rgyas kyi sa thob pa'i tshul). A translation of the topical outline is provided in Tauscher 2003: 239-243. 83 See sTan thun 24,10-26,14. Phya pa speaks in the first case of "defining as a substance" (rdzas su mtshon pa), in the second of "defining as a convention" (ldog pa de la ... tha sfiyad rten par mtshon). That is to say, implicative and non-implicative negations, per se, can be characterized either as relatively or absolutely true according to the cognition that apprehends them, but as an implicative negation or as a non-implicative negation, they will conventionally be designated as one or the other. Note,

The origin of the theory of definition

357

Phya pa has to defend his position against a number of objections which need not be detailed here. Two of them deserve our attention, however: 84 the first addresses the application of the definiendum "empty of a true nature" to the definitional basis "entity" (dfws po bden pa'i no bos ston bar mtshon pa), the second the application of the definiendum "knowable" to "emptiness" (ston iiid gial byar mtshon pa). These cases are none other than the ones that are dealt with in Phya pa's theory of definition, namely a) and bii) discussed above. Although in the sTon thun the problem is not identified in terms of an implicative and a non-implicative negation being undifferentiated elements, one recognizes among the objections the questions raised in Mun sel and 'Od zero Further analysis of Phya pa's other works might reveal other corresponding passages. I did not discover any such parallels, however, in my preliminary look at Phya pa's commentary on JiUinagarbha's Satyadvayavibhanga,85 which is one of the three treatises constituting the basis for the sTon thun. This commentary reveals a good deal of structural similarity with the sTan thun (for instance, in both works, the section on the distinction between the two truths follows the same outline).86 But the extensive discussion on the two truths closely follows JiUinagarbha's verses and commentary, and applies a hierarchical organization that does not echo

in the first case, Phya pa's use of the analogy of the appearance of two moons (cf. note 81). 84 See, respectively, sections 123.4 and 123.5, specifically 123.42 (sTan thun 51,19-54,22) and 123.52 (sTan thun 55,17-56,21). 85 Note that dBu ma bden gfiis kyi 'grel ba is the title adopted in the dkar chag. No title is given on the first folio. The colophon identifies the text as "dbu rna bden pa gfiis rnam par bsad pa yi ge fiun nus giun gsal bar byed pa." 86 See bDen gfiis 'grel 5b3ff. The definitional basis of relative truth is identified as an "implicative negation" (ma yin dgag), whereas that of ultimate truth is presented in this text in terms of "devoid of proliferations" (spros pa dan bral ba, Skt. ni'iprapafica) rather than as a "non-implicative negation" (med dgag). Cf. bDen gfiis 'grel7a7ff.

358

Pascale Hugon

the rest of the presentation in the sTan thun. 87 In particular, I did not find a specific section on the problems relevant to the definitional basis. Although issues arise that correspond to the questions that we examined above,88 they are not articulated in the framework of the undifferentiability of the definitional basis, the definiens and the definiendum. Is Phya pa's theory of definition intended to help solve a question that is specific to the two truths, or are the two truths brought into his epistemological works as an example of a problem inherent in the theory of definition? It looks as if, for Phya pa, the theory and its practical applications are not strictly distinguished from one another. The discussion occurring in Mun sel and 'Od zer about the mode of distinction of the elements of a definition, although it is presented in generic terms - solving cases where one element is an implicative negation, the other a non-implicative negation - deals exclusively with cases belonging to the key issue of the definition of the two truths, which, we have seen, are also dealt with in the sTan thun. Actually, the solution brought to the problem of definitions whose elements are respectively implicative and non-implicative negations only makes sense in the context of the discussion of the two truths. The issue of the distinction between the two truths and that of the distinction between the elements of definition were unquestionably linked for Phya pa; these were also linked, one can add, with the question of the status of implicative or non-implicative negation of the elements of inference. 89 A passage of the De kha na fiid bsdus

87

bDen gfiis 'grel 8a6ff. (rgyas par Mad pa).

88 See for example bDen gfiis 'grel 6b4ff., where Phya pa explains how "entity'1 can be substantially identical with "emptiness" without implying an "ultimate entity." Here again, the analogy of the appearance of two moons is mentioned. 89 In the context of the inference proving emptiness in the sTan thun, Phya pa defines implicative and non-implicative negation (sTan thun 85,14-87,12), and spells out the difficulties of combining in an inference a logical reason and a property to be proved that would be of different types (sTan thun 87,13-90,15). The first paragraph finds a close parallel

The origin of the theory of definition

359

pa tends to confirm that the two themes, namely definition and the two truths, were merged in discussions around Phya pa's time. In this passage, the author summarizes the problem ofthe undifferentiability of the elements of definition and Phya pa's solution in the following terms: the one element that is a non-implicative negation is absolutely true, and the one element that is an implicative negation is relatively true; admitting that these two are undifferentiated amounts to accepting that the two truths are "one nature" (no bo gcig pa), and that the differentiation comes from a distinction relative to the cognizing mind. 90 The author further mentions a thinker by the name of Byan chub skyabs (hitherto unidentified),91 who holds for his part that although the non-implicative negation (e.g., empty of causal efficiency that withstands analysis) and the implicative negation (e.g., pot) are undifferentiated, it is not the case that the two truths have a single nature. The argument attributed to Byan chub skyabs invokes in particular a half-verse of the Satyadvayavibhanga (9ab) in support of the claim that the "emptiin Mun sel 61b9-62a8 and a more approximate one in 'Od zer 104b3ff. Note that the status of the subject is not discussed in this context. Phya pa deals extensively in his epistemological works with the status of the property to be proved (see Mun sel4a8-6a2, parallel in 'Od zer 27a7-28b2). The question of the combination of elements of inference that represent different types of negation comes up for instance when discussing logical reasons that establish a convention (Mun sel40b9ff., see also 'Od zer 104a2-3 in the discussion on anupalabdhi). 90 See De kho na fiid bsdus pa 71-73, and in particular 72,18-73,4: des na phya pa na re dpyad bzod pa'i don byed nus pas ston pa ni don dam pa'i bden pa yin la I mtshan gii bum pa ni kun rdzob kyi bden pa yin pas bden pa gfiis no bo gcig pa yin te I dpyad bzod pa'i don byed pas ston pa gfiis gcig yin kyan ses no gfiis la ltos nas snon po mnon gyurd dan skad cig ma lkog gyur yin pa mi 'gal ba ltar I 'dir yan mthar thug dpyod pa'i rigs ses la ltos nas don dam pa'i bden pa dan I ma brtags ma dpyad pa'i blo la ltos nas tha sfiad kun rdzob kyi bden pa yin pa mi 'gallo ies zer ro II 91 This thinker is frequently mentioned in the section on definition (see n. 43). It remains to be studied whether he is identical with the Byail chub skyabs who authored an outline and explanation of the Sik~iisamuccaya and a commentary on the last chapter of the Siltriilankiira, three works published in the bKa' gdams gsun 'bum, vol. XXv.

360

Pascale Rugon

ness of causal efficiency that withstands analysis" is in fact not an instance of ultimate truth but, insofar as it is an object of mind, of relative truth. 92 The situation appears to have changed with Phya pa's student gTsan nag pa. Looking in comparison at the latter's presentation of the theory of definition in his bsDus pa, one notices that the above discussion, which covered more than four folios (recto-verso) in Phya pa's epistemological work, is reduced here to three lines!93 One recognizes in the objection the problematic case involving "emptiness of causal efficiency that withstands analysis," but gTsan nag pa's answer stays on a general, theoretical level, and does not mention the two truths or related notions. This is not to say that gTsan nag pa is not concerned about the two truths;94 rather, it would seem to go along with a general tendency in his epistemological writings: indeed, gTan nag pa hardly ever introduces Madhyamaka concerns into the bsDus pa. To give a simple illustration, whereas Phya pa's very first example in the theory of definition already involves the "emptiness of causal efficiency," it is the cow and its hump that figure as gTsan nag pa's first and main example. As the same is the case for authors following gTsan nag pa, such as mTshur ston gZon nu sen ge or Chu mig pa Sen ge dpal,95 it could also be a sign that
92 De kho na fiid bsdus pa 73,4-14, and in particular lines 6-11: dpyad bzod pa'i don byed pas stoTi pa blo'i yul du byas pa de'aTi kun rdzob kyi bden pa yin te I skye la sogs pa bkag pa yaTi II yaTi dag pa daTi mthun phyir 'dod II ces dbu ma bden gfiis su bSad pa'i tshul gyis spros pa phyogs re tsam bkag pa de ni mthun pa'i don dam ies pa'am rnam graTis kyi don dam pa ies miTi de ltar btags par zad kyi de ni kun rdzob kyi bye brag yin no II 93 See bsDus pa 14a4-7 (chos de dag Tio bo mi gcig pa dgag pa). 94 gTsan nag pa deals with the definition of the two truths, albeit concisely, in his commentary on the Bodhicaryavatara; see sPyod 'jug 36a3b5 (es par bya ba bden pa gfiis kyi raTi biin) (my thanks to Kevin Vose for this reference). 95 Concerning the latter, if he does not, so far as I know, include in his presentation of the theory of definition discussions involving the two truths, or the distinction between implicative and non-implicative negation, he does on the other hand include a whole section on the two truths

The origin of the theory of definition

361

once the theory of definition was incorporated into the framework of epistemological treatises, the discussions relative to specific applications were' reduced in favor of a more theoretical presentation. Conclusion This tentative sketch ofthe early development of the Tibetan theory of definition remains incomplete. In particular, it remains unclear to what extent and in which context Phya pa's predecessors discussed issues related to definition. Although a number of passages in rNog Lo's works and the citations attributed to his disciples in later works substantiate their knowledge of some definitional model, we have as yet no available evidence of a systematic presentation prior to Phya pa, either in a work on epistemology or as an excursus. rGya dmar pa's works on epistemology are not extant, and his commentary on the Bodhicaryiivatiira does not involve a discussion of definition.96 As for rNog Lo's student Gro lun pa, in his voluminous bsTan rim he makes occasional mention of the pairs mtshan fiid/mtshan gii, mtshan fiid/mtshon bya, and mtshon bya/ mtshon byed (but never, as far as I know, of a triad), but does not offer a discussion of these notions. The numerous Indian Buddhist works containing the terms "lak~a1J.a" and "lak~ya" are likely to have promoted discussions about these notions and their translation. Our as yet limited look in this direction, starting with the commentaries on the AA, has revealed that if passages involving the notions of "lak~a1J.a" and "lak~ya" provided the occasion for the application of a definitional model, they do not appear to have prompted theoretical digressions (the possibility remains that Sakya mchog Idan, when mentioning
from the point of view of the idealist school (sems tsam pa) when discussing the ontological status of phenomenal objects (gzUli yut). See rNam rgyal 6a7ff.

96 The only context in which rGya dmar pa uses definition-related terminology is his discussion of the definitions of the two truths (Byari chub sems dpa'i spyod pa la jug pa'i tshig don gsal bar Mad pa 60a2ff.). The model presupposed there is a two-term model, and no mention is made of a "definitional basis."

362

Pascale Hugon

the extensive developments on definition in the commentarialliterature on the AA, might be referring to sources that are not available to us). One might have expected such an excursus from rNog Lo; his commentary on theAA shows, however, a state in which mtshan gii and mtshon bya have not yet become distinct elements in a triadic model of definition, but stand as nuances in the expression of the counterpart of lak'fWla. It is possible that this commentary is earlier than rNog Lo's dKa' gnas, where an (at least rudimentary) triadic model is presupposed and applied. Although the Tibetan tradition does not point to Indian precedents,97 I have suggested that a plausible source for this basic model and several of its particularities can be found in Dharmottara's discussion of the "definitional logical reason" (a notion already present in Dignaga's works, but absent in those of Dharmaklrti), and that, in general, the parallelism with inference might have played a role in the development of the theory of definition. 98 This possible source of influence does not in the least rule out that the motivation for elaborating the theory of definition exceeded the scope of the theory of inference. If the context of discussions of definition prior to Phya pa remains unclear, for Phya pa himself, . although the systematization of the theory of definition occurs in the context of his epistemological treatises, an important motivation in its development appears to be the need to define the two 97 As mentioned earlier, Phya pa does not try to legitimize his theory by invoking Indian sources. But the fact that he assimilates the triad in MSA XI.36 to the definitional triad could be a sign that he nevertheless considered that this theory had an Indian background. 98 A passage defining mtshan fiid and mtshon bya attributed to Sankanlnandana by Glo bo mkhan chen could also, if it is indeed a citation and not a paraphrase in which Tibetan terminology left its mark, hint at another source of influence from the side of the Indian pramiilJa tradition. Cf. Rigs gter Ni ma 160,8-11: kha che'i rtog ge pa chen po bde byed dga' ba na re chos gcig fiid don gyi rnam par sar ba'i bios yul du byas pa'i cha nas mtshan fiid daft / tha sfiad sgra'i rnam pa sar ba'i bios yul du byas pa'i cha nas mtshon bya fiid du bied do /. I could not find any equivalent in the works of Sankaranandana available in Tibetan in the bsTan 'gyur.

The origin of the theory of definition

363

truths. While Phya pa's Madhyamaka works abound with epistemological terrrtinology, his epistemological works convey the converse impression: not only in the context of the "innovative" topic of definition, but also when dealing with universals, with relations or inference, one sees his discussion turning towards the questions of emptiness, of ultimate truth and the connected proofs thereof. This is not to say that, from Phya pa's perspective, Madhyamaka and epistemology, dbu ma and tshad ma, are undifferentiated: the works of Phya pa represent different facets of Buddhist scholastics in their respective orientation, but reveal a system in which they are not compartmentalized.

Bibliography
-. Aryalankavatarasutra. P.L. Vaidya (ed.). Darbhanga, 1963: The Mithila Institute (Buddhist Sanskrit Texts 3). Tib. in DI07, Ca 56al-191b7. -. bKa' gdams gsun 'bum phyogs bsgrigs thens dan po/gfiis pa. Ed. dPal brtsegs bod yig dpe mitt fib 'jug khan. Vols. I-XXX, Chengdu, 2006; vols. XXXI-LX, Chengdu, 2007: Si khron mi rigs dpe skrun khan. -. Mahavyutpatti. In Bon-zo-kan-wa Shiyaku-taiko Honyaku Myogi Taishu {Mahavyutpatti, Skt. Tib. Chin. Jap. juxtaposed}, Sakaki R. (ed.), 2 vols., Kyoto, 1916-25 (reprint Tokyo, 1965). -. sDe dge Tibetan Tripitaka bsTan 'gyur -preserved at the Faculty ofLetters, University of Tokyo (D). J. Takasaki, Z. Yamaguchi, Y. Ejima (eds.), Tokyo, 1981-1984. -. The Tibetan Tripitaka, Peking edition (P). D.T. Suzuki (ed.), Tokyo-Kyoto, 1955-1961.
Abhayiikaragupta. Arya~tasahasrikaprajfiaparamitavrttimarmakaumudr. Tib. in D3805, Da 1b1~228a4. Anonymous (attributed to Klon chen rab 'byams pa). Tshad ma'i de kho na fiid bsdus pa (De kho na fUd bsdus pa), Padma tshul khrims (ed.). Chengdu, 2000: Si khron mi rigs dpe skrun khan. Ar Byan chub ye ses. mNon par rtogs rgyan gyi 'grel ba rnam par 'byed pa. In bKa' gdams gsun 'bum, vol. II, 91-447. Asvabhiiva. Mahayanasutralankaratrka. Tib. in D4029, Bi 38b6-174a7. Avalokitavrata. Prajfiapradfpatrka. Tib. in D3859, Wa 1b1-287a7, ta 1b1338a7, Za 1b1-341a7.

364

Pascale Hugon

Bhaviveka. Prajiiapradfpa. Tib. in D3853, Tsha 45b4-259b3. Buddhapalita. Mulamadhyamakavrtti. Tib. in D3842, Tsha 158bl-281a4. CandrakIrti. Prasannapada. For the Skt. see Nagarjuna, Mulamadhyamakakarika. Tib. in D3860, 'k Ibl-200a7. Chu mig pa Sen ge dpal. gtan gyi phyogs thams cad las rnam par rgyal ba (rNam rgyal). In bKa' gdams gsun 'bum, vol. XLV, 11-163. D'Amato, Mario. 2003. "The Semiotics of Signlessness: A Buddhist Doctrine of Signs." Semiotica 147.1/4: 185-207.
Datp.~trasena. Aryasatasahasrikiipaiicavirrtsatisahasrika~tadasasahasrika

prajiiaparamitabrha!!fka. Tib. in D3808, Pha Ibl-292b3. Devendrabuddhi. Pramiir:taviirttikapaiijika. Tib. in D4217, Che Ibl-326b4. DharmakIrti. Nyiiyabindu (NB). In Nyayabindu with Dharmottara's Nyayabindu!fka and Durveka Misra's Dharmottarapradfpa, Dalsukh Bhai Malvania (ed.). Patna, 1955: Kashi Prasad Jayaswal Research Institute [Second ed. 1971]. - . PramalJavarttika (PV) and PramalJ-avarttikasvavrtti (PVSV). In The PramalJ-aviirttikam of Dharmakrrti: The First Chapter with the Autocommentary, R. Gnoli (ed.). Rome, 1960: Istituto Italiano per il Medio ed Estremo Oriente (Serie Orientale Roma XXIII). - . PramiilJ-aviniScaya (PVin). Chap. I and II in Dharmakfrti's PramalJ-aviniscayatt, Chapters 1 and 2, E. Steinkel1ner (ed.), Beijing/Vienna, 2007: China Tibetology Publishing House/Austrian Academy of Sciences Press (Sanskrit Texts from the Tibetan Autonomous Region No.2). Chap. ill Tib. in D4211, Ce 152bl-230a7. Dharmottara. Nyayabindu!tka (NHJ). See DharmakIrti, Nyayabindu. - . PramalJaviniscaya!fka (PVinT). Tib. in D4229, Dze Ibl-289a7 and D4227, Tshe Ibl-178a3. Dignaga. PramalJ-asamuccaya (PS) and PramalJ-asamuccayavftti (PSV). Tib. by Vasudharara~ita (V) in D4203, Ce Ibl-13a7 and D4204, Ce 14b185b7, by Kanakavarman (K) in P5700, Ce lal-13a3 and P5702, Ce 93bl177a5. Durvekamisra. Dharmottarapradfpa (DhPrJ. See DharmakIrti, Nyiiyabindu. Fukuda, Yoichi. 2003. "Shoki chibetto ronrigaku ni okeru mtshan mtshon gzhi gsum 0 meguru giron ni tsuite." Journal of the Japanese Association for Tibetan Studies 49: 13-25. 'Gos Lo tsa ba gZon nu dpal. Deb ther snon po. In The Blue Annals, reproduced by Lokesh Chandra. New Delhi, 1974: International Academy of Indian Culture (Sata-pitaka Series, vol. CCXII). 010 bo mkhan chen bSod nams lhun grub. sDe bdun mdo dan bcas pa'i dgons 'grel tshad ma rigs pa'i gter gyi 'grel pa'i rnam Mad rigs lam gsal ba'i iii ma (Rigs gter iii ma). In Tshad ma rigs gter gyi 'grel pa, rDo rje rgyal po

The origin of the theory of definition

365

(ed.), dByaJis can sen ge (gen. ed.), 1-262. Qinghai, 1991: Kruil go'i bod .kyi ses rig dpe skrun khan. Go rams pa bSodilams sen ge. sDe bdun mdo dan bcas pa'i dgons pa phyin ci ma log par 'grel pa tshad ma rigs pa'i gter gyi don gsal bar byed pa (Rigs gter gSal byed). In Sa skya bka' 'bum: The Complete Works of the Great Masters of the Sa skya Sect of the Tibetan Buddhism (sDe dge edition), compiled by bSod nams rgya mtsho, vol. XI, Kha lal-148a2. Tokyo, 1969: The Toyo Bunko. Gro lun pa Blo gros 'bymJ. gnas. bsTan rim chen mo'i stod cha/smad chao In bKa' gdams gsun 'bum, vol. IV, 35-735 and vol. V, 3-242. gSer mdog pal). chen Sakya mchog ldan. Tshad ma'i mdo dan bstan bcos kyi sin rta'i srol mams ji Itar byun ba'i tshul gtam bya ba iiin mor byed pa'i snan bas dpyod ldan mtha' dag dga' bar byed pa (dGd byed). In The Complete Works (gSun 'bum) of gSer mdog paT} chen Slikya mchog Idan, Kunzang Tobgey (ed.), vol. XIX, Dza lal-69a6. Timphu, 1975 [Reprint: Nagwang Topgyal, Delhi, 1988]. - . Tshad ma rigs pa'i gter gyi dgons rgyan rigs pa'i 'khor los lugs nan pham byed (fham byed). In The Complete Works (gSun 'bum) of gSer mdog paT} chen Slikya mchog ldan, Kunzang Tobgey (ed.), vol. X, Tha lal-294a4. Timphu, 1975 [Reprint: Nagwang Topgyal, Delhi, 1988]. gTsan nag pa btrson 'grus sen ge. Tshad ma mam par nes pa'i !i ka legs Mad bsdus pa (bsDus pa). Kyoto, 1989: Rinsen Book Co (Otani University Tibetan Works Series, Volume IT). - . sPyod 'jug gi mam Mad (sPyod 'jug). In bKa' gdams gsun 'bum, vol. XIII,487-579. gZon nu rin chen. mNon par rtogs pa'i rgyan gyi 'grel ba bcud kyi thig. In bKa' gdams gsun 'bum, vol. XXI, 15-303. Haribhadra. AbhisamaylilankliraklirikliSlistravivrti (AAKSV). In AbhisamayliIa'!lklira-klirikli-slistra-vivrti. Haribhadra's commentary on the Abhisamaylila'!lklira-klirikli-slistra edited for the first time from a Sanskrit manuscript, Koei H. Amano (ed.). Kyoto, 2000: Heirakuji-Shoten. Tib. in D3793, fa 78bl-140a7. - . Abhisamaylilanklirlilokli (AAA). In Abhisamaylila'!lklirlilokli Prajfilipliramitlivylikhyli (Commentary on A>taslihasrikli-prajfilipliramitli) by Haribhadra together with the text commented on, Unrai Wogihara (ed.). Tokyo 1932-1935: The Toyo Bunko. Tib. in D3791, Cha Ibl-341a7. Hugon, Pascale. 2008a. Tresors du raisonnement. Sa skya PaT}tj,ita et ses predecesseurs tibetains sur les modes de fonctionnement de la pensee et Ie fondement de ['inference. Edition et traduction annotee du quatrieme chapitre et d'une section du dixieme chapitre du Tshad rna rigs pa'i gter. Vienna: Arbeitskreis fUr Tibetische und Buddhistische Studien (Wiener Studien zur Tibetologie und Buddhismuskunde 69.1 and 69.2).

366

Pascale Hugon

- . 2008b. "Arguments by Parallels in the Epistemological Works of Phya pa Chos kyi sen ge." Argumentation 22: 93-114. (DOl 1O.1007/s10503007-9074-7) Jayananda. Madhyamakavataratfka. Tib. in D3870, Ra 1bl-365a7. linendrabuddhi. Pramaflasamuccayatfka (PST). Chap. I Skt. in Jinendrabuddhi's Visalamalavatf Pramaflasamuccayatfka, E. Steinkellner, R Krasser and R Lasic (eds.). Beijing/Vienna, 2005: China Tibetology Publishing House/Austrian Academy of Sciences Press (Sanskrit Texts from the Tibetan Autonomous Region No.1). Tib. in D4268, Ye 1b1314a7, P5755 Re 1-355a8. Ifianagarbha. Satyadvayavibhmiga. In Jfianagarbha's Commentary on the Distinction Between the Two Truths: An Eighth Century Handbook of Madhyamaka Philosophy, M.D. Eckel (ed. and transL). Albany, New York, 1987: State University of New York Press. Kamaiasiia. Tattvasarigrahapafijika (TSP). See Santarak~ita, Tattvasarigraha. Tib. in P5765, 'E 159b2-Ye 404b6. KarI,lakagomin. Pramaflavarttikasvavrttitfka (PVSVT). In Karflakagomin's Commentary on the Pramaflavarttikavrtti of Dharmakfrti, Rahula Sankrtyayana (ed.). Kyoto, 1982: Rinsen Book Co. (RBTS 1). Katsura, Sh5ryu. 1977. "lZ9fjJlIEEJ.r~~liTf~ [-]" {"A Study of the Nyayamukha I"}. The Hiroshima University Studies (Literature Department) 37: 106-126. Kellner, Birgit. 2004. "Why Infer and not just Look? Dharmaklrti on the Psychology ofInferential Process." In The Role of the Example (dr~tanta) in Classical Indian Logic, Sh. Katsura and E. Steinkellner (eds.), 1-51. Vienna: Arbeitskreis flir Tibetische und Buddhistische Studien (Wiener Studien zur Tibetologie und Buddhismuskunde 58). La Vallee Poussin, Louis de. 1971. L'Abhidharmakosa de Vasubandhu : traduction et annotations, 6 volumes, nouvelle edition preparee par E. Lamotte. Bruxelles: Institut beIge des hautes etudes chinoises (Melanges chinois et bouddhiques, voL XVI) [original edition: Paris, 1923-1931: Geuthner; Louvain: J.-B. Istas]. Lalou, Marcelle. 1953. "Les textes bouddhiques au temps du roi Khri-sronglde-bt~an." Journal Asiatique 241: 313-353. Maitreyanatha/Asanga. Abhisamayalarikara (AA). In AbhisamayalarikaraPrajfiaparamita-Upade.a-Sastra: The Work of Bodhisattva Maitreya, E. Obermiller and T. Stcherbatsky (eds.). Delhi, 1992: Motilal Banarsidass [First Indian edition; original edition: Leningrad, 1929]. Tib. in D3786, Ka Ibl-13a7. - . Mahayanasatralarikara (MSA). In Mahayana-SatralalJ1kara .. expose de la doctrine du Grand vehicule selon le systeme Yogacara, 2 vols., Sylvain Levi (ed. and trans.). Paris, 1907-1911: H. Champion (Bibliotheque

The origin of the theory of definition

367

de l'Ecole des Hautes Etudes. Section 4, Sciences historiques et philologiques 159 and 190). Tib. in D4020, Phi Ibl-39a4. rnTshur ston gZon nu sen ge. Tshad ma ses rab sgron ma (sGron ma). In mTshur ston gZon nu sen ge, Tshad rna ses rab sgron rna, P. Hugon (ed.). Vienna, 2004: Arbeitskreis fUr Tibetische und Buddhistische Studien (Wiener Studien zur Tibetologie und Buddhisrnuskunde 60). Niigiirjuna. Mulamadhyamakakiirikii. In Mulamadhyamakakiitikiis (Miidhyamikasutras) de Niigiirjuna avec la Prasannapadii, Commentaire de Candrakfrti, Louis de la Vallee Poussin (ed.). st. Petersburg, 1903-1913: Russian Academy of Sciences (Bibliotheca Buddhica 4) [Reprint: Delhi, 1992: Motilal Banarsidass]. Tib. in D3824, Tsa Ibl-19a6. Onoda, Shunzo. 1984. "rntshan nyid to rntshon bya ni tsuite." Indogaku bukkyogaku kenkyu 33.1: 92-95. Phywa pa Chos kyi sen ge. Theg chen mdo sde rgyan gyi legs Mad rgyan 'iii 'od gsal ba. In bKa' gdams gsun 'bum, vol. VII, 352-537. - . dBu ma bden giiis kyi 'gre! ba (bDen giiis 'grel). In bKa' gdams gsun 'bum, vol. VI, 185-250. - . dBu ma sar gsum gyi ston thun (sTon thun). In Phya pa Chos kyi sen ge, dBu rna sar gsurn gyi ston thun, H. Tauscher (ed.). Vienna: Arbeitskreis fUr Tibetische und Buddhistische Studien (Wiener Studien zur Tibetologie und Buddhisrnuskunde 43). - . Tshad ma rnam par nes pa'i 'grel Mad yi ge dan rigs pa'i gnad la 'jug pa'i ses rab kyi 'od zer (,Od zer). In bKa' gdams gsun 'bum, vol. VIII, 35-427. - . Tshad ma yid kyi mun pa sel pa (Mun sel). In bKa' gdams gsun 'bum, vol. VIII, 434-626. Prajfiiikiiragupta. PramiilJaviirttikiilankiira. In PramiilJaviirtikabhiishyam or Viirtikiilankiira!:t ofPrajiiiikaragupta, R. Siinkrtyiiyana (ed.). Patna, 1953: Kashi Prasad Jayaswal Research Institute. Tib. in D4221, Te Ibl-308a7. Ratniikarasiinti. Aryii~tasiihasrikiiprajiiiipiiramitiipaiijikiisiirottamii. Tib. in D3803, Ta 204a3-270a7. rGya drnar pa Byan chub grags. Byan chub sems dpa'i spyod pa la 'jug pa'i tshig don gsal bar Mad pa. In bKa' gdams gsun 'bum, vol. VI, 11-174. rNog Bio ldan ses rab. mNon par rtogs pa'i rgyan 'grel rin chen sgron me. In bKa' gdams gsun 'bum, vol. I, 125-201. - . mDo sde rgyan gyi don bsdus. In bKa' gdams gsun 'bum, vol. I, 207-253. - . Tshad ma rnam nes kyi dka' ba'i gnas rnam par Mad pa (dKa' gnas). Sun Wenjing (ed.), Qinghai, 1994: Krun go'i bod kyi ses rig dpe skrun khaJi. - . mTshan iiid kyi chos. In bKa' gdams gsun 'bum, vol. I, 369-409. Sa skya Pa~c;lita Kun dga' rgyal rntshan. Tshad ma rigs pa'i gter (Rigs gter). In Tshad ma rigs pa'i gter gyi ran gi 'grel pa, Nor bran 0 rgyan (ed.). Lhasa, 1989: Bod Ijons rni drnans dpe skrun khan.

368

Pascale Hugon

Sakyabuddhi. PramiilJaviirttikatzkii (PV'!'). Tib. in D4220, Je 1bl-328a7 and He 1bl-282a7. Santarak~ita. Tattvasafigraha (TS). In Tattvasmigraha ofAciirya Siintarak~ita with the Commentary Pafijikii of Srz Kamalasfla, 2 vols., Svaml Dvarikadas SastrY (ed.). Varanasi, 1981: Bauddha Bharati. sGros riiin pa 'Od zer mgon po. mNon par rtogs pa'i rgyan gyi rnam par 'byed pa'i 'grel ba. In bKa' gdams gsuri 'bum, vol. XV, 27-322. Stearns, Cyrus. 2001. Luminous Lives: The Story of the Early Masters of the Lam 'bras Tradition in Tibet. Boston: Wisdom Publications (Studies in Indian and Tibetan Buddhism). Tauscher, Helmut. 1999. "Phya pa Chos kyi seng ge's Opinion on PrasaIiga in His Dbu ma'i shar gsum gyi stong thun." In Dharmakfrti's Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of the Third International Dharmakfrti Conference. Hiroshima, November 4-6, 1997, Sh. Katsura (ed.), 387-393. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften. -. 2003. "Phya pa Chos kyi seng ge As a Svatantrika." In The SviitantrikaPriisarigika Distinction. What difference does a difference make? G. Dreyfus and S. McClintock (eds.), 207-255. Boston: Wisdom Publications (Studies in Indian and Tibetan Buddhism). van der Kuijp, Leonard W.J. 1983. Contributions to the Development of Tibetan Buddhist Epistemology from the Eleventh Century to the Thirteenth Century. Wiesbaden: Franz Steiner Verlag. - . 2003. "A Treatise on Buddhist Epistemology and Logic Attributed to Klon chen rab 'byams pa (1308-1364) and Its Place in Indo-Tibetan Intellectual History." Journal of Indian Philosophy 31: 381-437. Vasubandhu. Abhidharmakosa (AK) and Abhidharmakosabhii~ya (AKBh). In Abhidharmakoshabhii~ya of Vasubandhu, Pradhan (ed.). Patna, 1967 [Reedition 1975]. Tib. in D4089, Ku Ibl-25a7 and D4090, Ku 26bl258a7 and Khu Ibl-95a7. VinItadeva. Nyiiyabindutfkii (NBTv). Tib. in D4230, We 1bl-36b2.

Phywa pa's argumentative analogy between factive assessment (yid dpyod) and conceptual thought (rtog pa)
Jonathan Stoltz

In order to understand how knowledge is obtained, it is important first to understand the workings of the mind. Thus, it is not surprising that when Tibetan Buddhist philosophers began composing epistemological "summaries" (bsdus pa) in the eleventh or twelfth century, they placed a discussion of mental states - or, more accurately, mental episodes - at the very beginning of those texts. Because the authors of these texts held that all cognitive episodes take some object as their content, part of this investigation into cognitive episodes involves explaining the relations that these episodes bear to those objects. In this paper, I will be delving into one particular topic within this Buddhist theory of cognition. I will examine a single argument by Phywa pa Chos kyi sen ge (henceforth, Phywa pa; 1109-1169) contained within his famous epistemology text, the Tshad rna yid kyi rnun sel, drawing out the philosophical implications that this argument has on his theory of cognition and his account of ontological dependence. 1 Specifically, I will make
1 I use the term "dependence" to translate the Tibetan term 'breI ba. Many other writers translate 'breI ba as "relation," which is what the term means in its colloquial usage. In philosophical contexts, however, the term clearly has the stronger sense of dependence - of one object depending on another. One further consideration that motivates my decision to translate the term 'breI ba as "dependence" is the following:

a. Dependence (x depends on y) is an asymmetric relation [i.e., if x depends on y, it needn't necessarily be the case that y depends on x] b. Relatedness (x is related to y) is a symmetric relation [i.e., if x is related
Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 369-386

370

Jonathan Stoltz

the case that Phywa pa's argument fails to explain adequately the nature of the relation between certain cognitive episodes and the contents of those episodes. In addition, I will show that Phywa pa is forced to accept an arguably dubious version of externalism about mental content. Mental operations and their contents Before evaluating Phywa pa's argument, it is important to clarify a number of points about the nature of cognitive states/episodes. In so doing, it will be possible both to see more clearly what issue Phywa pa is grappling with in his argument, and to better understand some of the potential problems Phywa pa's position might face. Following the lead of his predecessor rNog BIo ldan ses rab (1059-1109), Phywa pa develops an account of cognition in which it is maintained that there are seven distinct kinds of cognitive episodes. 2 What exactly does it mean, however, to be a "kind of cognitive episode," and what sorts of features distinguish cognitive episodes as being of different kinds? All cognitive states and cognitive episodes are, at the very least, some sort of mental operation on some content. To take an example from contemporary philosophy, consider a mental state such as believing that Jones will win the lottery. Here, the mental operation is that of believing, and the content of this operation is the proposition that Jones will win the lottery. If we change the content to some other proposition - such as the proposition that Jones will lose the

to y, it necessarily follows that y is related to x]. The Tibetan term 'breI ba, when used in its philosophical sense in epistemology treatises, is clearly an asymmetric relation (for example, while smoke [du ba] causally depends on fire [me], the opposite is not the case; fire does not depend on smoke). 2 See rNog (2006), 9b5 and following, and Phywa pa (2006), 8a6 and following. The seven kinds of cognitive episodes are: valid perception (mfwn sum tshad rna), inference (rjes dpag), indeterminate appearing (snan. la rna lies pa), subsequent cognition (bead pa'j yul ean), factive assessment (yid dpyod), false cognition (log ses), and doubt (the tshom).

