Vous êtes sur la page 1sur 540


Journal of the International

Association of Buddhist Studies

Volume 31 Number 12 2008 (2010)

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. JIABS is published twice yearly. Manuscripts should preferably be submitted as e-mail attachments to: editors@iabsinfo.net as one single le, complete with footnotes and references, in two dierent formats: in PDF-format, and in Rich-Text-Format (RTF) or OpenDocument-Format (created e.g. by Open Oce). Address books for review to: JIABS Editors, Institut fr Kultur- und Geistesgeschichte Asiens, Prinz-EugenStrasse 810, A-1040 Wien, AUSTRIA Address subscription orders and dues, changes of address, and business correspondence (including advertising orders) to: Dr Jrme Ducor, IABS 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: +41 21 692 29 35 Subscriptions to JIABS are USD 55 per year for individuals and USD 90 per year for libraries and other institutions. For informations on membership in IABS, 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. Print: Ferdinand Berger & Shne GesmbH, A-3580 Horn


Journal of the International Association of Buddhist Studies
Volume 31 Number 12 2008 (2010)

Obituaries Jonathan A. SILK In memoriam, Erik Zrcher (13 Sept. 1928 7 Feb. 2008) . . . . . . 3 Articles Diwakar ACHARYA Evidence for Mahyna Buddhism and Sukhvat cult in India in the middle period Early fth to late sixth century Nepalese inscriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Early Chinese Buddhist translations
Contributions to the International Symposium Early Chinese Buddhist Translations, Vienna 18 21 April, 2007

Guest editor: Max Deeg Max DEEG Introduction 79

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Max DEEG Creating religious terminology A comparativeapproach to early Chinese Buddhisttranslations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Hubert DURT Early Chinese Buddhist translations Quotations from the early translations in anthologies of the sixth century . . . . . . . . . . . . . 119 Toru FUNAYAMA The work of Paramrtha: An example of Sino-Indian crosscultural exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141


Andrew GLASS Guabhadra, Boyn, and the Sayuktgama . . . . . . . . . . 185 Paul HARRISON Experimental core samples of Chinese translations of two Buddhist Stras analysed in the light of recent Sanskrit manuscript discoveries . . . . . . . . . . . . . . . . . . . . . . . . . 205 Elsa I. LEGITTIMO Reopening the Maitreya-les Two almost identical early Maitreya stra translations in the Chinese Canon: Wrong attributions and text-historical entanglements . . . . . . . . . . . . . . . . . . . . 251 Jan NATTIER Who produced the Da mingdu jing (T225)? A reassessment of the evidence . . . . . . . . . . . . . . . . . . . . . . 295 Jungnok PARK () A new attribution of the authorship of T5 and T6 Mahparinirvastra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 Jonathan A. SILK The Jifayue sheku tuoluoni jing Translation, non-translation, both or neither?. . . . . . . . . . . . . . . . . . . . . . . . 369 Stefano ZACCHETTI The nature of the Da anban shouyi jing T 602 reconsidered . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421 ZHU Qingzhi On some basic features of Buddhist Chinese . . . . . . . . . . . 485 Book review Tsunehiko SUGIKI David B. Gray, The Cakrasamvara Tantra (The Discourse of r Heruka): A Study and Annotated Translation. . . . . . . . . . . . . . . . 505 Notes on the contributors
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


In memoriam Erik Zurcher

(13 September 1928 - 7 February 2008)

Jonathan A. Silk

Erik Zurcher was born in Utrecht, in the center of the Netherlands, where he was educated through secondary schooP Although he originally intended to study Egyptology, when he came to the University of Leiden he began the study of Sinology with Gan Tek Chiang, later curator for the Chinese department of the National Museum of Ethnology (Rijksmuseum voor Volkenkunde) in Leiden. Zurcher was soon invited to join more advanced classes with Jan Julius Lodewijk (J. J. L.) Duyvendak (1889-1954). During this time his interests in art led him to Sweden, where he worked with Osval Siren (1879-1966), this resulting in one of his first publications, "Imitation and Forgery in Ancient Chinese Painting and Calligraphy," Oriental Art (1956): 141-156. Later he was to publish a few other papers also concerned with art and material culture, although this never became a major research interest. At Leiden
1 I have been fortunate to be able to make use of a variety of materials including Stephen Teiser's Foreword to the third edition of Zurcher's Buddhist Conquest ("Social History and the Confrontation of Cultures"), Tim Barrett's "Erik Zurcher, 1928-2008: Buddhism and the European Understanding of China" (The China Quarterly 196 [December 2008]: 919-923), the memorial note by my colleague Barend ter Haar found at http://www.hum.leidenuniv.nl/medewerkers/forum/index-l08/im-zurcherengl-108.html, and the remarks of Wilt Idema in Levensberichten en herdenkingen 2009 of the Koninklijke Nederlandse Akademie van Wetenschappen (Amsterdam: Koninklijke Nederlandse Akademie van Wetenschappen,2009): 100-108.

Journal of the International Association of Buddhist Studies Volume 31 Number 1-2 2008 (2010) pp. 3-22

Jonathan A. Silk

Zurcher worked under Duyvendak's successor, tp.e historian Anthony Francois Paulus Hulsewe (1910-1993), who served as his doctoral supervisor. In 1961 Zurcher himself took up what had been the chair of Colonial History, renamed "Far Eastern History, in particular the contacts between East and West." Important formative influences included Zurcher's study in Paris with Paul Demieville (1894-1979) and his friendship, if not rivalry, with Jacques Gernet, whose interests were so very similar to his own in many respects. From 1976 to 1992 Zurcher was co-editor together with Getnet of T'oung Pao, which had always been a joint Leiden-Paris effort, and remains the oldest continuously published sinological journal. In the preface to his Buddhist Conquest of China (see below), Zurcher also mentioned his appreciation of his "honoured friends Et[ienne] Balazs ... and P[iet] van der Loon" and his "commilitones A[rthur] F. Wright ... and L[eon] Hurvitz." These connections with the most excellent ranks of Sinologists and students of Chinese Buddhism were clearly important for his trajectory as a scholar. Zurcher was, among other things, a member of the Royal Netherlands Academy of Arts and Sciences (from 1975) and Correspondant etranger de l'Academie des Inscriptions et Belles-lettres in Paris (from 1985), and his close connections with the tradition of French sinology are thus obvious in mUltiple dimensions. It is impossible to come to terms with any full scholarly life, certainly that of a great scholar like Zurcher, in the few words alloted for a memorial, and the bare facts of his career offer little hint to Zurcher's impact on his chosen fields of study. One thing to be made clear is that, despite the contributions discussed below, Zurcher himself would not have characterized his field as Buddhist Studies pur sang, but rather perhaps as Chinese History, with a focus on the integration and naturalization of the foreign into Chinese culture. In particular, at least in so far as they are preserved in the form of written documents, Zurcher's scholarly interests were almost equally split between (earlier) Chinese Buddhism and Christianity in China. Here I will attempt nothing more than a brief appreciation of his contributions to the field of Chinese Buddhism. 2 The

For an appraisal focused on Zurcher's work on Christianity in

Erik Ziircher (13 Sept. 1928 - 7 E::b. 2008)

importance of these contributions may be capsulized by saying that they were, in a much overused but here entirely apt characterization, seminaL They are represented, however, relatively sparsely: one monograph, a small general book, a volume of translation and some thirty papers. But oh, what a monograph! This study, with which he launched his career, is of course the monumental The Buddhist Conquest of China, being his doctoral thesis, first published in 1959 and reprinted in 1972 and 2007. 3 The main theme of Zurcher's research was cultural interaction, which in the case of Buddhism meant the ways in which this foreign religion found, or forged, a home in China. Or perhaps this is the wrong way to put it for Zurcher, since he was not interested in seeing the process - pace the provocative title of his book - from the point of view of the foreign invaders, as it were, but on the contrary squarely from the side of the Chinese themselves. 4 As will be men-

China, see Nicolas Standaert, "Erik Zurcher's Study of Christianity in Seventeenth-Century China: An Intellectual Portrait." In the press in China Review International 15/4 (2010): 472-502. 3 These are referred to by the publisher as new editions, but in fact are virtually identical as far as content is concerned, only the typography having been updated: in the second edition Zurcher's elegant but poorly reproduced calligraphy was replaced with type-set Chinese characters; in the third edition the whole text is reset and the romanization changed to Pinyin, but the pagination is retained. It is a pity that this recent reprint appears not as the original in two volumes, with text and notes separated, but in a single binding, making reference to the extensive notes laborious. 4 From another point of view, see also Arthur Wright's comment in his review of Buddhist Conquest (see 2d in the list of publications): "The title is surely a misnomer, for this is a close and careful study of two cultures - Chinese and Buddhist - interacting with one another, with neither, at the period's end, 'conquering' the other. I doubt that military metaphor is ever applicable to studies of culture contact and acculturation. It certainly is not here." Zurcher was not altogether unaware of this imbalance of his study. In the preface to the 1972 reprint, he wrote: "The reader may feel that in describing the process of acculturation I have somewhat overstressed the Chinese side. The reader is right: it takes two to acculturate. More stress