Phywa pa's argumentative analogy

371

lottery - the particular mental state will no doubt be different, but it will not (all other things being equal) be a different kind of mental state. On the either hand, if one were to change the original mental operation to one of desiring, then the resultant mental state, desiring that Jones will win the lottery, would be different in kind from the original. This suggests that it is the particular operation that determines what type of mental state one is in. By contrast, changes in content do not necessarily entail changes in the type of mental state. 3 That said, as will be explained below, it is possible that certain facts about the content of the mental state could play a role in determining the type of mental state one is in or the type of mental operation that occurs. Putting this in the context of Phywa pa's epistemological program, we can now articulate more clearly what is meant when he speaks of there being seven types of cognitive episodes (blo bdun). Each of the seven kinds of cognitive episodes is a distinct kind of mental operation. For instance, perceiving a fire on a hill is different from inferring the existence of a fire on a hill. Both of these cognitive episodes take a fire on a hill as their object of engagement ('jug yut), but the mental operations - those of perceiving and inferring - are different in kind. One additional point of note is that while each of the seven types of cognitive episodes described by Phywa pa is a distinct kind of mental operation, it may still be the case that these seven types of cognitive episodes share in common more basic mental operations. 4
3 It is possible, however, that changes in the type of content could necessitate changes in the type of mental state one is in. For example, a mental operation that takes propositions as its content will likely be different in kind from a mental operation that takes some non-propositional object as its content. In the Tibetan tradition, one of Phywa pa's typologies of cognitive episodes - his three-fold typology of rtog pa, rtog med ma 'khrul pa, and rtog med 'khrul pa - is one where cognitions are divided into three different kinds by way of the types of objects appearing in those episodes. See Phywa pa (2006), 4alff.
4 By analogy, while the operations of doubling (J(x) = 2x) and tripling (J(x) = 3x) are distinct mathematical operations, both could be said to

be instances of the operation of multiplying. More generally, even when

372

Jonathan Stoltz

Specifically, all seven kinds of cognitive episodes share in common the operation of 'affirming their object(s)' (don yons su gcod pa). Thus, while perceiving (mnon sum) a fire is a different kind of cognitive episode from falsely cognizing (log ses) a fire, both of these types of cognitive episodes are instances of the mental operation of cognitively affirming a fire. The moral to draw from this is that just because two cognitive episodes are instances of the same general mental operation on the same content, it does not necessarily follow that they are the same type of cognitive episode full stop, for they may differ with respect to being instances of other, more specific, mental operations.
The argument

Among the seven kinds of cognitive episodes described by Phywa pa, one, yid dpyod - which I translate into English as "factive assessment" - is portrayed as a conceptual mental episode similar to inference that, by definition, has novel, true content, but that falls short of being an episode of knowledge due to its not being grounded in adequate evidence. 5 For example, if, upon seeing people carrying water up a hill, I were to form the belief that there is a fire burning on the top of that hill (due to my mistakenly thinking . that whenever people carry water up a hill there must be a fire on the hilltop), my resulting judgment, if true, would be an instance of factive assessment. I make a true judgment through an inferencelike process, but the mental episode is not genuinely inferential, because it is based on faulty reasoning. After providing his definition of factive assessment and offering a nuanced typology of the various subtypes of this mental epiA and B' are two different types of operations, there may still be some operation C, such that performing operation A entails performing operation C and performing operation B entails performing operation C. 5 For more on the role played by factive assessment in bKa' gdams Tibetan epistemology, see my (2007). In that article I address some of the broader ways in which Phywa pa's argument bears on our understanding of Tibetan views of cognition. Phywa pa discusses factive assessment most extensively in his (2006), lOa7ff.

Phywa pa's argumentative analogy

373

sode, Phywa pa goes on to defend his account of factive assessment against possible criticisms. In his defense of factive assessment against one of these criticisms, Phywa pa provides an argument by analogy, comparing the relationship between factive ass~ssment and its object to the relationship between conceptual thought and its object - concepts. 6 By examining this argument, much can be learned about Phywa pa's overarching theory of cognition and cognitive objects. The argument by analogy that I will be focusing on is just one component strand within a larger argument that Phywa pa takes up in this particular passage. The more general thrust of the argument relates to the issue of how evidence is used in inferential reasoning, and whether the definition of factive assessment requires us to accept a new type of evidence: evidence lacking dependence. More specifically, Phywa pa argues that while episodes of factive assessment logically entail the possession of a true object, it does not follow that one can appeal to an occurrence of factive assessment as evidence for inferring that there is a true object, for one can ascertain that one is having an episode of factive assessment only after already identifying the object as true. While keeping in mind that much of his argument deals with that issue, I have framed the argument by analogy in such a way as to focus exclusively on the issue of cognitive episodes and their contents.

6 Our understanding of Phywa pa's use of analogical reasoning has been greatly enhanced through recent research by Pascale Hugon (2008). She has shown in intricate detail how analogical reasoning is employed in Phywa pa's writing, and has speculated on what these appeals to analogies can tell us about Phywa pa's theory of argumentation and inference. In the remainder of this paper I will focus on just one of Phywa pa's arguments that employs analogical reasoning. Instead of looking at this analogy with the intent of understanding Phywa pa's theory of argumentation, however, I will dissect the analogy in the hopes of learning more about Phywa pa's philosophical views concerning the nature of mental content.

374

Jonathan Stoltz

The argument is presented as a back and forth dialectic between Phywa pa and a hypothetical philosophical antagonist? The dialectic contains four parts:
1. The antagonist begins by raising the problem: Since all episodes of factive assessment (yid dpyod) must have true contents (yul bden pa), it would follow either that all such cognitive episodes depend on their contents, or that there is a type of evidence (one that could be used to draw inferences) that is not grounded in any sort of dependence relation (' brel med kyi rtags). But neither of these possibilities is acceptable. In short, the criticism is that if all episodes of factive assessment have true contents, unacceptable consequences follow.

2. Phywa pa replies by posing an analogous problem. Specifically, he constructs a case parallel to the antagonist's, dealing with the relation between conceptual thought (rtog pa) and concepts (don spyi). Just as in the original case, Phywa pa points out that, by analogy, if all episodes of conceptual thought entail the presence of concepts, then an unacceptable consequence follows. 3. The antagonist provides a solution to Phywa pa's analogous case, showing that there is really nothing problematic about the relation between conceptual thought and concepts, for this is a case where, while conceptual thought entails the presence of concepts, conceptual thought cannot be evidence for the presence of concepts. This is because one could never be in the epistemic position of knowing that one is having an episode of conceptual thought without already knowing that concepts are present. 4. Phywa pa applies the antagonist's solution to the original case, shqwing that, similarly, there is no problem making sense of
7 Phywa pa (2006), lOb6-9. The text of the argument and an English translation are included in the appendix. This same argument is spelled out by one of Phywa pa's successors, the author of the Tshad ma'i de kho na fiid bsdus pa, a work which has been, likely mistakenly, attributed to Klon chen pa. See Klon chen (2000), pp. 58-59. For more on the authorship of the above text see van der Kuijp (2003).

Phywa pa's argumentative analogy

375

the relation between factive assessment and its object for the same relationship holds. Factive assessment and true content This argument by analogy highlights a number of important philosophical issues. I find the most interesting of these issues to be that of how mental episodes are related to the contents of those episodes. Phywa pa, like other bKa' gdams epistemologists, maintains that all cognitive episodes (blo) have some sort of object (yul). That is, all cognitive episodes are cognitions of something. In the above argument, Phywa pa is focusing in on the relation between one particular kind of cognitive episode - factive assessment (yid dpyod) - and the objects or contents of those cognitive episodes. With his argument by analogy, Phywa pa is attempting to defend factive assessment against the criticism that there is something illegitimate about its postulated existence. Given its definition, all episodes of factive assessment have a true object (yul bden pa or don bden pa). Put in another way, episodes of factive assessment are truth entailing: if someone factively assesses P, it follows that P is a true object. Provided that it is of the nature of episodes of factive assessment that they are truth entailing, one might presume that this entailment is ensured by some sort of deep metaphysical relationship between this kind of mental episode and its content. That is, one might think that episodes of factive assessment depend on their objects. Since, for example, I cannot factively assess that there is a fire on the other side of the hill unless there actually is a fire on the other side of the hill, one might think that my having this mental episode depends on the existence of that fire. Phywa pa cannot accept any such dependence, however. As is stated in his argument, Phywa pa supports the now standard Buddhist view that there are only two types of dependence grounded in the nature of entities (ran biin gyi 'breI pa, svabhiivapratibandha), and neither of these types of dependence is compatible with the

376

Jonathan Stoltz

relationship between factive assessment and its object. 8 Episodes of factive assessment cannot essentially depend (bdag gcig tu 'brei ba) on their objects, for essential dependence is a relation that requires the two relata (ordinarily; properties) to share the same ontological substratum. In this case that would require factive assessment to be physical matter, since it must be of the same nature as its object. Nor can the relation between factive assessment and its object of engagement be one of causal dependence (de byun 'brei ba), for if an extra-mental object did cause a person's cognition of it, such a cognition would actually be a form of knowledge (a pramiilJa) akin to perception and inference. 9 As such, Phywa pa seeks to accept the following two theses: (a) All episodes of factive assessment have true objects. (b) Episodes of factive assessment do not depend on their objects. In order to support his contention that (a) and (b) can both be true, Phywa pa draws an analogy in which he appeals to the relationship between episodes of conceptual thought and their phenomenal contents, concepts. Without explicitly endorsing it, he is presenting the view that the relationship between factive assessment (yid dpyod) " and its object is analogous to the relationship between conceptual thought (rtog pa) and its object - concepts (don spyi). Phywa pa takes it to be obvious that, (a*) All episodes of conceptual thought take concepts as their objects, and yet, (b*) Episodes of conceptual thought do not depend on concepts.

8 The view that there are just two forms of dependence (bdag gcig tu 'brei ba and de byun 'brei ba, Skt. tadatmya/tadutpatti) becomes standard with DharmakIrti (see Dunne 2004, p. 152). 9 The reasoning here is that if factive assessment causally depended on its object, it would not arise in the absence of its object (med na mi 'byun ba, avinabhava), which would be a sufficient condition for knowledge (tshad ma, pramalJa).

Phywa pa's argumentative analogy

377

The first claim (a*), just like the parallel claim (a), is true by definition. On Phywa pa's analysis of conceptual thought, it is precisely the class of mental states that take concepts (don spyi) as their phenomenal objects (gzuri yul). The second claim (b*) is true because were the relationship between conceptual thought and concepts to be one of dependence, it would have to follow that concepts are real, causally effective things (drios po). Neither Phywa pa nor his hypothetical antagonist, however, take concepts to be real objects. The insinuation here is that just as there is nothing problematic about asserting the conjunction of (a*) and (b*), so too, by analogy, there is nothing problematic about asserting the conjunction of (a) and (b). Now, of course, Phywa pa's argument hinges on the idea that the relationship between factive assessment and its object is analogous to the relationship between conceptual thought and its object. It is thus important to look more closely at the parallels between these two cases. In order to determine whether Phywa pa's analogy is cogent, we must focus on whether the relationship between conceptual thought and concepts is sufficiently similar in structure to the relationship between factive assessment and its objects. On the one side there are two types of mental episodes: factive assessment and conceptual thought. While each of these is, in Phywa pa's epistemological system, a type of cognitive episode, conceptual thought is a much broader category than is factive assessment.lO In fact, all episodes of factive assessment are episodes of conceptual thought (but not vice versa). What is odd about the analogy, however, are the structural dissimilarities between the two objects of these mental episodes. In the case of conceptual thought, a point is being made about the mental episodes' phenomenal objects (gzuri yul), which are the objects that directly appear to the mind. Yet, in his remarks about factive assessment, when he speaks of "true objects" (don bden pa or yul bden pa) Phywa pa is certainly not making a point

10 In Phywa pa's system, factive assessment is one of the seven cognitions within his blo bdun typology, while conceptual thought is one of the three cognitions within his blo gsum typology.

378

Jonathan Stoltz

about the mental episodes' phenomenal objects.u Rather, he is apparently talking about the objects of engagement ('jug yul). This is no small difference, for in the case of conceptual mental episodes the relation between a mental episode and its phenomenal object (gzun yul) is fundamentally different from the relation between a mental episode and the cognition's object of engagement ('jug yul). The phenomenal objects of conceptual thought - concepts (don spyi) - are mentally constructed entities that dawn (,char ba) in one's mind. In Phywa pa's claims about factive assessment, however, the objects about which he is most concerned are ones that are not typically mentally constructed. Rather, in many cases they are held to be external objects (phyi rol gyi don), that are causally efficacious (don byed nus pa) and hence real (dnos po). For example, when factively assessing - from incorrect evidence - that there is a fire on the hill, the object at issue is a real, external thing. Given the differences between these two types of objects - (1) the phenomenal objects of conceptual thought and (2) the objects of engagement in episodes of factive assessment - one might have suspicions about whether Phywa pa's analogy is really adequate. In particular, given that the analogous case is grounded on claims about phenomenal objects whereas the original case is not, there is reason to think that the explanation for why conceptual thought entails the existence of concepts will be quite different from the explanation for why factive assessment entails the existence of a true object. That a given mental episode can entail the existence of some mentally constructed entity is not surprising at all. What would be surprising is a mental episode entailing the existence of an independently existing external object. Allowing for such mental episodes is tantamount to accepting externalism with respect to mental content. In essence, what this means is that which particular cognitive episode a person is having is not intrinsic to the mind. Part of what determines a person's cognitive episodes are features of the world external to and independent of the mind.
11 We know this because Phywa pa's definition of "true phenomenal object" (gzun yul bden pa) is "being able to perform functions" (don byed nus pa), which is a feature that the phenomenal objects of factive assessment (concepts) are known not to possess. See Phywa pa (2006), 9al-2.

Phywa pa's argumentative analogy

379

Externalism
In order to better understand what I am speaking of when I say that Phywa pa is committed to externalism, let us look at a couple of examples. First, consider two situations, a and ~. Situation ais one in which a person, upon glimpsing something looking like a cloud of smoke on the top of a hill- though it is actually a huge swarm of flies - mistakenly believes that it is smoke. As a result of believing there is smoke on the hilltop, this person illicitly "infers" that there is a fire on the top of the hill. Moreover, in situation a there really is a fire on the top of the hill. As such, in this situation, the person's resultant cognitive episode is, by Phywa pa's standards, one offactive assessment, for the person forms a true judgment, but does so without relying on good evidence. Situation ~ is, mutatis mutandis, identical to a except that in this case there actually is smoke rising from a fire on the top of the hill. In this latter case then, the person correctly draws an inference from the presence of the smoke to the presence of a fire on the hill. Thus, in situation ~, the person's resultant cognitive episode is one of inference. In short, the person in a has a different kind of cognitive episode than the person has in ~, despite it being the case that the facts about the objects in both situations are identical. That is, in both situations, there really is a fire on the hill. What the above two cases show is that the difference between episodes of inferential knowledge and episodes of factive assessment must be located in the way or manner in which the mental operations occur (blo'i 'dzin stails), and not in differences between the objects cognized. Despite having the same objects, the mental operation that occurs in a genuine case of inferential reasoning is different from the operation that occurs in factive assessment. In contrast to the above, consider a third situation y. This third situation is, mutatis mutandis, identical to situation a, except that in y there is no fire on the hill. In this third situation, the resultant cognitive episode, by definition, cannot be one of factive assessment. Instead, it would be labeled by Phywa pa as an episode of false cognition (log ses). We thus have two cases, a and y, that are very similar, but that differ in the kind of cognitive episode that

380

Jonathan Stoltz

results: in a, the person has an episode of factive assessment, and in y the person has an episode of false cognition. Situations a and yare identical, however, with respect to the manner in which the cognitive episodes occur. That is, the causal geneses of the cognitive episodes are identical in the two situations, as are all other relevant aspects contributing to the occurrence of the cognitive episodes. Yet, in the one situation the cognitive episode is of one type - factive assessment - whereas in the other it is an episode of a different type - false cognition. The only difference between the two cognitive episodes is a difference between the contents. In case a the resulting cognitive episode is one whose object has the property of being true, while in case y the resulting cognitive episode is one whose object has the property of being false. But since, in these cases, the content or object of the cognitive episodes does not seem to be relevant to their occurrence, it is difficult to see why the episode in situation a should be of a different type than the episode in situation y. And yet, as Phywa pa and other bKa' gdams (and dGe lugs) epistemologists contend, in these cases it is precisely a characteristic of the object of the cognitive episode - its being true or false - and only this characteristic, that determines what kind of cognitive episode a person is having. It is precisely in this way that . Phywa pa and his followers are committed to externalism about mental content. We can also see from this example why Phywa pa's claims about dependence are so puzzling. While being forced to maintain that what determines the type of cognitive episode a person is having - what determines whether it is an episode of factive assessment or false cognition - is a fact of the matter about the object of the cognition, he still holds the view that having an episode of factive assessment does not depend on the object of the cognition being true. How is this possible?

Dependence
Phywa pa's response in both the case of factive assessment's relation to its object and in the case of conceptual thought's relation to concepts is to maintain that while having such a cognitive episode entails the existence of a certain kind of object, it still isn't the case

Phywa pa's argumentative analogy

381

that the cognitive episode depends on that object. While this stance makes it possible for Phywa pa to show the reader more clearly why there needn't be 'any additional forms of evidence (rtags) - forms of evidence that are not rooted in a dependence between the involved entities - his position still does not remove the tension noted at the end of the preceding section. Phywa pa is committed to the view that part of what determines the kind of cognitive episode a person is having is the truth status of the object of the episode, but also that this determination relation doesn't imply dependence. Although such a position is not contradictory, it is still philosophically tenuous. In the paragraphs that follow, I will offer two ways in which the tensions in Phywa pa's account could be resolved. As should be obvious, one way in which Phywa's account could be made more tenable is simply by biting the bullet and granting that the relationship between factive assessment (yid dpyod) and true content (yuZ bden pa) is in fact one of dependence. At least in terms of what Phywa pa is ultimately trying to accomplish in this analogical argument - viz., to show that the existence of factive assessment does not require us to accept an additional form of evidence (rtags) capable of being used in inferential reasoning - there is no loss associated with granting that these two items do stand in a relation of dependence. Nevertheless, such a move is not something Phywa pa is in favor of, as it conflicts with the traditional Buddhist view that there are just two forms of dependence: essential dependence (bdag gcig tu 'brei ba) and causal dependence (de byUli 'brei ba). Since the relation between factive assessment and true objects fails to fit into either of these two categories of dependence, Phywa pa assumes that episodes of factive assessment can't depend on having true objects. The broader philosophical point at issue is just what we take the nature of dependence (,brei ba) to be. There are two general approaches one could take. On the one hand, one can hold that what it is for one entity to depend on another just is for the one entity either to essentially depend or causally depend on the other. Inasmuch as Phywa pa takes it for granted that factive assessment does not depend on its having a true object simpiy because it neither essentially nor causally depends on its true object, his position

382

.Jonathan Stoltz

is most similar to this first approach. On the other hand, one could hold that there is a more general concept of dependence - e.g., x depends on y iff x could not exist unless y exists - within which essential and causal dependence are merely two salient sUbtypes. This latter approach would leave open the possibility of denying the central Buddhist assumption that those are the only two forms of dependence. My own view is that it is this second approach that results in a more philosophically tenable position, especially given the influence of externalism present in Phywa pa's overarching account of cognition. One final possibility, one that requires no changes to be made to the Buddhist tradition's understanding of dependence, would be to conclude that there is something fundamentally wrong with the notion of factive assessment. Put simply, one could reject Phywa pa's claims about the relationship between factive assessment (yid dpyod) and its object (yuZ). Interestingly enough, something like this position is adopted by Sa skya Pa1).<;iita in his Tshad rna rigs gter. He rejects the claims made by supporters of factive assessment on the grounds that they illegitimately assume the existence of a kind of entity that simply does not exist. Sa skya Pa1).<;iita maintains that there are fundamental problems with bKa' gdams Tibetan accounts of factive assessmentY On Sa skya Pa1).<;iita's view, there is no such type of cognitive episode as factive assessment. All of the cognitions that bKa' gdams writers consider to be episodes of factive assessment, Saskya Pa1).<;iita classifies as episodes of doubt (the tshorn)P In summary, in accepting Phywa pa's definition of factive assessment one is confronted by a set of unappealing consequences.

Sa skya Pal).~ita (1989), pp. 59-64. 13 Sa skya Pal).~ita would be critical of the other half of Phywa pa's analogy as well (dealing with the relation between conceptual thought and concepts [don spyiD. At the very beginning of his Tshad rna rigs gter Sa skya Pal).<;iita passionately argues that concepts do not genuinely exist (1989, pp. 40-43). In particular, he argues this point by showing that hypothetically speaking - if concepts were to exist, they would be mind dependent entities incapable of being grasped by other cognitive agents.
12

Phywa pa's argumentative analogy

383

First, his account of factive assessment requires a strong version of externalism about mental content, a version of externalism that is highly counterintuitive. The consequences are even less intuitive and more controversial if one goes along with Phywa pa and denies that episodes of factive assessment depend on their having true content. If, on the other hand, one admits that these cognitive episodes do depend on their having true objects, one faces the interpretive difficulty of squaring that position with the view that there are only two types of dependence relations. In my own estimation, the difficulties here are best dealt with by denying that the definition of factive assessment describes a genuine type of cognitive episode. 14 Philosophically speaking, there is something quite bizarre about claiming that episodes of factive assessment represent a distinct type of cognitive episode. While Sa skya Pm;u;iita does not express his criticisms in the same manner as I have above, he is on the right track in his criticisms of factive assessment.

Appendix Translation of Phywa pa's argumentative analogy between factive assessment (yid dpyod) and conceptual thought (rtog pa): A: Well, if it is the case that factive assessment entails having an ascertaining consciousness with true content, then in requiring that it entails true content is there (1) entailment by essential [dependence], (2) entailment by causal [dependence], or (3) entailment without dependence? If (1) then [mental episodes of] factive assessment would become physical matter, since they would share the same essence as physical matter. If (2) then [factive assessment] would become a type of knowledge, since the cognition wouldn't arise in the absence of the affirmed object. If (3) then the acceptability of this en-

14 Of course, this is not to deny the existence of cognitive episodes that, without relying on good evidence, take true, novel objects as their contents, it is just to deny that these factors are essential to the nature of the cognitive episodes themselves.

384

Jonathan Stoltz

tailment without any form of dependence would imply the possibility of "evidence (rtags, logical reason) [that proves a property to be proved] without depending [on it]." P: Well then, such is the case also for conceptual thought entailing concepts as their phenomenal objects. Is that entailment due to (1) essential [dependence], (2) causal [dependence], or is it (3) without dependence? If (1) or (2) then insofar as concepts would be either materially identical with conceptual thought or the cause of conceptual thought, concepts would be real things. If (3) then since that would bring about entailing and entailed items (khyab byed and khyab bya) that are [nonetheless] not dependent [on each other], it would be possible for there to be "evidence [that proves a property to be proved] without depending [on it]." A: But since I don't accept [that the link between conceptual thought and concepts is based on] either of these two forms of dependence, concepts would not become real things. And although there are entailing and entailed items that don't depend on each other, this doesn't amount to evidence (rtags), because there is no inference that can be drawn - since in ascertaining a cognitive episode as conceptual, one [must already have] ascertained its object, concepts, as well. Thus, this wouldn't lead to the possibility of "evidence [that proves a property to be proved] without depending [on it]." P: So too for factive assessment; since I don't assert either of the two forms of dependence on objects, I suffer neither of the first two faults. And although these are entailing and entailed items that don't depend on each other, if one is to ascertain a cognitive episode as being factive assessment, one must already have ascertained that it had a true object. Because this rules out inference [i.e., it rules out the possibility of inferring that a cognition has a true object due to its being an episode of factive assessment], it wouldn't lead to the possibility of "evidence [that proves a property to be proved] without depending [on it]."

Phywa pa's argumentative analogy

385

Phywa pa's argument (Phywa pa 2006, .lOb6-9):

'0 na yid dpyod la yul bden pa nes pas khyab na yul bden pas kyan khyab dgos na bdag eig pas khyab pam de las byun pas khyab pa 'am 'brei pa med kyan khyab I dan po ltar na yid dpyod yul bems po dan bdag Gig pas bems por 'gyur la I gfiis pa ltar na blo de yons su bead pa'i don med na mi 'byun pas tshad mar 'gyur ba dan I gsum pa ltar na 'breI med la khyab pa 'thad pas 'breI med kyi rtags srid par 'gyur ro ie na I '0 na rtog pa la gzun pa don spyis khyab pa'an bdag Gig pas sam de las byun nas sam 'breI pa med kyan khyab I dan po gfiis ltar na don spyi rtog pa dan rdzas eig pa 'am rtog pa'i rgyu yin pas dnos por 'gyur la gsum pa ltar na 'brei med kyi khyab bya dan khyab byed byun pas 'breI med kyi rtags srid par 'gyur ro I gal te 'breI pa gfiis mi 'dod pas dnos por yan mi 'gyur la I 'breI med kyi khyab bya dan khyab byed yin yan rtags ma yin te blo rtog par nes pa na yul don spyi'an nes pas dpag tu med pa'i phyir ro II des na 'breI med kyi rtags srid par mi 'gyur ro ie na I yid dpyod kyan yul dan 'breI pa gfiis khas mi len pas fies pa dan po gfiis med la I 'breI med kyi khyab bya dan khyab byed yin yan blo yid dpyod du nes pa na yul bden pa'an nes zin te dpag tu med pa'i phyir 'breI med kyi rtags srid par yan mi 'gyur ro II

Bibliography
Dunne, John. 2004. Foundations of Dharmakfrti's Philosophy. Boston: Wisdom Publications. Hugon, Pascale. 2008. ''Arguments by Parallels in the Epistemological Works of Phya pa Chos kyi seng ge." Argumentation 22: 93-114. Klon chen rab 'byams (attributed). 2000. Tshad ma'i de kho na fiid bsdus pa. Chengdu: Sichuan People's Press. van der Kuijp, Leonard. 2003. ''A Treatise on Buddhist Epistemology and Logic Attributed to Klong chen rab 'byams pa (1308-1364) and its Place in Indo-Tibetan Intellectual History." Journal of Indian Philosophy 31: 381-437. rNog BIo ldan ses rab. 2006. Tshad ma rnam par lies pa'i dka' ba'i gnas rnam par Mad pa (Tshad ma rnam lies kyi 'grel ba). In bKa' gdams gsuli 'bum phyogs bsgrigs, vol. 1, 420-705. Chengdu: Sichuan People's Press.

386

Jonathan Stoltz

Phywa pa Chos kyi sen ge. 2006. Tshad ma yid kyi mun sel. In bKa' gdams gsun 'bum phyogs bsgrigs, vol. 8, 434-626. Chengdu: Sichuan People's Press. Sa skya P3.I).gita Kundga' rgyal mtshan. 1989. Tshad ma rigs pa'i gter gyi rtsa ba dan 'grel pa. Qinghai: Tibet People's Press. Stoltz, Jonathan. 2007. "Gettier and Factivity in Indo-Tibetan Epistemology." The Philosophical Quarterly 57: 394-415.

Pa tshab and the origin of PrasaIigika

Georges Dreyfus and Drongbu Tsering

One of the unresolved questions of the history of Tibetan scholasticism is the origin of the Svatantrika-PrasaIigika distinction (thaI ran gi khyad par). A lot has been written about the importance of this distinction but scholars, traditional and modern, have found it difficult to pin down its historical origin. Some have argued that it is an artificial Tibetan creation without much precedent in the Indian tradition. Others have responded that although this distinction only gradually emerged through centuries of debates, it is well grounded in classical Indian sources, particularly Candraklrti's response to Bhavaviveka's critique of Buddhapruita. 1 I do not have the pretension in this short essay2 to settle once and for all this difficult question but I believe that we are now in measure .to advance our understanding by examining some of the new sources that throw light on the early developments of PrasaIigika in Tibet.

1 For a discussion of various views on this topic, see Dreyfus and McClintock 2003. 2 This essay is born of a close collaboration between myself and Professor Drongbu Tsering. We worked together on the manuscript, particularly on its philological and historical aspects. Hence, it is fitting that the authorship of this article be attributed to both of us. Many thanks to all those who have supported this project, particularly Williams College and the Tibetan Academy of Social Sciences with its very helpful director and staff. Thanks are also due to many others who have helped with their comments or contributions: Chizuko Yoshimizu, Pascale Hugon, Kevin Vose, Jay Garfield, Tom Tillemans and many others who will have to remain anonymous for the time being.

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 387-417

388

Georges Dreyfus and Drongbu Tsering

In an earlier work, 3 Karen Lang drew a suggestive picture of these early developments by sketching out the life and work of Pa tshab Ni rna grags (1055-1145?) through the few sources available at that time. But the lackof direct textual sources prevented Lang from drawing clear conclusions as to the origin of the PrasaIigika tradition and the role of Pa tshab therein. The recent and on-going publication of a number of early bKa' gdams manuscripts by the Peltsek Institute for Ancient Tibetan Manuscripts (dPal brtsegs bod yig dpe rfiili fib 'jug kha7i) has changed the situation. In particular, its publication of some works attributed to Pa tshab has put us in a position to advance significantly the question of the origin of the Prasangika tradition. 4 In her essay, Lang already emphasized the importance of Pa tshab in the development of Tibetan scholasticism in general and Madhyamaka in particular. I believe that Lang was right and that Pa tshab was one of the key figures in the incredible burst of creativity that took place during the second transmission of Buddhist learning to Tibet (phyi dar),S almost on a par with the great translator rNog Blo ldan ses rab (1059-1109), the pivotal figure in the early stages of the development of Tibetan scholasticism in general and of the gSan phu tradition in particular. 6 Pa tshab's contribution was first and foremost that of a translator, who produced impressive translations of Candraklrti's works. He was also, however, an important teacher, who started the Madhyamaka lineage that has come to dominate the landscape of Tibetan philosophical thinking.

Lang 1990. bKa' gdams gSUli 'bum, vol. 11, pp. 29-203. 5 For a description of some of the aspects of this revival, particularly as they pertain to the esoteric domain of tantras, see Davidson 2005. 6 The monastery of gSan phu Ne'u thog was founded in 1073 by rNog Blo ldan ses rab's uncle rNog Legs pa'i ses rab, one of AtIsa's direct disciples. Under the impulsion of rNog BIo ldan ses rab and later that of Phywa pa Chos kyi sen ge (1109-1169), it became the main center of Tibetan scholasticism in Central and Western Tibet and remained so for several centuries until it was replaced in that role by the great dGe lugs centers of Central Tibet and Tsang.
3
4

Pa tshab and the origin of PrasaIigika

389

Up to now, however, we had no textual evidence of his role as a commentator and hence were not in a position to appreciate more precisely his contribution to the formation of Tibetan Madhyamaka. The publication of some works attributed to him has changed this situation and put us in a position to understand more fully hIs role and his ideas. It also gives us a fascinating glimpse of a late Indian Buddhist tradition, which had already been hinted at by Lang as the likely source of the Tibetan Prasangika tradition. In this essay, I pursue the work started by Lang to flesh out our view ofPa tshab's role and ideas by looking at the texts attributed to him. I first examine their authenticity, an obvious and yet perilous prerequisite to any serious discussion. I also consider the historical significance of these texts, showing how they strongly suggest that the origin of the Prasangika tradition has to be found in India rather than in Tibet. I then move to examine the doctrinal content of these texts, focusing on two of the central questions that have agitated Tibetan scholars for many centuries: can Madhyamikas have any thesis (khas len, pak~a) and is there a place for valid cognition (tshad ma, pramafJa) within their tradition? In the process I compare Pa tshab's view to that of other Tibetan Madhyamikas (particularly his students Zan Than sag pa and rMa bya) and locate his place within the history of the Tibetan Madhyamaka tradition. I conclude with a few remarks attempting to locate Pa tshab's ideas within the conceptual landscape of Tibetan Madhyamaka, showing that he occupies a rather unique place. It should remain clear, however, that the conclusions presented here are merely provisional. They are based on a partial reading of the texts, which I have yet to explore fully. Moreover, further discoveries are likely to bring about changes that cannot be anticipated. Nevertheless, I present this essay as a contribution to the history of early Madhyamaka in Tibet with the hope of eliciting more research in the early parts of this tradition, an area that has just started to receive the attention that it deserves. Is the manuscript authentic? To assess critically Pa tshab's contribution to the formation of Tibetan Madhyamaka we need first to face the obvious and yet deli-

390

Georges Dreyfus and Drongbu Tsering

cate question of the authenticity of the manuscript itself. Are the texts that have been published in the eleventh volume of the series of bKa' gdams Writings (bKa' gdams gsun 'bum) by the Peltsek Institute really Pa tshab's? As is the case for many historical questions, an. answer to this question is bound to remain tentative but I believe that a certain number of elements lend credence to the claim that these texts are Pa tshab's, or are, at least, closely related to him and his time. The texts published by the Peltsek Institute are facsimiles of a manuscript found in the library of the Drepung monastery? Its title page is quite remarkable in that it contains two contradictory statements. The title reads: A Madhyamaka Commentary Composed by Candrakfrti (sloon {= slob dpon} zla ba grags pas mdzad pa 'i dbu ma'i 'grel pa) suggesting that we are dealing with a translation of an Indian text. Under this title, however, we find this annotation: "this [text] was not composed by CandrakIrti but by the translator Pa tshab" (' di slob dpon zla grags kyis mdzad pa min par pa tshab 10 tsas mdzad pa yin' dug). A quick look at the text confirms that this is not Candraldrti's text but a text originally written in Tibetan. Obviously, this hardly proves that it is Pa tshab's but it shows that the compiler or copier of this manuscript thought that it is Pa tshab's. And since, as we will see, the manuscript is quite old, this lends some support to the view that it is authentic. The manuscript is written in cursive script (dbu med). It is well preserved and is not particularly difficult to read. Its typographic disposition is, however, unusual. Rather than being written in a single line, the text is divided into two columns. Does that indicate that the manuscript is quite archaic? I would assume so, though I am not sure and hope that more expert opinions will throw some light on this matter. A better proof of the ancient nature of the manuscript is found in the way in which the text is written. For

7 This manuscript is reproduced in vol. 11 of the bKa' gdams gsun 'bum, which contains another short work entitled bti brgya pa'i bsdus don attributed to Pa tshab in its colophon. The latter was not examined in the present study, which focuses on the three texts that are part of the first manuscript.

Pa tshab and the origin of Prasatigika

391

example, the manuscript does not spell numbers but leaves them as numerals. This is hardly surprising, given that it is a manuscript and that its wtiter has adopted a number of shortcuts. More remarkable is the spelling. Many words appear. at first misspelled. "Thing" (dnos po) is spelled rnos po, the verb "to need" is written rgos pa instead of dgos pa, and the negative mi/med is rendered as myi/myed. These spellings are quite systematic throughout the text, suggesting that these are not mistakes. In fact, the text appears to reflect that archaic spelling of the first centuries of the second spread of Buddhism in Tibet, prior to the reform of the Tibetan language carried out during the Sa skya rule. A similar impression of archaism can be found in the frequent use of expressions derived from Sanskrit such as buta (for the Skt. buddha) and bode (for the Skt. bodhisattva) rather than their Tibetan translations (sans rgyas and byan chub sems dpa'). All this suggests that the manuscript is quite old or, at least, reflects the writing of an older manuscript that most likely dates from the twelfth century, a period quite close to Pa tshab's time. The manuscript is a collection of three texts perhaps put together by the writer of the annotation on the title page. The first text is a commentary on Nagarjuna's Mulamadhyamakakarika entitled A Commentary on [Nagarjuna's] Mulamadhyamakakarika, the Lamp that Reveals the Treatise (dbu' ma rtsa ba'i ses rab kyi ti ka / bstan bcos sgron ma gsal bar byed pa). It occupies fiftytwo folios and its colophon indicates that it is the writing down of the tradition of explanation of the pandit Hasumati (pan rjJ" ta ha su mati'i Mad lugs bris pa).B It should be noted that the colophon does not claim that this text was written by Pa tshab. This claim is made by the compiler of the text in his annotation to the title of the text. What the colophon does, however, is to attribute clearly the content of the text to Hasumati, who appears to be none other than Mahasumati, the scholar with whom Pa tshab worked while studying during the twenty-three years that he is said to have spent in

B bKa' gdams gsuli 'bum, vol. 11, p. 132 (52b.II.l1). Lang gives Hasumati as the alternative name of Mahiisumati, whom she identifies as Parahita's student. Lang 1990: 137.

392

Georges Dreyfus and Drongbu Tsering

Kashmir. 9 This strengthens the case for the attribution of this text to Pa tshab since who else but him (or somebody very close to him) would be inclined to lay down the views of Mahasumati, a rather obscure figure?lO The second and third texts are explicitly attributed to Pa tshab by their colophons or the notes added to these colophons. The first of these two texts is a short explanation of two folios about the relation between the chapters of Nagarjuna's Mulamadhyamakakiirikii. Its colophon gives it the title of Pa tshab's Instructions on the Relation Between the Chapters [of Niigiirjuna's Mulamadhyamakakiirikii] (Ie 'brei pa tshab kyi man nag).H The third and last text is an "explanation ofthe difficult points" (dka' ba Mad pa), in thirty-four folios, on Candraklrti's Prasannapadii. Its colophon reads: Explanation of the Difficult Points in Master Candrakfrti's Prasannapadii Composed <by Pa tshab> in Dependence on the Instructions of Pandit Merchant (slob dpon zla ba grags pa'i tal sna nas kyis sbyar ba / tshig gsal ba'i dka' ba Mad pa / bla ma tshon dpon pan i ta'i gdam nag la brten <pa tsab kyis> sbyar ba'o).12 The words "by Pa tshab" in the colophon are added below the line in small characters (hence my use of pointed brackets), raising some doubt about the authorship. It should also be noted that the colophon is curiously written and somewhat ambiguous as it can be read as indicating that the text is a composition of the Master Candraklrti. This seems to be why the small letters "by Pa tshab" (pa tsab kyis) were added, perhaps by the compiler or the writer of the manuscript. Finally, one cannot but wonder who the Pandit Merchant is, on whose instructions Pa tshab is said to have depended to compose this exegetical text. Is he Mahasumati or another Indian scholar?

Lang 1990: 133. Obviously, this is hardly an iron clad argument. Somebody could have written a text and claimed that it reflects Mahasumati's views. But given the fact that this scholar is a rather obscure figure, the likelihood of such a false attribution is small. 11 bKa' gdams gSU1i 'bum, vol. 11, p. 136 (54b.II.4). 12 bKa' gdams gsun 'bum, vol. 11, p. 203 (88a.I.9-1O).
9 10

Pa tshab and the origin of Prasangika

393

All three texts focus mostly on exegetical matters and do not offer as much philosophical insight as one might have hoped for. Nevertheless, the first parts of the first and third texts offer some interesting explanation of the Svatantrika-PrasaIigika distinction and provide indications of Pa tshab's own PrasaIigika philosophy. The exegetical nature of these texts is not surprising if one considers that Pa tshab was known more as a translator than as an original thinker. There are reports that he composed an original text entitled List of the Essential Points of the Madhyamaka View (dbu ma'i Ita ba gnad tho).13 None of the three texts found at Drepung have this title or have a content that corresponds to this title. Hence, it is clear that what we have here are texts of a different nature, not original expositions of the Madhyamaka view, but exegetical glosses with a few doctrinal discussions. The first text consists mostly of a gloss on Nagarjuna's Mulamadhyamakakiirikii but its first part offers a well-delineated articulation of the Svatantrika-PrasaIigika distinction. As stated above, it presents itself as the writing down of the explanations of the Indian scholar Mahasumati. We do not know very much about this figure except that he was Parahitabhadra's student and was well versed in logic and epistemology, an interest reflected by the first part of the text, a sustained refutation of the idea that the notion of valid cognition (pramii1}a, tshad ma) has a role to play in Madhyamaka thinking. We also know that he collaborated with Pa tshab to produce the translation of Candraklrti's Prasannapadii. The claim that this text reflects the views of Mahasumati is, however, of considerable historical significance. For if we believe its colophon and its attribution of authorship, this suggests that the Svatantrika-PrasaIigika distinction was not created by Pa tshab, as many informed scholars, both traditional and modern, have assumed, but pre-existed in India and that Pa tshab recorded, clarified and intensified a distinction that was already made by some late Indian scholars. Obviously, when Pa tshab provides the first historically known mention of PrasaIigika, speaking of "the explanation of Candraklrti's presentation as our own system according to which the master
Khu byug 2004: 133.

13

394

Georges Dreyfus and Drongbu Tsering

Nagarjuna's thought is only Prasangika,"14 we have to wonder. Is he making an original claim or is he reporting Mahasumati's explanations, thus demonstrating the existence of a Kashmiri Prasangika school? This question goes to the heart of a persisting mystery of Madhyamaka studies, the emergence of the Svatantrika-Prasangika distinction within Madhyamaka and the role of Candraklrti therein. Although this distinction has played an enormous role in Tibet and although there are clear indications of differences between Bhavaviveka and Candraklrti in their interpretations of Nagarjuna, it does not seem that these differences were very significant for several centuries in India. As Kevin Vose has shown, it was only during the tenth and eleventh centuries that Candraklrti emerged as an important Madhyamaka interpreter in India and that his ideas started to be taken seriously.ls This emergence does not prove, however, that there was a Prasangika school in India at that time. Hence, we have to wonder what is the source for Pa tshab's clear assertion of the preeminence of Prasangika? Was there a Prasangika school in Kashmir that Pa tshab tapped into or did he articulate a set of ideas that were in the air but were not clearly formulated? It is hard to answer this question since we know very little about the Buddhism that existed in Kashmir at that time. One intriguing fact is that the translator rNog BIo ldan ses rab studied logic and epistemology with Parahitabhadra, Mahasumati's teacher, and came out of his extended stay in Kashmir favoring the three MMhyamikas from the East (dbu rna sar gsurn) , Santarak~ita, KamalasIla and Jfianagarbha, thinkers identified by later Tibetan doxographers as Svatantrikas. Does that mean that Parahitabhadra and Mahasumati differed in their Madhyamaka interpretations? Vose suggests that the Kashmiri Buddhism that rNog BIo ldan ses rab and Pa tshab encountered was rather eclectic, with scholars able to, comment on a variety of texts and provide multiple points

14 Pa tshab, Ses rab ti ka, p. 48 (10b.II.6): slob dpon klu grub kyi dgOlis pa thai 'gyur pa yin pa Rid du zla grags ran gi lugs rnam par biag pa stan pa. lS Vose 2009. My thanks to him for his helpful comments and suggestions on this question of the Indian origin of the Prasailgika tradition.

Pa tshab and the origin of Priisangika

395

of view. 16 If we follow his suggestion, 'we may surmise that when Pa tshab speaks of Prasailgika, he may not necessarily be defending the view of an exclusive or sectarian school with clear affiliations, but that he is at least reflecting the interpretative line he received from Mahasumati. We obviously cannot know whether Mahasumati himself exclusively subscribed to this interpretive line or whether it is Pa tshab who chose to focus on this approach. Hence, we are not quite in a position to decide whether there was a Kashmiri Prasailgika school at that time, but what seems reasonable to assume is that there was at least a clear interpretive line that favored Candraldrti over the proponents of the Svatantra approach. Pa tshab received this line of interpretation from Mahasumati and recorded it in the three texts examined here, thereby clarifying and intensifying a pre-existing distinction. The question of whether this line of interpretation was already described as "Prasailgika" or whether he coined the term himself is a question that must remain unresolved for now. It remains the case, however, that if I am right, we have here a clear Indian antecedent for the Prasailgika interpretation that was to dominate the landscape of Tibetan Madhyamaka. The depiction of the views of this text as being Mahasumati's raises another question. Does this text represent Pa tshab's own view or does it merely record Mahasumati's opinions? It also raises obvious questions about Pa tshab's originality as a thinker and his place within the tradition. To a certain extent, Pa tshab's Prasailgika as it appears in the first text may be less his than Mahasumati's. Nevertheless, inasmuch as Pa tshab seems to endorse Mahasumati's views and to expand on them in the third text, we will discuss the ideas found in all three texts as being Pa tshab's, regardless of whether these ideas were originally his or not. As to Pa tshab's originality, it appears that his role may have been more that of a translator and transmitter than an original thinker. This should come as no surprise since it is in fact how he is presented traditionally. What are we to conclude from this examination of the manuscript? As often with historical matters, it is hard to reach solid
Vose 2009: 48-49.

16

396

Georges Dreyfus and Drongbu Tsering

conclusions and we are often left with conjectures and likely explanations rather than definitive truths. Nevertheless, I would suggest that here the preponderance of the evidence suggests that these texts were written by Pa tshab himself or by one of his close disciples during his time. The antiquated spelling definitively suggests an ancient manuscript, most likely written before the thirteenth century, and the colophon suggests that we are dealing here with notes taken by Pa tshab himself. All in all, I think that we can proceed on these assumptions and assume that we are dealing here with Pa tshab's texts and ideas, with the obvious caveat that further research will be needed and that new evidence can change this provisional conclusion.
Pa tshab's Prasangika

If we assume that we are dealing with Pa tshab's works, what can we say about the ideas that they reflect? As I already indicated, the texts considered here mostly focus on exegetical matters and are not philosophically very sophisticated. Nevertheless, they provide an interesting and thought-provoking picture of a consistent skeptical approach to Madhyamaka as well as an historical docu.ment about the development of the Prasailgika interpretation. The gist of the Prasangika approach as understood by Pa tshab is that Madhyamikas should not try to establish emptiness through a reasoning demonstrating that phenomena are empty of intrinsic nature. Rather, they should take as their targets particular views asserting various possible candidates for intrinsic nature and show the internal contradictions these views lead to through statements of consequences (prasmiga, thai 'gyur). Thus, Madhyamikas should not attempt to prove a general thesis but should stick to the refutation of opinion& relating to particular topics on the basis of their adversaries' assumptions, as Nagarjuna did in his Mulamadhyamakakiirikii. This is the proper procedure that Madhyamika should follow and this is what distinguishes the correct approach, the Prasailgika interpretation, from that of the Svatantrika. Whereas the former appropriately denies that emptiness can be established as true through logically compelling demonstration, the latter attempts to do so, suggesting that emptiness is the correct philosophical view.

Pa tshab and the origin of Prasat:gika

397

Let us further explore this topic and appreciate its importance for our appraisal of skeptical Madhyamaka interpretationsY The question of the ways in which emptiness is established may seem arcane but it is not, for it connects to central philosophical issues. For Pa tshab, the Svatantrika approach is not just mistaken in determining the dialectical tools appropriate to Madhyamaka, but ignores and even threatens Nagarjuna's central insight that emptiness is not a view but the suspension of all views. 18 Understanding emptiness does not entail the adoption of a doctrine that asserts some essential truth and rejects other views as mistaken. Rather, it is an insight that frees us from the compulsion to look for such truth. Hence, it is completely self-defeating for Madhyamikas to attempt to demonstrate the truth of emptiness through reasoning, for this reinscribes the very essentialism that Madhyamaka seeks to overcome. It is the very project of following the Middle Path that avoids any extreme, that is, dogmatic position, which is threatened by the attempt to set emptiness as the right view. To make his point, Pa tshab proceeds through a lengthy analysis ofthe various ways in which the doctrine of emptiness could be validated, showing how each possibility leads to contradictions. Words cannot establish the Madhyamaka view that all things lack a real essence as true, for the simple fact of stating such a view does not make it true. Otherwise the thesis that things have a real essence would be also true since it can be stated. Can then inference establish emptiness as true, like it does in the case of other theses and as Svatantrikas argue? For Pa tshab, this is not possible for an inference requires a probative argument (sbyor ba, prayoga) whose terms (the subject, the reason and the pervasion, khyab pa) are established by valid cognitions in common by the two parties taking part in the argument. This is what is called an autonomous inference (ran rgyud kyi rjes dpag, svatantra-anumiina), that is, an

Pa tshab, Ses rab ti ka, p. 38. I am referring to several well-known passages, the most famous being MuZamadhyamakakarikii xiii.8 where Nagarjuna declares that emptiness is the relinquishing of all views and that the one who takes emptiness to be such a view is incurable. See Garfield 1995: 212.
17 18

398

Georges Dreyfus and Drongbu Tsering

inference supported by an argument whose terms are established in common by the two parties independently of the particular ways in which they view things. For Pa tshab, such an inference is not possible within the context of a discussion of the ultimate nature of things, for if the terms of the argument were established by a valid cognition they could not be refuted or undermined by any other valid cognition. The requirement that the terms of the inquiry should be established in common by both parties and should not be undermined by the conclusion may seem unexceptional, but it creates a particular problem within the context of Madhyamaka inquiry. This is so because this inquiry is based on the investigation of the ultimate nature of things. In such an investigation, one looks for the essence of the examined phenomenon among, for example, its parts. Upon not finding it, one comes to realize that such a phenomenon lacks any true essence, that is, that its ultimate nature is that it lacks such a nature. For Pa tshab, the non-findability of a particular phenomenon is not just a failure to find the phenomenon through a particular mode of analysis (as it is for Tson kha pa, for example), but, rather, it shows that the phenomenon itself cannot be validated. It invalidates (gnod) or undermines the phenomenon itself. Hence, Madhyamikas cannot use probative arguments to demonstrate the absence of ultimate essence without committing themselves to the very essentialist view that they are attempting to take apart. The Svatantrikas reply that the establishment of the terms of the reasoning by a valid cognition does not entail that they exist ultimately. The terms of the reasoning are established, they argue, by conventional valid cognitions that do not differentiate whether things exist ultimately or not, but merely ascertain them according to how they appear ordinarily. Hence, there is no problem in finding a common subject in reference to which Madhyamikas and their adversaries can argue. 19 Madhyamikas can deploy their argu-

19 Pa tshab, Ses rab ti ka, p. 43 (8a.I.7-9): chos can lastsogs pa <tshad mas ma grub na tshad ma ran rgyud myi 'thad / grub na tshad mas grub la / tshad ma gian gyis myi gnod pas na chos can la stsogs pa> yan dag du grub par 'gyur bas ran biin med pa fiams ces pa'i skyon mi 'on ste chos

Pa tshab and the origin of Prasarigika

399

ments against their adversaries in reference to conventional phenomena that are established on the basis of common appearances, much like Buddhists demonstrate to Vaise~ikas that sounds are impermanent on the basis of the common understanding of sound, despite the fact that both parties have a different understanding of the nature of sound. For Pa tshab, this example is inconclusive since in the case of the ascertainment of the impermanence of sound, the common conception of sound is not negated by the realization that it is impermanent, whereas the insight into the lack of essence of the subject invalidates the subject itself. This insight comes as a result of the search for the ultimate nature of the subject, a process through which one decides whether the subject exists or not. One then comes to the provisional conclusion that it does not exist. This conclusion does not mark, however, the end of the inquiry, for one should also realize that the subject does not not-exist either.20 Nevertheless, it remains the case that it is the very subject of the inquiry that is negated in the Madhyamaka inquiry, not some other entity. Hence, if the subject were to be established by a conventional valid cognition, its existence would be confirmed and it could not be invalidated by the search for its inherent nature. Hence, it would have to exist ultimately. 21 The conclusion is then that emptiness cannot be established by an inference. Could it then be established by a yogic perception (yogipratyak$a, mal 'byor miwn sum)? No, replies Pa tshab, for any yogic perception is born from the habituation to and the enhancement of the insight that one gains through inference. But

canji ltar snan ba tsam po gan gis kyan khyad par du ma byas la tha sfiad pa'i tshad mas grub pa de la gtan tshigs gcig dan du ma bead pa tsam yan tha sfiad pa'i tshad mar mnon3 {= mnon sum} ede {= la sags?} pas grub / 20 I am here obviously alluding to Nagarjuna's famous tetralemmic approach. For a classical treatment of this question, see Ruegg 1977. 21 Pa tshab, Ses rab ti ka, p. 43 (8a.II.5-6): chos can snan ba dan gtan tshigs tshad mas grub na / tshad mas grub pa la tshad ma gian kyis mi gnod pas na chos can lastsogs pa don dam pa'i bden par grub par thaI bas na ran biin med par myi 'thad do /

400

Georges Dreyfus and Drongbu Tsering

since emptiness is beyond the scope of inference, it cannot be made into an object of yogic perception. 22 Hence, emptiness cannot be established by either inference or perception. How can then Madhyamikas proclaim emptiness if it cannot be established as true? Pa tshab's answer is that emptiness is not established as true. MMhyamikas do not have any thesis to establish, view to defend or position to eliminate. They merely proceed by consequences exposing the contradictions to which the views of their adversaries lead. MMhyamikas are not in the game of demonstrating truth or falsity. They do not need to defend their position as true and criticize their adversaries as being mistaken, for they do not believe that it makes sense to make pronouncements about how things are. Since nothing can be found under analysis, no statement can be established as true or shown to be false upon being analyzed from an ultimate perspective. Even emptiness is not findable under analysis and hence the statement that things are empty cannot pretend to be true, since it sets up an object (emptiness) that fails to be confirmed by further analysis. Hence, even when they offer the proposition that phenomena are empty, Madhyamikas should not be understood to hold this proposition to be true and to decry the opposite proposition to be false. Similarly, for Pa tshab, Madhyamikas should not take their "conclusions" to derive logically from unassailable arguments. The claim that phenomena are empty is not a conclusion that logically derives from the contradictions contained in their opponents' views. Rather, it is merely a provisional stance, a kind of slogan that aims at showing opponents the way to get out of the contradictions entailed by their own commitments. As for Madhyamikas, they do not have any commitment to any view about how things are. Hence, they do not assert that the realist view is logically inconsistent. and that logic requires them to adopt an anti-realist position, for this would assume that they are committed to the rules of logic and that their arguments better conform to the canon of sound reasoning than those of their opponents. But this is ignoring, argues Pa tshab, that Madhyamikas merely argue on the basis of their oppo-

22

Pa tshab, Ses rab ti ka, p. 48.

Pa tshab and the origin of Prasangika

401

nents' own ground and rules but have no other commitment over and above those necessary to the pursuit of the conversation. It is precisely this lack of commitment that the realist opponents seem to miss when they argue that the Madhyamaka argument is self-stultifying. For if it is true that phenomena lack any essence and have merely conventional existence, is it not the case that the truth of this very statement becomes merely a matter of convention and lacks any bite? Nagarjuna's famous answer is well known and illustrates the skeptical attitude that Pa tshab sees as being at the core of Madhyamaka:
If I had any position, I thereby would be at fault. But since I have no position, I am not at fault at all. 23

The realist objection would be true if Madhyamikas were in the business of establishing some truth (the fact that phenomena lack essence and have only nominal existence), but since they are not, it just misses the point. Madhyamikas are neither in the business of defending true positions nor in that of exposing wrong ones. Hence, they cannot be assailed on the ground of being inconsistent since they do not hold to any view whatsoever. In Tibet, this "thesislessness" (khas len kun bral) has excited the verve of numerous commentators, who have debated about whether it should be taken literally or not. Pa tshab is often credited by later authors such as Go rams pa bSod nams sen ge (1429-1489) with the view that although Madhyamikas do not entertain positive theses, they hold negative ones. 24 But this does not appear to be Pa tshab's position here. His assertion that Madhyamikas do not hold any thesis elicits objections from an opponent, who argues that although MMhyamikas have no positive thesis (YOlis geod kyi bsgrub bya), they must have negative theses (rnam bead dgag pa tsam gyi bsgrub bya) since they refute the view of their adversaries. To this Pa tshab

23 Niigiirjuna, Vigrahavyiivartanf29 quoted by Pa tshab in Ses rab ti ka, p. 49 (l1a.I.1). For a thorough examination of this statement, see Ruegg 2000: 115-156. 24 Ruegg 2000: 48, 159-160.

402

Georges Dreyfus and Drongbu Tsering

responds unambiguously that negative theses are not different from positive ones and hence both are to be rejected. Pa tshab states:
Objection: Although you do not accept positive theses, you must accept negative theses, i.e., mere negations. Hence, you must also accept reasonings establishing them. Answer: I do not even accept negative theses, for if the essence of that which is to be negated does not exist, its negation cannot hold either. 25

As we can see here quite clearly, there is no ambiguity. Pa tshab rejects both types of thesis, for both presuppose that one can make sense of the idea of understanding how things are or, at least, how things are not. But for Madhyamikas, this is not possible for it would presuppose that the objects of some statements about how things are can be confirmed by ultimate investigation. Since this is not possible, no statement can claim any privileged status as far as deciding how things are. In a certain sense, from the ultimate point of view, every statement is false since its object cannot resist investigation. But to say that all statements are false creates an obvious problem of self-reference. Is the statement "all statements are false" true or false? For Pa tshab, this self-referential paradox is not a problem, for it points to the problems that one gets into when one asserts or negates some view about how things are. What Madhyamikas should do is refuse to play the game of negating or affirming any view about reality and merely use their opponents' views to point out their contradictions without committing themselves to any standpoint.

Pa tshab, Zan Than sag pa and the evolution of Tibetan Prasangika

Now that we have some sense of Pa tshab's PrasaIigika, we are in positipn to locate Pa tshab within the overall evolution of Tibetan

25 Pa tshab, Ses rab ti ka, p. 49 (lla.I.8-9): khyed yons gcod kyi bsgrub bya mi 'dod kyan / rnam gcad dgag pa tsam gyi sgrub bya 'dod dgos pas de sgrub pa'i gtan tshigs lastsogs pa khas len rgos {= dgos} so ie na / ned la dgag pa tsam gyi bsgrub bya'an 'dod pa med de / dgag bya'i ran biin ma grub pas na bkag pa yan mi 'thad ste /

Pa tshab and the origin of Prasangika

403

Madhyamaka. Lang already delineated his role as the starting point . of the Prasangika tradition in Tibet, listing his four main disciples among whom Zan Than sag pa Ye ses 'byUll. gnas and rMa bya Byan chub brtson 'grus were the most important in the, development of this tradition. 26 This is particularly true of the former, who appears to have developed the kind of Madhyamaka interpretation that came later to be identified as the "Madhyamaka free from any extreme" (mtha' bral dbu ma) and was adopted by most Tibetan thinkers outside of the dGe lugs tradition (with the exception of the proponents of extrinsic emptiness whose view will be briefly examined below). This description as "Madhyamaka free from any extreme" raises some difficulties. Any credible Madhyamaka interpretation is bound to claim to be free from elaborations (spros bral) and free from extremes (mtha' bral) since it is based on Nagarjuna's deconstruction of the four extremes (catu~koti, mtha' bii).27 Hence, any genuine Madhyamaka interpretation is bound to lay claim to this label. Nevertheless, inasmuch as this line of interpretation sticks to a very literal understanding of Nagarjuna's refutation of the four extremes and avoids the use of an operator to make this refutation logically consistent, its description as being freefrom extremes andfreefrom elaborations makes sense. This view of emptiness as freedom from extreme and freedom from elaboration is quite close to Pa tshab's own interpretation. Throughout his text, Pal tshab takes literally Nagarjuna's rejection of categories such as existence, non-existence, causes, etc. and argues for a view of emptiness as being beyond any possible description. For Pa tshab, as for Zan Than sag pa, any description of reality, either positive or negative, is deceptive and hence to be dispensed with, except as a concession to worldly usages. The ultimate nature of things is utterly beyond any description and hence radically ineffable. This is so not because ultimate truth is too full to be captured within our limited conceptual schemas but, on the contrary, because there is not enough there for conceptuality to latch upon. Ultimately, things lack any core features that could be

26

Lang 1990: 137.

27

For a masterful discussion the four extremes, see Ruegg 1977.

404

Georges Dreyfus and Drongbu Tsering

seized upon by descriptions in any meaningful way. Any description can only concern the provisional and the relative domain of worldly transactions. Hence, language radically fails when it attempts to signify the ultimate. All that language can do there is to provide helpful therapies to cure the habit of reifying phenomena and metaphors pointing in the direction in which to proceed. This understanding of the radical inability of language to signify reality and the consequent view of Madhyamaka as therapeutic rather than systematic philosophy highlight the similarity between Pa tshab's and Zan Than sag pa's interpretations. But does this mean that they agree on the other important issues? This is where we need to be somewhat nuanced and to consider other aspects of Pa tshab's view. One of the more intriguing suggestions that we found in the first text we examined, the one that purports to be notes from Mahasumati's teachings, concerns the status of valid cognitions in Madhyamaka. There, Pa tshab seems to take a radical stance, rejecting all notions of valid cognition, a view that does not seem to be accepted by Zan Than sag pa and his later followers. We recall that in his attack on the Svatantrika view Pa tshab had argued that the use of valid cognition is self-defeating within the context of the investigation of the ultimate since it entails a use of the very notion of essence that the Madhyamaka claims to deconstruct. But what about other contexts? Can Madhyamikas use valid cognitions within the context of transactional usages or are they committed to being thorough skeptics? Zan Than sag pa argues that the rejection of autonomous inference concerns only the ultimate context. The use of inference is not possible when one is engaging in determining the ultimate nature of things but it is allowable in other conventional contexts. 28 Hence, for him, the notion of valid cognition has a clear though limited place within the Madhyamaka, as also asserted by most of his later followers. But what about Pa tshab? Does he agree with his student's view or does he offer a more thorough skepticism? This is perhaps the place to remind our readers of the provisional nature of my conclusions. I have yet to explore fully the three texts and new textual material
Yo shimizu 2005: 133.

28

Pa tshab and the origin of PrasaIigika

405

may appear. But, as far as I can make it, it would seem that Pa tshab's view is more radical than that of his disciple and that he is committed to a thorough skepticism concerning valid cognition. Pa tshab says:
As for the first point concerning the answer to five questions about valid cognitions, when asked: if you don't believe that there are objects of valid cognitions (gial bya, prameya), do you accept that there are valid cognitions or not? Answer: I do not accept valid cognition, since both valid cognitions and their objects have been refuted [by N::igarjuna] in [his] Vigrahavyavartanf. This is so because valid cognitions require objects but no such objects exist. Some hold both that conventional valid cognitions realize [their objects] in accordance with what is renown and that the ultimate is realized [by ultimate valid cognitions]. This is incorrect for inasmuch as what is accepted as a conventional valid cognition is valid with regard to the object that it realizes, it cannot be invalidated by another valid cognition. Hence, [its object] would have to be real and it would be impossible to assert conventional valid cognition as a second kind. In so far as a valid cognition is established, its object has to be true. Hence, no valid cognition can ever exist and we do not accept [this notion]. Objection: If there are no valid cognitions, the distinction between the two truths cannot be? Answer: I accept the consequence that ultimately the distinction between the two truths does not exist. It is not taught from the point of view of reality but has been uttered by Buddha and the Noble Ones (or the two Noble Ones, i.e., Nagarjuna and A.ryadeva) conventionally for some [particular] need. 29

29 Pa tshab, Ses rab ti ka, p. 49 (lla.II.2-7): gfiis pa tshad ma la dri pa lna'i Ian gdab pa las dan po ni khyed gial bya khas mi len na tshad ma khas len nam mi len ie na / ned tshad ma khas mi len te tshad ma dan gial bya gfiis ka rtsod zlog du bkag pas ste I tshad ma la gial bya yod rgos {= dgos} pa la gial bya med pa'i phyir ro II kha Gig tha sfiad pa'i tshad ma grags tshod du jal pa dan I don dam jal ba gfiis yod zer na I myi 'thad de tha sfiad pa'i tshad mar 'dod pa des rtogs pa'i don la tshad ma yin phan cod tshad ma gian kyis mi gnod pas I de yan dag du 'gyur bas na tha sfiad pa'i tshad ma ces pa dan gfiis med kyi / tshad ma gGig grub phan cad de'i don bden par 'gyur bas na tshad ma gan yan med ste khas mi len to II tshad ma med na bden pa gfiis gyi dbye ba med par 'gyur ro ie na I bden pa gfiis kyi dbye ba don dam par med par thaI pa 'dod thog yin pa la I don

406

Georges Dreyfus and Drongbu Tsering

This passage seems to reject any notion of valid cognition and thus to entail a thorough skepticism. Obviously we can wonder whether it should be taken literally or should be interpreted as referring to the context of the ultimate. It is well accepted among many Tibetan thinkers that a PrasaIigika cannot accept the validity of autonomous inference within the context of the ultimate. This is so because, as we saw above, the investigation into the ultimate invalidates the very nature of the investigated phenomena. Hence, as Pa tshab argued earlier, there cannot be any subject established through a valid cognition for the Madhyamika. But Pa tshab seems to go much beyond this point to argue that the rejection of the notion of valid cognition is not just limited to the context of the investigation of the ultimate but is global. It is the very notion of knowledge (or its Indian rough equivalent, valid cognition)30 that seems to be at stake here. Inasmuch as knowledge presupposes that one can make claims about how things are, Madhyamikas cannot accept any conception of knowledge without reintroducing the very essentialism that it seeks to undermine. Thus, the depiction of Pa tshab as holding the view that PrasaIigikas reject only objectively grounded (vastubalapravrtta, ditos po stobs tugs) valid cognitions and accept consensually acknowledged (lokaprasiddha, 'jig rten grags pa) valid cognitions does not appear to be borne out by the texts examined here. 31 It is not just the idea of objectively grounded knowledge that is rejected by Pa tshab but the very notion of valid cognition. As we can imagine, the wholesale rejection of any idea of truth has important consequences for the doctrine of the two truths, which is often construed as central to Madhyamaka. For Pa tshab, this doctrine has no intrinsic validity. It is not meant to capture the

du mi 'chad bka' tha sfiad du rgas {= dgas} pa'i dbmi gis bu ta dan 'phags pa dag gsun bas sa II 30 For a discussion of the differences between the Western concept of knowledge and the Indian conception of pramii:(ta, see Dreyfus 1997: 287-288. 31 As reported, for example, by 'Jam dbyails Mad pa in his Great Tenets (Grub mtha'i rnam bad chen ma 102b-103a). See Ruegg 2000: 47.