Jonathan A. Silk

tioned below, he found the contrast with the case of Christianity in China interesting and challenging. Zurcher approach~d both sets of questions of acculturation (or 'inculturation,' although he did not use this missiological term, as far as I have noticed) entirely from the side of reception. He paid careful attention to early ChInese translations of Buddhist scriptures, for example, but even in his grammatical discussions almost never entered into considerations . of the sources the Chinese (or Central Asian) translators were attempting to render, although he could not avoid such obvious 'issues as the introduction of mid-sentence vocatives, previously unknown in Chinese and appearing only as a calque on Indic sentence order. This concentration on the Chinese reception of Buddhism yields many advantages, but at least as far as translation goes, one cannot escape the conclusion that a rather great deal may be learned by studying both the input as well as the output of the process, as it were, as recent work by Seishi Karashima and Stefano Zacchetti, for instance, attempts to do. It is worth noting that both of these scholars benefitted from Zurcher's advice. Although he did touch upon later Chinese Buddhism in some publications, the lion's share of Zurcher's attention was devoted to the earlier periods, with the fifth century a tacit upper limit (and he more than once explicitly limited his interest to the period between the first and fifth centuries). The primary thrust of Zurcher's research was to build up, stone by stone, as comprehensive a mosaic of early Chinese Buddhism as possible. He began this effort with his Buddhist Conquest of China which was, however, as he acknowledged, largely concerned with literate, socially and politically prominent elites. In his Foreword to the 2007 reprint, Stephen Teiser wrote (p. xv):
The most important sources come from two classes of Chinese Buddhist writing. One class consists of the early biographies of famous monks and nuns and a history of the formation of the Chinese Tripitaka. The second class is what ZUrcher terms "early apologetic

could have been given to the 'donor' side - the way in which the foreign missionaries consciously or unconsciously responded to the Chinese public and its demands."

Erik Zurcher (13 Sept. 1928 - 7 Feb. 2008)

and propagandistic literature," that is, works written by Buddhist devotees, both lay and monastic, designed to defend the faith from the criticisms of its cultured despisers .... [W]e should pause to note what Zurcher is leaving out and to appreciate the weight of the unannounced tradition that he is arguing against. Zurcher intentionally ignores the great number of texts in the Chinese Buddhist canon that were translated during this period from Sanskrit and other Indian languages. As Zurcher writes elsewhere, the canon is an embarrassment of riches; its sheer volume seems to suggest how well it represents Chinese Buddhism. Teiser's points here are, first, that Zurcher's sources belong to the small slice of elites at the top of the Buddhist pyramid, second that the texts in question are mostly self-consciously propagandistic, and last that he turns away for the most part from translations in favor of native compositions. It could certainly be argued that to a very great extent these choices alone strongly determined the kind of picture Zurcher was able to paint. Teiser goes on to point out hOw reliant Zurcher is on the work of the Chinese scholar Tang Yongtong ~ffl~ and his History of Buddhism during the Han, Wei, Two lin, and Northern and Southern Dynasties (Han Wei Liangjin Nanbeichao fojiao shi >lft mElmjc'A{~fj(~), published in 1938. But he concludes, interestingly (p. xix): In the end, I believe that Zurcher's reliance on Tang's work is in fact a strength. As an in-depth, modern guide through the complexities and problems of the primary sources, Tang's book remains the indispensable starting point for any serious work in early Chinese Buddhism. Zurcher's use of Tang is a testament to the interconnections between two great traditions of modern scholarship, both a sign of the past and an augur for the future. 5

5 It should further be pointed out that unlike the work of Tang or the Japanese scholar Tsukamoto Zenryfi ~*~~, Zurcher was compelled not to quote his sources in the original Chinese but to offer an interpretation in the medium of translation, a far from trivial task with such difficult materials. From this point of view, even a mere translation of Tang's study, along the lines perhaps of Hurvitz's rendering of Tsukamoto's Chugoku Bukkyo Tsushi $1lI{~~lm~ (1968) as A History

Jonathan A. Silk

Teiser further points out that "since 1959 no original work in a western language broadly covering the same period of Chinese Buddhist history has been published." This is certainly due to several factors - the excellence of Zurcher's book, a growing recognition of the availability of previously unexplored sources (niuch of this awareness in its turn thanks to Zurcher's own further studies), and an appreciation of the difficulty of such comprehensive surveys. For, tackling big questions requires big theoretical assumptions, which are out of favor in some quarters these days, especially among the more philologically minded who dare to delve into the very difficult old materials which provide the fodder for such research. Teiser devotes a number of pages of his Foreword to criticism of Zurcher's book, some of which concerns these very theories. In this respect, one point of interest is that although Teiser notes in his bibliography Arthur Wright's review of Buddhist Conquest, implicitly noted by Zurcher himself in his 1972 Preface when he avers that he would not again use the term 'gentry,' for example, Teiser does not anywhere actually refer to Wright's review, nor, as far as I know, did Zurcher himself explicitly acknowledge the sometimes detailed critiques his book evoked from Wright and others. He did write in the 1972 edition: "It goes without saying that a new version would bear the marks of beneficial criticism, made by masters and colleagues in reviews and personal correspondence," and the reprint contains two pages of corrections of Zhou Yiliang (hidden after the index). But G. E. Sargent's corrections of Zurcher's translations, for instance, are passed over in silence by all concerned. This, however, certainly does not mean that Zurcher considered the work begun in Buddhist Conquest completed by that work. In fact, he evidently regarded this study as a mere beginning, in part because of the range of materials which it considered; he more than once characterized the types of information available from

of Early Chinese Buddhism, From Its Introduction to the Death of Huiyuan (Tokyo: Kodansha International, 1985), would have constituted a substantial contribution. What Zurcher accomplished is, however, much more than this.

Erik Zurcher (13 Sept. 1928 - 7 Feb. 2008)

written sources as not only biased but distorted out of all proportion. But such "official" sources of information are not in fact the only ones available. A similar distinction has been pointed out vividly for a later period by the lamented Antonino Forte in hi.s review of Stanley Weinstein's Buddhism Under the T'ang, in which Forte praises Weinstein for exhausting traditional sources, but critiques him for, inter alia, ignoring other materials,6 many of which Forte himself mined to such brilliant ends. For the earlier periods under Zurcher's lens, however, even such unofficial materials are rarely available; how can one learn to see the invisible? His thesis seems to have taken the wind out of Zurcher's sails, and through the 1960s and most of the 1970s he published very little on Buddhism, save a couple of general surveys. He was not idle, however, and the next decades reveal the fruit of his work (and of course, he was extremely active on other fronts during this period, including initiating the highly successful Documentation and Research Center for Contemporary China). In 1980 Zurcher published his survey "Buddhist Influence on Early Taoism: A Survey of Scriptural Influence," followed quickly by "Eschatology and Messianism in Early Chinese Buddhism" and '''Prince Moonlight': Messianism and Eschatology in Early Medieval Chinese Buddhism." These papers reveal more than Zurcher's vast reading in the Daoist canon (of which he seems to have made little use thereafter). While they certainly stand as a contribution to Daoist Studies, I read them differently. I think they mark the first sustained effort to try to overcome the horrible imbalance Zurcher lamented in his evidence about early Chinese Buddhism. The problem to which he returned again and again is how to squeeze from sources which do not explicitly deal with Buddhism information nevertheless relevant to its reception in China. In early texts of Daoism he thought he found a way to backlight, as it were, the types of concerns which could only have shone or reflected onto Daoist surfaces from otherwise invisible Buddhist faces. In other words, what he looked for in seeking out Buddhist

Toung Pao, 2nd Series, 75/4-5 (1989): 317-324.