Pa tshab and the origin of Prasangika

407

validity that worldly practices have but is merely a skillful means. Asked by an opponent why the two truths are propounded, Pa tshab enumerates alist of wrong views to be eliminated such as the denial that past and future lives exist, the view that things are permanent, etc. He then concludes: "[The two truths] were spoken to refute these views, not because they are established by valid cognitions."32 Thus, what we seem to have here is a wholesale rejection of the notion of knowledge. To assert that a cognition is true is to assert that its content is true and hence cannot be invalidated by another cognition. But since the content of every cognition can be invalidated within the context of the search for the ultimate nature of things, no cognition can claim to be true. Hence, it makes no sense to talk about truth except as a concession to the usages of the world. The use of the notion of truth is purely pragmatic and appears to be without any normative consequence. We may, however, wonder about the understanding of emptiness. Is it not true? Pa tshab's answer to this question seems to be quite clear. Even the direct realization of emptiness is not a valid cognition, for its object, emptiness, is invalidated when it is further inquired into. Hence, even yogic perceptions cannot be asserted to be true. One can wonder, however, how we are to take this radical denial of the validity of knowledge. Does it apply to the Buddha as well? And if so, how are we to understand his omniscience? In the third text Pa tshab gives a very brief indication that seems to confirm that the denial of validity does not concern solely ordinary beings but the Buddha as well. He says:
As for the nature of conventional truth it is the appearance to infantile beings and the grasping to form and so on under the power of the passions and the obstructions to full knowledge. It is also the appearance, free from grasping, to the perceptions following [the direct realization of emptiness] of the bodhisattvas who have entered the great stages [of the A.rya bodhisattva] due to their not having abandoned the obstructions to full knowledge. The appearance to the omniscient mind is the ultimate truth. This term "appearance to the omniscient mind" should be understood to refer merely to the cessation of the stream of

32 Pa tshab, Ses rab ti ka, p. 50 (lIb.!.l): Ita ba de dag dgag pa'i don du gsulis kyi tshad mas grub nas biag pa ma yin no /

408

Georges Dreyfus and Drongbu Tsering

consciousness and mental factors like the reality free from elaborations due to the absence of the way of looking of infantile beings with grasping and of that of bodhisattvas free from such grasping. This term should not be [understood to refer to] a seeing by taking something as an object, much like the double moon and the failing hair appear to the one with ophthalmia despite their non-existence but not to the ones free from such disease. 33

Pa tshab follows Candraklrti literally in his understanding of buddhahood as consisting of the elimination of all the obstacles to the full knowledge of emptiness. The buddha's wisdom is not a panoramic omniscience in which all phenomena are perceived in their particularities. Rather, it is merely the actualization of the full potential of the realization of emptiness. It is not cognitive in the usual sense of the word, for it does not cognize any object, positive or negative. This wisdom is not a cognitively active state engaged in figuring particular objects but, rather, is the ceasing of any attempt to cognize reality.34 Although this description of the realization of emptiness is short, it reveals quite clearly Pa tshab's understanding of the implications of Buddhist soteriology for Madhyamaka. This passage comes at the beginning of the third text of the collection of the three texts we have been examining here. This third text is a commentary on Candraklrti's Prasannapadii that incorporates elements from 33 Pa tshab, Tshig gsal dka' Mad, p. 146 (59b.I.10-II.4); no bo ni fion mons pa dan ses bya'i sgrib pa'i dban gis gzugs lastsogs par zen cin byis pa la snan ba dan / sa chen po la zugs pa'i bo des ses bya'i sgrib pa ma span pas I rjes kyi ses pa la zen pa dan bral ba'i snan ba la yod pa ni kun rdzob kyi bden pa yin la I thaJ?Zd {= thams cad} mkhyen pa'i snan ba ni don dam pa'i bden pa ste I thaJ?Zd {= thams cad} mkhyen pa la snan ba zes bya ba'an byis pas zen pa dan bcas par mthon pa dan I bo des zen pa dan bral bar mthon ba de ltar mthon ba med pas chos fiid spros pa dan bral bzin du sems dan sems las byun ba'i 'jug pa rgyun chad pa tsam la tha sfiad gdags par zad kyi I yul du byas pa'i sgo nas mthon ba yod ba ma yin te I dper na rab rib la zla ba gfiis dan skra sad lastsogs pa med par snan gi I rab rib can ma yin pa la mi snan ba biin no II 34 For an excellent discussion of the contrast between Candraklrti and Santarak~ita, see McClintock 2000.

Pa tshab and the origin of Prasangika

409

the Madhyamakiivatiira of the same author. Hence, it discusses Madhyamaka soteriology and as such it is particularly revealing of Pa tshab's conception of the nature and place of the realization of emptiness. It is certainly true that Pa tshab's main concern in the works we have examined here is not Buddhist soteriology but the elimination of the reification of knowledge. He sees the Svatantrika as an example of how this reification can infect and corrupt the Madhyamaka tradition, which is supposed to provide an antidote to this disease. Hence, his main concern is to counteract this tendency toward reification, not to provide an analysis of the epistemic status of transcendent wisdom. But he does seem ready to draw radical consequences from his views and does not exclude transcendent wisdom from the deconstruction of knowledge. Even the direct realization of emptiness is not true in the sense that it does not have an object that can be confirmed by further investigation. Such wisdom is true, however, in the sense that it is the final realization of the futility of looking for such a form of knowledge. For Pa tshab, this is the only wisdom that is available. The attribution of a rather austere interpretation of Buddhist soteriology to Pa tshab is to some degree confirmed by the material examined by Kevin Vose in this volume. There we find the twelfth century Sa skya master bSod nams rtse mo, Phywa pa Chos kyi sen ge's student, differentiating the Prasangikas as the ones who hold that the continuum of wisdom is interrupted by the realization of emptiness, and the Svatantrikas, who hold that it is not. bSod nams rtse mo criticizes the Prasangikas for holding a nihilist view of the realization of emptiness as the complete cessation of any cognitive activity, a state of total blankness without any content whatsoever, and contrasts this view with the more positive Svatantrika account of the persistence of wisdom in the state of enlightenment. We obviously do not know whether bSod nams rtse mo is referring to Pa tshab or not, but it is clear that he has in mind somebody holding a view similar to the one expressed above. The weight of this critique also goes some way to explain some of the reasons why Pa tshab's students thought that they had to modify some of his more radical views. Faced with the accusation of nihilism and Phywa pa's

410

Georges Dreyfus and Drongbu Tsering

critiques of CandrakIrti,35 his students (Zan Than sag pa and rMa bya) and their followers, particularly gTsan nag pa brTson 'grus sen ge (?-after 1195), had to come up with more positive accounts of cognitive activity in general and of the enlightened state in particular. Being themselves Phywa pa's disciples, they were able to take to heart his criticisms and respond to them, thereby moderating Pa tshab's austere skepticism through the reintroduction of the notion of valid cognition. In this way, they provided a more positive account of knowledge in general and of the ultimate state of enlightenment in particular. Conclusion: Pa tshab's Priisaligika in a comparative perspective Now that we understand the radical nature of Pa tshab's Prasangika and some of the transformations that it later underwent we are in better position to locate him within the landscape of Madhyamaka in Tibet. Delineating the main lines of Madhyamaka interpretations has always been an exercise that has drawn the attention of scholars interested in this field. Traditional Tibetan thinkers have spent a great deal of effort mapping the various positions within the Madhyamaka tradition, and so have modern scholars. David Burton, for example, offers an interesting typology of Madhyamaka interpretations as falling into one of three lines. 36 1) The first view takes Nagarjuna's emptiness to entail a radical attack on knowledge itself, a thoroughly skeptical refutation of any knowledge claim as lacking any ground and hence being fundamentally undecidable. 2) The second approach agrees with this skeptical repudiation of the possibility of knowing only as far as the conventional is concerned but recoils from going all the way to a total rejection of any knowledge claim. For this view, the conventional may not be knowable with the degree of certainty required to differentiate knowledge from mere opinion but the ultimate can be known directly. This knowledge is not, however, conceptual since 35 See Tauscher, ed. 1999. 36 Burton 1999: 2-4.

Pa tshab and the origin of Prasatigika

411

the ultimate is utterly ineffable, completely beyond the purview of thought and language, being accessible only to direct experience. Hence, this view can be characterized as transcendentalist. It is trans-rational and mystical but does not need to be taken to assert an unknowable absolute reality. Rather, it simply denIes that thought and language can understand ultimate reality, a domain reserved to transcendent wisdom. 3) Finally, a third view holds that these two approaches lack nuance and fail to understand that Nagarjuna's refutation concerns a particular set of ontological presuppositions (the assumption that things have an essence), not the possibility of knowing simpliciter or even the possibility of knowing the conventionaL This distinction between three perspectives is useful to map out the landscape of Tibetan Madhyamaka. But to do so, we may need to add 4) a fourth view according to which only the ultimate exists, the conventional being utterly devoid of any essence. In this absolutist perspective, the conventional cannot be known because there is nothing to know. Only the ultimate, which exists absolutely, can be known by direct experience. Being the realm of essential and absolute being, it escapes the dichotomies on which thought and language are based and hence is completely ineffable. Three of these four views are well represented within the Tibetan tradition. The fourth absolutist position is at the core of the extrinsic emptiness (gian stan) tradition of Dol po pa Ses rab rgyal mtshan (1292-1361), who holds the view that only the ultimate is real, the conventional being only a product of the illusion to be destroyed by transcendent wisdomY This view of extrinsic emptiness has been embraced by the Jo nang school and by a number of thinkers, particularly, but not only, within the bKa' rgyud tradition. The third view argues for a nuanced view according to which Madhyamaka reasonings seek to undermine less ontology or epistemology per se than the reified sense of reality that functions as a central pragmatic presupposition of discourse and ordinary life. This model seems to capture quite well the gist of Tsoil kha pa's approach and his attempt to establish a clear and principled dis-

37

For a discussion of this view, see Hopkins 2006.

412

Georges Dreyfus and Drongbu Tsering

tinction between the conventional, which is valid though distorted, and the reificationist presuppositions of true or essential existence, which are to be thoroughly rejected. Inasmuch as this view holds to the validity of the conventional, it differs quite clearly from the second interpretation mentioned above, the transcendentalist view according to which reality is ineffable. We still do not fully understand the historical origin of this line of Madhyamaka interpretation but it may well be the case that rMa bya played an important role through his attempts to find a place for Dharmaklrti's logic and epistemology within Prasangika philosophy, as shown by Thomas Doctor's essay in this volume. The transcendentalist interpretation has been advocated by many thinkers in Tibet, particularly but not only Tson kha pa's critics, 00 rams pa, gSer mdog pa~ chen Sakya mchog Idan (14281509) and the Eighth Karmapa Mi bskyod rdo rje (1504-1557). All three thinkers argue that CandrakIrti's refutation of the use of autonomous reasoning within the Madhyamaka entails a rejection to a large extent of the validity of conventional knowledge, which they see as being invalidated by the search for the ultimate truth. They are not ready, however, to dispense entirely with the twin ideas of knowledge and truth at the transcendent level. When commenting on Candraklrti's discussion of how the mind ceases within the state of full enlightenment, the Eighth Karmapa argues that Candraklrti's view of wisdom as a cessation of mental activity should not be taken literally but as referring to the dissolution of dualistic consciousness within the uncontrived expanse of primordial wisdom. 38 Similarly, dOe 'dun chos 'phel (1904-1951), a recent and controversial critic of Tson kha pa who is sometimes presented as holding views quite close to Pa tshab's,39 is in fact closer to the 38 BrunnhOlzl 2004: 329. For an analysis of Go rams pa's view within a similar perspective, see Thakchoe 2007. 39 There is some similarity if one considers the skeptical arguments used by dGe 'dun chos 'phe1, particularly the one that impugns knowledge on the basis of its circularity. To know something, one must be able to relate it to other things. This in turn implies that one must know these other things. But how can one be sure that one knows that one knows these things? Similarly, Pa tshab attacks the view that to be valid, statements

Pa tshab and the origin of Prasangika

413

transcendentalist position than to real skepticism. He relentlessly critiques ordinary knowledge, exposing its limitations and weaknesses. This critique is only, however, a partial skepticism for it does not entail a thorough rejection of the twin notions of truth and knowledge. Rather, it uses a partial skepticism concerning conventional modes of knowing for the soteriological purpose of debunking the pretensions of worldly ways in order to exalt transcendent wisdom. This transcendentalist view is quite different from Pa tshab's position and seems to owe more to Zan Than sag pa than to him. Pa tshab's view differs from this transcendentalist perspective, for unlike the Eighth Karmapa or dGe 'dun chos 'phel, he does not seem to exclude wisdom from the thoroughly deconstructive enterprise that he sees as being essential to Madhyamaka. For him omniscience is not the absolute overseeing of all things in their particularities but the full realization of the soteriological potential of emptiness. It is not a form of knowledge in the usual sense of the word, but, rather, a wisdom born from the realization of the futility of any attempt to determine the nature of things. Hence, Pa tshab seems to come close to the thorough skepticism of Pyrrho for whom philosophy is a cure for our excessive tendency to constantly judge things by asserting them or denying them. Both attitudes are extreme and lead to dissatisfaction, restlessness and strife. Hence, they are to be overcome by the suspension of judgment, the refusal to assert or deny anything. This is not unlike Pa tshab's understanding of thesislessness (khas len kun bral), which also aims at the suspension of judgment, whether positive or negative. 40 If I am right in my understanding of his views (and if the attribution of the manuscripts' authorship is correct), Pa tshab would

require prior validation. For if this was the case, what would validate this prior validation? Pa tshab, Ses Tab ti ka, p. 50. For an excellent discussion of dOe 'dun chos 'phel's skepticism, see Lopez 2006: 128-136. 40 For a discussion of the philosophical implications of Pa tshab's stance and a more developed comparison with ancient skepticism, see my forthcoming "Can a Miidhyamika be a Skeptic? The Case of Patsab Nyimadrak."

414

Georges Dreyfus and Drongbu Tsering

occupy a rather unique position within the Tibetan tradition as a thorough skeptic ready to dispense almost entirely with the twin notions of knowledge and truth. Obviously this is just a preliminary conclusion and only further research will be able to confirm this view, but it looks likely that Pa tshab will emerge as a rather unique and isolated figure within the history of Tibetan Madhyamaka. 41 This is confirmed by the little we know of his life. Pa tshab seems to have been respected more for his work as a translator than for his Madhyamaka explanations. In fact, he seems to have had a remarkably hard time attracting disciples and sponsors. Coming back from a long stay of twenty-three years in Kashmir where he studied extensively Sanskrit and the Indian Madhyamaka tradition of his day, Pa tshab must have expected to be considered the great authority on these matters in Central Tibet. But far from this being the case, he seems to have been almost completely ignored. Seeing this, dGe bses Sa ra ba directed to him a number of his disciples thus ensuring his success and prosperity.42 We obviously do not know the reasons behind Pa tshab's lack of success. It may be that he was more gifted for languages and close philological work than for philosophical speculations and analyses. We may also speculate, however, that his thorough skepticism may have played a role in his lack of success. The depiction of the realization of emptiness as a mere suspension of judgments may not have been very

41 Skepticism covers a large range of positions, from Pyrrhonism to more moderate positions. Hence, Jay Garfield is quite right to argue for a large cross-cultural family of skeptical thinkers, including Pyrrho, Sextus, Hume, Wittgenstein, Nagarjuna, Candraklrti, and TSOIl. kha pa. See Garfield 2003. In this perspective, all Tibetan Madhyamikas would be skeptic of some kind, but it is clear that Pa tshab's view seems to be much more thoroughly skeptical than the other three strands mentioned here. In fact, he seems to come very close to Pyrrho, the founder of skepticism in the ancient Greek world. Finally, it should be clear that the skepticism discussed here has little to do with the modern radical skepticism used by Descartes in his Meditations. For a wonderful exploration of ancient skepticism that lends itself to a comparison with Madhyamaka, see Conche 1973. 42 Khu byug 2004: 131.

Pa tshab and the origin of Prasailgika

415

attractive to Tibetan thinkers intensely immersed in the transcendent perspective of the tantras and influenced by the more positive views of omniscience found in the Perfection of Wisdom literature propagated by rNog BIo ldan ses rab and the gSaIi phu tradition. Thus, far from being the first proponent of the interpretation of the Prasangika that has come to dominate non-dGe lugs traditions in Tibet, as he is often presented, Pa tshab may have been a rather unique and isolated, though pivotal, figure. It is true that he did champion the Prasangika interpretive line by his work as a translator and that this promotion had important consequences for the development of Tibetan Madhyamaka. It is also true, as we saw above, that he had important disciples such as Zan Than sag pa and rMa bya, who played decisive roles in the various formulations of the Prasangika that came to dominate the Tibetan tradition. But Pa tshab's actual views as they are reflected in the available texts were largely ignored or thoroughly reinterpreted. It is his disciples Zan Than sag pa and rMa bya who first formulated the interpretations of Madhyamaka that were going to be successful in Tibet by weakening Pa tshab's skeptical deconstruction of knowledge to emphasize the validity of knowledge and the transcendent perspective of enlightenment. Their views were then taken over by gTsan nag pa, who defended the newly formulated PrasaIigika against Phywa pa's attacks and established it as the line of Madhyamaka interpretation that came to dominate the landscape of Tibetan Madhyamaka. In this way, Pa tshab's critique became integrated into a dominant paradigm predicated on the critique of the ordinary worldview as a way to exalt the transcendent perspective of the enlightened state, a perspective more attuned to the Tibetan religious context than Pa tshab's austere skepticism.

References
bKa' gdams gsun 'bum phyogs bsgrigs=bKa' gdams gsun 'bum. Chengdu: dPal brtsegs bod yig dpe riiin iib 'jug khan, 2006. Brunnholzl, Karl. 2004. The Center of the Sunlit Sky: Madhyamaka in the Kagyii Tradition. Ithaca: Snow Lion.

416

Georges Dreyfus and Drongbu Tsering

Burton, David. 1999. Emtpiness Appraised: A Critical Study of Niigiirjuna's Philosophy. London: Curzon. Conche, Marcel. 1973. Pyrrhon ou l'apparence. Villers sur Mer: Editions de Megare. Davidson, Ronald. 2005. Tibetan Renaissance. New York: Columbia University Press. Dreyfus, Georges. 1997. Recognizing Reality: Dharmakfrti's Philosophy and Its Tibetan Interpretations. Albany: State University of New York Press. Dreyfus, Georges and McClintock, Sara. 2003. The Sviitantrika-Priisarigika Distinction: What Difference Does a Difference Make? Boston: Wisdom. Garfield, Jay. 1995. The Fundamental Wisdom ofthe Middle Way: Niigiirjuna 's Mulamadhyamakakiirikii. New York: Oxford University Press. Garfield, Jay. 2003. "Epoche and Shiinyata." In Empty Words: Buddhist Philosophy and Cross- Cultural Interpretation. Oxford: Oxford University Press, 3-23. Hopkins, Jeffrey. 2006. Mountain Doctrine: Tibet's Fundamental Treatise on Other-Emptiness and the Buddha Matrix. Ithaca: Snow Lion. 'Jam dbyaIis Mad pa. Grub mtha'i rnam Mad chen mo. In Ngawang Gelek Demo ed., The Collected works of 'Jam-dbywis-biad-pa'i-rdo-rje. New Delhi, 1974. Khu byug. 2004. Bod kyi dbu ma'i Ita ba'i 'chad nan dar tshul blo gsal mig 'byed. Lhasa: China Tibetological Publications. Lang, Karen. 1990. "Pa-tshab Nyi-ma-grags and the Introduction of PrasaIigika into Tibet." In L. Epstein & R. Sherburn ed., Reflections on Tibetan Culture. Lewiston: Mellen Press, 131-142. Lopez, Donald. 2006. The Madman's Middle Way. Chicago: University of Chicago. McClintock, Sara. 2000. "Knowing All through Knowing One." Journal of the Internal Association of Buddhist Studies 23.2: 225-244. Pa tshab. Ses rab ti ka=dBu' ma rtsa ba'i ses rab kyi ti ka bstan beos sgron ma gsal bar byed pa. In bKa' gdams gsuri 'bum, vol. 11, pp. 29-132. Pa tshab. Tshig gsal dka' Mad=Tshig gsal ba'i dka' ba Mad pa. In bKa' gdams gsuri 'bum, vol. 11, pp. 137-203. Ruegg, bavid Seyfort. 1977. "The Uses of the Four Positions ofthe Catu~koti and the Problem of the Description of Reality in Mahayana Buddhism." Journal of Indian Philosophy 5: 1-71. Ruegg, David Seyfort. 2000. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy. Vienna: Arbeitkreis fUr Tibetische and Buddhistische Studien.

Pa tshab and the origin of PrasaIigika

417

Tauscher, Helmut (ed.). 1999. Phya pa chos kyi sen ge: dBu ma sar gsum gyi ston thun. Vienna: Arbeitkreis flir Tibetische and Buddhistische Studien. Thakchoe, Sonam.. 2007. The Two Truths Debate: Tsongkhapa and Goramba on the Middle Way. Boston: Wisdom. Vose, Kevin. 2009. Resurrecting Candrakfrti. Boston: Wisdom. Yoshimizu, Chizuko. 2005. "A Tibetan Buddhist Text from the Twelfth Century Unknown to Later Tibetans." Cahiers d'Extreme-Asie 15: 127164.

In pursuit of transparent means of knowledge

The Madhyamaka project of rMa bya ByaIi chub brtson 'grus* Thomas Doctor

Based on our few extant sources it appears that from the 8th and well into the 11th century, Tibetan scholars considered the tension between Nagarjuna's Madhyamaka and the Vijiianavadin doctrine associated with Maitreya and Asailga to have been decisively resolved through Santarak~ita's sophisticated synthesis. Yet as Candraklrti's works, with their critique of central Vijiianavadin principles, became available in Tibetan, the success and relevance of such an interpretation could no longer be left unquestioned. In fact, as we now know from recently available source material, for a number of influential Madhyamikas it was soon a central concern to be able to explain correctly what it is that distinguishes Candraklrti's approach from that of Santarak~ita or Bhavaviveka, and so draw what became known as "the *Prasailgika/*Svatantrika distinction" (dbu rna thaI ran gi khyad par).1 This paper will examine the approach of rMa bya Byail chub brtson 'grus (12th century)2 who according to several sources

Many thanks are due to Tom Tillemans and Mark Siderits for their input and critique. Thanks also to Pascale Hugon and Kevin Vose for many helpful suggestions and hints.
1 See DREYFUS and MCCLINTOCK (eds.) 2003 for a rich discussion of this hermeneutical issue that over the centuries spurs tremendous philosophical activity in Tibet.
2 rMa bya and his Madhyamaka are discussed in WILLIAMS 1984 and 1985, YOSHIMIZU 1993, RUEGG 2000, CABEZON 2003, and VOSE 2005.

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 419-441

420

Thomas Doctor

was among the foremost students in the circle surrounding the pramiiIJa ("valid means of cognition," Tib. tshad ma) theorlst Phya pa Chos kyi sen ge (1109-1169).3 Yet contrary to Phya pa, who is known for his thoroughgoing criticism of the newly imported texts, rMa bya embraced the teaching of the Prasannapadii and the Madhyamakiivatiira. Having studied with the translator Pa tshab Ni ma grags pa and/or (depending on our sources) the Indian pal.:H;lita Jayananda, rMa bya taught extensively and produced several Madhyamaka treatises and commentaries, thereby coming to playa central role in the development of Candraklrti-inspired Madhyamaka in Tibet. 4 Beginning with a presentation of his hermeneutical framework with respect to Yogacara and Madhyamaka, we will here seek to delineate some of the central features in rMa bya's comprehensive treatment of the two truths and the means for knowing them. This will lead us to raising a question as to how far rMa bya's account of mere appearance (snail tsam), which is at once critical of and inspired by the pramiiIJa tradition, may be compatible with his carefully non-committal approach to ontology and metaphysics. In the process it should become apparent that rMa bya Byan chub brtson 'grus exercised a decisive influence on the formation of Tibetan Madhyamaka thought. Our sources for the discussion will be rMa bya's newly discovered dBu ma rigs pa'i tshags kyi rgyan de kha na fiid snail ba (with auto-commentary)5 and his commentary to the Mulamadhyamakakiirikii. 6

3 The identity of rMa bya Byan chub brtson 'grus remains somewhat enigmatic. See RUEGG 2000: 50-51 and WILLIAMS 1985. 4 Authors such as Klon chen pa, Tson kha pa, Sakya mchog Idan, Go ram pa, and the eighth Karma pa, Mi bskyod rdo rje, testify, both explicitly and indirectly, to rMa bya's status as a principal exponent among the "early generations" (sna' rabs pa) of Tibetan Madhyamikas. See my forthcoming study of rMa bya's Madhyamaka. 5 Hereafter De fiid snan ba. My forthcoming study of rMa bya's Madhyamaka will include an English translation of the Tibetan text. 6 dBu rna rtsa ba ses rab kyi 'grel pa 'thad pa'i rgyan, hereafter 'Thad rgyan. English translation in rMa bya 2011 (forthcoming).

In pursuit of transparent means of knowledge

421

rMa bya on the three dharmacakras

rMa bya Byan chub brtson 'grus is a Madhyamika who concurs with Candraklrti that liberation may only be gained through the path shown by Arya Nagarjuna. For rMa bya, this achievement depends on being able to existentially negate (med par dgag pa, prasajyaprati~edha) all mental constructs (spros pa, prapaiica), and so the Middle Way is discovered through a universal dismissal of both positive and negative phenomena. It is only through this realization of complete freedom from constructs (spros bral, ni~prapaiica) that the liberation of a sriivaka, a pratyekabuddha, or a bodhisattva can be won. 7 Whether the mind holds onto crude entities of self and other, whether it believes in non-dual cognition, or whether it adheres to emptiness itself, all such grasping is, in the end, detrimental to the pursuit of liberation, 8 and rMa bya thus sees the Prajiiiipiiramitiisutras and Nagarjuna's "collections of reasoning" (rigs tshogs) as the culmination of the Buddhadharma. 9
See, e.g., De iiid snan ba 812-813. 8 Cf., e.g., De iiid snan ba 767 and 812. 9 Under the category of treatises that explain the intent of the Prajiiiipiiramitiisutras, rMa bya distinguishes between two categories (,Thad rgyan 8): "Second are [the treatises] that clarify the meaning of the siltras on the absence of characteristics, within which there are two categories. One is constituted by the master Nagarjuna's Six Collections of Reasoning and the treatises that adhere to them. These are works that use reasoning to explain what is taught explicitly by the siltras [of the Prajiiiipiiramitii], namely the natural emptiness of all phenomena that is observed by knowledge, or by the view. The second category contains the noble Maitreya's Ornament ofManifest Realization (Abhisamayiilamkiira) and the treatises that adhere to it. Distinguishing between eight manifest realizations, these works clarify the implicit teaching of the siltras. They are thus concerned with the subject and the stages of manifest realization that comprise the path and fruition." (giiis pa mtshan iiid med pa'i mdo'i don gsal bar byed pa la giiis te / mdos dnos su brjod pa ses rab bam Zta ba'i dmigs pa chos thams cad ran biin ston pa iiid rigs pa'i sgo nas gsal bar byed pa slob dpon klu sgrub kyi dbu ma rigs pa'i tshogs rjes su 'brans pa dan bcas pa dan / sugs la tshig gi don du bstan pa yul can lam dan 'bras bus bsdus pa mnon par rtogs pa'i rim pa mfwn rtogs brgyad du phye
7

422

Thomas Doctor

In rMa bya's interpretation it is, however, not only the Prajfiiipiiramitii literature that expresses the definitive meaning (nftiirtha). The third and final dharmacakra (rMa bya appears to refer to the Lankiivatiira as an example of the texts included in this category) reveals as well the intent of the Buddha directly, for these texts resolve the apparent conflict between the declaration of existent but personless dharmas in the initial dharmacakra on the four truths of the Aryas (bka' dan po bden pa bii'i chos 'khor) and the subsequent denial of the same in the Prajfiiipiiramitii. 10 It is, rMa bya explains, in relation to this third dharmacakra that the Cittamatra and Madhyamaka commentarial traditions emerge. 11 Whereas Cittamatra seeks to achieve the necessary reconciliation between the first two dharmacakras by interpreting the Prajfiiipiiramitii's
nas ston pa 'phags pa byams pa'i mrwn par rtogs pa'i rgyan rjes su 'brans pa daft bcas pa'o /) 10 See note 12. 11 'Thad rgyan 8-9: "The third [category of treatises] explains the intent of the final teachings of definitive meaning. Some rely upon the view of the Mind Only School to explain numerous principles associated with the view and conduct. In this category we find the noble Maitreya's Distinguishing the Middle from Extremes (Madhyantavibhaga), Ornament of the Siitras (Mahayanasiitralarhkara), and Distinguishing Phenomena from their Intrinsic Nature (Dharmadharmatavibhaga). This also includes the treatises that adhere to these three, such as the master AsaJiga's five-fold Grounds of Yogic Practice (Yogacarabhiimi) and his two compendia, as well as the master Vasubandhu's eight-fold prakaralJa. There are also treatises that explain the intent of the final set of teachings to be the Middle Way. These include the noble Maitreya's Supreme Continuity (Mahayanottaratantra) and the master Candrakirti's Entering the Middle Way (Madhyamakavatara)." (gsum pa bka' tha ma nes pa don gyi dgofts 'grella Ita ba sems tsam la brten nas Ita ba daft spyod pa'i tshul ci rigs su gsal bar byed pa 'phags pa byams pa'i dbus dan mtha' rnam par 'byed pa daft / mdo sde rgyan daft / chos can [sic] daft chos fiid rnam par 'byed pa gsum dan / de dag gi rjes su 'brafts pa slob dpon thogs med kyi mal 'byor spyod pa'i sa sde lfta sdom gfiis daft bcas pa dan / slob dpon dbyig gfien gyi pra kar lJa sde brgyad la sogs pa'o / bka' tha ma'i dgofts pa dbu mar 'chad pa 'phags pa byams pa'i rgyud bla ma daft / slob dpon zla ba grags pa'i dbu ma la 'jug pa la sogs pa'o /)

In pursuit of transparent means of knowledge

423

declaration of the non-existence of all dharmas as pertaining only to what is of imaginary nature (kun brtags, parikalpita), "the learned followers of noble Nagarjuna employ," according to rMa bya, "scripture and reasoning to explain the intent of the siitras without error." Thus the Madhyamikas understand that while the initial dharmacakra considers "the relative, deluded perspective," the Prajiiiipiiramitii's universal negation of all dharmas is made "in accordance with the ultimate, logical perspective that is free from de1usion."12
12 'Thad rgyan 6: "For those individuals with a potential to be inclined towards a diversity of vehicles, the Buddha turned the final wheel of Dharma on the definitive meaning at locations such as Lal!kapuri. In terms of content, the conduct presented in these teachings consists of working for the welfare of others on a vast scale. In terms of view, they address a qualm that is based on a perceived contradiction between the initial and intermediate teachings. The initial teachings explain that, while there is no person, the mere phenomena of the five aggregates, or those associated with the four truths, do exist. The intermediate teachings, meanwhile, claim that even these phenomena are nonexistent. Addressing this problem, the followers of the Yogic Practice School consider the intent of the third set of teachings to be the Mind Only [position]. When stating that these phenomena are nonexistent, they explain, the second set of teachings intends to show the non-existence of the imaginary object-subject duality. This does not, however, mean that the dependent, non-dual cognition that is the support for the affiictions and their purification is also nonexistent. Therefore, they say, these two do not contradict one another. Nevertheless, the learned followers of noble Nagarjuna employ scripture and reasoning to explain the intent of the sutras without error. They resolve the message of the first and second sets of teachings in the following way: In the initial set of teachings, the aggregates, elements, and sense sources (or the mere phenomena associated with the four truths, the dependently originated phenomena that are devoid of a person and subsumed by affiiction and purification) are said to exist. This statement is made in consideration of the relative, deluded perspective. On the other hand, in the second set of teachings these are all said to be nonexistent. This statement is in accordance with the ultimate, logical perspective that is free from delusion. Hence, they say, there is no contradiction." (gduZ bya thegs pa sna gtsogs Za mos pa'i rigs can Za yuZ Zan ka bu ri la sogs par brjod bya spyod pa gian don rgya chen po dan /

424

Thomas Doctor

Two truths based on the present appearance

The Madhyamaka paradigm of two truths, one relative (kun rdzob, samvrti) and the other ultimate (don dam, paramartha), is, for rMa bya, entirely mind dependent. Just as the same thing may be classified as either long or short depending on our point of reference, the truth is twofold simply because whatever appears to us can be related to by either a mind that suffers from delusion or one that is free from distorting influences. Both truths are thus distinguished based on "precisely that which appears" (ji ltar snati ba'i tio bo 'di fiid).13 That something does appear is nothing but the result of delusion,14 and yet there is, rMa bya appears to explain, no ultimate truth to be discovered beyond the present appearance:
All that appears is unestablished, both from the perspective of the non-conceptual wakefulness of a noble one's equipoise, as well as from the viewpoint of a consciousness that has employed reason to analyze and investigate. The appearance itself is neither existent nor nonexistent, neither true nor false. It entirely lacks any essential establishment, such as being an object of cognition, expression, or evalua-

lta ba bka' dan par gan zag gis stan pa'i phun po lna'am bden bii'i chos tsam yod par biad pa dan / gfiis par de yan med par biad pa gfiis 'galla sfiam du dogs pa la rnal 'byor spyod pa ba dag bka' gsum pa'i dgons pa sems tsam du byas nas bka' gfiis par de dag med par Mad pa ni kun brtags gzun 'dzin gfiis med pa la dgOlis pa yin gyi fion mons rnam byan gi rten gian dban gfiis med kyi ses pa'an med par stan pa ma yin pas de gfiis mi 'gal ies 'chad mod kyi lun dan rigs pas mdo'i dgons pa phyin ci ma log par 'chad pa 'phags pa klu sgrub kyi rjes su 'bran pa'i mkhas pa rnams ni bka' dan par gan zag gis stan pa'i phun po khams dan skye mched dam / fion mons rnam byan gis bsdus pa'i rten 'grel bden pa bi(i chos tsam yod par gsuns pa ni kun rdzob 'khrul pa'i no dan / gfiis par de dag med par gSUJi pa ni don dam pa rigs pa'i ses pa ma 'khrul pa'i nor yin pas mi 'gal lo ies stan pa bka' tha ma nes pa don gyi chos kyi 'khor lo ste bka'i rim pa gsum dan/) 13 De fiid snan ba 755. 14 Cf. De fiid snan ba 813ff.

In pursuit of transparent means of knowledge

425

tion. It is, hence, beyond all conceptual constructs. The conventional term "ultimate truth" is applied in consideration of this alone. 1s

Mark Siderits' has summarized Madhyamaka antirealism in the statement, "the ultimate truth is that there is no ultimate truth,"16 and while rMa bya stops short of saying just that, such formulation of the niJ:tsvabhava doctrine is much in tune with his discourse. For rMa bya, any mention of such things as defining characteristics (mtshan fiid, lak~m:/.a), definienda (mtshon bya, lak~ya) and bearers of characteristics (mtshan gii)17 in the context of the ultimate is simply meant as a corrective against the mind's tendency to construct such features of an ultimate reality. Thus one sets forth a defining characteristic of ultimate reality precisely because it is necessary to show that there could not be any:
Unless such things as defining characteristics are taught by way of superimposition, it will not be understood that defining characteristics, etc., have no bearing on the ultimate truth. And unless that is realized, it will not be possible to achieve liberation from the attachments to an ultimate truth that has defining characteristics, etc. Therefore, so that those who wish for liberation may gain such realization, it is, at the level of convention, necessary to teach through the superimposition of such things as defining characteristics. 18
1S 'Thad rgyan 40: ji ltar snan ba 'di ilid 'phags pa'i milam gzag mi rtog pa'i ye ses sam / brtags zin dpyad nas 'jug pa'i ses pa'i no na yod med dam bden rdzun nam ses brjod gzal bya la sogs pa'i ci'i no bor yan grub pa med cin spros pa'i mtha' thams cad dan bral ba tsam fa don dam pa'i bden pa zes tha silad tsam du byas par zad de / 16 SIDERITS 2003. 17 On this triad, see Pascale Hugon's discussion included in this volume. 18 De ilid snan ba 760: mtshan ilid la sogs pa'i sgo nas sgro btags nas ma bstan na don dam pa mtshan ilid la sogs pa med pa'i tshul gyis mi rtogs la de ltar ma rtogs na yan don dam mtshan ilid dan bcas pa la sogs par zen pa las thar pa mi thob pa'i phyir / thar pa 'dod pas de ltar rtogs par bya ba'i don du tha sfiad du mtshan fiid la sogs pa'i sgo nas sgro btags la bstan dgos pa yin ste / See likewise 'Thad rgyan 38 and De fiid snan ba 762. The latter states: "From the perspective of reasoning one fails entirely to find any phe-

426

Thomas Doctor

The denial of relative particulars Once this complete absence of anything ultimate has been acknowledged, it will, rMa bya explains, as well be seen that whatever appears is not even conventionally related to any fact of the matter at all. Preceding Tsori kha pa by more than two centuries, rMa bya repeatedly draws a distinction that will provide the framework for the perhaps most elaborate debate in the history of Tibetan Madhyamaka. Here is an example:
Arising, cessation, and so forth are like dreams and illusions. They are merely apparent from the perspective of the deluded mind. Particular characteristics, such as those that the Middle Way Autonomists (*svatantrikamadhyamika) believe in, do not even exist conventionally.19

What distinguishes the flawless Madhyamaka from the Svatantrika approach (contrary to Pa tshab, but similar to gTsari nag pa,20 rMa bya does not use the term thal 'gyur ba in juxtaposition with ran rgyud pa, but speaks instead of dbu chen, "the Great Madhyamaka," or simply chos 'di, "this doctrine") is the fact that the latter buys
nomenon that supports true characteristics, and so it is established that the bearers of such characteristics are beyond the realms of language and mind. Therefore, given this absence of a support for them, it certainly follows that there are no characteristics either. When earlier we nevertheless set forth characteristics [of the ultimate truth] the explanation was given in terms of superimposition, so that it may be realized that there are no characteristics." (rigs nor bden pa'i mtshan fiid rten pa'i chos 'ga' yan ma rfied pas mtshan gii sgra blo'i yullas 'das par grub la / de fiid kyis na rten med pas mtshan fiid kyan med par nes pa yin no / 'on k)jan snar mtshan fiid smros pa ni mtshan fiid med par rtogs par bya ba'i don du sgro btags nas bstan par yin no /)
19 'Thad rgyan 27: skye dgag la sogs pa'an rmi lam sgyu ma ltar ses pa 'khrul pa'i no na de ltar snan ba tsam yin gyi ran rgyud du smra ba'i dbu ma pas brtags pa Ita bu'i ran gi mtshan fiid du gyur pa ni kun rdzobs du yan med lao 20 See Kevin Vose's discussion, included in this volume, of this prolific Madhyamika, who like rMa bya is reported to have left Phya pa to pursue Candrakirti studies.

In pursuit of transparent means of knowledge

427

into that reification of phenomena, which is the shared tendency of all deluded beings. Although Svatantrikas restrict the relevance of "particular characteristics" (ran gi mtshan fiid, svalakCiaTJa) to conventional reality, they thereby reveal their cryptorealist ontology, which, as we shall later see, goes hand in hand with an unrealistic expectation of perfect epistemological certainty. rMa bya lets a Svatantrika defend their position as follows:
Indeed, as they cannot be found upon examination by means of logical cognition, there are no ultimately real entities. Hence we do not say either that "things are ultimately real"! Our position is merely that while double moons, etc., are not even real entities in terms of the relative, a single moon, etc., is, at the level of the relative, a real thing by virtue of its particular characteristics. The latter sort of thing will not be invalidated on account of not being observed by logical cognition, because such things are not the object of logical cognition. Speech is, similarly, not invalidated by virtue of not being observed by the visual cognition. 21

The opponent believes that he has fully acknowledged the Madhyamaka critique of real existents, for he accepts that no thing can have ultimate reality. Yet it is, for him, just as obvious that, as a follower of Nagarjuna, one must differentiate between what is correct and incorrect conventionally based on the presence or absence of relative substance. The Madhyamaka arguments therefore target only entities that are held to feature ultimate reality. If we are to duly appreciate the two truths of the Middle Way, we must, he will tell us, recognize both ultimate emptiness and the relative existence of an objectively established world. Negating the latter on account of the former would be similar to rejecting the existence of sound because of its not being perceptible to the eyes. What we want, in other

21 De iiid snmi ba 774: ci ste rigs ses kyis dpyad na ma riied pas dftos po don dam par bden pa khegs mod kho bo cag don dam par dftos po rnams bden no i.es ni mi smra'i / 'on kyaft zla giiis la sogs pa ni kun rdzob tu yaft dftos por mi bden fa / zla ba gcig fa sogs pa ni kun rdzob tu yaft raft gi mtshan iiid kyi dftos po bden pa yin i.e 'di tsam i.ig smra ba yin no / de ni rigs ses kyis ma dmigs pas mi khegs te rigs ses kyi yuf ma yin pa'i phyir mig ses kyis smra ba ma dmigs pas mi khegs pa bi.in no /

428

Thomas Doctor

words, is equal recognition of the two truths, not some prejudiced favoring of one particular cognitive perspective. This seemingly prudent assertion of relative, particular characteristics is, rMa bya argues, nevertheless far from innocent, and in fact antithetical to both the Madhyamaka philosophical enterprise and the pursuit of liberation. rMa bya therefore retorts:
Suppose that our failing to find them by means of logical cognition would then not be sufficient to invalidate relative reality in the form of particularly characterized entities. This would also mean that our failing to find them when we are awake would not be sufficient to refute that what we experience during dreams are particularly characterized entities, because dream-experience is, similarly, not the object of the waking consciousness. Such a consequence can, however, not be accepted, because it would turn any object into a true entity, and all apprehending minds would end up free from delusion. 22

The opponent wanted to restrict the scope of Madhyamaka negation, and so created hedges whereby the refutations would apply only in terms of ultimate reality (dgag bya la don dam gyi khyad par sbyar ba). Yet what he achieves is not secure ontic grounds for distinguishing between real and unreal relative phenomena. Contrary to such Svatantrika wishes and expectations, he ends up, rMa bya argues, forced to accept objects that are clearly illusory, such as those that we seem to encounter in dreams, as being perfectly real. Likewise, given that its object is real, the mind that apprehends such illusions will have to be veridical. This quite paradoxical consequence follows because all hedging against the Madhyamaka critique can be done only along lines that are entirely arbitrary.23 We may for some reason (its being perceptible, efficacious, undeniable, etc. - the list can include any given reason)

22 De tiid snan ba 774: gal te rigs ses kyis ma rtied pas / kun rdzob tu ran gi mtshan tiid kyi dnos por bden pa mi khegs na / sad pa'i ses pas ma rtied pas kyan rmi lam du ran gi mtshan tiid kyi dnos po mi khegs par 'gyur te rmi lam sad pa'i ses pa'i yul ma yin par mtshun pas so / de ltar yan 'dod par mi nus te yui thams cad dnos po bden pa dan de 'dzin bio thams cad 'dzin stan ma 'khrul bar 'gyur ba'i phyir ro / 23 Cf. 'Thad rgyan 54-56 where this point is argued extensively.

In pursuit of transparent means of knowledge

429

decide that the vase before us is only refutable at the level of the ultimate, and hence objectively real as a relative entity. Yet what, rMa bya's argument goes, is then to prevent someone from convincing us that the horned donkey and furred turtle, which we may have dreamt of encountering, are unreal only in the sense that they cannot be found in the world we experience when awake, and that despite this specific lack of establishment, they are still undeniably real and to be reckoned with as inhabitants of the world of dream? If the present experience ofthe vase is sufficient grounds for ascribing to it a particular, restricted level of undeniable reality, then why would not the same have to follow in the case of that which appears to us in dreams, optical illusions, etc.? What the Svatantrika presented as sound Madhyamaka - i.e., antirealist metaphysics and epistemology, properly safeguarded against nihilist pitfalls - turns out to be realism with a vengeance. As they end up allowing that anything can be real in some sense, the Svatantrikas have denied themselves the very possibility of absence and nothing. On rMa bya's argument, hedging in terms of conventional/ultimate reality does not bring secure epistemological foundations but opens up the floodgates for the irrational. rMa bya's anti-realism considers, we may at this point conclude, any unequivocal ontological classification of apparent reality a severe error, and in the end even the perception of appearance as such is but a delusion that is to be overcome through sustained cultivation of the path. Yet, as we shall next see, rMa bya's approach is far from quietist. Madhyamaka epistemology While indeed the Madhyamaka could not hold any ultimate view, it is, for rMa bya, just as obvious that in accepting this ultimate absence of any position, the Madhyamaka is necessarily committed to the conventional truth of the proposition whereby this absence is expressed. Responding to those who take Vigrahavyiivartanf 29 to be saying that as a Madhyamika one does not have any view of one's own, rMa bya enquires:

430

Thomas Doctor

Self and other are mutually dependent, so if there is no position of one's own there is no position of others either. What, then, is it that you wear yourself out refuting?24

Given the interdependent nature of such notions, the very idea of "extreme positions," "positions of others," etc., necessarily involves their opposites. Hence, if there is a wrong view there is also a right . view, and "why then would one not, as mere convention, assert the path of the Madhyamaka, free from extremes and originating in dependency?"2S Indeed, conventionally the teaching of the Madhyamaka is dependent origination free from prapafica, just as shown in the Mulamadhyamakakarika's opening praises. Hence, as there is a view of the Madhyamaka, understanding and developing this view will, rMa bya insists, require prama1}as, but prama1}as that carry just as little metaphysical baggage as that appearance of dependent origination to which they are applied. A substantial part of rMa bya's De fiid snali ba consists of a detailed criticism of Indian and Tibetan prama1}a thinkers, whom rMa bya rejects for their claim that prama1}as apply to their objects "by the power of fact" (dlios po'i stobs tugs, vastubalapravrtta).26 Thus rMa bya considers the definitions of what constitutes a pramana found in the works of Dharmottara, SaJikaranandana, Prajfiakaragupta, Devendrabuddhi, and Sakyabuddhi only to point out that they, in their attempt at grounding prama1}a in facts, all inadvertently end up denying any possibility of reliable means of knowledge. As facts are simply not available, any epistemology that assumes otherwise has already failed. For rMa bya, a Svatantrika is a Madhyamaka who falls into this trap of accepting the self-defeating paradigm of *vastubala24 De fiid snan ba 746: ran gian ltos grub yin pas rait phyogs ni II med pas gian phyogs yod min de yi phyir II khyed ni ci Zig 'gog nas nal bar 'gyur II 25 De fiid snan ba 746: mtha' bral rten cin 'brel ba'i dbu ma'i lam II kun rdzobs tsam du ji ltar khas mi len II 26 De fiid snan ba 768-777. Note the close resemblance to the Tarkamudgara (Toh. 3869, sde dge bstan 'gyur, dbu ma, vol. ya, 374b) by rMa bya's reported teacher, Jayananda.

In pursuit of transparent means of knowledge

431

pravrttapramiilJa, i.e., the notion that the mind can a) "newly re-

alize and experience its object free from superimposition, just as it objectively- is" and b) "accurately realize its object by means of evidence [apprehended in a way] that is free from delusion."27 The Svatantrika's belief in such fact-based cognition ties in with his mistaken ascription of particular characteristics to conventional reality, an ascription that, in turn, equally involves a reification of emptiness (ston pa fiid yan dag par sgro 'dogs pa). For the Svatantrika, objective pramiilJas hence perform the dual role of establishing conventional entities as well as identifying the ultimate facts of the emptiness of such entities. When this reifying tendency is allowed to go unchecked, it develops into the position of aston pa mtshan mar smra ba, "someone who teaches emptiness as if it were a mark (Skt. nimitta)."28 Adhering to the idea that emptiness exists, indeed must exist, as an object of rational cognition, these realists in Madhyamaka guises miss the point of freedom from mental construction (spros bral, ni~prapafica) and so end up denying themselves liberation. 29
27 De fiid snan ba 768: ses bya ran gis sgro ma btags su dnos po fa gnas pa ltar snar ma rtogs pa gcig myon pa dan mi 'khrul pa'i rtags las ji ltar gnas pa ltar rtogs pa la dnos po'i stobs kyis iugs pa'am 'dzin stans mi 'khrul pa'i tshad ma ies bya la / 28 De fiid snan ba 745. 29 De fiid snan ba 767: "There are some Madhyamikas who believe that emptiness is a mark. For them, emptiness, as the negation of any sense of personal and phenomenal self, exists as the object of logical cognition. Were it not so, they argue, the [cognition of emptiness] would be lacking a cognizable object and an object of evaluation (Skt. prameya), and so it would neither be a rational nor a valid cognition (Skt. pramaIJa). Furthermore, what is meant to be negated, namely true entities, would instead end up established. Also this position, which holds that emptiness is real, must be refuted, because as long as one harbours such attachments liberation will be impossible. Moreover, when the object [i.e., emptiness conceived of as reality] is examined logically, we do not find it, and so it can [indeed] be negated." (ston pa mtshan mar Ita ba'i dbu ma pa kha cig gan zag dan chos kyi bdag thams cad bkag pa'i ston fiid rigs ses kyi yul na yod pa yin te / de ltar ma yin na ses bya dan gial bya med pas rigs ses pa dan tshad ma ma yin par 'gyur iin dgag bya bden dnos grub par yan

432

Thomas Doctor

As a general observation, we may note that much of rMa bya's criticism of traditional Prama:t:lavada targets the idea that constructed kinds and classes can be meaningfully associated with real particulars. The realist Prama:t:lavadin and the cryptorealist Svatantrika both think it possible to determine properties that although constructed and insubstantial can nevertheless somehow be used to develop a map of the real. For rMa bya it is, however, impermissible in this way to turn back from the anti-realist path that the Prama:t:lavadins/Svatantrikas have set out on with their critique of universals. 30 The full consequences of their insight must be acknowledged, and so we cannot, rMa bya argues, even conventionally continue to accept the notion of objectively established properties. When the analysis of the universal is carried through to its Madhyamaka consequences, no thought or perceptual appearance can retain any association with facts. Now whereas Pa tshab dismisses the notion of pramii]Ja across the board,31 rMa bya insists that such means of knowledge are a natural and undeniable element in any successful enterprise, whether mundane or spiritual. 32 The dependently originating experience of ourselves, others, and the world indeed calls for the develop'gyur ro zes smra'o I de Itar stan nid yan dag par 'dod pa 'di yan dgag bya yin te I de la zen pas kyan thar pa mi thob pas dgag dgos pa'i phyir TO I yul rigs pas dpyad na mi rned pas dgag nus pa'i phyir ro !) See also 804-806. Note the resemblance between rMa bya's portrayal of the stan pa mtshan mar smra ballta ba and the position that Phya pa defends in his dBu rna sar gsum gyi stan thun (see, e.g., TAUSCHER [ed.] 1999: 16-22). 30 In the pramiilJa tradition of Dignaga and Dharmakirti only particulars count as real entities. Universals are mere abstractions and as such fictional in the final analysis. See, e.g., DREYFUS 1997 and DUNNE 2004. 31 dBu rna rtsa ba'i 'grel pa, folios 40-55. In their article in the present volume Georges Dreyfus and Drangbu Tsering present and discuss the three texts contained under this title. The total rejection of pramiilJa is presented in the first of these, the rTsa ba ses rab kyi ti ka bstan beos sgron rna gsal bar byed pa, which does not explicitly claim to be a work by Pa tshab. See Dreyfus's and Tsering's discussion of the question of the authorship of the three texts. 32 Cf. De nid snan ba 777-782 and 'Thad rgyan 41-45.

In pursuit of transparent means of knowledge

433

ment of an account of trustworthy cognition. Having dismissed the very possibility of concepts corresponding to facts, valid means of knowledge can, for rMa bya, then only be that which is accepted as such by the world ('jig rten na grags pa, lokaprasiddha). Having seen that the two-tiered ontology of the Pramat;lavada is hopelessly flawed, the Madhyamika will have no qualms about rejecting the logicians' parsimonious typology, and so can instead confidently follow Candraklrti in expanding the array of pramalJaS in a way that better describes our actual experience of epistemological success. 33 Yet no pramalJa can, on rMa bya's account, even conventionally lead to a "certainty" (lies pa, niscaya) that could never be proven wrong. In both the 'Thad rgyan and the De fiid snali ba, rMa bya develops an argument against an idea that he terms ldog pa gcig pu'i khyab pa,34 the notion that two or more things may share the same property, and that the identification of such shared properties can thus be relied on in the drawing of definitive inferences leading to infallible conclusions. rMa bya here argues that the unreal nature of universals renders any such Svatantrika claim to infallible judgment false. The Svatantrika's epistemological enterprise stands and falls with arguments of "definitive pervasion" (khyab pa lies pa), yet Dharmaklrti himself has tefuted any actual existence of pervasive properties. Hence, in the absence of such properties, there cannot be any infallible inferences either. Indeed, we can have qualified expectations, so qualified that in the given situation any doubt about our judgment may be entirely unreasonable. Yet the notion of infallible certainty is in the end just as baseless as the belief in particular characteristics (rali mtshan, svalak~alJa). If there were, rMa bya argues, such a thing as *vastubalapravrttapramalJa all the claims and practices of the Svatantrikas would be sound and unob-

33 'Thad rgyan 41-45 and De fiid snmi ba 780-782. In the Prasannapada Candraklrti rejects the perception/inference paradigm of the Dignaga school and instead appears to endorse a fourfold enumeration (perception, inference, analogy, and scripture) similar to the one employed by the non-Buddhist Naiyayikas. 34 'Thad rgyan 57-59 and De fiid snali ba 796-798.

434

Thomas Doctor

jectionable, but since not even conventionally are such pramiilJas available all that the Svatantrikas say and do "is simply pathetic."35 There are indeed, rMa bya appears to be saying, such things as knowledge, reason, and certainty, yet they are not what the realist imagines. In a discussion that to me is very telling about his stance on inference,36 rMa bya argues that karmic causality throughout successive lives can be proven simply by considering the overwhelming evidence that we have for concluding that one moment of mind leads to another. Without being able to identify any counterexample, our experience3 informs us that this is indeed how it ? has always been. Hence, we can conclude that also the first moment of mind in the present life must have been preceded by a similar moment in a previous life. If, rMa bya asks, we cannot consider that type of evidence conclusive, then how can we allow ourselves to, for example, infer the presence of fire based on the perception of smoke? In the latter case we do not have anything else to go by than the perception of regularity either. Certainty does not, we may conclude for rMa bya, get any better than this. In his view, a Madhyamika clearly knows both certainty and conviction, but simply as undeniable cognitive events that take place within the unexamined world of practical transactions. Upon examination all apparent natural laws will however inevitably fall to the wayside, and so there is, as rMa bya tersely states, upon examination no certainty at all. 38 In rMa bya's view, the Svatantrika's requirement of incontestable certainty is, we might surmise, not just ultimately unachiev35 De fiidsnait ba 768: tshad ma de Ita bu ni dpyad na tha sfiad du'ait mi 'thad pas 'di dag thams cad gyi na'o / 36 De fiid snait ba 783-784. 3? to be precise, rMa bya explains that we know this based on selfawareness (svasa'!lvedana, ran rig). For rMa bya, this principle, which is central to traditional Buddhist Pramfu;1avada, can be accepted (conventionally only, of course) once the Svatantrikas' realist assumptions about it have been dismissed. 38 Cf. De iiid snait ba 796: chos gait yait dpyad na ites pa kun rdzob tu yan mi 'dod.

In pursuit of transparent means of knowledge

435

able. It is, as well, a distorting notion that distracts us and carries us away from acknowledging and exploring those instruments of knowledge that we encounter in the world of practical transactions. It is those metaphysically transparent means of knowledge that are already part of our cognitive horizon, which can as well be cultivated and relied on for an unproblematic approach to emptiness - an approach that will not seduce us into reifying either of the two interdependent truths. What is or is not a means of knowledge depends on the context, and the sense of certainty that ensues from the application of the merely apparent, Madhyamaka-style pramii1}-as must therefore be appreciated simply as a reliable and undeniable factor for guidance and perspective; beyond that it cannot support any categorical mapping of the world. Hence, if I am justified in making these suggestions, proper means of knowledge are on rMa bya's Madhyamaka interpretation just that which we already know them to be: time-tested means for practical success. This renders the pramii1}-as metaphysically quite neutral and - as it would seem, given the absence of infallible conclusions - in principle unfixed and open for development. 39 The word "scientific" is perhaps not an entirely inapt description of an approach that while deliberately avoiding all claims to metaphysical truth, in this way treats knowledge as a vital but very practical issue. 4o

39 We may here further note that rMa bya's approach clearly does not facilitate the import of any of the traditionally enumerated three nonMadhyamaka tenet systems to account for the saf!1vrtisatya. Whereas Vaibha~ikas posit a pratftyasamutpilda that bears characteristics such as arising and cessation, rMa bya's dependent origination can conventionally be determined as displaying the properties of precisely the absence of such characteristics. Likewise, while Sautrantikas of the Dharmaklrtian orientation draw a sharp distinction between imagined endurance and impermanent reality, rMa bya prefers granting, as we shall see just below, the same ontological status to the lasting and the momentary. Finally, rMa bya sees, as one might expect given his Candraklrtian inspiration, mind and matter to be on an equal phenomenological footing, hence precluding any conventional alignment with a Cittamatra system. 40 Leaving aside the issue of what scientists mayor may not think themselves, it is obvious that not all philosophers of science would subscribe

436

Thomas Doctor

The status of appearance We have seen rMa bya as a champion of reason and pramiiIJa, yet all within a metaphysically non-committal, Madhyamaka atmosphere where, upon examination, not even conventional marks of reality can be determined. It is, we may conclude, the hallmark of rMa bya's Madhyamaka to combine a doctrine of freedom from mental construction (spros bral, ni~prapafica) as the single means for deliverance with a sophisticated cultivation of valid means of cognition (tshad ma, pramiiIJa). Such a systematic unification of Madhyamaka and pramiiIJa may obviously make us think of Santarak~ita and Kamalaslla, yet rMa bya's account of the relative reality (sa1'[lvrtisatya) is just as obviously not analytical in the same way as theirs. The general ambiguity of the word snali ba does, as far as I can see, lead to two somewhat different accounts of sa1'[lvrti in rMa bya's works. On the one hand conventional reality is simply whatever appears to be the case from the perspective of deluded consciousness. At this level, rMa bya suggests expanding the inventory of relative reality to include also "permanent things" (rtag pa'i dlios po), for while permanent and impermanent entities are in reality equally void, they are to the same extent apparent from the perspective of a mind that has not undertaken any Madhyamaka enquiry.41 rMa bya thus here appears happy to endorse, as conventionally relevant, even such anathematic (for the PramaI}.avada) notions as permanent entities, simply because "they seem to be

to such an understanding of what it may mean to be "scientific." Yet with its emphasis on observation as the final court of appeals and its "strong dissatisfaction with claims that proceed, even if called for, by postulation" (VAN FRAAS SEN 2002: 37), central elements in the empiricist tradition seem tq me resonant with rMa bya's antirealist epistemology. The fact that both antirealist and realist philosophers of science generally agree that even our most successful scientific theories may at some point be proven false, and that any account of scientific knowledge must therefore incorporate a form of fallibilism, is likewise suggestive of a conceptual convergence. In my forthcoming rMa bya study I seek to explore further the feasibility and restraints of such a comparison. 41 See De fiid snan ba 767-768.

In pursuit of transparent means of k.TJ.owledge

437

there." That which appears, i.e., the object of perceptual cognition, is on the other hand for rMa bya to be classified as svabhava-Iess, dependent origination. Turning Dignaga and Dharmaklrti on their heads, rMa bya explains: The mind of confined perception sees this outer and inner dependent origination that is without particular characteristics. [Based on that,] it conceives of the objects that are established to the unimpaired six faculties as if they were entities of particular characteristics, and it likewise believes the [perceiving] subject to be free from delusion. 42 In this striking example of his almost pirating appropriation of the Prama1favada apparatus in his declared Candraklrtian Madhyamaka interpretation, rMa bya apparently wishes to reserve a peculiar Miidhyamika version of epistemological accuracy to non-conceptual perception. In rMa bya's account, the sense consciousnesses experience illusory dependent origination free from svalak~Wtas, and it is, we may surmise, only by subjecting this near-perfect (from the Madhyamaka perspective) epistemic situation to conceptual construction that the mind really errs against the abiding way of its object (ses bya'i gnas lugs). Rather than acknowledging an otherwise evident absence of nature (raft biin med pa, niJ:tsvabhava),43 the mind instead proceeds to superimpose the notion of real particular characteristics (raft mtshan, svalak~Q/:ta) so forcefully that it ends up getting the impression that such objective facts are actually accessible to it.

42 De fiid snan ba 756: tshur rol mthon ba'i blo ni ran gi mtshan fiid du med pa'i phyi nan gi rten 'breI 'di mthon zin de yan gnod med kyi dban ses drug gi yulla ran gi mtshan fiid kyi dnos po dan yul can la ma 'khrul pa yan zen fa / 43 rMa bya on several occasions explains that barring the equipoise of the Aryas (phags pa'i mfiam bzag) and the "logical perspective" (rigs ses), niJ:tsvabhtiva (ran biin med pa) in the sense of emptiness as a paryudtisa (ma yin dgag) must indeed be considered established in positive determination (yons gcod du grub pa). See, e.g., De fiid snan ba 788: kun rdzob

du ran biin med pa'an yons gcod dan der snan ba dan tshogs pa'i ma yin dgag bsgrub byar byed pa yin no /

438

Thomas Doctor

Yet although emptiness (stan pa fiid, sunyatii) can in this sense be seen to be experienced in perception, rMa bya nevertheless contends that in the end all of whatever may appear will be nothing but delusion. 44 Appearance as such is therefore finally to be transcended upon the attainment of Buddhahood,45 and this relinquishment of appearance (snan ba) is moreover only possible as the result of gradual and sustained practice on the Madhyamaka path. Hence it is, rMa bya explains, necessary to distinguish that which is to be negated by reasoning (rigs pa'i dgag bya) from what is to be eliminated by means of the path (lam gyi dgag bya).46 The former expresses itself as conceptual grasping (zen pa), the cause of suffering, which is to be overcome by the realization of the absence of nature (ran biin med pa, niJ:tsvabhiiva). The latter object of negation (dgag bya, prati/fedhya) endures on the other hand throughout the entire path of cultivation (bsgam lam, bhiivaniimiirga), and while not preventing the realization of emptiness this factor is actually necessary for the bodhisattva's accomplishment of virtues (snali bcas kyi dge ba)Y Constituted by mere imprints of ignorance (ma rig pa'i bag chags), it is this subtle obscuration that is responsible for all the bodhisattva's perceptions (snali ba). For rMa bya, we might then conclude, all appearance is necessarily empty, yet emptiness itself is no appearance. Now if he accepts such natural laws that govern the manifestation and disappearance of snan ba, we may wonder whether rMa
44 Note that rMa bya elsewhere (De nid snan ba 809-810) qualifies his account of empirical sunyata cognition with the words: "[Dharmatii] cannot be realized and cultivated by positive determination. When it is taught [that it can be so realized and cultivated] then that is in order to facilitate the beginner's partial relinquishment of afflictions." (de yons gcod d,u rtogs pa dan bsgom pa med do / de ni las dan po pa la non mons pa'i phyogs re span ba'i phyir bstan pa yin la!) 45 See, e.g., De nid snail ba 762-763. 46 Also in his treatment of the "object of negation" (dgag bya) rMa bya may be providing an important conceptual framework for Tson kha pa's later discussion. See, e.g., De nid snail ba 762: dgag bya dilos ma bzuil bar ni dgag par mi nus te ... 47 De nid snail ba 763 and 818.

In pursuit of transparent means of knowledge

439

bya, despite all efforts to the contrary, has not himself in the end been seduced by realist intuitions. If, in short, appearance is one thing and emptiness another (as it would seem given the assertion that the former must fall away upon the completion of the path) it is hard to think of a reason why it should not be granted that conventionally appearances are objectively characterized in terms of being delusion, obscuration, and the like. If the Svatantrika strays from the Great Madhyamaka by asserting a sense of objectivity that is merely conventional, would not this idea of appearance as a negandum, invariably linked with a metaphysical force termed "delusion," just as much be a betrayal of Madhyamaka antirealism? Likewise, while rMa bya elsewhere forcefully describes the ruinous consequences of a reification of emptiness, we may wonder whether the utterly transcendent version of sunyatii that has now emerged in his account would not itself fall under the category of such debilitating constructs. Taking the general thrust of his argumentation into account, I believe that if faced with such concerns rMa bya could respond by reiterating his earlier statement that, "the appearance itself is neither existent nor nonexistent, neither true nor false."48 The convention "the ultimate truth" is thus, he might remind us, still simply applied in consideration of the way that the present appearance defies all extremes of mental construction (spros pa, prapaiica). Seen in this light, the account of appearance as a negandum is then simply to be taken as a description of the way this realization of freedom from constructs (spros bral, ni~prapaiica) may occur to the mind, without thereby becoming entangled in any metaphysical claims about the Being of appearance, the Emptiness of that, and a relationship between the two. Conventional and ultimate truths are mutually dependent notions that ensue in relation to simply that which appears Ui ltar snail ba tsam). With this, rMa bya might argue, all discussion ends, and yet it is also here that we find our point of departure. Given the increasing availability of texts, we have now sufficient sources to conclude that rMa bya Byan chub brtson 'grus developed 48 See ll. 15.

440

Thomas Doctor

a complex, comprehensive, and original Madhyamaka interpretation. In this little essay I have sought to highlight some of its central features. rMa bya's Indian sources, his debates with contemporaries, and his influence on posterity are certainly all promising issues for Tibetan Madhyamaka research, but for practical reasons I have not been able to consider any of these here. If, as we have reason to propose, rMa bya was the first to draw a distinction between those Madhyamakas who accept conventional "particular characteristics" (ran mtshan, svalak~a1:za) and those who do not, it goes without saying that his influence on Tibetan Madhyamaka debates remains perceptible up to the present day. rMa bya's efforts to establish and develop a Jig rten la grags pa'i tshad rna, a pramiilJa system that can find its place within Candraklrtian Madhyamaka, are another important precursor to the enterprises of Tsoil kha pa et al. Given the recently available texts, I believe that we may soon hear more on rMa bya.

Bibliography
CABEZON, JOSE 2003. "Two Views on the Svatantrika-Prasarigika Distinction in Fourteenth-Century Tibet." In DREYFUS and MCCLINTOCK 2003: 289315. DREYFUS, GEORGES 2004. Recognizing Reality - Dharmakfrti's Philosophy and Its Tibetan Interpretations. New York: SUNY Publications. DREYFUS, GEORGES and MCCLINTOCK, SARA (eds.) 2003. The SvatantrikaPrasangika Distinction - What Difference Does a Difference Make? Boston: Wisdom Publications. DUNNE, JOHN D. 2004. Foundations of Dharmakfrti's Philosophy. Boston: Wisdom Publications. RMA BYA BYAN CHUB BRTSON 'GRUS 1975. dBu ma rtsa ba ses rab kyi 'grel pa 'thad pa'i rgyan. Rumteg: rGyal ba Karma pa. -2006a. dBu ma rigs pa'i tshogs kyi rgyan de kho na nyid snan ba'i rtsa ba. In DPAL BRTSEGS BOD YIG DPE RNIN ZIB 'JUG KHAN 2006, vol. 13: 745-748. -2006b. dBu ma rigs pa'i tshogs kyi rgyan de kho na nyid snan ba. In DPAL BRTSEGS BOD YIG DPE RNrn ZIB 'JUG KHAN 2006, vol. 13: 754-820. -2011 (forthcoming). Ornament of Reason - The Great Commentary to Nagarjuna's Root of the Middle Way. Trans. DHARMACHAKRA TRANSLATION COMMITTEE. New York: Snow Lion.

In pursuit of transparent means of knowledge

441

PA TSHAB NI MA GRAGS PA 2006. dBu ma rtsa ba'i 'grel pa. In DPAL BRTSEGS BOD YIG DPE RNIN :lIB 'JUG KHAN 2006, vol. 10:29-203. DPAL BRTSEGS BOD YIG DPE RNIN :lIB 'JUG KHAN 2006. bKa' gdams gsun 'bum phyogs bsgrigs. Chengdu: Si khron dpe skrun tshogs pa, Si khron mi rigs dpe skrun khan. RUEGG, DAVID SEYFORT 2000. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy. Vienna: Arbeitskreis fUr Tibetische und Buddhistische Studien, Universitiit Wien. SIDERITS, MARK 2003. Personal Identity and Buddhist Philosophy. Hampshire and Burlington: Ashgate Publishing. TAUSCHER, HELMUT (ed.) 1999. Phya pa chos kyi sen ge: dBu ma sar gsum gyi ston thun. Vienna: Arbeitskreis fUr Tibetische und Buddhistische Studien, Universitiit Wien. VAN FRAASSEN, BAS C. 2002. The Empirical Stance. New Haven and London: Yale University Press. VOSE, KEVIN A. 2005. The Birth of Prasangika: A Buddhist movement in India and Tibet. Ph.D. thesis, University of Virginia. Charlottesville: University of Virginia. WILLIAMS, PAUL 1984. "Review Article (on CHRISTIAN LINDTNER 1982. Nagarjuniana: Studies in the Writings and Philosophy of Nagarjuna, Indiske Studier 4, Copenhagen: Akademisk Forlag)." Journal of Indian Philosophy 12: 73-104. -1985. "Rma bya pa Byan chub brtson 'grus on Madhyamaka method." Journal of Indian Philosophy 13: 205-225. YOSHIMIZU, CHIZUKO 1993. "The Madhyamaka Theories Regarded as False by the dGe lugs pas." Wiener Zeitschrift for die Kunde Siidasiens und Archiv fur Indische Philosophie 37: 201-227.