Jonathan A. Silk

influences on Daoism were Buddhist elements in contemporary Chinese society which left no palpable trace elsewhere. Almost as with the fossilized impression of a dinosaur's skin left in mud, Zurcher sought in these papers to learn about Buddhism by studying the impressions it made on another object, in this case, the formative thought of Daoism. Few efforts have been made to follow up this methodological insight. Despite his rather clear, if tacit, presupposition of the nature of correCt and proper normative Buddhism, Zurcher is especially interested in what he does not find reflected in Daoist texts: the "complete absence of typically scholastic terminology" indicates "a very low level of doctrinal sophistication" (Buddhist Influence p. 119). "Taoism," he writes further, "was not influenced by 'professional' Buddhism, but through the distorting and simplifying filter of lay Buddhism; we must assume that the human contacts which formed the channel of transmission must not be sought in the monasteries or the ch'ing-t'an salons where learned monks were present to explain the doctrine ... but rather in lay society where Taoists and Buddhist devotees met. ..." (p. 143). This leads him to conclude (p. 146) that "the selection of Buddhist ideas, particularly at the level of complex borrowings that we find in Taoist literature, gives a very valuable clue as to what ideas were the 'focal points' in Buddhism with the strongest appeal - so strong that they could influence Chinese thought beyond the limits of the Buddhist community and be accepted by its greatest riva1." He immediately continues:
But if we agree with the conclusion drawn above, that Taoism in fact got its Buddhist impulses from lay Buddhism, the information is even more valuable. We actually know very little about that sector of Buddhist religion in mediaeval China .... It could well be that a more detailed analysis of Buddhist complexes mirrored in Taoist literature could teach us much about contemporary lay Buddhism, in spite of all misunderstandings and distortions. But perhaps such misunderstandings and distortions were also widely spread among the simple Buddhist believers themselves. Perhaps we are - as so often happens - handicapped by the fact that we can only observe Buddhism and Taoism at the very highest level, that of the religious "professionals" and their written texts - the tops of two pyramids. We may consider the possibility that at a lower level the bodies of the pyramids merged

Erik Zurcher (13 Sept. 1928 - 7 Feb. 2008)


into a much less differentiated lay religion, and that at the very base both systems largely dissolved into an indistinct mass of popular beliefs and practices.

In his studies on eschatology, Zurcher directed his attention to materials which either fell below the radar of the official arbiters of Buddhist norms, or which were actively suppressed by them. He linked these in a number of cases to the Buddho-Daoist substrate which he postulated to run beneath, as it were, the high traditions as a common river. And this pattern can be detected in other studies as well, although perhaps not in exactly the same manner. For example, Zurcher devoted a number of studies to the earliest translations and translators. On the one hand, this topic involves the court and official sanction or canonization of translations. At the same time, much translation, and other scripture production, as withcertain eschatological texts, took place outside of and alongside official channels. Zurcher's interest in these translations extended to the language in which they were composed, and while it is hard to say which came first, this interest in the translations as preserving evidence of the early sources of Chinese Buddhism also proved to be a key to thinking about early vernacular Chinese language. The way in which texts were rendered from foreign tongues by those outside the educated literate elite allowed Zurcher to speculate that it is possible to discern traces of "the living language of second century Loyang" among these earliest works by An Shigao and a few others. The problems were naturally not only linguistic; as Zurcher says in ''A New Look at the Earliest Chinese Buddhist Texts" (p. 278), "The question will be to what extent, and in what ways, these archaic translations can be made to yield information about the intellectual and social context of the very first stage of Chinese Buddhism." I mentioned above that despite few publications on Buddhism during this period, Zurcher was evidently not idle during the 1960s and 1970s. One fruit of this period remains almost unknown, and might even appear to be a non-scholarly product. That is his Ret leven van de Boeddha (Life of the Buddha) of 1978. Published in a popular series, without a single Chinese character in evidence, this is nevertheless a monument of learning, an integral translation of


Jonathan A. Silk

the Xiuxing benqi jing ~~1'T.2/=m~ (T. 184) and Zhpng benqi jing c:p.2/=m.i[ (T. 196), the two-part hagiography dating to the second century. That this work remains nearly unknown is due, without doubt, entirely to the fact that it is written in Dutch. It is impossible to discuss all of Zurcher's papers here, of course.? Nevertheless, it is good to draw attention to his important remarks on "Buddhism and Education in T'ang Times" and "Buddhist Art in Medieval China: The Ecclesiastical View," subjects which are hardly noticed by other scholars, or at least not by scholars sufficiently equipped to deal with them. (Recent publications on Chinese Buddhist art, at least by scholars writing in English, seem all too frequently to highlight the vast chasm separating those who specialize in visual culture from those trained to read written sources. Scholarship by the fonner in particular sometimes contains statements that make a textual scholar cringe.) I referred above to Zurcher's parallel interest in Christianity in China. These two interests were clearly not distinct for him and, on the largest scale, they form two poles, as it were, of a common problem. In fact, Zurcher explicitly confronts this issue in a short but extremely interesting paper translated into English as "The Spread of Buddhism and Christianity in Imperial China: Spontaneous Diffusion Versus Guided Propagation." Here Zurcher attempts to understand why it is that the foreign religion Buddhism succeeded in implanting itself in Chinese soil, and why Christianity, in the form of Roman Catholicism, failed. Zurcher's conclusion is as follows:
We find a whole range of contrasts and oppositions. Spontaneous infiltration versus guided introduction. Unprepared roaming monks versus well-trained missionaries. Monastery versus church and mission house. Free Buddhist laity versus bound Christian converts. Pluriformity versus uniformity. Indigenous sources of income versus external funding. Homogenous status versus a broken, dissonant role

7 A more comprehensive version of the present short appreciation will be included as the introduction to the volume I am now editing in which most of these papers will be reprinted; it will be published from Brill in the near future.

Erik ZUrcher (13 Sept. 1928 - 7 Feb. 2008)


pattern. Many oppositions, that yet have one thing in common: they all are related to the character of the Jesuit mission as a guided process. And that is the great paradox. Planning and guidance were factors of weakness, whereas Buddhism was strengthened by the very absence of planning and central guidance, by its spontaneous and totally uncoordinated development.

There are naturally some topics which Zurcher intended to address but never did. In his "Eschatology and Messianism," for example (p. 42) he promises to examine theories of mofa ('decline of the teaching'), a project of which we hear no more. Perhaps the most disappointing loss is the apparent disappearance of a draft grammar of Kumarajlva's translation of the Lotus Sutra. I have been assured of its one-time existence by my colleague Barend ter Haar, but am so far unable to locate a copy. It must come as something of a surprise to realize that Zurcher only directed two doctoral theses on Buddhism, those of Barend ter Haar (published as The White Lotus Teachings in Chinese Religious History [1992]), and Valentina Georgieva, "Buddhist Nuns in China from the Six Dynasties to the Tang" (2000, regrettably still unpublished; the advisor was Tilmann Vetter, with Zurcher as co-advisor). Through his published writings, however, he leaves a much greater number of students around the world, who join with his more direct disciples in mourning his passing.


Jonathan A. Silk

Erik ZUrcher (Xu Lihe ~t:E.lfD /



1. "Zum Verhaltnis von Kirche und Staat in China wahrend der

Frtihzeit des Buddhismus." Saeculum 10/1: 73-81. 2. The Buddhist Conquest of China: the Spread and Adaptation of Buddhism in Early Medieval China. Sinica Leidensia 11 (Leiden: E. J. Brill: 1959; 2nd ed. 1972; 3rd ed. 2007, with a new Foreword by Stephen Teiser, pp. xiii-xxxvii). . 2a. Japanese: Eriku Churuhya x- I) '/' T::L Iv 1:: -t' - , Bukkyo no Chugoku Denrai {b~G')r:pOO{~*. Trans. Tanaka Sumio ES r:pMl5lj, Watarai Akira 5M<~Ji, Naruse Yoshinori f&~.l! B':tf, Tanaka Fumio ESr:p:X:ME. (Tokyo: Serika shobo 1 l)iJ~i:fm, 1995). With a new preface (jobun ff:X:) by Zurcher dated Dec. 1993, pp. 9-16. Reviewed: Hirai Yukei 3f#1i~, Buzan gakuhO ':'Wq:& 40 (1997): 186-195. 2b. Chinese: Fojiao zhengfu Zhongguo: Fojiao zai Zhongguo zhonggu zaoqi de chuanbo yu shiying {~~1jHa r:p ~: {~~ ttr:p~r:pc!l!f.Mi'l'j{itl~~H!. Trans. Li Silong *l21lj~, Pei Yong ~~, et al. Haiwai Zhongguo yanjiu congshu 54llY'~r:pOO tiH%~i:f. Nanjing ijg: Jiangsu renmin chubanshe 5Ii*A~ II!1lH, 1998,2003,2005. 2c. Partial translations into Korean by P'yo ChOnghun *J&lb and Choi Yun-sik remain unpublished. 2d. Reviews: Hans Steininger, "Der Buddhismus in der chinesischen Geschichte (Zu den Arbeiten von E. Zurcher, 1 Gernet und A. F Wright)." Saeculum 13 (1962): 132-165 (esp. 133-143); Paul Demieville, Toung Pao, 2nd Series 47/3-5 (1959): 430-435; Edward Conze, The Middle Way 34 (1960): 173-176; Leon Hurvitz, Journal of the American Oriental Society 80/3 (1960): 277-279; Giuseppe Tucci, East and West 11/4 (1960): 295-296; Jacques Gernet, Journal Asiatique 248 (1960): 408-410; Arthur E Wright, Journal of Asian Studies 20/4 (1961): 517-520; Janusz Chmielewski, Rocznik Orientalistczny 25/2 (1961): 140-144; ConradM. Schirokauer,American Historical Review 67/1 (1961): 139-140; K[atherine] pro] K[an] Whitaker, Bulletin of the School of Oriental and African Studies 24/3 (1961): 599-600; T[imoteus] Pokora, Archiv