Zan Than sag pa on theses (dam bca', pratijiiii) in Madhyamaka thought*

Chizuko Yoshimizu

Zan Than sag pa 'Byun gnas ye ses (alias Yes ses 'byun gnas, active in the first half of the 12th century)l is reported to have been one of the chief disciples of Pa tshab Ni rna grags (lOSS-ca. 1145), who translated Candrakirti's main treatises into the Tibetan language. 2 One composition by Zan Than sag pa is now available to us in the form of a handwritten manuscript: a complete commentary on Candraki"rti's Prasannapada (Pr) entitled dBu rna tshig gsal gyi
* My special thanks are due to Kevin Vose, who corrected the English of the present paper, and Pascale Rugon for their valuable suggestions and comments. 1 Zan Than sag pa is supposed to have belonged to the Zan clan and resided in Than sag monastery, which he himself built in 'Phan yul, north of IRa sa, where Pa tshab was born and based after his return to Tibet. Presumably Zan Than sag pa learned from Pa tshab after the latter had returned to Tibet from Kashmir around 1101 (cf. Yoshimizu 2005: 128 n. 2). According to Chos 'phel (2004: 166f.), the Than sag monastery was converted from bKa' gdams pa to dGe lugs pa at the time of the 5th Dalai Lama. It is therefore to be identified with the current dGe lugs temple dGa' ldan chos 'khor dgon, the foundation of which Ses gfien tshul khrims (2001: 203) ascribes to Zan Ye ses 'byun gnas in the 13th century. I am indebted to Maho Iuchi for the information about Than sag monastery. Iuchi (2007: 62) presents the list of the bKa' gdams pa temples surrounding IRa sa. 2 Pa tshab translated into Tibetan the Prasannapada, the Madhyamakavatara and its Bha~ya as well as the CatuJ:zsatakatfka. In some biographical literature, Pa tshab is also given the clan name Zan. Cf. Seyfort Ruegg 2000: 45 n. 89.

Journal of the International Association of Buddhist Studies

Volume 32 Number 1-2 2009 (2010) pp. 443-467

444

Chizuko Yoshimizu

ti ka. This manuscript is definitely a discovery of great importance for the study of Tibetan scholasticism in the 11-12th centuries. 3 As I have previously discussed, Zan Than sag pa held the Madhyamaka theory of "freedom from extremes" (mtha' bral dbu ma) in terms of "neither existence nor non-existence" (yod min med min), which is the exact view that the later Sa skya scholar Go rams pa bSod nams sen ge (1429-1507) ascribed to him.4 With regard to Candraklrti's defense (in the first chapter of the Prasannapada) of Buddhapalita's statements, against Bhaviveka's attacks on them,S unlike later Tibetans Zan Than sag pa never expresses the idea that the Madhyamaka school thereby divided into two branches, the ThaI 'gyur ba (*Prasangika) and the Ran rgyud pa (*Svatantrika). He is rather of the opinion that a ran rgyud pa, such as Bhaviveka, who uses an autonomous inference (ran rgyud rjes dpag, svatantranumana) is unqualified to claim to be a dbu ma pa 6 and that between a ran rgyud pa and a dbu ma pa there is no common establishment (mthun snan = ubhayasiddha) of the subject of debate.? Following Candraklrti, Zan Than sag pa does
3 For detailed features of the manuscript, the authorship, as well as the historical figure of Zan Than sag pa, cf. Yoshimizu 2005 and 2006, which include an edition of the 18th chapter and of some parts of the first chapter. This manuscript is not included in the recently published bKa' gdams gSUli 'bum. 4 ITa ba'i san 'byed 8a4-8b1 and 17b3f.; cited in Yoshimizu 2005: 130 n. 10. For Go rams pa's detailed discussion, cf. Matsumoto 1999: 205ff. and Cabez6n and Dargyay 2007: 203ff. S For a detailed investigation of Zan Than sag pa's interpretation of Buddhapalita's statements, Bhaviveka's criticism thereof, as well as Candrakirti's rejoinder focusing on the reading of Pr 18,5-19,7, see Yoshi~izu 2006: 87ff. (English summary on 114). 6 See dBu ma tshig gsal gyi ti ka 21bl-4 edited in Yoshimizu 2006: 102f. (Text 1) and translated in ibid.: 81. Zan Than sag pa does not use the appellation ThaI 'gyur ba (*Prasangika) as a branch name referring to the Buddhapalita-Candraklrti lineage, although he calls them the advocates of prasmiga reasoning (thai 'gyur smra ba) (cf., e.g., 6b5 cited below in n. 8). ? Therefore, in Zan Than sag pa's view, dbu ma ran rgyud pa is not

Zan Than sag pa on theses

445

not accept an autonomous inference as a means of e~tab1ishing ultimate reality, 8 since, as taught by Nagarjuna and Aryadeva, 9 the MMhyamika(dbu rna pa) disowns any doctrinal position (phyogs, pak~a) or thesis (dam bea', pratijiia) to be proven on his own account. Thus, on one hand strictly rejecting formal probative reasoning, Zan Than sag pa on the other hand adopts in his own discussion plenty of logical terms such as reason (he du, hetu), pervasion (khyab pa, vyapti), argument ('thad pa, upapatti) and the like. 10 The flourishing of pramalJa studies in his time well accounts
established. See dBu ma tshig gsal gyi ti ka 21bl-4 edited in Yoshimizu 2006: 102 (Text 1) and translated in ibid.: 81. 8 Zan Than sag pa states that the Madhyamika may use an autonomous inference if not investigating ultimate reality. See dBu ma tshig gsal gyi ti ka 6b5 (cited and translated in Yoshimizu 2005: 132): don dam spyod (read: dpyod) pa'i skabs min pas I (sic) rali rgyud byas kyali 'gal ba med II de phyir thai 'gyur smra ba la II 'gal ba'i fies pa mi bsam mo II Go rams pa quotes the first half of this verse in his dBu ma'i spyi ston (105alf.) as a statement by Zan Than sag pa without specifying the source (see Yoshimizu 2005: 132). For citations from the dBu ma tshig gsal gyi ti ka, I retain orthographic peculiarities of the manuscript, which I have listed in Yoshimizu 2005: 138. 9 Cf. Vigrahavyavartanf (VV) 29-30 cited in Pr D6a3f., P6b5f.: gal te lias dam beas 'ga' yod II des na lia la skyon de yod II lia la dam bca' med pas na II lia la skyon med kho na yin II gal te mlion sum la sogs pa'i II don gyis 'ga' zig dmigs na ni II sgrub pa 'am (D 'ali) bzlog par bya na de II med phyir lia la klan ka med II = L. de La Vallee Poussin ed. (LVP) 16,710: yadi kacana pratijfia syiin me tata eva (VV Johnston & Kunst 1978 reads e~a) me bhaved do~al,L I nasti ca mama pratijfia tasman naivasti me do~al,L II yadi kiY(!cid upalabheyaf!! pravartayeyaY(! nivartayeyaY(! va I pratyak~adibhir arthais tadabhiivan me 'nupiilambhaJ:t II *Lak~a1Jatfka (L'O edited by Yonezawa 2004: 132; Catul,Lsataka (CS) 6.25 cited in Pr D6a2f., P6b4f.: yod dali med dali yod med ces II phyogs ni gali la 'ali yod min pa II de la yun ni rin po na'ali II klan ka brjod par nus ma yin II = LVP 16,4f.: sadasatsadasac ceti yasya pak~o na vidyate I upiilambhas cire1Japi tasya vaktuY(! na sakyate II Cf. further 'the investigation of these verses in Seyfort Ruegg 2000: 115ff. 10 For instance, Zan Than sag pa terms the three verses cited above (see

446

Chizuko Yoshimizu

for his broad knowledge of the Buddhist logico-epistemological system.l1 To a certain extent, indeed, Zan Than sag pa admits Madhyamikas' commitments to logical argumentation. It is wellknown that Candrakirti has called "theses" (pratijiia) Nagarjuna's negative statements in Miilamadhyamakakarika (MMK) 1.1,12 Zan Than sag pa also calls "theses" (dam bca') Buddhapalita's commentarial statements on MMK 1.1 as will be seen below, as well as several of Candrakirti's statements including Madhyamakavatara (MA) 6.8cd, which is cited in the Prasannapada.13 How should one
n. 9), viz., CS 6.2S and VV 29-30 "three arguments for the Madhyamika's lacking probandum and argument" (dBu ma tshig gsal gyi ti ka lSa3: dbu ma pa la sgrub bya dan 'thad pa med pa'i 'thad pa gsum). Rejecting the MMhyamika's use of probandum and argument, he thus gives arguments for the rejection. 11 It is, however, unknown whether Zan Than sag pa received education at gSan phu sNe'u thog monastery, which was the centre of pramalJ-a studies at that time. We cannot exclude the possibility that he learned logic from Pa tshab and Kanakavarman, for both of them supposedly inherited the tradition of Buddhist logic that flourished in Kashmir. Moreover, Pa tshab's first collaborator Mahasumati is described as a great logician in the colophon of the Prasannapada (see Yoshimizu 200S: 133 n. 19). 12 Madhyamakavatarabha~ya (MABh) 81, 17ff.: de biin du dam bea' bag sum ehar la yan sbyar bar bya'o II dam bea' ba bii po de rjes su brjod nas rigs pas sgrub pa'i phyir Mad pa I Pr LVP 13,2f.: tatas eaivarrt sarrtbandhalJ, naiva svata utpanna jatu vidyante bhavalJ kva eana ke eana I evarrt pratijtiatrayam api yojyarrt II Cf. MMK 1.1 (cited in Pr LVP 12,13f.): na svato napi parato na dvabhyarrt napy ahetutalJ I utpanna jatu vidyante bhavalJ kvaeana keeana II Cf. also dBu ma tshig gsal gyi ti ka lOaIf.: de bii ear las skye bar mi 'thad do stiam du nes pa ni 'phags pa klu grub la mna' nas dam bea' mdzad do II 13 dBu ma tshig gsal gyi ti ka llbSf.: skye bar gyur pa slar yan skye bar rigs pa 'an ma yin tiid (= MA 6.8cd) ees pa ste dam bea' tsam mo 1/ Cf. MA 6.8 (cited in Pr LVP 13,7-8): tasmad dhi tasya bhavane na gUlJ-o 'sti kaseij I jatasya janma punar eva ea naiva yuktam II Cf. also dBu ma tshig gsal gyi ti ka 14aS, where Zan Than sag pa terms dam bea' Candraklrti's refutation of Bhaviveka's assertion: tshig de dag gis ni ties pa brjod par mi rigs ees dam bea' bstan te I = "[Candraklrti] presents a thesis by saying '[it is] not proper' [that Bhaviveka] has indicated failures by those words

Zan Than sag pa on theses

447

distinguish these "theses" which Indian Madhyamaka masters are supposed to have advocated from those theses which they all have rejected? In the present paper, I would like to clarify how Zan Than sag pa confronted this most controversial problem in the history of the Indo-Tibetan Madhyamaka tradition, which his contemporary as well as later Tibetan scholars were also destined to encounter.14 I will first examine Zan Than sag pa's arguments for distinguishing between acceptable theses and unacceptable theses for Madhyamikas. The focus will be on his interpretation of the kind of thesis that is grounded in the logical system of Dignaga and Dharmaklrti. Then, I will analyze his ontology-based defense of the theory that for the Madhyamika negation is not what is to be proven (sgrub bya, siidhya), confirming an essential link between this theory and the theory of neither existence nor non-existence (yod min med min), according to which the negation should eventually be negated as well. Our final aim is to gain a clearer perspective of the historical development of Tibetan Madhyamaka, which has proceeded in a close relation to the Buddhist logico-epistemological tradition. Zan Than sag pa's arguments for and against the Madhyamika's use of theses Insofar as the statements of MMK 1.1 are concerned, the fact that Candrakirti has referred to them as theses (pratijfiii) requires an explanation from later interpreters, for it is obviously contradictory to his own statement that Madhyamikas have no thesis of their own. is

[in Buddhapalita's commentarial statements on MMK 1.1]." 14 Seyfort Ruegg (2000: Section II, especially 115ff. and 219ff.) has provided a detailed investigation of this problem in the Indo-Tibetan Madhyamaka tradition. is See Pr LVP 23,3: niismiikar[L svapratijiiiiyii abhiiviit / Accepting the ambiguity of Candraklrti's usage of the word pratijiiii, Matsumoto (1997: 372f., 383) has inferred that CandrakIrti just followed Bhaviveka in calling MMK 1.1 theses while insisting that the Madhyamika has no the-

448

Chizuko Yoshimizu

Zan Than sag pa solves this problem by explaining that Candrakirti applies the word pratijfiii or dam bca' in "mere transactional usage" (tha sfiad tsam). Zan Than sag pa also describes such a thesis as "mere thesis" (dam bca' tsam) or "mere name" (mili tsam) and opposes it to a "genuine thesis" (dam bca' mal ma),16 notably to the thesis or the position (pak~a)17 defined by Dignaga. Zan Than sag pa says as follows:
[1] [Each statement in MMK 1.1 is] thus called a root thesis (rtsa ba'i dam bea') but [this is] mere transactional usage (tha sfiad tsam), for [this thesis] is not a [thesis] like that which has five characteristics (mtshan fiid Ina), namely: 1) [stated in its own] form ([ran gi] no bo, [sva]rupa), 2) alone (kho na, eva), 3) being intended ('dod pa, i~ta), 4) [by him]self (bdag fiid, svayam), 5) [and] unopposed (ma bsal ba, anirakrta), [which are regarded as] the defining characteristics of the thesis (dam bca'i mtshan fiid).18 sis of his own (Le., according to Matsumoto, svatantra pratijfia). Seyfort Ruegg (2000: 129f.) has proposed to distinguish between a pratijfia as a philosophical statement or thesis by Madhyamikas that does not posit any substantial self-existence and a thesis that posits a substantial selfexistence, which Madhyamikas reject. 16 Cf. dBu ma tshig gsal gyi ti ka llb5f. cited above in n. 13, which comments on MA 6.8, and lObif. cited below in n. 34. 17 The words pratijfia (dam bea'), pak~a (phyogs) and siidhya (sgrub bya) are generally used as equivalents in both Indian and Tibetan logical traditions. Dharmakirti employs in his verse commentary on PS 3.2ab the word sadhya for pak~a, which is to bear the fourfold characteristic of the thesis (PV 4.28, cited and translated in Tillemans 2000: 48: gamyarthatve 'pi sadhyokter asa"f!lmohaya lak~wzam I tae caturlak~w:za"f!l rupanipate~tasvaya"f!lpadaiJ:z II). Tillemans (2000: 4, n. 16) indicates that Dignaga uses the terms anumeyanirdesa,pak~avacana and sadhyanirdesa as synonyms. Those equivalent terms, pratijfia, pak~a and sadhya, are employed as such by Madhyamikas too (cf. Seyfort Ruegg 2000: 13 Of.). Zan Than sag pa himself admits that they are equivalents (dBu ma tshig gsal gyi ti ka 13a3f.: dam bca' dan sgrub bya dan phyogs ni rnam grans so). 18 dBu ma tshig gsal gyi ti ka lObI: de ltar na rtsa ba'i dam bca' zes pa yan tha sfiad tsam yin te II dam bca'i mtshan fiid no bo kho na 'dod pa bdag fiid ma {bsal} (Ms. gsal) ba ees pa mtshan fiid Ina ldan Ita bu ma

Zan Than sag pa on thes.;:s

449

The five defining characteristics of the thesis enumerated here can be identified with those which Dignaga stated in PramiifJasamuccaya CPS) 3.2 as pak!falak!fafJa. Dharmaklrti, let us note, only considered the first four as such. 19 Let us see PS 3.2: . [PS 3.2] [A valid thesis] is one which is intended (i~ta) by [the proponent] himself (svayam) as something to be stated in its proper form alone (svarupelJaiva) [i.e., as a sadhya]; [and] with regard to [the proponent's] own subject, it is not opposed (anirakrta) by perceptible objects, by inference, by authorities or by what is commonly recognized. 20 Now it is clear that Zan Than sag pa differentiates the "genuine thesis" fulfilling the Buddhist logicians' definition from the set of theses acknowledged by CandrakIrti as "mere transactional usage." Accordingly, one could offer as a hypothesis that the Madhyamika may advocate a thesis if it is not intended as a

yin no II
19 See PV 4.28-29 (cited and translated in Ono 1986: 849 and Tillemans 2000: 48f.) and Tillemans ibid.: 49. The enumeration of these five characteristics appears in PV 4.85 (ibid.: 116: svayarrznipatarupakhya vyatirekasya badhikaf:i, I sahanirakrtene~tasrutir avyiiptibadhanf II), where Dharmaklrti explains that Dignaga's definition of the thesis serves to eliminate the faults of over- and non-pervasion (ativyapti, avyiipti). 20 PS 3.2: svarupelJaiva nirdeiyaf:i, svayam i~to 'nirakrtaf:i, I pratyak~a rthanumiiniiptaprasiddhena svadharmilJi II The English translation and the Sanskrit reconstruction follow Tillemans 2000: 47. The Tibetan version runs (Tillemans 2000: 47 n. 166, cf. Kitagawa 1973: 471f.): ran gi no

bo kho nar bstan II bdag 'dod ran gi chos can la II mnon sum don dan rjes dpag dan II yid ches grags pas ma bsal ba'o II Cf. also Nyayabindu (NB) 3.38: svarupelJ-aiva svayam i~to 'nirakrtaf:i, pak~a iti II NM 1: svayarrz sadhyatvenepsitaf:i, pak~o viruddharthanirakrtaf:i, I cited in PramalJavarttikavrtti (PVV) 320,16 [443,3:ff.]; PramalJavarttikabha~ya (PVBh)

51O,23f.; and Tillemans 2000: 117f. Dignaga refutes in PS 3.3 and 3.5 respectively the Nyayasutra's definition of the thesis as a presentation of the probandum (siidhyanirdeia) and that in Vasubandhu's Vadavidhi as a statement of the probandum (sadhyabhidhana). See Tillemans ibid.: 39 n.145.

450

Chizuko Yoshimizu

probandum (siidhya) by the Madhyamika himselF1 This distinction between the Madhyamika's use of theses and that advocated by Buddhist logicians is supported by the thesis-definitions of Dignaga and Dharmaklrti in Nyiiyamukha (NM) 1 and PramiilJaviirttika (PV) 4.86ab, which say respectively: "The thesis (or the position) is what is intended by [the proponent] himself as the probandum" (svayaf!l siidhyatvenepsitaJ:!. pak:jo ... ) and "What is accepted as the probandum is the defining characteristic of the thesis" (siidhyiibhyupagamaJ:!. pak:jalak:jalJaf!l).22 In light of those logicians' definitions, the Madhyamika's principle, in turn, would clearly emerge in contrast, that the Madhyamika or the follower of the middle way is one who neither intends to prove something nor accepts what is to be proven (siidhya) as his own. 23 The Madhyamika's rejection of autonomous inference (svatantriinumiina) also results from his non-acceptance (anabhyupagama) of siidhya, as Candraklrti explains in his Prasannapadii. 24 It
21 As for the fifth characteristic, "unopposed by perceptible objects, etc.," it is unclear whether the Madhyamika is totally free from this condition. However, the fact that Candraldrti did not ultimately accept Dignaga's theory of valid cognition (pramal}a) (cf. Yoshimizu 1996) may allow us to conjecture that this condition is acceptable for Madhyamikas only in the case of investigating common sense objects. 22 Cf. n. 20 above and Tillemans 2000: 117f. 23 In this regard, Zan Than sag pa expressly defines the Madhyamika as the one who has no doctrinal position (phyogs med pa) of his own and hence excludes Bhaviveka, who sets forth a probandum or a doctrinal position (sgrub bya 'am phyogs), from the lineage of Nagarjuna. Cf. dBu ma tshig gsal gyi ti ka 14b8: dbu ma pa ces pa ni phyogs med pa la zer la legs ldan khyod ran rgyud byed na ni dbu ma pa ma yin tin klu'i rjes su mi 'bran pa zig ste I sgrub bya 'am phyogs 'dod pa'i phyir TO /1 24 Pr LVP 16,2: na ca madhyamikasya svataJ:z svatantram anumana1!l kartu1!l yukta1!l pak:;antarabhyupagamabhavat I (The Tibetan translation D6a2, P6b3f., omits svataJ:z) = "For the Madhyamika, it is not appropriate to formulate an autonomous inference on his own account, because [for him] there is no acceptance of the positions alternative [to the position that things arise from themselves, for instance, the position that they arise from an other]." Although Candraklrti is here rejecting the four

Zan Than sag pa on these~

451

would therefore be natural that whoever has no intention to prove something positively neither states a thesis on his own account nor formulates an inference independently of his opponent's assertion. ''Autonomous'' (svatantra) can be interpreted as synonymous with "intended by the proponent himself" (svayam i.JtaJ:t) in terms of Dignaga, even if it is unclear whether Candrakirti knew Dignaga's thesis-definition. Zan Than sag pa makes this point clear in his own definition of autonomous inference as well as in his commentary on Candraklrti's refutation of autonomous inference:
[Definition:] Autonomous inference is the proof of a probandum (sgrub bya, sadhya) that has a [genuine] characteristic (mtshan fiid dan ldan pa, *lak.JWlavatllak.JWlika) by means of [a logical reason fulfilling] the three conditions (tshul gsum, trirupa[linga]) that are established for both proponent and opponent by a valid means of cognition (tshad ma, pramaIJ-a).25 [2] The thought [expressed by Candraklrti] here [in Pr LVP 16,2] is [the following]: The formulation of an autonomous logical reason (ran rgyud kyi he tu) [i.e., an autonomous reasoning or inference] entails (khyab) that the probandum (sgrub bya, sadhya) has a [genuine] characteristic (mtshan fiid dan ldan pa). It is not appropriate for the Madhyamika to state an autonomous logical reason [i.e., an autonomous reasoning] because for him there is no probandum, which would be entailed (khyab byed du gyur pa) [by an autonomous logical reason].26

alternative positions of the tetralemma (catu.Jkoti, mu bzi) enumerated in MMK 1.1, he has undoubtedly extended the scope to all possible doctrinal positions. Pr LVP 18,5f. cited below in n. 27 is also often referred to as Candraklrti's refutation of autonomous inference. For his criticism of autonomous inference, cf. further, e.g., Yotsuya 1999: 47ff., Seyfort Ruegg 2000: 129ff. and Yoshimizu 2003: 269ff. 25 dBu ma tshig gsal gyi ti ka 14b8f. edited in Yo shimizu 2006: 104 (Text 2): ran rgyud ni sgrub bya mtshan fiid dan ldan pa la tshul gsum rgol phyir rgol gfiis ka'i tshad mas grub pa zig gis sgrub pa'o II 26 dBu ma tshig gsal gyi ti ka 15alf.: de'i bsam pa ni ran rgyud kyi he du byed pa la sgrub bya mtshan fiid dan ldan pas khyab la I dbu ma pa la khyab byed du gyur pa'i sgrub bya med pa'i phyir ran rgyud kyi he du

452

Chizuko Yoshimizu

While Zan Than sag pa does not say what he has in mind by the expression "a probandum that has a characteristic" (sgrub bya mtshan fiid dan ldan pa), the most plausible reading is to take it as a probandum that has the fivefold characteristic of a genuine thesis in accordance with the aforementioned logicians' thesis-definitions. In short, it is a probandum "intended by the proponent himself" (i.e., svayam i~ta). In this manner, Zan Than sag pa rules out both the genuine theses of the Buddhist logicians and autonomous inferences from the Madhyamika's logical scene. Consequently, whatever inferential statement Madhyamaka masters have properly made must be either a prasanga reasoning or an other-acknowledged inference (paraprasiddhiinumiina, gian grags rjes dpag), the probandum of which is not intended by the Madhyamika himself. Zan Than sag pa actually reads Buddhapalita's commentarial statement on MMK 1.1 twofold as a prasanga as well as an other-acknowledged inference. 27 What is interesting for our discussion is that in both interpret abrjod par mi rigs so II 27 In this respect, one should note that Zan Than sag pa interprets Pr LVP 18,5-19,7, which has recently aroused discussion among modern scholars, to deal with the question of whether the MMhyamika should state a logical reason and examples of an other-acknowledged inference. This interpretation consents with that proposed by MacDonald 2003: 167f. According to Zan Than sag pa, Candraklrti there eliminates the need for such an inference, but then in Pr LVP 19,8-21,7, Candraklrti shows that Buddhapalita has presented an other-acknowledged inference. See dBu ma tshig gsal gyi ti ka 16b3-17b2 cited and translated in Yo shimizu 2006: 95ff., Text 7 [109-112] and 114 ad Pr LVP 18,5ff.: athapi syiit I miidhyamikiiniilJ'l pak~ahetudr~tiintiiniim asiddheIJ svatantriinumiiniinabhidhiiyitviit svata utpattiprati~edhapratijiiiirthasiidhanalJ'l (according to MacDonald 2003: 167; LVP: -pratijiiiitiirthasiidhanalJ'l) mii bhad ubhayasiddhena viinumiinena parapratijiiiiniriikaralJalJ'l ... II (= D6blf., P7a3ff.) = "[It may be granted], since Madhyamikas do not state an autonomous inference owing to the non-establishment of the position, the logical reason, and the example, that there should be neither a proof of the content of the thesis (pratijiiiirthasiidhana) when negating the origination from self, nor a refutation of the opponent's thesis by means of an inference established for both [parties]. ..."

Zan Than sag pa on theses

453

tions he terms it a "thesis" (dam bca'). More precisely, he terms Buddhapalita's statement, "Things do not arise from themselves (na svata utpadyante bhavas .. .)," a "pseudo-thesis" (ltar snaJi dam bca', *pratijfiabhasa) when he interprets it as a prasaJiga reasoning,28 and when he interprets it as an other-acknowledged inference, he renders it as a thesis properly attended by a logical reason, exampIes and pervasion. 29 But yet, in his view, Buddhapalita's statement, 28 One should note that in the manuscript the expression "pseudo-thesis" (ltar snan dam bca') is a replacement for "negative thesis" (bkag pa'i dam bca'). It is clear that the letters bkag pa'i have been deleted and the letters ltar snan have been inserted instead. Either Zan Than sag pa himself or the scribe made this correction, which I think proper and necessary, because a prasanga reasoning is not a thesis that establishes negation. Moreover, as will be discussed, Zan Than sag pa's final position is that negation (bkag pa) is also to be negated by the Madhyamika. See dBu ma tshig gsal gyi ti ka 12alf.: 'gal brjod kyi tshul Mad pa I dnos po rnams bdag las skye ba med de I [-bkag pa'i- deleted] [ltar snan inserted] dam bca' ste I dnos po rnams bdag las skye ba bkag pa tsam dam bca'i tha siiad du byed pa'o II = "[The following is] the explanation of [Buddhapalita's] way of indicating contradiction [in the Sarp.khya theory of the origination from self]: [Buddhapalita says] 'Things do not arise from themselves.' [This is] a {pseudo} thesis. [Buddhapalita] made the mere negation of things' origination from self in a transactional usage of thesis." 29 dBu ma tshig gsal gyi ti ka 17b6f.: skyans kyi nag gis yan lag Ina pa'i sbyor ba bstan lugs Mad pa dnos po rnams bdag las skye ba med de ces pas dam bca' ste I dnos po rnams ni chos can no II grans can gyi lugs kyis chos can giiis te I bdag las gian pa'i dnos po rnams dan I mnon par gsal ba'i bum pa las gian pa'i dnos po rnams chos can gyi don to II bar 'dir yod pa'i phyir ces pa'i he du kha bskan no II dpe' bstan pa de dag gi skye ba don med pa iiid du 'gyur ba'i phyir dan ces pa'o II sin tu thaI bar 'gyur ba'i phyir dan ies pa yan dper bya'o II = "The explanation of the way in which a formal probative argument (prayoga) with five members is stated by Buddhapalita's words is [as follows]: a thesis [is given] by saying 'Things do not arise from themselves;' 'Things' are the subject (or "the property possessor;" chos can, dharmin). According to the Sarp.khya tenet, the subject is twofold: 'things other than self' and 'things other than a directly perceptible pot' are the subject matter. Between these [thesis and examples], the reason (hetu), 'because [they]

454

.Chizuko Yoshimizu

whether it is read as a prasaJiga or a paraprasiddhtinumtina, is safe from being a genuine thesis in terms of Buddhist logicians, because it is not intended as what is to be proven by Buddhapalita himself. Moreover, all those Madhyamaka teachers, Nagarjuna, Buddhapalita and Candrakirti, have solely negated others' positions without intending to prove something positively as their own position. 30 Therefore, Zan Than sag pa maintains, their statements cannot be genuine theses but must be regarded as theses only in transactional usage. To these arguments for sanctioning the Madhyamikas' use of theses, however, the following objection may well arise: The negation itself can be construed as a probandum that the Madhyamaka masters have intended to prove. By negating the origination from self, for instance, they would intend to establish the non-origination from self as their own thesis, even though they insist that it is mere transactional usage. 31 Zan Than sag pa indeed deals with this

exist,' is to be added. [Buddhapalita] shows an example (dpe, d!~tiinta) by saying 'because their origination would be pointless and ... .' [His subsequent statement] 'because [it] would be over-extended and' is also stated as an example." Also, ibid.: 18a3: dnos po rnams bdag las skye ba med de dam bea' / yod pa'i phyir he du / khyab pa ni dpe' dan geig / dam bea' slar brjod 'jug sdud do / "Thesis: Things do not arise from themselves. Reason: Because they exist. The pervasion is the same as the examples [show]. The conclusion is the restatement of the thesis." As for the details of Zan Than sag pa's interpretation of the pointlessness and the over-extension of the origination from self, given in ablative forms (tadutpiidavaiyiirthyiit atiprasangado~iie ea) as the example or the proof of pervasion, see Yoshimizu 2006: 89-94 and Yoshimizu 2008. 30 Cf. Pr LVP 34,5: parapratijfiiini~edhaphalatviidasmadanumiiniiniim/ 31 The possibility should also be precluded that the negation here in question might be an implicative type of negation (i.e., paryudiisa, ma yin dgag) that affirms the contrary position, that things arise from others. Both Bhaviveka and Candraklrti take the negations in MMK 1.1 as non-implicative negations (i.e., prasajyaprati~edha, med dgag). See Prajfiiipradfpa (Prajp) D48b6f., P58a6f.: bdag las ma yin ies bya ba'i dgag pa 'di ni med par dgag pa'i don du Ita bar bya ste / dgag pa gtso ehe ba'i phyir dan / 'di ltar rtogs pa (P rtog pa) ma Ius pa'i dran dgag pas

Zan Than sag pa on theses

455

objection by introducing mainly ontology-based arguments. That will be our next subject of consideration.