Erik Zurcher (13 Sept. 1928 - 7 Feb. 2008)


Orientalnf 29 (1961): 172-174; G[alene] E[ugene] Sargent, Orientalische Literaturzeitung 57/1-2 (1962): 82-85; Andre Bateau, Bulletin de rEcole jranfaise d'Extreme-Orient 50/2 (1962): 564-569; J[aroslav] Prusek, Central Asiatic Journal 8/1 (1963): 67-70; 2nd edition: W[erner] Eichhorn,' Orientalisthe Literaturzeitung 71/5 (1976): 509-510; Chuang Shen, Journal of the Hong Kong Branch of the Royal Asiatic Society 13(1974): 175-178; Zhuang Shen iff$, Xianggang Zhongwen daxue zhongguo wenhua yanjiusuo xuebao W~r:p-x.*'!!f:.r:p 00 -x.{bliJf~p.lT'!!f:.~ / Journal of the Institute of Chinese Studies of The Chinese University of Hong Kong 8/1 (1974): 365-367; 1975 Tzevi Shiprin, HaMizrah hehadas / New East, Quarterly of the Israel Oriental Society 1-2 (97-98) (1975): 128-129. 1961 3. Het Boeddhisme: ontstaan en verbreiding in kaart, woord en beeld. Beknopte geschiedenissen van wereldgodsdiensten. Dutch trans. by Max Schuchart (Amsterdam: De Brug-Djambatan). 3a. English: Buddhism: its origin and spread in words, maps and pictures (London: Routledge & K. Paul, 1962). 3b. Danish: Buddhismen. Dens oprindelse og udbredelse i tekst, kort og billeder (Copenhagen: Munksgaard, 1962) 3c. Swedish: Buddhismen: ursprung och utbredning (Stockholm: Natur och Kultur, 1962) 1964 4. "Inleiding," "China" and "Vertaling van enkele Zen-teksten." In: F. Vos and E. Zurcher, Spel zonder snaren. Enige beschouwingen over Zen (Deventer: Uitgeverij N. Kluwer). 5. "Buddhism in China." In: Raymond Dawson (ed.), The Legacy of China (Oxford: Clarendon Press): 56-79. 5a. Dutch: "Het boeddhisme in China." In: Raymond Dawson (ed.), De Chinese Traditie. Chinese Biblioteek. Studies en Teksten 3 (Amsterdam: Uitgeverij de Arbeidserspers, 1973): 68-90. Trans. C. E. van Amerongen-van Straten. 1968 6. "The Yiieh-chih and Kani~ka in the Chinese Sources." In: A.L. Basham, ed., Papers on the Date of Kani~ka, Submitted to the Conference on the Date of Kani~ka, London, 20-22 April, 1960.


Jonathan A. Silk Australian National University Centre of Oriental Studies, Oriental Monograph Series 4 (Leiden: E.J. Brill): 346-390.

1977 7. "Late Han Vernacular Elements in the Earliest Buddhist Translations." Journal of Chinese Language Teachers Association 12: 177-203. 7a. Chinese: "Zuizao de fojing yiwen zhong de Donghan kouyu chengfen" m!:!f.S'j{5I!3~Ilfg':X:CPB/;J*5~D~RX::51-. Trans. Jiang Shaoyu Mi~t?~. Yuyanxue luncong ~~~~fi 14 (1987): 197-225. 8 1978 8. Het leven van de Boeddha. De Oosterse bibliotheek 10 (Amsterdam: Meulenhoff). Illustrations credited to SjefNix (Zurcher himself). Translation of Xiuxing benqi jing {~1'T*!lli~ and Zhong benqi jing CP*!lli~. 1980 9. "Buddhism in a Pre-modern Bureaucratic Empire: the Chinese experience." In: A.K. Narain, ed., Studies in the History of Buddhism. Papers Presented at the International Conference on the History of Buddhism at the University of Wisconsin, Madison, WIS, USA, August, 19-21, 1976 (Delhi: B. R. Publishing Corporation): 401-411. 10. "Buddhist Influence on Early Taoism: A Survey of Scriptural Influence." Toung Pao 66: 84-117. 1981
11. "Eschatology and Messianism in Early Chinese Buddhism." In: Wilt Idema, ed., Leyden Studies in Sinology: Papers Presented at the Conference Held in Celebration of the Fiftieth Anniversary of

8 See: Chen Xiulan ~*j'M, "Dui Xu Lihe jiaoshou 'Zuizao de fojing yiwenzhong de Donghan kouyu chengfen' yiwen de jidian buchong"

!T*t:f"fD~~ {m!:!f.S/;J{5I!3~g':X:CPS/;J*D~RX::51-> -:x:s'j~!MjJE,

in Guhanyu yanjiu cj~liJf% 2 (1997); 55-57. More recently, Zhang Chunxiu 5K~j', "Dui Xu Lihe jiaoshou 'Zuizao de fojing yiwenzhong de Donghan kouyu chengfen' yiwen de zaibuchong" :x]i,f:f"fD~~ {m!:!f.S/;J {5I!3~~:x:cps/;J*&Di.gRX::51-> -:x:s/;Jf!fW3'E, in Hechi xueyuan xuebao 5EJ 5fu~~J'G~f~ 28/1 (2008).

Erik ZUrcher (13 Sept. 1928 - 7 Feb. 2008)


the Sinological Institute of Leyden University, December 8-12, 1980 (Leiden: E. J. Brill): 34-56.

1982 12. "'Prince Moonlight': Messianism and Eschatology in Early Medieval Chinese Buddhism." T'oung Pao 67: 1-75. 13. "Perspectives in the Study of Chinese Buddhism." Journal of the Royal Asiatic Society 1982: 161-176. 14. "Chinese Chan and Confucianism." In: H. Brinker, R. P. Kramers, C. Ouwehand, eds., Zen in China, Japan, East Asian Art. Swiss Asian Studies. Research Studies 8 (Bern: Peter Lang): 29-46. 1984 15. '''Beyond the Jade Gate': Buddhism in China, Vietnam and Korea." In: Heinz Bechert and Richard Gombrich, eds., The World of Buddhism (New York: Facts on File. Reprint: NY: Thames and Hudson, 1995): 193-211. 15a. German: "Buddhismus in China, Korea und Vietnam." Der Buddhismus: Geschichte und Gegenwart (Munich: c.ll. Beck, 1989): 215-251. Trans. Siglinde Dietz.
1985 16. "Maha-CIna: la n~interpretation bouddhique de l'histoire de la Chine." Academie des Inscriptions et Belles-Lettres, Comptes Rendus des seances de l'annee 1985 Juillet-Octobre, 477-492.

1987 17. "Buddhism in China." In: Mircea Eliade, ed., The Encyclopedia of Religion (New York: Macmillan): 2.414a-421a. 17a. Reprinted in Joseph M. Kitagawa and Mark D. Cummings, eds., Buddhism in Asian History (New York: Macmillan, 1989): 139-150. 17b. "II buddhismo in Cina." In: Giovanni Filoramo, ed., Storia delle religioni 4. Religioni dell'India e dell'Estremo Oriente (Bari: Editori Laterza, 1996): 369-410. This in its tum was reprinted in Giovanni Filoramo, ed., Buddhismo (Bari: Laterza, 2001): 185-236. This appears to be an augmented Italian translation the encyclopedia entry. 18. '~mitabha." In: Mircea Eliade, ed., The Encyclopedia of Religion (New York: Macmillan): 1.235a-237b. 2nd ed., Lindsay Jones, ed., (Detroit: Thompson Gale, 2005) 1.291a-293a.


Jonathan A. Silk 19. "Buddhism, Schools of: An Overview." In: M}rcea Eliade, ed., The Encyclopedia of Religion (New York: Macmillan): 2.440a444a. 20. "Buddhist Missions." In: Mircea Eliade, ed., The Encyclopedia of Religion (New York: Macmillan): 9.570a-573b.

1988 21. "China, boeddhisme en christendom: spontane en geleide expansie." Streven 55: 913-925. 1989 22. "Buddhism and Education in T'ang Times." In: Wm. Theodore De Bary et aI., eds., Neo-Confucian Education: The Formative Stage (Berkeley: University of California Press): 19-56. 23. "The Impact of Buddhism on Chinese Culture in an Historical Perspective." In: Tadeusz Skorupski, ed., The Buddhist Heritage. Papers Delivered at the Symposium of the Same Name Convened at the School of Oriental and African Studies, University of London, November 1985 (Tring: Institute of Buddhist Studies): 117-128. 1990 24. "Han Buddhism and the Western Regions." In: W. L. !dema and E. Zurcher, eds., Thought and Law in Qin and Han China: Studies Dedicated to Anthony Hulsewe on the Occasion of His Eightieth Birthday. Sinica Leidensia 24 (Leiden: E.J. Brill): 158-182. 24a. Chinese: "Handai fojiao yu xiyu" ~{-t{~~:Wj2t~. Trans. Wu Xuling ~~~~. Guoji Hanxue ~~~~~ 2 (1998): 291310. 25. "Bouddhisme et christianisme." In: E. Zurcher, Bouddhisme, Christianisme et societe chinoise (Paris: Julliard): 11-42. 25a. Dutch: See 21 above. 25b. English: "The Spread of Buddhism and Christianity in Imperial China: Spontaneous Diffusion Versus Guided Propagation." In: China and the West (Proceedings of the International Colloquim Held in the Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van Belgie, Brussels, November 23-25, 1987). Brussels: Paleis der Academien, 1993: 9-18.

Erik ZUrcher (13 Sept. 1928 - 7 Feb. 2008)


26. "Religieuses et couvents dans 1'imcien bouddhisme chinois." In: idem, 43-94. 1991 27. "A New Look at the Earliest Chinese Buddhist Texts." In: Koichi Shinohara and Gregory Schopen, eds., From Benares to Beijing: Essays on Buddhism and Chinese Religion in Honour of Prof Jan Yiin-hua (Oakville, Ontario: Mosaic): 277-304. 27a. Chinese: "Guan yu chuqi hanyi fojing de xin sikao" ~m~*)] :!tIj~i~{~Mm~!FJf}(SI,~. Trans. Gu Manlin Rm:)jl0#. Hanyushi yanjiujikan ~~g1:liJf3E$fU 4 (2001): 286-312. 28. "Buddhismus in China: Die Grenzen der Innovation." In: S.N. Eisenstadt, ed., Kulturen der Achsenzeit I (Frankfurt: Suhrkamp): 199-232. 1995 29. "Obscure Texts on Favourite Topics: Dao'an's anonymous scriptures." In: Helwig Schmitt-Glintzner, ed., Das andere China: Festschrift flir Wolfgang Bauer zum 65. Geburtstag (Wiesbaden: Harrasowitz): 161-181. 30. "Aliens and Respected Guests: The Role of Foreign Monks in Early Chinese Buddhism." Transactions of the International Conference of Eastern Studies 40: 67-92. 31. "Buddhist Art in Medieval China: The Ecclesiastical View." In: Karel R. van Kooij and H. van der Veere, eds. Function and Meaning in Buddhist Art: Proceedings of a Seminar Held at Leiden University 21-24 October 1991 (Groningen: Egbert Forst): 1-20. 1996 32. "Vernacular Elements in Early Buddhist Texts: An attempt to define the optimal source materials." In: Vernacularism in Medieval Chinese Texts. Sino-Platonic Papers 71: 1-31. 32a. Chinese: "Zaoqi fojing zhong de kouyu chengfen - queding zuijia yuancailiao de changshi" !f.:!tIj{~~CP8~D~~5t-Jii JE~{~~'@:;M*ot8~tf~, published as an appendix in the Renming Daxue dissertation of Zhu Guanming *ffitBA, Mohesengqilii qingtaidongci yanjiu ~~OJ{'I!r~a1.f'fJf~liJgiiJliJf3E (Beijing: Zhongguo xiju chubanshe cp~~~Ut:BM&H, 2008). 33. Portions of "Ideologies and the First Universal Religions" (pp. 56-63), "Introduction" (pp. 483-484) and, with Fukui-Bunga Fu-


Jonathan A. Silk mimasa, "Early Medieval China" (pp. 508-517), in Joachim Herrmann and Erik ZUrcher, eds., History of Huma~ity: Scientific and Cultural Development. Volume III: From the Seventh Century BC to the Seventh Century AD (Paris: Unesco).

1997 34. "China." In: Karel R. van Kooij and Pauline Lunsingh Scheurleer, eds., A Companion to Buddhist Art. An Introductory Course in Buddhist Art in Asia at Leiden University (Leiden: Werkgroep: . Niet-Westerne Kunst en MaterieIe Cultuur): 50-57. 1999 35. "Buddhism Across Boundaries: The Foreign Input." In: Collection of Essays 1993: Buddhism across Boundaries: Chinese Buddhism and the Western Regions. Incorrectly listed as "by E. ZUrcher, Lore Sander and others." (Sanchung, Taiwan: FoGuang Cultural Enterprise Co.): 1-59. The volume is more correctly: Buddhism across Boundaries Chinese Buddhism and the Western Regions, Collection of Essays 1993. Edited by John McRae and Jan Nattier (Taipei: Fo Guang Shan Foundation).
2001 36. "Xu Guangqi and Buddhism." In: Catherine Jami, Peter M. Engelfriet and Gregory Blue, eds., Statecraft and Intellectual Renewal in Late Ming China: The Cross-Cultural Synthesis of Xu Guangqi (1562-1633) (Leiden: Brill): 155-169. 2002 37. "Tidings from the South: Chinese Court Buddhism and International Relations in the Fifth Century AD." In: Antonino Forte and Frederico Masini, eds., A Life Journey to the East: Sinological Studies in Memory of Giuliano Bertuccioli (1923-2001) (Kyoto: Italian School of East Asian Studies): 21-43. 2006 38. "Buddhist Chanhui and Christian Confession in SeventeenthCentury China." In: Nicolas Standaert and Ad Dudink, eds., Forgive Us Our Sins: Confession in Late Ming and Early Qing China. Monumenta Serica Monograph Series 55 (Nettetal: Steyler Verlag): 103-127.

Erik Zurcher (13 Sept. 1928 - 7 Feb. 2008)


Important Reviews
1977 Richard Robinson, Early Madhyamika in India and China (1967). Indo-Iranian lournal19: 122-124. 1978

Melanges de sinologie offerts a Monsieur Paul Demieville. II (1974). T'oung Pao 64: 114-115, 117-120, 122-124.
1996 Timothy Brook, Praying for Power: Buddhism and the Formation of Gentry Society in Late-Ming China (1993). Journal of the Economic and Social History of the Orient 39/4: 468-471. 1997 Kuo Li-ying, Confession et Contrition dans Ie bouddhisme chinois du Ve au Xe si(xle (1994). T'oung Pao 83/1-3: 207-212. 1998 Antonino Forte, The Hostage An Shigao and his Offspring: An Iranian Family in China (1995). T'oung Pao 84/1-3: 173-177.

. Volumes offered in Zurcher's honor

China's Modernisation: Westernisation and Acculturation. Kurt Werner Radtke and Tony Saich, eds. (Mtinchener Ostasiatische Studien 67) (Stuttgart: Franz Steiner, 1993). Conflict and Accommodation in Early Modern East Asia: Essays in Honour of Erik ZUrcher. Leonard Blusse and Harriet T. Zurndorfer, eds. (Sinica Leidensia 29) (Leiden: Brill, 1993). Words from the West. Western Texts in Chinese Literary Context. Essays to Honor Erik ZUrcher on his Sixty-Fifth Birthday. Lloyd Haft, ed. (CNWS publications 16) (Leiden: Centre of Non-Western Studies, 1993). From ClassicalfLI to "Three Inches High": Studies on Chinese in Honor of Erik ZUrcher. IC.P. Liang and R.P.E. Sybesma, eds. (Leuven-Apeldoom: Garant, 1993). Development and Decline of Fukien Province in the 17th and 18th Centuries. Eduard B. Vermeer, ed. (Sinica Leidensia 22) (Leiden: E.I Brill,


Jonathan A. Silk

1990). [Proceedings of a seminar held in 1986 to celebrate the 25 th anniversary of Zurcher's appointment to the chair in Chinese history.] Function and Meaning in Buddhist Art: Proceedings of a Seminar Held at Leiden University 21~24 October 1991. Karel R. van Kooij and H. van def Veere, eds. (Groningen: Egbert Forsten, 1995). [Dedicated "To the happiness and well-being of Erik Zurcher."]