Zan Than sag pa's defense and his final position that the Madhyamika does not even accept negation (bkag pa) as a probandum
A crucial point is how to interpret the value of the negation of origination, which brings a logical and doctrinal determination for the Madhyamika himself. Zan Than sag pa first discusses the issue just after the aforementioned statement [1]. 32 Let us see his argument, which is based on an ontological observation:
[3.1] [Objection:] Granted that [your] intended thesis ('dod dam bca') is mere name (min tsam), it is established with regard to entities (dnos po) as a genuine thesis (dam bca' mal ma), for non-origination is established by [your] negating origination, because [these] two, i.e., origination and non-origination, are [respectively] positive determination (yons gcod) and negative determination (rnam gcod), and because according to the principle of direct contradiction (dnos 'gal) the negation of one results in the establishment of the other. 33 Thus, the thesis (dam bca') that there is no origination is indirectly established.
rnam par mi rtog pa'i ye ses ses bya'i yul ma Ius pa dan ldan pa 'grub par dgons pa'i phyir ro II and Pr LVP 13,4ff.: nanu ca, naiva svata utpanna, ity avadharyamalJe parata utpanna ity ani~taJ?1 prapnoti I na prapnoti, prasajyaprati~edhasya vivak~itatvtit parato 'py utpadasya prati~etsya manatvat I 32 See statement [1] cited above in n. 18 from dBu ma tshig gsal gyi ti ka lObI. He further develops it after statement [2], cited above in n. 26 from dBu ma tshig gsal gyi ti ka 15alf., as will be discussed below. 33 The principle of direct contradiction conforms with the principle of the excluded middle, which holds in any logical discussion. The Madhyamaka masters, Santarak~ita and KamalaSIla, also apply it to their argument for negating self-nature. Cf., e.g., MadhyamakalaJ?1karavrtti (MAlv) ad MadhyamakalaJ?1kara (MAl) 1 D56b7f. (Ichigo 1985: 22): ran biin yod par gyur na ni gcig pa'am Gig sos las mi 'da'o II de dag ni phan tshun spans te gnas pa'i mtshan fiid yin pas ph un po gian sel bar byed do II Madhyamaktiloka (MA) D191a4f. (cited in TSOIi kha pa's rTsa se tik chen 25b4f.): phan tshun spans te gnas pa'i mtshan fiid kyi chos dag ni cig

456

Chizuko Yoshimizu

[Reply:] The error that [you] have stated does not exist, for, if we negated the origination that is unexamined [by an analytical reasoning] (ma brtags pa'i skye ba), non-origination would be established, but we negate the origination postulated [by the Sa:rpkhya] (btags pa'i skye ba), so that the non-origination that is unanalyzed [by a reasoning] (ma dpyad pa'i skye med) does not come to be established, since [these] two, viz., examined origination (brtags pa'i skye ba) and unexamined non-origination (ma brtags pa'i skye med), are not directly contradictory (dnos 'gal).34

Here Zan Than sag pa's wording is puzzling, but in effect, there is no substantial difference between the notions "unexamined" (rna brtags pa) and "unanalyzed" (rna dpyad pa), which both refer to the origination that is conventionally accepted by the world (i.e., kun rdzob, sa'!1Vrti), without being examined by a reasoning which analyzes ultimate existence (i.e., don dam, paramiirtha). 35 Since the
sos dgag pa I gian sgrub pa med na med pa yin pa'i phyir gfii ga ma yin pa'i phyogs su rtog pa yan rigs pa dan ldan pa ma yin no II 34 dBu ma tshig gsal gyi ti ka lOblff.: gal te 'dod dam bca' ces pa min tsam du 'dod kyan dnos po la dam bca' mal mar grub ste I skye ba bkag pas skye myed grub par 'gyur te I skye ba dan skye med gfiis rnam gcod yons gcod yin pa'i phyir dan I dnos 'gal gyi lugs kyis cig bkag pa cig gi sgrub byed du 'on pa'i phyir ro I de ltar na sugs la skye ba med pa ies bya ba'i dam bca' grub po ce na II brjod pa skyon de ni med de II kho bo cag gis ma brtags pa'i skye ba bkag na I skye med de grub par thaI ba bden mod kyi I kho bo cag ni btags pa'i skye ba 'gog pas ma dpyad pa'i skye myed grub par mi 'gyur te I brtags pa'i skye ba dan ma brtags pa'i skye med gfiis dnos 'gal ma yin pa'i phyir ro II 35 The establishment of things without examination or analysis could be adopted from the definition of the conventional (kun rdzob) by Santarak~ita in MAl 64-65 that the conventional is agreeable and acceptable,only as long as it is not examined (see Ichigo 1985: CXXV, tr. CXLII): ma brtags gcig pu fiams dga' iin II skye ba dan 'jig pa'i chos can pa II don byed pa dag nus rnams kyi II ran biin kun rdzob pa yin rtogs II brtags pa ma byas fiams dga' ba'an II bdag rgyu sna ma sna ma la II brten nas phyi ma phyi ma yi II 'bras bu de 'dra 'byun ba yin II Phya pa Chos kyi sen ge also makes use of the notions "unexamined" and "unanalyzed" in his Sar gsum ston thun. See, e.g., his definition of the conventional as that which is true in the perspective of non-analytical thinking (16,4f.:

Zail Than sag pa on theses

457

origination "postulated" (btags pa) by the Saq1khya opponent does not exist even conventionally, the negation of this kind of origination cannot result in establishing conventional non-origination from the Madhyamaka point of view. Then, Zan continues his discussion as follows:
[3.2] [Objection:] Although that failure does not exist, it still follows that the non-origination that is examined [by an analytical reasoning] (skye med brtags pa) is established by [your] negating the origination that is examined [by an analytical reasoning] (skye ba brtags pa). [Reply:] No, it doesn't. The establishment of imagined non-origination (skye med kun brtags) does not result in establishing a thesis (dam bca', pratijtijj), for none [of] the characteristics of the thesis (dam bca'i mtshan tiid) are observed, because the origination postulated [by the Sarp.khya] (btags pa'i skye ba) and the unexamined non-origination that is imagined (ma brtags pa'i skye med kun brtags) are nothing but names (min tiid). Or alternatively (rnam pa gcig du na),36 [according to the principle of the excluded middle one could posit that] nonorigination is established by virtue of negating origination. By this alone, however, no thesis comes to be established, for a thesis intends a state of affairs (don, artha) as something to be proven (sgrub bya, siidhya), and yet we do not even intend non-origination as something to be provenY
ma dpyad pa'i bsam nor bden pa kun rdzob kyi bden pa'i mtshan tiid do). In the next passage cited in the body of the present paper, however, Zan Than sag pa subsumes the unexamined conventional establishment of non-origination under the concept "imagined." I am deeply indebted to Kevin Vose for the interpretation of the notions "unexamined" (ma brtags pa) and "unanalyzed" (ma dpyad pa), and the information about Phya pa's use thereof. 36 The phrase rnam pa gcig du na can be identified with the phrase rnam pa gcig tu na, which is used in canonical texts as the translation of the Sanskrit phrase atha vii. In the manuscript of the dBu ma tshig gsal gyi ti ka, the letter du often appears in the place where according to Tibetan orthography the letter tu should appear. 37 dBu ma tshig gsal gyi ti ka lOb3f.: gal te ties pa de med mod kyi 'on kyan skye ba brtags pa bkag pas skye med brtags pa grub par thalIa ce na / ma yin te skye med kun brtags grub pas dam bca' grub par mi 'gyur te / btags pa'i skye ba dan ma brtags pa'i skye med kun brtags ni min i'lid

458

Chizuko Yoshimizu

On my reading, the opponent is presumably pointing out the possibility that the substantially real or ultimate non-origination that bears "examination" (brtags pa) might be established by negating the origination that is examined and purportedly established as real by the Sii:q1khya. In his reply, Zhan Than sag pa rejects this objection by defining the Madhyamika's establishment of non-origination as "imagined" (kun brtags, parikalpita). Thus eliminating the establishment of ultimate non-origination, he likewise characterizes the conventional non-origination that is unexamined as "imagined" (ma brtags pa'i skye med kun brtags).38 To sum up, Zan Than sag pa is arguing that the non-origination from self is, be it ultimate or conventional, merely imagined and constructed by name (min) or verbal transaction. 39 From the viewpoint of Candraklrti's Madhyamaka, the negation of origination from self gains neither ultimate nor conventional ontological establishment.
yin pa'i phyir dam bca'i mtshan fiid mi dmyigs so // rnam pa gcig du na skye ba bkag pas skye med grub kyan de tsam gyis dam bca' 'grub par mi 'gyur te // dam bca' ni don sgrub byar 'dod pa yin la / kho bo cag ni skye med sgrub byar yan mi 'dod pa'i phyir ro 1/ 38 I prefer leaving ma brtags pa'i skye med kun brtags as it appears in the manuscript and not emending it to brtags pa'i skye med kun brtags, since the latter sounds like a tautology. My solution, moreover, suggests that in Candraklrti's Madhyamaka system even conventional non-origination or the negation of origination is regarded as being imagined. 39 The underlying idea is, in my view, that the imagined characteristic (kun brtags pa'i mtshan fiid, parikalpitalak~W:ta) is non-substantial with regard to characteristics (lak~w:tanil:tsvabhavata) and therefore not substantially existent but merely postulated by names, as taught in Sayt'ldhinirmocanasiltra (SNSu) chapter 7 (Lamotte 1935: 67f., tr. 194): don dam yan dag 'phags de la chos rnams kyi mtshan fiid no bo fiid med pa fiid gan ie na / kun brtags pa'i mtshan fiid gan yin pa'o / de ci'i phyir ie na / 'di ltar de ni min dan brdas rnam par biag pa'i mtshan fiid yin gyi / ran gi mtshan fiid kyis rnam par gnas pa ni ma yin pas de'i phyir de ni mtshan fiid no bo fiid med pa fiid ces bya'o // Tson kha pa evidently shares this idea, for he forms the concept ran gi mtshan fiid kyis grub pa to describe substantial existence on the basis of the three kinds of nonsubstantiality taught in the Sayt'ldhinirmocanasiltra, as I have extensively discussed (cf., e.g., Yoshimizu 1993).

Zan Than sag pa on theses

459

For the purpose of ruling out the possibility that the Madhyamika could establish the non-origination from self as his own thesis, Zan Than sag pa adverts in the last portion of his reply again to the logicians' thesis-definition and indicates that a genuine thesis establishes a probandum that has substantial reality, saying, "for a thesis intends a state of affairs (don, artha) as something to be proven." The condition that both the reason (hetu) and the probandum (sadhya) in an inference must be a real state of affairs (artha), and not imagined, can be found in Dharmakirti's PV 4.13 on Dignaga's PS 3.1ab (pararthanumana1J1 tu svadr~tarthaprakasanam).40 Zan Than sag pa means that since non-origination is not such a real state of affairs nor is it substantially existent, it cannot be a genuine thesis. His argument ends with the emphasis that the Madhyamika does not even intend the negation of origination as a probandum. Interestingly, he confronts the same problem elsewhere, too, after the aforementioned statement [2].41 There he clarifies the value of the negation in Candrakirti's system in contrast with that in the system of the three masters from the East (sar gsum pa). Replying to the objection that his statement [2] involves acceptance (khas len, abhyupagama) of reason, pervasion and the like,42 Zan Than sag pa argues as follows:
40 PS 3.1 (Tillemans 2000: 9): pararthiinumana1J1 tu svadr~tartha prakasanam I tatranumeyanirdeo hetvarthavi~ayo mata/:t II = (Tillemans tr.) "An inference-for-others, however, elucidates the state of affairs which [the proponent] has understood himself. There, the presentation of the inferendum is held to have the goal of the reason as its object." PV 4.13 (Tillemans 2000: 24f.): tad arthagrahalJaYfl sabdakalpanaropitatmanam I aliligatvaprasiddhyartham arthiid arthaprasiddhita/:t II = (Tillemans tr.) "This word 'state of affairs' [in Dignaga's definition of an inference-forothers, i.e., svadr~tarthaprakasana] is designed to establish that things whose natures are verbally and conceptually superimposed are not [valid] reasons, for [one] state of affairs [viz., the sadhya] is established from [another] state of affairs [viz., the reason]." For further analysis of these verses, cf. Tillemans ibid. 41 See statement [2] cited above in n. 26 from dBu ma tshig gsal gyi ti ka 15alf. 42 dBu ma tshig gsal gyi ti ka 15a2: gal te dbu' ma pa khyed khas len

460

Chizuko Yo shimizu

[4] [Reply:] No, it doesn't. Whereas the three [texts of the masters from] the East [i.e., JiUinagarbha, Santarak~ita and KamalasIla] intend the negation (bkag pa zig) when they negate entities, having negated those entities which are [supposed to be] either one or many by [the argument] that they are neither one nor many (geig dandu bral, ekiinekaviyoga), this Madhyamika [i.e., Candraklrti] does not intend even such a thing as the negation. Because [he] intends that it is nothing (ci yan ma yin), he has no acceptance at all. Or alternatively [one could accept the negation but] by a mere acceptance [of the negation] there occurs no thesis. There would occur a thesis if one accepted [the negation] as what is to be proven (sgrub bya, siidhya), but there is no [such] failure because the Madhyamika does not accept [the negation] as what is to be proven. 43

Here Zan Than sag pa draws the same conclusion as that in the aforementioned reply [3.2], that there is no thesis for the Madhyamika because he does not accept any probandum, but from a slightly different observation. Describing Candraklrti's intentions as "it is nothing" (ci ymi ma yin), Zan Than sag pa seems to be resorting to the theory of "freedom from extremes" (mtha' bral) in terms of "neither existence nor non-existence" (yod min med min), acmed par 'dod kyan khas len dan bcas pa fiid de I ran rgyud kyi he du brjod pa la ran rgyud kyi sgrub bya mtshan fiid can yin pas khyab ees khyab bya khas blans so II sgrub bya khas blans pa med pa'i phyir ces he du khas blans so II de'i phyir khas len can du 'gyur ro ze na I = "[3] [Objection:] Although you Madhyamikas assert to have no acceptance, [your statements] definitely involve an acceptance, for you accept the pervasion in saying, 'the formulation of an autonomous logical reason [i.e., an autonomous reasoning or inference] entails that the probandum of the autonomous inference has a [genuine] characteristic.' [And] you accept the reason in saying, 'because for him the asserted probandum does not exist.'" 43 dBu ma tshig gsal gyi ti ka 15a2f.: ma yin te sar gsum pa gcig dan du bral gyis geig dan du ma'i dnos po bkag nas dnos po bkag pa'i bkag pa zig 'dod la I dbu ma pa 'di ni bkag pa de de lta bu yan mi 'dod de I ei yan ma yin par 'dod pas khas blans gan yan med do II rnam pa geig du na khas blans pa tsam gyis dam bear mi 'gyur te I sgrub byar khas blans na dam bear 'gro la I dbu' ma pas sgrub byar khas ma blans pa'i phyir fies pa med do II

Zan Than sag pa on theses

461

cording to which even the negation should be counted among what is to be abaridoned. 44 And the non-acceptance of the negation as a probandum "is Zan Than sag pa's final argument to ensure the Madhyamika's disowning of a thesis. He thereby completely differentiates Candraklrti's system from that of those who intend a negative determination, such as the non-existence of intrinsic nature (niJ:tsvabhiivatii), by means of an autonomous proof.45 In this regard, it is significant that Zan Than sag pa defines the nature of non-implicative negation (med dgag, prasajyaprati~edha) specific to the Madhyamaka system and sets it apart from the traditional definition thereof, which both Buddhist logicians and their followers, viz., Bhaviveka, Santaraksita and KamalasIla, are supposed to have acknowledged. 46 Zan Than sag pa states as follows: 44 As I have closely discussed, Zan Than sag pa demonstrates this theory in his commentary on the 18th chapter of the Prasannapada without mentioning the designations mtha' bral or yod min med min. See dBu ma tshig gsal gyi ti ka 74b6 (cited and translated in Yo shimizu 2005: 136): theg pa gsum gyi rigs can ston fiid kyi sa bon smin pa la bdag med pa dan bdag bkag pa'i bkag pa yan med ces so II One should note that the logical rule of double negation, in the sense that the negation of the negation of a position affirms the position, is inapplicable to the theory of freedom from extremes. For the rule of double negation, cf., e.g., Pramal!aviniscaya (PVin) chapter 3 in the passage preceding the head of the Sanskrit fragment edited by Matsuda and Steinkellner (1991). See PVin chapter 3 D224b7, P323b4: mthun pa'i phyogs ma yin pa fiid la med do ies bya bas ni 'di mthun pa'i phyogs fa yod par brjod pa yin te I dgag pa gfiis kyis (emended: DP kyi) mal ma go ba'i phyir ro I Manuscript A64a2: asapak~a eva nastfti casya sapak~e 'stitocyate, prati~edhadvayena prakrtagamanat I Cf. also ibid. D225a2, P323b6: de dgag pa bkag pa'i no bo ni sgrub pa'i ran biin yin pa'i phyir ro II Manuscript A64a4, Matsuda and Steinkellner 1991: 142: prati~edhani~edhasya vidhiinariipatvat I lowe this information to Tom Tillemans and Pascale Rugon. 45 The masters from the East maintain that the Madhyamika establishes the negation of superimposed ultimate intrinsic natures. Cf., e.g., MA D179b5-180a2 cited and translated in Keira 2004: 3lf. 46 dBu ma tshig gsal gyi ti ka lla3: sgra'i dban du byas na bya ba dan 'breI ba don gyi dgag pa ston pa ste med dgag ste I dper na sid zan mi bza' ces pa ste bza' ces pa bya ba yin I de dan dgag tshig mi gfiis sbyar bas

462

.Chizuko Yoshimizu

[5] The non-implicative negation (med dgag) by Buddhist logicians (tshad ma pa) is also regarded as implicative negation (ma yin dgag) from the Madhyamaka [viewpoint]. They implicitly intend the property of negation [or the property of being negated] (dkag pa'i chos iig) after negating the blue. Madhyamikas do not even intend the negation, because [they] intend that it is nothing (ci yari ma yin). [They] intend neither the collection of the property and the property-possessor (chos dan chos can gyi tshogs don), nor the property-possessor, nor the single property to be proven (sgrub bya'i chos), [i.e.,] the non-existence of intrinsic nature (ran biin med).4J

In this way, Zan Than sag pa throughout insists on negating the negation with regard to Madhyamikas' negative statements. 48 Obviously, he opposes Candraklrti's system to that of those adherents of logic who had been regarded as authentic Madhyamikas in Tibet until CandrakIrti's works were introduced. In fact, Zan
mi bza' ces so II des na sid zan za ba bkag pa tsam sgrub pas med dgag go I It also deserves attention that Zan Than sag pa defines med dgag in general as the "establishment of mere negation [e.g.,] of the eating of food offerings" (sid zan za ba bkag pa tsam sgrub pa lla3). This interpretation is different from the well-known definition of prasajyaprati:jedha in Tarkajviilii (TJ) D59b5f.: med par dgag pa ni dnos po'i no bo iiid tsam iig 'gog par zad kyi de dan 'dra ba de ma yin pa gian gyi drios pos grub par mi byed pa ste I dper na bram zas chan btun bar mi bya'o ies bya ba de tsam iig 'gog par zad kyi de las gian pa'i btun ba bturi rio ie'am mi bturi no ies mi brjod pa Ita bu'o II Cf. further Phya pa's definition cited below in n. 49. As for the two kinds of negations, viz., paryudiisa and prasajyaprati:jedha, in the Indian Buddhist tradition, see, e.g., Kajiyama 1973. 47 dBu ma tshig gsal gyi ti ka lla4: tshad ma pa'i med dgag kyan dbu' ma pa la [tos nas ma yin dgag du 'gro ste I snon po bkag nas bkag pa'i chos iig sui du 'dod pa'o II dbu' ma pa ni bkag pa yan mi 'dod te I ci yari ma yin par 'dod pa'i phyir ro II chos dan chos can gyi tshogs don yan mi 'dod [chos can yan mi 'dod inserted] sgrub bya'i chos [rari biin med inserted] rkyari pa yan mi 'dod zer I 48 Cf. Nagarjuna's statements in his VV 63 that he negates nothing and that there exists nothing to be negated. He also describes in VV 23 the nature of negative statements as an illusion which stops another illusion (cited and translated in Seyfort Ruegg 2000: 119f).

Zan Than sag pa on theses

463

Than sag pa's contemporary Phya pa Chos kyi sen ge (1109-1169) defines non-implicative negation (med dgag) in the system of the masters from-the East as apprehending "sole negation" (i.e., dgag pa 'ba' zig par blos zen par bya ba).49 It is further interesting to recall that Tson kha pa (1357-1419) expressly reiterates the view that non-implicative negation (med dgag) is an establishment of negation. so Now one can properly assume that Tson kha pa thereby tried to integrate the view of the masters from the East into Candraklrti's Madhyamaka system. Go rams pa later made a sharp rejoinder to Tson kha pa, defending the theory of "neither existence nor nonexistence."sl

49 Sar gsum stoli thun 87,9f.: sgrub pa 'ba' iig dali dgag sgrub tshogs pa ma yin dgag yin fa dgag pa 'ba' iig pa myed dgag yin pas med dgag gi mtshan fiid ni ldog pa de kha yar lies pa na dgag pa 'ba' iig par bios ien par bya ba yin la I It is unclear whether Zan Than sag pa and Phya pa actually knew each other. 50 Cf., e.g., rTsa se tik chen 25a1ff.: dam bca' bii med dgag tu bied pa'i phyir rali biin 'gog pa'i skabs thams cad du med dgag rtags kyi bsgrub byar bya'o II des na tshig gsalfas I rjes su dpag pa dag ni gian gyi dam bca' ba 'gog pa tsam gyi 'bras bu can yin pa'i phyir TO II ies gsulis pa ymi rali biin yod pa rnam par bead tsam iig sbyor ba rnams kyis sgrub kyi de las gian pa'i chos gian mi sgrub pa'i don yin pas rali biin yod pa 'gog gi med pa mi sgrub ces pa min no II Quoting this passage, Matsumoto (1997: 32lf.) has described it as an "astonishing statement," meaning that it is completely different from the thought of non-implicative negation introduced by Bhiiviveka into the Madhyamaka tradition. The fact is, however, that Tson kha pa's statement shows full agreement with the interpretation which Zan Than sag pa has ascribed to Buddhist logicians. Matsumoto has carefully compared Tson kha pa's view with Bhiiviveka's concerning non-implicative negation and detailed the unique characteristics of TsoIi kha pa's Madhyamaka thought. It will become clear upon reading earlier Tibetan treatises to what extent Tson kha pa owes his thought to early masters. 51 Regarding the controversy between Tson kha pa and Go rams pa et aI., cf., e.g., Matsumoto 1997: 288ff., Seyfort Ruegg 2000: 195-266, Cabez6n 2003, Cabez6n 2007, and Yo shimizu 2005: 137ff.

464

Chizuko Yoshimizu

Concluding remarks
The problem of whether the Madhyamika should propound theses of his own has developed in Tibet into the controversy as to whether the Madhyamika should establish or adopt the negation of intrinsic nature as his own probandum. This controversy first took place between the followers of the Santarak~ita-KamalasIla lineage of Madhyamaka and the followers of the newly introduced Madhyamaka of Candraklrti (i.e., between the so-called Svatantrika and Prasangika). Due to Tson kha pa's integration of the logical methods shared by the Buddhist logicians and Santarak~ita-KamalasIla into Candraklrti's Madhyamaka system, however, this debate shifted to a matter for dispute among the adherents of Candraklrti in the dGe lugs and Sa skya schools. Considered from the historical perspective, one should review the value and the nature of the thesis for the Madhyamika in wider scope, including the Buddhist logical tradition. The influence of the Madhyamaka thought of Santarak~ita and KamalasIla, in particular, needs to be reexamined.

Bibliography

Primary sources
dBu ma tshig gsal gyi ti ka: Zan Than sag pa 'BYUIi gnas ye ses, dBu ma tshig gsal gyi ti ka. Manuscript. dBu ma'i spyi ston: Go rams pa bSod nams sen ge, rGyal ba thams cad kyi thugs kyi dgolis pa zab mo dbu ma'i de kho na fiid spyi'i liag gis ston pa lies don rab gsa!. In The Complete Works of the Great Masters of the Sa skya Sect of the Tibetan Buddhism, 12, edited by bSod nams rgya mtsho, 348-451 (Ca 1-208a). Tokyo: The Toyo Bunko, 1969. CS: Aryadeva, Catuf:tsataka. V. Bhattacharya, editor. The Catuf:tataka of Aryadeva. Calcutta: Visva Bhiiratl, 1931. D: sDe dge Tibetan Tripitaka bsTan 'gyur - preserved at the Faculty of Letters, University of Tokyo. J. Takasaki, Z. Yamaguchi, Y. Ejima, editors. Tokyo, 1981-1984. *LT: Dharma grags (Dharmaklrti), *Lak~mJatfkii. See Yonezawa 2004. MA: Candraklrti, Madhyamakiivatiira. See MABh.

Zan Than sag pa on theses

465

MABh: Candraklrti, Madhyamaki1vatiirabhii~ya. L. de La Vallee Poussin, editor. Madhyamakiivatiira par Candrakfrti: Traduction tibetaine. St. Petersburg: Russian Academy of Sciences, 1907-1912. Mal: Santarak~ita, Madhyamakiilal?lkiira. D3884, P5284, Ichigo 1985. MAlv: Santarak~ita, Madhyamakiilal?lkiiravrtti. D3885, P5285, Ichigo 1985. MP..: KamalasIla, MadhyamakiiIoka. D3887, P5287. MMK: Nagarjuna, Malamadhyamakakiirikii. J.W. de Jong, editor. MaIamadhyamakakiirikii/:l. Madras: Adyar Library and Research Center, 1977. NB: Dharmaklrti, Nyiiyabindu. PaI}.qita Dalsukhbhai Malvania, editor. Pa1J4ita Durveka Misra 's Dharmottarapradfpa, Being a Sub- Commentary on Dharmottara's Nyiiyabindutfkii, a Commentary on Dharmakfrti's Nyiiyabindu. Patna: Kashi Prasad Jayaswal Research Institute, 1971. NM: Dignaga, Nyiiyamukha. Taish5 1628. P: The Tibetan Tripitaka, Peking edition. D.T. Suzuki, editor. Tokyo-Kyoto, 1955-1961. Pr LVP: Candraklrti, Prasannapadii. L. de La Vallee Poussin, editor. MaIamadhyamakakiirikiis de Niigiirjuna avec Ia Prasannapadii. St. Petersburg: Russian Academy of Sciences, 1903-1913; D4860, P5260. Prajp: Bhaviveka, Prajfiiipradfpamalamadhyamakavrtti. D3853, P5253. PS 3: Dignaga, Pramii1Jasamuccaya, chapter 3. See Kitagawa 1973. PV 4: Dharmaklrti, Pramii1Javiirttika, chapter 4. See Tillemans 2000. PVBh: Prajfiakaragupta, Pramii1Javiirttikabhii~ya. R. Sankrtyayana, editor. Pramii1Javiirttikabhii~ya or ViirttikiiIal?lkiira of Prajfiiikaragupta. Patna: Kashi Prasad Jayaswal Research Institute, 1953. PVin 3: Dharmaklrti, Pramii1Javiniscaya, chapter 3. D4211, P5710. PVV: Manorathanandin. S.D. Shastri, editor. Pramii1Javiirttikavrtti. Pramii1Javiirttika of Acharya Dharmakfrti with the Commentary 'Vrtti' of Acharya Manorathanandin. Vanarasi: Bauddha Barati, 1968. Sar gsum ston thun: H. Tauscher, editor. Phya pa Chos kyi sen ge, dBu ma sar gsum gyi ston thun. Vienna: Arbeitskreis fUr Tibetische und Buddhistische Studien Universitat Wien, 1999. SNSi1: Sal?ldhinirmocanasatra. See Lamotte 1935. TJ: Bhaviveka, Madhyamakahrdayavrtti TarkajviiIii. D3856, P5256. ITa ba'i san 'byed: Go rams pa bSod nams sen ge, ITa ba'i sen 'byed theg mchog gnad kyi zia zer. In The Complete Works of the Great Masters of the Sa skya Sect of the Tibetan Buddhism, 13, edited by bSod nams rgya mtsho, 1-24 (Ca 1-47b). Tokyo: The Toyo Bunko, 1969.

rTsa se tik chen: Tson kha pa BIo bzan grags pa'i dpal, rTsa se tik chen. dBu ma rtsa ba'i tshig Ie'ur byas pa ces bya ba'i rnam Mad rigs pa'i rgya mtsho. In The Collected Works of Tson kha pa BIo bzan grags pa'i dpal,

466

Chizuko Yoshimizu

23 (bKra sis lhun po blocks), 1-566 (Ba 1-283b). New Delhi: Ngawang Gelek Demo, 1975. VV: Nagarjuna, Vigrahavyiivartanf. E.H. Johnston & A. Kunst, editors. The Dialectical Method of Niigiirjuna. Vigrahavyiivartanf. Delhi: Motilal Banarsidass, 1978.

Secondary literature
Cabez6n, Jose Ignacio, 2003. "1\vo Views on the Svatantrika-Prasatigika Distinction in Fourteenth-Century Tibet." In The Sviitantrika-Priisaitgika Distinction - What Difference Does a Difference Make? edited by G.B.J. Dreyfus & S.L. McClintock, 289-315. Boston: Wisdom Publications. Cabez6n, Jose Ignacio and Geshe Lobsang Dargyay, 2007. Freedom from Extremes - Gorampa's "Distinguishing the Views" and the Polemics of Emptiness. Boston: Wisdom Publications. Chos 'phe1, 2004. lHa sa sa khul gyi gnas yig. Garis can bod kyi gnas Mad lam yig gsar ma 2. Beijing: Mi rigs dpe skrun khaIi.. Ichigo, Masamichi, 1985. Madhyamakiilal'{tkiira of Siintarak#ta - With His Own Commentary or Vrtti and with the Subcommentary or Pafijikii of Kamalasfla. Kyoto: Buneido. Iuchi, Maho, 2007. Kodenki no Tibet Bukkyo Sekai - bKa' gdams pa wo chushin to shite (Tibetan Buddhist Universe in the phyi dar Period, Centering on bKa' gdams pa). Dissertation, Otani University. Kajiyama, Yiiichi, 1973. "Three Kinds of Affirmation and 1\vo Kinds of Negation in Buddhist Philosophy." Wiener Zeitschrift fur die Kunde Sudasiens und Archiv fUr Indische Philosophie 17: 161-175. Keira, Ryusei, 2004. Miidhyamika and Epistemology. A Study ofKamalasfla's Method for Proving the Voidness of All Dharmas. Vienna: Arbeitskreis fUr Tibetische und Buddhistische Studien Universitat Wien. Kitagawa, Hidenori, 1973. Indo koten ronrigaku no kenkyu, Jinna (Digniiga) no taikei (Study of Indian Classical Logic, System of Digniiga). Tokyo: Suzuki Gakujutsu Zaidan. Lamotte; Etienne, ed., 1935. L'Explication des Mysteres. Louvain: Universite de Louvain. MacDonald, Anne, 2003. "Interpreting Prasannapadii 19.3-7 in Context - A Response to Claus Oetke." Wiener Zeitschrift fur die Kunde Sudasiens und Archiv fUr die Indische Philosophie 47: 143-195. Matsuda, Kazunobu, and Steinkellner, Ernst, 1991. "The Sanskrit Manuscript of DharmakIrti's PramiiT}aviniScaya." Wiener Zeitschrift fUr die Kunde Siidasiens undArchiv fUr die Indische Philosophie 35: 139-149.

Zan Than sag pa on theses

467

Matsumoto, Shiro, 1999. Tibet bukkyo tetsugaku (Tibetan Buddhist Philosophy). Tokyo: Daizo-shuppan. Ono, Motoi, 1986. "DharmakIrti ni okeru shuchomeidai no teigi ni tsuite" ("On the Definition of Pak~a as Amended by Dharmaklrti"). Indogaku-Bukkyogaku-Kenkyu (Journal of Indian and Buddhist Studies) 34/2: (109)-(112). Ses gfien tshul khrims, 2001. lHa sa'i dgon tho. Lhasa: Bod Ijons mi dman dpe skrun khan. Seyfort Ruegg, David, 2000. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy. Studies in Indian and Tibetan Madhyamaka Thought, Part 1. Vienna: Arbeitskreis flir Tibetische und Buddhistische Studien Universitat Wien. Tillemans, Tom J.F., 2000. Dharmakfrti's PramiilJavilrttika: An Annotated Translation of the Fourth Chapter (pariirthiinumiina). Vol. 1 (k. 1-148). Vienna: Verlag der Osterreichischen Akademie der Wissenschaften. Yonezawa, Yoshiyasu, 2004. "*Lak~alJatfkil, Sanskrit Notes on thePrasannapadii (1)." Naritasan bukkyo kenkyusho kiyo (Journal of Naritasan Institutefor Buddhist Studies) 27: 115-154. Yoshimizu, Chizuko, 1993. "On ran gi mtshan fiid kyis grub pa III Introduction and Section I." Naritasan bukkyo kenkyasho kiyo (Journal of Naritasan Institute for Buddhist Studies) 16: 91-147. -1996. Die Erkenntnislehre des Priisaligika-Madhyamaka. Vienna: Arbeitskreis flir Tibetische und Buddhistische Studien Universitat Wien. - 2003. "Tsong kha pa's Reevaluation of Candraklrti's Criticism of Autonomous Inference." In The Sviitantrika-Priisaligika Distinction What Difference Does a Difference Make? edited by G.B.J. Dreyfus & S.L. McClintock, 257-288. Boston: Wisdom Publications. - 2005. "A Tibetan Text from the Twelfth Century Unknown to Later Tibetans." Cahiers d'Extreme-Asie 15: 125-163. - 2006. "Indo-Tibet Chiigan-shiso-shi no saikouchiku ni mukete" ("Toward Reconstructing the History of Indian and Tibetan Madhyamaka Thought - The Launch of the Study of dBu ma tshig gsal gyi ti ka I"). Tetsugakushiso ronsha (Studies in Philosophy) 32 (Tsukuba University): 73-114. - 2008. "Zhang Thang sag pa's Reevaluation of BuddhapiHita's Statement of Consequence (prasaliga)." Tetsugaku-shiso ronsha (Studies in Philosophy) 34 (Tsukuba University): 81-99. Yotsuya, Kodo, 1999. The Critique of Svatantra Reasoning by Candrakfrti and Tsang kha pa. Tibetan and Indo-Tibetan Studies 8. Stuttgart: Franz Steiner Verlag.