Evidence for Mahayana Buddhism and . SukhavatI cult in India in the middle period
Early fifth to late sixth century Nepalese inscriptions* Diwakar Acharya

During the last three decades, the perception of Indian Buddhism in the middle period has drastically changed. A few scholars have significantly contributed to bring about this change, and Gregory SCHOPEN is the foremost of them. He has surveyed and analysed large bodies of textual and epigraphical data. He has singled out inscriptions significant for the history of Indian Buddhism in India in the period from the beginning of the Common Era to the fifth/ sixth century, reflected upon them carefully, and matched the inscriptional evidence with textual evidence. In this way, he has convincingly demonstrated that "it is virtually impossible to characterise Indian Buddhism in the middle period ... as in any meaningful sense Mahayana" (p. 12).1 As he remarks, "the Mahayana in India,
An abridged version of this paper was presented as a special lec-. ture under the title "Mahayana Buddhism and SukhavatI Cult in Ancient Nepal" at the 14th biennial conference of the International Association of Shin Buddhist Studies held at Ryokoku University, Kyoto, in June 2009. I am grateful to Paul HARRISON, Shoryu KATSURA, Werner KNOBL, Jan NATTIER, Vincent TOURNIER, and Yuko YOKOCHI for their comments and valuable suggestions on earlier drafts of this article. I would like to thank Arlo GRIFFITHS for improving my English and making valuable remarks . on the final draft.
1 If not specified otherwise, all references to Schopen are from his 2005 collection Figments and Fragments of Mahayana Buddhism in India.

Journal of the International Association of Buddhist Studies Volume 31' Number 1-2 2008 (2010) pp. 23-75


Diwakar Acharya

however, appears to have continued very much on the margins" (p. 11), and "however mainstream the early Mahayana was in China, it was in India constituted of a number of differentially marginalized minority groups" (p. 17). This clearly suggests that we need to pay proper attention to the Buddhist communities livIng in the marginal areas, including Nepal, while dealing with the history of Indian Buddhism of this period. However, Licchavi inscriptions from Nepal, many of them Buddhist, have not been carefully studied, though they have been published several times. SCHOPEN himself refers to two undated Nepalese inscriptions from the seventh century but misses other important ones. 2 So, in this article, I will present some inscriptions from the early fifth to the late sixth century that have not been rightly read and interpreted until now, and make a few observations here and there, attempting to analyse the data in the light of textual evidence. The earliest inscription from India which clearly refers to Amitabha Buddha is the Govindnagar inscription from the time of Huvi~ka, dated 26 of the Kani~ka era (equivalent to 104 or 153 CE), inscribed on the pedestal of an image of Buddha Amitabha. 3 This

2 After the publication of Dhanavajra VAJRACHARYA'S Nepali book on Licchavi inscriptions, all books on the topic are unoriginal; they rely on him for the reading and interpretation of inscriptions. REGMI (1983) who published these inscriptions with an English translation and notes has heavily relied on him. RICCARDI (1980) has tried in an article to study all available Licchavi inscriptions which have to do with the history of Buddhism but, materials being muddled, his study reveals very little and confuses more. Recently LEWIS (2004) has published a study on traces of the Sukhavatl cult in Newar society but, his starting point being SCHOPEN'S conclusions, the historical aspect of Sukhavatl has remained beyond his scope. In the same way, while writing the entry on Nepal in Encyclopedia of Buddhism, he has relied on earlier publications. 3 Though published several times, this inscription was not edited and interpreted properly before SCHOPEN. He reread and translated it in his 1987 paper (now included in his 2005 collection, pp. 247-277). In 1999, FUSSMAN published his own reading of the inscription with a translation which is different in a few places. FUSSMAN'S understanding of the date of the inscription is better than SCHOPEN'S; unlike the latter, he has not

Mahayana Buddhism and SukhavatI cult in India


is the first indirect evidence to the early phase of the Mahayana,

ignored 'va' before '2,' in the first line, and has rightly interpreted it as an abbreviation for var:;iimiisa, the rainy season. Otherwise, I find SCHOPEN'S reading more accurate. However, I am bothered with one thing in the second line of the inscription: the reading p[i]t[-x](~)[-x] and its interpretation as an instrumental singular of pitr. As SCHOPEN has stated, the upper parts of the ak:;aras are broken, leaving only the consonants certain, but the vowel sign on top of the first ak:;ara is still partially visible. So, SCHOPEN has conjectured the first ak:;ara as pi and suggested to read the word as pitn1ii. He himself, however, has noted a negative point against his suggestion: "pitrlJii in epigraphical sources has generally been interpreted as genitive plural" (p. 252). FUSSMAN has tried to get rid of this negative point by suggesting another reading, pite~a, keeping the meaning unchanged. On logical, contextual and palaeographical grounds, I see problems in accepting either one of these conjectures. I find it less likely that the donor is first introduced as the father of his son, and then as the grandson and son of his ancestors. We do not have any parallel for such a description. Instead, what is logically likely is that he is introduced as someone's great-grandson, grandson and son successively. We have parallels for such a description even from one of the Ku~al).a inscriptions from Mathura (Liiders 1961: 194-195, 162). However, this parallel is not from an inscription in Buddhjst Sanskrit but standard Sanskrit, and so, it does not help us to conjecture the word we need. Nevertheless, I propose that the donor is not the father of Sax-caka/Sancaka but a great-grandson. If we look at the undamaged pi in line 3, we can see that the sign of i starts on the top of the frontal bar of pa, extends to the right, curves in, and rises up turning counterclockwise and making a shape resembling to a swan's neck. Now if we look at the proposed pi in line 2, what we see is a stroke starting at the frontal bar of pa and extending to the left without rising up. This sign is very close to the sign of 0, so the ak:;ara at issue must be po. I present here both of these ak:;aras:

Altogether the word in the inscription seems to be potre~a which only means grandson, but the proper term in the language of the inscription should be papotrelJa. I see two possibilities: either we have a case of haplography, I mean, the first pa is dropped, or potra- itself is used here to


Diwakar Acharya

which is not referred to by name in Indian inscripti9ns until the late fifth and early sixth centuries. 4 According to SCHOPEN, "the earliest known [inscriptional] reference to Amitabha prior to the Govindnagar inscription occurred in a fragmentary slab inscription from Saiki ... dated to the end of the seventh century" (p. 247). This is not true. About a century before the Sanci inscription, a Nepalese inscription refers not only to Amitabha in Sukhavati but also to his attendant Bodhisattvas LokeSa and Mahasthamaprapta. Following the proper chronological order, I will deal with this inscription in detail as the last item in this article.

A lady wants to get rid of her female nature In front of the Dha:rp.do caitya/Bhagavanthan in Chabahil (Kathmandu), there lies an important inscription which contains some clues hinting at the nature of Buddhism practiced in Nepal at the very beginning of the fifth century. This is the first half of an original inscription inscribed on the lotus baseS pedestal of a lost image of Mahamuni. 6 Unfortunately, the other half of the lotus is missing.
mean great-grandson, when its original meaning is conveyed by another equivalent term niittika. 4 Schopen 2005: 11. However, in a mixed Indian and Chinese context such an inscriptional reference is found one century earlier (ibid. 13). S This inscribed base, which was placed earlier facing downward, serving as a support to a stone pillar used for offering lamps, in front of the west face of the caitya, is now turned into the right position since 2003, the time of renovation of the caitya. Now that the base was turned into the right position, it is possible to see part of a lotus rising above the base which was under the ground before. See photo on p. 27. 6 Mahi'imuni is generally regarded as an epithet of the historical Buddha, but the situation might be different in our inscription, and it might have been used as an independent substantive, like Sakyamuni, referring to the historical Buddha. When some donation is made to a newly consecrated temple and recorded in an inscription, the proper designation of the de-

Mahayana Buddhism and SukhavatI cult in India


It contained the other half of the inscription with the second half of each line including the year of the religious gift as well as the name of the then ruling king? On palaeographical grounds, VAJRACHARYA

ity in that temple is used, not an adjective. So, there is a high chance that the image of the Buddha referred to in our inscription was worshipped as Mabamuni. It is noteworthy that the second Bahubuddhasiitra from the Mahavastu records Mahamuni as one of the Buddhas (SENART 1897: 230). Also in the versified core of the Dasabhiimika section of the Mahavastu, the name Sakyamuni is used in a similar way, where Sakyamuni is used only once but Mabamuni 15 times. See also fn. 31. 7 VAJRACHARYA relates this inscription to the lime-washed white caitya in front of which it is currently placed. However, the inscription itself does not speak of any caitya/stiipa but of an image of Mahamuni and a community of the nobles (aryasmigha). It appears clear to me that the lime-washed caitya surrounded by several votive caityas and more than one Buddha image was in the south end of a larger vihara complex. The vihara in an inverted V-shape opening to the caitya, which must have suffered damage and got repaired several times in later periods, is now occupied by the Pashupati Mitra High School. A narrow motorable road separates the caitya and present-day school. The school has built new buildings in place of the old ones on one side and rebuilt the old buildings with additional floors on the other sides. Hopefully the original foundation is not yet completely destroyed. The complex also suffered loss to the east side by the construction of the Ring Road; at that


Diwakar Acharya

makes this inscription the first inscription in his book of Licchavi inscriptions arranged in chronological order. And more, following late chronicles, he suggests that it can be dated to the time of Manadeva's great grandfather Vf$adeva, who is described as 'siding with Buddhism' (sugatailsanapak~apiltl) in an eighth century inscription of King Jayadeva and late chronicles. There are in fact some clues in the inscription itself which can help us to guess at its time. First, donative formulas in Licchavi inscriptions after King Manadeva's time never begin' with the expression asyilryl divasapurvilyill?l.B So, this can be taken as one clue to assign it in or before the period of Manadeva. Second, this inscription refers to a Jovian year with the atypical expression milghavar~e kille, but such a reference is not found again in any other Licchavi inscription. This system was abandoned in North India earlier than in the rest of India, though it was still in use in the south until the beginning of the sixth century.9 References to Jovian years appear in Gupta inscriptions only between 475-528 CE lO where we find them in a standardised expression - a month name prefixed with mahil- and compounded with sal?lvatsara. Two more references appear also in Kadamba inscriptions of about the middle of the fifth century, but there the expression is not standardised. 11 The expression in our inscription is still different but is closer to those found in Kadamba inscriptions. Therefore, it is save enough to place it before Manadeva, but there is no proper

time, as local people recall, some votive caityas on the track of the road were pushed inside the caitya complex and minor objects were destroyed. 8 Even during Manadeva's time, it appears only twice, in inscriptions dated to Saka419 (VAJRACHARYA 1973: no. 15, p. 65) and 425 (VAJRACHARYA 1973: no. 16, p. 67). 9 DIKSHIT 1888: 316, fn. 16. 10 See, FLEET 1888, DIKSHIT 1888.
11 The expression pau~e SaT!1vatsare occurs in one of the Halsi grants of Mrgesvaravarman dated in his third regnal year (line 8), and vai.akhe saYJ1vatsare in the other dated in his eighth regnal year (line 10). FLEET (1888: 334, fn. 9) relates the use of the prefix maha- to the heliacal-rising system and absence of it to the mean sign system.

Mahayana Buddhism and Sukhavati cult in India


ground to say that this inscription really belonged to the time of . Manadeva's great grandfather V:r~adeva (circa 410 CE) as VAJRACHARYA suggestedY The first available inscription of Manadeva is dated Saka 381 (459/460 CE) and it does not contain a reference to a Jovian year. Before this date, the Jovian year of Magha fell in Saka 371 (449/450 CE), 359 (437/438 CE), 347 (425/426 CE), and 335 (413/414 CE)P So, the image of Mahamuni with this inscription must have been installed in one of these years.14 The inscription, except the last line, is composed in twelve Anustubh stanzas. The metre has helped me to determine the number ~f missing ak~aras in each line.
(1) durddharair indriyai~ k!tsnii viihyate yair iyam prajii

dasavat tani sandhiiryya k!payii paripfya ta[J?lF 5 [1]

(1) danafla<k~amiivfryadhyanaprajfiiini~eva~at>16

12 In the mediaeval period, the caitya in front of which our inscription is found was called Dharp.do caitya. This has prompted some scholars to relate the caitya with Manadeva's father King Dharmadeva. I think this is a very weak argument in the light of the fact that any stupa/caitya can be named after dharmaldharmariija, and we have a few examples of such names, like the Dhammekha sU7pa in Sarnath and the Dharmarajika stL7pa in Taxila. No doubt, Dharp.do can be imagined as a Newar rendering of Dharmadeva, but it is much more likely that as a name of caitya it refers to the Dharma-god, the Buddha. 13 My calculation of these years with the Jovian year of Magha is based on KETKAR'S table (1923: 195, table 20). 14 An allusion to the Buddha's identity as a Bodhisattva in our inscription can be considered as yet another clue for assigning it to a relatively early date. As Buddhologists and historians have noted, early Buddhist cult images are overwhelmingly referred to as Bodhisattva in their accompanying inscriptions, even when they iconographically represent Buddhas. See SCHOPEN 2005: 116. 15 VAJRACHARYA reads ta and interprets that as a plural, obviously assuming that the visarga has been dropped by irregularly observing sandhi between two verses. 16 The acts of the Buddha are described here incorporating the essential components of the Bodhisattva path: restraint of the senses, cultivation


Diwakar Acharya
+ + + + + + + + + + + + + + + +17 [2] (2) samprapyanuttaraTfljiianaTflpraja duf:tkhat pramocita pramocya sarvvaduf:tkhebhyo yo 'sau santaTfl padali gataf:t [3] [ma]<hamunif:t>



(3) salikhidya sucirali kalam bhavanam bhavavicchidaf:t kinnarfjatakakfn}1}ali 19 nanacitravirajitam [5]

of compassion and the six perfections, attainment of the ultimate knowledge, release of all people from sorrow, and departure. The Mahavastu describes it and says that these acts of the Buddha are purposeful: kalpako(isahasra1}i aprameyam acintiya I carito bhoti arthaya sarvajiio dvipadottamo II danaTfl SflaTfl ca k:fanti ca dhyanani ca nisevita I prajiia ca carita parvamkalpakotisataTfl bahaTfll1 (SENART 1890: 296). 17 The language of this inscription is colloquial and structurally loose. In the third stanza, when two successive actions are stated in two versehalves, the first action stated with a finite verb form in the first half is narrated in concatenation in the other half with an absolutive form together with its object. The writing style suggests that the same was true in the lost second half of the second stanza and the first half of the third. The latter, which has survived, states the second action 'released people from the sorrow' narrating the previous action in absolutive 'having obtained the ultimate knowledge.' Therefore, the last pada of the second stanza can be reconstructed as <praptaTfl jiianam anuttaram>, on the basis of the narrating phrase in the next stanza. ~ 1B The 9th stanza below tells us that the Buddha image the lady donated was named Mahiimuni, and we can judge from the context that stanzas 1-4 are dedicated to praise the inaugurated Buddha, the Mahamuni. Whether these opening verses were written in the form of veneration of the Buddha or blessing to the folks, the name of the god is expected here, most likely in the nominative case like in the first verse of Manadeva's ChiingunarayaI,la inscription (VAJRACHARYA 1973: inscription no.2). Another possibility is the dative case. In any case it is most likely that the name of Mahiimuni appeared here. 19 Normally it should be kfr1}1}an. In Nepalese manuscripts and sometimes even in Licchavi inscriptions a homorganic nasal before a sibilant is written as gutturalli, but gutturalli before nasal is a rarity. VAJRACHARYA misses to record this irregularity.