Roil zorn Chos bzail on Mahayoga and Madhyarnaka

Heidi Koppl

For Ron zorn Chos bzan, a primary authority in the rNin rna school who was active during Tibet's turbulent 11th century, the view (darsana, Ita ba) of the Mahayoga tantras is clearly superior to that of the slitras. Ron zorn hence compares the view of Mahayoga with a Madhyamaka that for him is inferior since it reifies relative truth (saf[lvrti, kun rdzob), thereby leading to an irreconcilable separation of the two truths. This article seeks to evaluate Ron zorn's vehement criticism of the notion of an objectively established relative reality against his declaration of "relative truth" as "pure appearance" (dag pa'i snan ba). Exploring the relationship between his radical critique of the notion of "nature" (svabhiiva, ran biin) and his arguments for appearances being divine (snan ba Ihar sgrub pa), we will contemplate the feasibility of locating Ron zorn's views on a realistantirealist scale. Finally, we will also consider the startling claim of Ron zorn's self-professed follower, 'Ju Mi pham (1846-1912), that Candraklrti and Ron zorn share the same view of emptiness (siinyatii, stan pa fiid).
Roil. zorn's view of Madhyamaka

Ron zorn firmly believes that the view of the tantras is superior to that of Madhyamaka. According to him the "view of Mantra" (snags kyi ita ba) is the realization of "the inseparability of the two truths" (bden pa gfiis dbyer med) in the Mahayoga context; Ron zorn specifies this by referring to the relative truth of "great purity" (dag pa chen po) and the ultimate truth of "great equality"
Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 469-481

470

Heidi KoppI

(mnams pa chen po). Now, when Ron zorn asserts that the view of Mantra is superior to that of Madhyamaka, it becomes of course important to know what Madhyamaka means to him. Although no extensive exegesis dn Madhyamaka by Ron zorn remains extant, we find numerous texts in which Ron zorn contrasts the view of Madhyamaka with that of Tantra, and it is therefore these that we will have to rely on as sources for his view of Madhyamaka. In The Great Memorandum on Views (lTa ba'i brjed byan chen mo) Ron zorn describes the Madhyamaka and its subschools in the following way:
Although [Madhyamikas] concord in their view of the ultimate, [the school] is, with respect to [its view] regarding relative [truth], divided into the Sautrantika-Madhyamaka and the Yogacara-Madhyamaka. 1 A few pages below, the same text further divides the Madhyamaka view into "those who assert [everything to be] thoroughly nonabiding" (rab tu mi gnas pa) and "those who assert [everything to be] like illusion" (sgyu ma Zta bU).2 Ron zorn describes Madhyamaka as a school that divides the relative into authentic and inauthentic with reference to the presence or absence of "causal efficacy" (arthakriyiisamartha, don byed nus pa) and does not consider Candraklrti's approach of referring to the perceiving subject, rather than the object, for establishing the relative. 3 He moreover expends a considerable amount of energy criti-

1 Ron zom chos bzan gi gsun 'bum, vol. II, 18: de yan don dam pa'i Ita ba mthun yan kun rdzob kyi bye brag gis I mdo sde dbu ma dan I mal - 'byor spyod pa'i dbu ma gfiis su gyes (sic) so I The same classification

is found in anonymous textual fragments found at Thnhuang (see Lang 1990: l30 and n. 12). Ye ses sde (ca. 800) in his Differentiation of Views (ITa ba'i khyad par) mentions Bhavaviveka and Santarak~ita as representatives of these two classifications (see Tauscher 1995: 6, n. 7). 2 See Ron zom chos bzan gi gsun 'bum, vol. II, 20: dbu ma'i ita ba mdor bsdus na I ma ya ste sgyu ma Ita bu 'dod pa dan I a bra ti sti te rab tu mi gnas par 'dod pa'o II Ruegg (1981: 58) mentions that Advayavajra (11th c.) similarily divided Madhyamaka into Mayopamadvayavadins and
Sarvadharmaprati~thanavadins.
3

Cf. Madhyamakavatara, chapter six, verse 23: dnos kun yan dag pa'i

Ron zorn ehos bzanon Mahayoga and Madhyamaka

471

cizing Sautdintika- and Yogacara-Madhyamaka for distinguishing relative truth on the basis of objective efficacy and the lack of it. For instance, in his Entering the Way of the Great Vehicle (Theg pa chen po'i tshulla 'jug pa) Roil zorn argues that both a vase and its reflection are individually able to perform their distinct functions and so one must, according to Ron zorn, accept that the division of the relative into authentic and inauthentic aspects based on efficacy is simply mistaken. 4 Arguing against any attempt at ascribing a limited, i.e., not ultimately valid, sense of authenticity to certain aspects of the relative, Roil zorn employs several pointed anecdotes. 5 One such story in Entering the Way of the Great Vehicle 6 tells of a king who was able to expand his dominion beyond any earthly ruler, and so even came to share the throne of Indra, king of gods. Yet throughout all of his successes he was driven by greed, and when sitting on the celestial throne he still remained dissatisfied. Upon wishing to have the throne all to himself, he was struck by disaster, and so he fell back into the realm of men. As the former king realized what he had lost, he was overcome by despair; destroyed by his own greed, he died of grief. This insatiable craving Roil zorn uses to illustrate the detrimental consequences of believing in an authentic relative truth:
If, since all phenomena are ultimately pacified beyond all mental constructs, one sees no probandum whatsoever to be proven and yet at the same time still perceives certain characteristics of an authentic relative [reality] that contains substances to be abandoned or accepted, that is an extremely inappropriate way of thinking and a cause for amazement.?

brdzun pa mth01i ba yis / dnos rfied no bo gfiis ni 'dzin par 'gyur / yan dag mthon yul gan yin de fiid de / mthon ba brdzun pa kun rdzob bden par gsuns / 4 Ron zorn chos bzan gi gsun 'bum, vol. I, 422. 5 See Koppl 2008: 44-49. 6 Ron zorn chos bzmi gi gsun 'bum, vol. I, 424. 7 Ron zorn chos bzan gi gsun 'bum, vol. I, 424: chos thams cad don dam par spros pa thams cad fie bar ii ste / bsgrub par bya ba gan yan mi

472

Heidi Koppl

In brief, Ron zorn seems quite bewildered by the Miidhyamika's tendency on the one hand to claim that ultimately everything is nonexistent, but still to insist that there are, relatively speaking, objectively established entities that can be accepted or rejected accordingly. To Ron zorn, it must be a feeling of greed and dissatisfaction that causes the Madhyamika to make the claim that relative truth has its own distinct status. A Madhyamika is dissatisfied with the unborn empty nature of things and therefore greedily insists on an artificial division between authentic and inauthentic relative truths based on a criterion of causal efficacy. Since the "real" relative truth is thus ascribed a sort of objective existence, the two truths, the ultimate and the relative, end up, for Ron zorn, separated at the level of ontology. If we take a further look at the image of the greedy king, it seems that for Ron zorn the Madhyamika is a rather impulse driven person, governed by emotions. This is a somewhat surprising portrayal. The above described bifurcation of the relative is associated with the approach of those Madhyamaka masters who became known as Svatantrikas, who draw this distinction in an appeal to reason and rationality, seeking to provide solid grounds for sound and careful epistemic practices. In other words, the distinction that they draw is part of an elaborate endeavor to pay proper attention to relative truth, and by all means to avoid denigrating it. We would associate this approach therefore with much more likeable qualities, such as cautiousness and intellectual discipline, rather than, as Ron zorn would have it, greed driven impulse. Ron zorn in fact appears quite disturbed by the Madhyamika's tendency on the one _ hand to proclaim that everything is "empty" (siinya, ston pay while still struggling to come up with objective properties that are to be accepted and rejected. Given his contemptuous dismissal of the perceived attempt to reify relative phenomena, then why, we may ask, is Ron zorn so concerned with proving that phenomena are pure, or even divine, on the relative level? This is, after all, the

sgrub par Ita biin du I yan dag pa'i kun rdzob kyi mtshan fiid span ba dan blan bar bya ba'i rdzas yod par 'dzin pa de ni I sin tu mi 'tsham pa 'dzin pa ste no mtshar ba'i gnas yin no II

 Ron zorn Chos bzaIi on Mahayoga and Madhyamaka

473

main project in his Establishing Appearances as Divine. If phenomena lack characteristics on both levels of truth, then why would he not settle for an acknowledgement of just that; why, in other words, be so greedy? Drawing the dividing line between Madhyamaka and. Mantra Before we consider a reply to this question let us explore a bit further Roil zorn's view of Madhyamaka and his way of distinguishing this view from that of Mantra. The Madhyamaka is, for Roil zorn, characterized by a failure to acknowledge the inseparability of the two truths. In his Entering the Way of the Great Vehicle, Roil zorn thus remarks:
Madhyamikas assert that [all] phenomena are indeed without nature (svabhiiva, ran biin) ... but since they do not abandon the concept of two truths, their view does not qualify as non-dual. S

Throughout Roil zorn's writings, we find ample evidence that for him a separation of the two truths, in the sense of conceptually isolating them from each other and ascribing to them each their own ontological status, is an error in the extreme. In Entering the Way of the Great Vehicle, Roil zorn says:
If one asserts that there is something to be established in reality, then all the particularities of the relative - as many as one may enumerate - are likewise suited to be established. If, [however], one does not assert that there is anything to be established in reality, then all the particularities of the relative will be on the same level. 9

Thus Roil zorn argues that both levels of truth, the relative and the ultimate, are of an equal ontological status. He again and again
S Ron zom chos bzan gi gsun 'bum, vol. I, 460: dbu ma pa chos rnams ran biin myed pa iiid yin par 'dod pa la /. .. 'on kyan bden pa giiis kyi blo mi 'dar bas / giiis su myed par lta ba'i grans su mi chud de / 9 Ron zom chos bzan gi gsun 'bum, vol. I, 422: gan gi tshe yan dag par bsgrub par bya ba zig yod par 'dod pa de'i tshe ni I kun rdzob kyi bye brag ji siied du dbye ba thams cad kyan de bf.in du bsgrub tu run bar 'gyur la I gan gi tshe yan dag par sgrub par bya ba zig bsgrub tu myed par 'dod pa de'i tshe I kun rdzob kyi bye brag thams cad 'go miiam par 'gyur TO II

474

Heidi Koppl

points out the futility and absurdity involved in considering everything to be ultimately unborn while at the same time dividing the relative into the dualism of authentic and inauthentic, thereby objectifying relative phenomena by perceiving them as entities that are to be accepted or rejected. In the Black Snake Discourse (sErul nag po'i ston thun)10 Roil zorn delineates the views and approaches of the outer tantras (snags phyi pa), the inner tantras (snags nan pa), and finally the Great Perfection (rdzogs pa chen po). He explains that it is only with the view of the tantras that the inseparability of the two truths begins to dawn, and that moreover this realization is only fully perfected within the Great Perfection. When an imaginary opponent subsequently inquires as to why such a view of inseparability would not be set forth by Madhyamaka, Roil zorn states that as long as one treats the truths as pertaining to separate states of affairs, one will, when examining phenomena, never succeed in discarding the dualistic mindset. In this way, one may propose that ultimately the two truths are inseparable, yet without giving up the idea that illusory phenomena exist on the relative level. Thus, even when attempting to establish the "intrinsic nature" (dharmatii, chos fiid) as nondual, one will still retain a dualistic perspective. The .person who falls into this trap, Roil zorn continues, might argue that all that exists by way of imputation is ultimately free from mental constructs and therefore similar to illusion, yet he will not assert the same with regard to the relative level, for he cannot accept that phenomena should lack substantial efficacy even conventionally. Thus, even while examining the characteristics of substantial existence and concluding that indeed such existence is ultimately unestablished, one will still not have mentally discarded the two ontological modes. Roil zorn illustrates his point by referring to "appearances" (snan ba): as long as one assumes that there is an appearance that can be taken as the "subject" (dharmin, chos can) or used as an "illustration" (mtshan gii) in argument, and as long as one considers this subject or illustration to be "free from mental constructs" (ni~prapafica, spros bral) and mere illusion, one will

10 Ron zorn chos bzan gi gsun 'bum, vol. II, 66-68.

 Ron zorn Chos bzan on Mahayoga and Madhyamaka

475

still not have gone beyond conceptualizing the characteristics of appearances. The view of great equality remains, therefore, still to be accomplished, for by identifying a particular appearance and then concluding that this indeed is beyond mental constructs, the "appearance aspect" (snan ba'i cha) and the "empty aspect" (ston pa'i cha) will have become temporarily separated. Let us in this context investigate a statement from Ron zorn's commentary to the Garland of Views as Oral Instructions (Man nag lta ba'i phren bay in which Ron zorn further objects to viewing the relative as "impure" (ma dag pa).
Madhyamikas adhere to relative truth as being the experiential domain of completely impure characteristics. l1

In the Great Memorandum on Views (lTa ba'i brjed byan chen mo) Ron zorn likewise remarks, when elucidating the view of Mantrayana:
The assertion that ultimately there is no birth and no cessation is the same as in Madhyamaka; the view that in terms of the relative the skandhas, dhiitus, and iiyatanas are mere illusion also concords. [But] in particular, [the view of Mantrayana] distinguishes itself in regarding even the characteristics of illusion as the characteristics of complete divine purity, and also in seeing the two truths as inseparable. 12

Therefore, according to Ron zorn the dividing lines between SUtra and Mantra can be drawn based on a) whether, as in the case of Madhyamaka, relative truth becomes reified and the two truths thus become separated and b) whether, on the relative level, appearances are perceived as purity (Mantra) or impurity (Sutra). For Ron zorn,

11 Ron zom chos bzan gi gsun 'bum, vol. I, 327: dbu ma pa ni kun rdzob kyi bden pa lyons su ma dag pa'i mtshan fiid kyi spyod yulla mnon par zenpa'o I 12 Ron zom chos bzan gi gsun 'bum, vol. II, 20-21: don dam par skye 'gag med par 'dod pa ni dbu ma dan mthun no I kun rdzob tu phun po khams dan skye mched sgyu ma tsam du Ita ba'an mthun no I bye brag tu na sgyu ma'i mtshan fiid de dag kyan yons su dag pa'i lha'i mtshan fiid du Ita ba dan I bden pa gfiis kyan dbyer med par Ita bas bye brag tu byas pa'o II

476

Heidi Koppl

we may conclude, there appears to be a connection between objectification of the relative and the view of impurity. Antirealism as a prerequisite for purity Let us now return to the question as to how somebody as profoundly antirealist as Ron zom can insist that phenomena, despite lacking any establishment, must be pure on the relative level. For Ron zom, a tantrika, it appears that the universal lack of establishment, even on the conventional level, carries with it a sense of "purity" (dag pa). Hence Ron zom sees Madhyamaka, with its separation of ultimate and relative truths and the division of the relative into authentic and inauthentic aspects, as disparaging the central esoteric notions of inseparability, equality, and purity. The Madhyamika who sets forth an authentic relative truth (ascribing to it a mode of existence that is more genuine than what must otherwise be dismissed as mistaken relative truth) becomes, we may interpret, on Ron zorn's view blind to the ultimate truth of great eqUality. This ultimate equality taught in the Mahayoga is moreover not, as the Madhyamikas would have it, realized by a negation of relative phenomena. Ultimate equality is, as the nature of all phenomena, itself inseparable from relative appearances, and since "these appearances are all equally non-established, Ron zorn can proceed to qualify them as pure, in the sense that they do not need to be negated or, indeed, purified. This approach becomes quite evident in the opening lines of Establishing Appearances as Divine:
The Vajra Vehicle of Secret Mantra states, "all mundane and supramundane phenomena, without any distinction, are primordially enlightened as the maI}.<;lala of vajra-like body, speech and mind. Thus [the maI}.<;lala] is not accomplished through a path." Here one may object, "All those phenomena that appear within the field of experience of sentient beings are not as they seem to be. They are delusion, and thus they cannot be primordial enlightenment." Yes, this is nothing but delusion. Besides this delusion of sentient beings, there is nothing else that might be posited and shown. These phenomena, known as deluded appearances, are all that there is. . .. In this way, all apparent phenomena are nothing but delusion and there is, moreover, no freedom from delusion to be achieved by dispelling

- Ron zorn Chos bzan on Mahayoga and Madhyamaka

477

delusion. Delusion is, by its own essence, completely pure and, hence, enlightenment. All phenomena are, in this way, primordial, full, and complete enlightenment. 13

Hence, when confronted with an objection to his thesis of primordial enlightenment, questioning how this may be reconcilable with the teaching of universal delusion on the part of sentient beings, Roil zorn appears simply to concede that there is nothing but delusion to be found anywhere. It may here be rewarding to once more briefly reflect on Roil zorn's utter dislike for a separation of the two truths as previously described. Indeed, had Roil zorn here appealed to the common Buddhist notion of two levels of truth, and so simply said that "ultimately all is primordially enlightened, yet on the relative level delusion is a fact" somebody accustomed to thinking in terms of relative versus ultimate reality might accept his thesis more readily. As he does not make this move, it appears that Roil zorn's unwillingness to confine either enlightenment or delusion to a single separated level of truth may be a fundamental feature of his approach. While a full application of the two truths model would have implied, in terms of the path, a gradual transformation of the elements of delusion into the nature of enlightenment, Roil zorn instead bluntly states that this all-pervasive experience of delusion is itself enlightenment, and that no other enlightenment could possibly be achieved through a process of purification. For Roil zorn,
13 Ronzom chos bzan gi gsun 'bum, vol. I, 517: gsan snags rdo rje theg pa'i tshullas 'jig rten dan 'jig rten las 'das pa'i chos thams cad dbyer med par sku gsun thugs rdo rje lta bu'i dkyil 'khor du lye nas sans rgyas pa yin pas da lam gyis sgrub pa lta bu ni ma yin no II ies 'byun ba la I 'dir brgal ba I da ltar 'gro ba rnams kyi spyod yul du snan ba'i chos 'di dag ni I ji ltar snan ba de ltar ma yin te I 'khrul pa yin pas lye nas saris rgyas pa ni yin du mi run no ie na I 'khrul pa yin du zad de I sems can gyi 'khrul ba las ma gtogs pa'i chos gian ci yari biag cin bstan du med de I 'khrul snan du grags pa'i chos 'di tsam ies bya ba yin no If. .. de bas na snari ba'i chos 'di dag thams cad 'khrul ba yin du zad de I de yan 'khrul ba bsal nas ma 'khrul pa iig bsgrub tu med de I 'khrul pa rio bo ilid kyis rnam par dag pas saris rgyas pa yin te I de bas na chos thams cad ye nas mrion par rdzogs par sans rgyas pa'o II

478

Heidi Koppl

appearances need not be negated or dispelled; they are pure. We have here emphasized the sense of lack of establishment that comes with the notion of purity, yet in the Mahayoga "purity" (dag pay is indeed associated with the principles of "luminosity" ('od gsal) and "divinity" (lha) as well. Nevertheless, I suggest that as Ron zorn sees it, pure divinity is to be established by taking a strong antirealist stance. In his commentary to the Guhyagarbhatantra Ron zorn states that relative truth is to be viewed as a result, while the ultimate is seen as a cause. 14 This may give us a further hint at Ron zorn's reason for repeatedly objecting to a reified relative truth. According to Madhyamaka, "conventions" (vyavahiira, tha sfiad) are the "means" (upiiya, thabs) for realizing the ultimate; hence, failing to pay the right kind of attention to the conventional is a severe mistake that will cause one to miss the truth of the ultimate. In Mahayoga the sequence seems to be turned around, and so the realisation of "suchness" (tathatii, de biin fiid), or at least an "approximation" (rjes mthun pa) of it, is required from the very beginning ofthe Mahayoga path. From within the state of either actual or approximate realization of suchness, appearances manifest. These need not be investigated, evaluated, or discarded in any way, for they are indivisible from emptiness and therefore pure. Clinging, at this point, to a relative truth that has its own ontological status would for Ron zorn be utterly pointless as well as detrimental to the pursuit of buddhahood: a miserable and completely unnecessary situation, reminiscent of that of the fallen, greedy king. Could Ron zorn be described as a Prasangika-Madhyarnaka? Now, taking into account his disdain for the distinction between two relat~ve truths based on the criterion of causal efficacy, we can finally pose the question of whether Ron zorn might have felt more at ease with Candrakirti's Madhyamaka account of a subjective version of the relative. 'Ju Mi pham (1846-1912) appears to think

14 Ron zorn chos bzan gi gsun 'bum, vol. I, 48: don dam pa rgyur blta ba dan / kun rdzob 'bras bur blta ba dan /

Ron zorn Chos bzan on Mahayoga and Madhyamaka

479

so, for he lists Ron zorn and Klon chen Rab 'byams (1308-1362) as his central sources of reference and inspiration for his exegesis on the Svatantrika/prasangika distinction. 1s In his Beacon of Certainty (Nes ses rin po che'i sgron me) Mi pham even goes so far as to declare that Ron zorn and Candrakirti share the same view of emptiness:
In India the glorious Candraklrti And in Tibet Ron zorn Chos bzan both With one voice and one intent Established the great emptiness of primordial purity. Because these phenomena are primordially pure, Or because they are primordially without intrinsic nature, They are not born in either of the two truths.16

As we have seen, Ron zorn criticized and objected vehemently to any objectively established reality and dismissed any sense of truth that would involve impure characteristics. In this way, we could say that he professed a view in which all phenomena are, from the very outset, established as emptiness. Candrakirti does, according to Mi pham's exegesis, likewise advocate a suddenist path of transcending conceptual extremes and constructs without ascribing any privileged status to certain aspects of relative truth. 1? For Mi pham, Candrakirti is a philosopher who, reluctant to separate the two truths, does not "qualify the object of refutation with the term 'ultimate'" (dgag bya la don dam pa'i khyad par sbyar ba),18 hence denying nature (svabhiiva, rali biin) on both levels of truth. 19
1S Cf. Doctor 2004: 76-78, 116 and 604-606. 16 Nes ses rin po che'i sgron me, 5: 'phags yul dpalldan zla ba dan / bod na ron zom chos bzan gfiis / dgons pa gcig dan dbyans gcig gis / ka dag ston pa chen po bsgrubs / chos 'di ka nas dag pa'am / gdod nas ran biin med pa'i phyir / bden pa gfiis char ma skyes pas / 1? Doctor 2004: 84. 18 Ses rab kyi le'u 'grel pa nor bu ke ta ka, 3b. 19 Considering that the rN-in rna school was often criticized for being similar to Chinese Ch'an Buddhism because of its suddenist approach, Mi pham's explicit alignment of Ron zorn with Candraklrti might have

480

Heidi K6ppl

If we take into account Ron zorn's explicit disdain for a Svatantrika type of distinction between two relative truths, we might speculate that indeed Ron zorn would have been more at ease with Candraklrti's approach. Yet when reading Ron zorn's works I see him primarily as an exponent of the Guhyagarbha and the Great Perfection. With his repeated emphasis on the two truths of the Mahayoga being clearly superior to those of Sutra, I would imagine that any Madhyamaka, be it Santarak~ita's or Candraklrti's, would remain subject to his criticism. Likewise, given his criticism of the Madhyamaka's view of impure relative reality, and his insistence that appearance is pure, it seems unlikely that Ron zorn would be fully satisfied with Candraklrti's account of the relative (which indeed bears little evidence of any application of a purity paradigm). Hence, unlike Mi pham, who to some extent attempted to harmonize the view of Madhyamaka with the tantras, I do not find any evidence that Ron zorn sought to justify his tantric views by appealing to a more readily acceptable Madhyamaka perspective, and I do not think that his approach can be seen as apologetic in that sense. Perhaps we can suggest that while Ron zorn primarily criticized the Madhyamaka view as imperfect, his critique itself was to inspire a type oftantric interpretation ofPrasangika-Madhyamaka. Hence, when Mi pham interpreted Candraklrti in the 19th century, he developed a Madhyamaka aimed at avoiding those very features that Ron zorn had been objecting to in the 11th.

Bibliography Tibetan sources

CandrakYrti. Madhyarnakavatara (dBu rna la 'jug pay. sDe dge 3861. Berkeley: Dharma Publishing, 1981. been an attempt to thwart such criticism. Demonstrating a predilection for the suddenist path in the works of Candraklrti - who in Tibet had become almost universally celebrated as a paramount philosopher among the masters of India - would provide authoritative support to the suddenist approach.

 Ron zorn Chos bzan on Mahayoga and Madhyamaka

481

Mi pham rgya mtsho. Nes ses rin po che'i sgron me. In Kunzang Palden, Commentary to the Beacon of Certainty (Nes ses sgron me rtsa 'grel). Chengdu: Si k)J.ron mi rigs dpe skrun khan, 1997. - . Dbu ma rgyan gyi rnam bSad 'jam dbyafts bla ma dgyes pa'i ial luft. Trans. in Doctor 2004. - . Ses rab kyi le'u 'grel pa nor bu ke ta ka. In Sde-Dge Dgon-Chen Prints of the Writings of 'Jam-Mgon 'Ju Mi-Pham-Rgya-Mtsho, vol. XIV, 1-96. Kathmandu: Shechen Monastery, 1990. Ron zorn Chos kyi bzan po. Raft zom Chos bzaft gi gsuft 'bum, vols. I and II. Chengdu: Si khron mi rigs dpe skrun khan, 1999.

Modern research and translations

Doctor, Thomas (trans.). 2004. Speech of Delight: Mipham's Commentary on Santarak:jita's Ornament of the Middle Way. Ithaca: Snow Lion Publications. Koppl, Heidi. 2008. Establishing Appearances as Divine: Rongzom Chozang on Reasoning, Madhyamaka, and Purity. Ithaca: Snow Lion Publication. Lang, Karen. 1990. "Spa-tshab Nyi-ma-grags and the Introduction of Prasangika Madhyamaka into Tibet." In L. Epstein and R. Sherburne, eds., Reflections on Tibetan Culture: Essays in Memory of Turrell V. Wylie. Studies in Asian Thought and Religion, 12. Lewiston/ Queenston/ Lampeter: The Edwin Mellen Press, 127-141. Ruegg, David Seyfort. 1981. The Literature of the Madhyamaka School of Philosophy in India. History of Indian Literature, vol. VII, fasc. 1. Wiesbaden: Otto Harrassowitz. Tauscher, Helmut. 1995. Die Lehre von den zwei Wirklichkeiten in Tsang kha pas Madhyamaka-Werken. Vienna: Arbeitskreis flir Tibetische und Buddhistische Studien.

Notes on contributors

James A. BENN (PhD, UCLA 2001) is Associate Professor of Buddhism and East Asian Religions at McMaster University. He studies Buddhism and Taoism in medieval China. To date, he has focused on three major areas of research: bodily practice in Chinese religions; the ways in which people create and transmit new religious practices and doctrines; and the religious dimensions of commodity culture. He has published on self-immolation, spontaneous human combustion, Buddhist apocryphal scriptures, and tea and alcohol in medieval China. He is the author of Burning for the Buddha: Self-immolation in Chinese Buddhism (University of Hawai'i Press 2007) and is currently completing a second book, Tea in China: A Religious and Cultural History. James A. Benn Department of Religious Studies McMaster University University Hall, Room 105 Hamilton, Ontario L8S 4Kl Canada Thomas DOCTOR has pursued Buddhist studies at Ka-Nying Shedrub Ling Monastery in Kathmandu and at the University of Copenhagen, where he received his BA and Master's degrees. He has been affiliated with Kathmandu University's Centre for Buddhist Studies since 2001. He is currently employed at the University of Lausanne, where he is preparing a doctoral dissertation on rMa bya Byan chub brtson 'grus's Madhyamaka interpretation. Thomas Doctor Anthropole 4115 Universite de Lausanne Dorigny 1015 Lausanne Switzerland Georges DREYFUS is the Jackson Professor of Religion at Williams College. He is the author of Recognizing Reality: Dharmakfrti and

Journal of the International Association of Buddhist Studies Volume 32 Number 1-2 2009 (2010) pp. 483-487

484

Notes on contributors

its Tibetan Interpretations (1997) and of The Sound of Two Hands Clapping: the Education of a Tibetan Buddhist Monk (2003).
Georges Dreyfus Stetson Hall Williams College Williamstown, MA 01267 USA

Vincent ELTSCHINGER completed his PhD at the university of Lausanne (2003) and has been a research fellow at the Institute for the Cultural and Intellectual History of Asia (Austrian Academy of Sciences, Vienna) since then. After a series of studies dedicated to the Buddhist epistemologists' polemics against Brahmanical orthodoxy, his research interests have shifted to the religious background and apologetic programme of late Buddhist philosophy in India. His publications include monographs on Buddhist polemics against the caste system (2000), Dharmaklrti's views of mantras (2001) and the Buddhist epistemologists' critique of the Mlmarp.sa doctrine of the authorlessness of the Veda (2007).
Vincent Eltschinger Austrian Academy of Sciences Institute for the Cultural and Intellectual History of Asia Apostelgasse 23 1030 Vienna Austria

Frances GARRETT is Associate Professor of Tibetan Buddhist Studies at the University of Toronto. Her research considers the intersections between tantric practice, ritual and occult knowledge and medical theory, and what these tell us about the processes of institutional and ideological change in Tibet. She is the author of Religion, Medicine and the Human Embryo in Tibet (Routledge 2008).
Frances Garrett Department for the Study of Religion University of Toronto 170 S1. George Street, Room 302 Toronto, ON M5R 2M8 Canada

Andrew McGARRITY is Lecturer in South Asian and Indo-Tibetan Studies at The University of Sydney. His specific research interests include Indian Madhyamaka and its Tibetan interpretations, and his general research areas are in the History of Ideas in India and Tibet, especially the hermeneutic issues involved in the transmission of systems of thought and in interactions between tradition and

Notes on contributors

485

modernity. Currently, he is researching theories of ethics and personhood and their implications for Indian and Buddhist traditions.
Andrew McGarrity Department of Indian Sub-continental Studies School of Languages and Cultures' Brennan-MacCallum. A18 University of Sydney' NSW 2006 Australia

Pascale HUGON is a research fellow at the Institute for the Cultural and Intellectual History of Asia of the Austrian Academy of Sciences. Her recent publications include a study on Sa skya Pa1!<:lita's Tshad ma rigs pa'i gter (Tresors du raisonnement, Vienna, 2008) and a number of articles dealing with the Indo-Tibetan Buddhist epistemological tradition.
Pascale Hugon Austrian Academy of Sciences Institute for the Cultural and Intellectual History of Asia Apostelgasse 23 1030 Vienna' Austria

Sarah H. JACOBY is Assistant Professor in the Department of Religious Studies at Northwestern University. She researches Indo-Tibetan Buddhist literature, specializing in Tibetan Buddhist revelation, Tibetan auto/biography, and Eastern Tibetan area studies. Her most recent publications include a co-edited volume titled Buddhism Beyond the Monastery (Brill 2009) in which she published an article titled "To be or not to be Celibate: Morality and Consort Practices According to the Treasure Revea1er Se ra mkha' 'gro's (1892-1940) Auto/biographical Writings." She is currently writing a monograph about the life and works of the Tibetan visionary Se ra mkha' 'gro tentatively titled Love Revelations: The Auto/ biographical Writings of a Tibetan Buddhist I)iikinf. Additionally, she is preparing a full annotated translation of Se ra mkha' 'gro's autobiographical writings.
Sarah Jacoby Department of Religious Studies Northwestern University Crowe Hall 5-179 1860 Campus Drive Evanston, IL 60208-2164 USA

Kazuo KANO is Assistant Professor at the Faculty of Letters, Koyasan University in Wakayama, Japan. He specializes in the Buddha-nature doctrine in India and Tibet, as well as Sanskrit and

486

Notes on contributors

Tibetan old manuscripts transmitted to Tibet. He published articles on these subjects in a number of journals, including the Wiener Zeitschrift fur die Kunde Sudasiens, the Journal of the American Oriental Society, and Tantric Studies. Kazuo Kano Koyasan University Department of Esoteric Buddhism 385 Koyasan, Koya-cho, Ito-gun Wakayama 648-0280 Japan Heidi I. KOPPL has worked as translator at the Ka-Nying Shedrub Ling Monastery in Nepal for more than a decade, and has been a faculty member at the Kathmandu University Centre for Buddhist Studies since the Centre's inception. She is currently doing research for a doctorate degree in Buddhist Studies at the University of Lausanne, Switzerland. Heidi 1. Koppl Anthropole 4115 Universite de Lausanne Dorigny 1015 Lausanne Switzerland Jan-Ulrich SOBISCH is Associate Professor for Tibetan Studies at Copenhagen University. He holds a Dr. phil. from Hamburg University. His research focuses on the reception, acculturation, and transmission of tantric Buddhism among bKa'-brgyud-pas and Saskya-pas. His publications include Three-Vow Theories in Tibetan Buddhism (2002), Life, Transmissions, and Works of A-mes-zhabs Ngag-dbang-kun-dga'-bsod-nams (2007), and Hevajra and Lam'bras Literature of India and Tibet (2008). Jan-Ulrich Sobisch Asian Department Institute of Cross-Cultural and Regional Studies University of Copenhagen Leifsgade 33 2300 Copenhagen S Denmark Jonathan STOLTZ is Assistant Professor of philosophy at the University of St. Thomas in St. Paul, Minnesota. He specializes in contemporary philosophy of logic, as well as Tibetan Buddhist epistemology. He has published articles on Buddhist epistemology in a number of journals, including The Philosophical Quarterly,

Notes on contributors

487

Philosophy East & West, and The Journal of the International Association of Buddhist Studies.
Jonathan Stoltz' University of St. Thomas, JRC 241 2115 Summit Avenue' Saint Paul, MN 55105 USA

Drongbu TSERING is Professor at the Tibetan Academy of Social Sciences in Lhasa. He is the author of several historical works and of the first manual of Western philosophy written in Tibetan.
Drongbu Tsering Tibetan Academy of Social Sciences 4, Sera Road' 850000 Lhasa, AR Tibet PR China

Kevin VOSE is Assistant Professor of Religious Studies at the College of William and Mary. He is the author of Resurrecting Candraklrti: Disputes in the Tibetan Creation of PrasaJigika (Wisdom Publications' Studies in Indian and Tibetan Buddhism, 2009). His current research examines late-Indian and early-Tibetan Buddhists' attempts to integrate models of authority and epistemology with the Madhyamaka view.
Kevin Vose Dept. of Religious Studies' College of William and Mary P.O. Box 8795 Williamsburg, VA 23187 USA

Chizuko YOSIDMIZU is Associate Professor at Tsukuba University in Japan. She studied Indology, Tibetology, and Buddhism at Tokyo and Vienna Universities, and has lectured at several universities in Japan, Europe and the USA. Her publications on Buddhist philosophy include numerous articles and a book entitled Die Erkenntnislehre des Prasangika-Madhyamaka nach dem Tshig gsal ston thun gyi tshad ma'i rnam Mad des 'Jam dbyans biad pa'i rdo rje (Vienna 1996).
Chizuko Yoshimizu Section of Philosophy Graduate School of Humanities and Social Sciences Tsukuba University Tsukuba-shi Tennodai 1-1-1 305-8571 Japan

The International Association of Buddhist Studies

Prof. Oskar von Hintiber (Freiburg) President Prof. Cristina Scherrer-Schaub (Paris) Vice-President Prof. Tom Tillemans (Lausanne) General Secretary Dr. Jerome Ducor (Lausanne) Treasurer Regional Representatives: Prof. Janet Gyatso (Cambridge, MA.) Prof. Kazunobu Matsuda (Kyoto) Prof. David Seyfort Ruegg (London) Mr. Peter Skilling (Bangkok) Directors at large: Dr. Nalini Balbir (Paris), Prof. Georges Dreyfus (Williamstown, MA.) Prof. Robert M. Gimello (Notre Dame, IN), Prof. Paul Harrison (Stanford) Dr. Helmut Krasser (Vienna), Dr. Li-ying Kuo (Paris) Prof. Ulrich Pagel (London), Prof. Richard Salomon (Seattle) Prof. Lambert Schmithausen (Hamburg), Mr. Peter Skilling (Bangkok) Prof. Jikido Takasaki (Tokyo)
The International Association of Buddhist Studies, founded in 1976, is devoted to promoting and supporting scholarship in Buddhist Studies in all its aspects, past and present, around the world. Membership is open to scholars of all academic disciplines. Membership dues are USD 55 for full members, USD 30 for student members, USD 1300 for life members. Subscriptions to the JIABS for libraries and other institutions are USD 90. Dues and subscriptions may be paid in the following ways: 1. NEW: Online by Paypal (http://iabsinfo.net/index.htm#payment) 2. by direct bank to bank transfers in US dollars or in the equivalent amount in Euros or Swiss francs. The transfers should be to the lABS accounts at the Banque Cantonale Vaudoise, PI. St-Fran<;:ois, Ch-1001 Lausanne, Switzerland (SWIFT CODE: BCVLCH2L, clearing no. 7677), account numbers, 983.51.04 for US dollars, 5042.09.82 for EUR and 983.51.02 for Swiss francs. 3. by Visa or Mastercard 4. by cheque payable to the "Association Internationale d'Etudes Bouddhiques (lABS)." There will be no supplementary charges for cheques drawn in Swiss currency on Swiss banks. Otherwise, please add 10 dollars or 15 Swiss francs to cover our processing charges. 5. by Euro cheque in Swiss francs. Prospective members from developing countries may contact the Treasurer concerning subsidized membership rates. Dues and subscriptions are payable per calendar year by December 31 of the previous year. Payments other than direct bank transfers should be sent to Dr. Jerome Ducor, Dept of Oriental Languages and Cultures, Anthropole, University of Lausanne, CH-1015 Lausanne, Switzerland.