Mahayana Buddhism and SukhavatI cult in India

+20 ++ ++ + + + + + ++ + + + + + + + + + + + + + + + + + + + + [6]


(4) catvarif[liat sapaficeha yatra dhanyasya manikaJ:t var~e var~e 'tha jayante k~etran tat tadrian dadau [7]

(5) bhayaJ:t smighasya bhaktartthaf[l pajartthaii ca mahamuneJ:t k~etran dattan taya hy atra a~tavif[liatimanikaJ:t [9]
+ + + + + + + + + + +21 + + + + + + + + + + + + + + + + + + + + + [10]

(6) vicitraf[l deyadharmmam me karayitveha yac chubham strfbhavaf[l hi viragyahaf[l puru~atvam avapya ca 22 [11] iokakamamayat pa<r,zkat>23 + + + + + + + +

<samvat> + + + (7) maghavarse kale a~arJhaiudiva 10 224 asyan divasaparvvayam bhattarakamaharajairf+ + + + + + + + + + + +


20 VAJRACHARYA reads irfhere which I cannot see on the stone or photo. 21 The context asks for an expression meaning 'of paddy are produced every year' in the lost part, something close to dhanyasya var~e var~e 'tha jayante as in stanza 7. 22 The usage of viragya here is noteworthy. This peculiar form is attested in the Daiabhamika, and other forms of the denominative verbal stem viragay are found also in other Mahayana satras. See EDGERTON, s.v. viragayati. 23 The ligature of tpa is rather clearly visible but VAJRACHARYA does not read pa. I have completed the word by supplying <fikat>. In the A~tasahasrika, all those Bodhisattvas who reach the land of Abhirati are said to 'have gone across the mire' (uttfrr,zapafikaJ:t). For this passage, see below, pp. 62-63. 24 VAJRACHARYA misses the symbol of 2 and takes the day as the 10th. 25 If we wanted to guess at the lost part of this line, adapting to the formula found in the inscription of Saka 425 mentioned before and using the possible names of the King V!~adeva and the donor Carumatl, it would be something closer to this: ovr~adevasya sagraf[l var~aiataf[l samajfia-


Diwakar Acharya Like a slave,26 having restrained the hardly re~trainable senses by which all these people are carried away - [and] having closely embraced them, [i.e. the people,] with compassion, <through the cultivation of> charity, good conduct, <perseverance, valour, meditation, and wisdom> ... <he obtained the supreme knowledge>; after obtaining the supreme knowledge, <he> freed the people from sorrow; [and] after freeing them from all sorrows, he attained the place of peace. That Ma<hamuni> ... 27 [ll.l-2=vv.l-4] Taking a lot of trouble 28 for quite a long time, [she built] the abode of the destroyer of the worldly existence, [i.e. Mahamuni,] which

payatalJ carumatya sthapito bhagavan mahamunilJ. 26 The comparison 'like a slave' can logically be associated either with Mahamuni or the senses, respectively the subject and object. I feel that our inscription is alluding, here too, to a specific Buddhist concept like in v.l1 below. Therefore, I am associating the comparison with the subject following the description of one of the arthacaryas in the BodhisattvabIn/mi. There, a Bodhisattva, though he is abiding in the best and foremost state of success, is said to be fulfilling the purpose of the beings, like a slave, with his mind lowered (in kindness), and his vanity, pride and ego destroyed: punar bodhisattvalJ pravarayam agryayam api saJTlpadi vartamano dasavaf pref!yavad vasyaputravac canaladarakavan nfcacitto nihatamadamanahm?1karalJ (WOGIHARA 1936: 225 reads nihita O == 'laid aside') sattvanam artham acarati (DUTT 1966: 154). Following Arlo GRIFFITHS' suggestion, I present the following alternative translation of the first verse: Having restrained them - the senses by which all these people are carried away, and having squeezed these [people], as [one oppresses] a slave, [but] with compassion (rather than stringency) .... The root paripf literally means 'to squeeze properly from all sides' or more negatively 'to oppress in all ways.' As I need something quite positive for the interpretation I favoured, I have taken it in its figurative sense, 'to embrace closely.' 27 The statement might have concluded with something like 'that Mahamuni excels all' or 'that Mahamuni may show us/you the way.' 28 The literal meaning, 'being deeply depressed' or 'having forced properly,' does not work well here. So I take it figuratively with positive implications.

Mahayana Buddhism and SukhavatI cult in India


is brilliant with many depictions illustrating [scenes] from the Kinnarfjiltaka ... [l.3=vv.5-6] Here [in the same locality] she donated such a piece of land where every year 45 Manikas of paddy are produced29 .. [1.4=vv.7-8] Again, for the purpose of [providing] food for the Community and also for the purpose of [financing the daily] worship of Mahamuni, another piece of land is donated by her where 28 Manikas <of paddy are produced every year.> ... [1.5=vv.9-1O] Whatsoever merit I have by making here this wonderful religious gift (deyadharma), <by that> I may lose attachment to womanhood and attain manhood, and <get out of> this <mire> consisting of sorrow and longing, ... 30 [1.6=vv.11-12] <The year>... , the time of 'the year of Magha,' the bright half of A~ac,lha, the 12th day. On this day the lord great king Illustrious ... [1.7]

The inscription mentions that the Kinnarljiitaka was depicted on the walls of the temple of Mahamuni. The likely candidate for this reference is the Kinnarljiitaka of the Mahiivastu. There is another version of this narrative in the Bhai~ajyavastu of the Miilasarviistiviidavinaya, which seems to be followed later in the Divyiivadiina. But in that version, unlike in the Mahiivastu, the narrative is not named "Kinnarljataka" and the character of the KinnarI is not highlighted. 31
29 Since the next piece of land is allocated for sustenance of the Community and daily worship of Mahamuni, it can be said that this piece of land with a larger amount of income was allocated for maintenance and repair of the abode, and probably to finance the annual ceremony (var,'favardhana) which is known from many Licchavi inscriptions. 30 Following the parallels from the Mahavastu and the A,'ftasahasrika, we can say that the next thing our lady donor is expected to wish is her rebirth in one of the bodhisattvabhumis, if not yet in the peaceful abode of Mahamuni. See below, p. 34. 31 I am aware of the fact that the Mahavastu is a composite text and the Kinnarfjataka might not have been part of it from the beginning. However, my supposition is that this jataka was already integrated in the Mahavastu by the time of our inscription.


Diwakar Acharya

It is interesting that this inscription praises the Buddha as Mahamuni, alluding to the path of the Bodhisattva; and it is almost certain that the inscription makes a reference to the six paramitas: two of them appear in the beginning of a compound, and the metre easily allows us to include the rest in the proper order in the same compound. Again, the six paramitas are present in early Mahayana texts and also in the Mahavastu. 32 The lady donor of the image of Mahamuni with this inscription first wishes to lose her attachment to womanhood and become a man by the merit of this donation. A woman on the Bodhisattva path is expected to change her gender and become a man at some point prior to the attainment of Buddhahood. Early Buddhist texts indeed hold a strict view on the spiritual limitations of women. Also the Mahavastu implies this in the Dasablutmika section, though quite vaguely, when it states that those Dharma followers who are in any of the ten stages are all men, not born again as a woman. 33 This idea is found in many of the Mahayana siitras including the A~tasahasrika, where Sister Gangadeva is predicted to become a man and reach the land of Ak~obhya to undertake the Bodhi-

32 If, as I suggested in fn. 6 above, a cult which worshipped the Sakyamuni Buddha as Mahiimuni had existed, that possibly had a link with the school of MahasaIighikas whose offshoot, the Lokottaravadins, preserve the Mahtivastu in their Vinaya. Our inscription relates Mahamuni and the Kinnarfjtitaka of the Mahtivastu. The name Mahamuni appears 27 times in the Mahtivastu, more than in any other text (the Saddharmapw:u;larfka comes second with 11 occurences). It is worth mentioning here that the presence of the MahasaIighikas in Nepal in the subsequent period has been considered to follow from a fragmentary inscription ascribed to the middle of the seventh century. This is a two line inscription damaged on the right side, first published by Levi (no. 17, plate 18). It reads the following preceded by an auspicious sign: (1) deyadharmo yaY[! srfdharmartijikamtitya-su[pa] II (2) stil!!ghikabhik~usaYf1ghasa II (LEVI does not read pa.). Unfortunately, the prefix mahti- is missing, which limits the importance of the inscription. 33 SENART 1882: 103: atha khalu sarvtisu dasabhami~u puru~ti bhavanti sarvtiYf1gapratYGJ?lgopetiif:t avikalendriyti[f:t]. (The edition omits visarga, probably because of yas ca in the following.)

Mahayana Buddhism and SukhavatI cult in India


sattva vow there, and become finally the Buddha SuvarI,1apu~pa.34 Even the wording in our inscription reminds us of the phrase in the
A~tasiihasrikCi. 35

It is known that Buddhists were present in Nepal before Manadeva, i.e. the early fifth century CE, but how strong they were in the society is not known well. No Buddhist inscription has yet been discovered from the time of Manadeva. 36 However, I would like to draw 34 This idea is found also in the nineteenth chapter of the A~tasahasrika
1935: 745): seyam ananda gangadeva bhaginf strfbhavaT]'t vivartya puru~abhavaT]'t pratilabhya itas cyutva ak~obhyasya tathagatasyarhataJ:zsamyaksaT]'tbuddhasya buddhak~etre abhiratyaT]'tlokadhatav upapatsyate. 35 The wording of the inscription, strfbhavaT]'t hi viragyahaT]'t puru~atvam avapya ca, is very close to the A~tasahasrika wording: strfbhavaT]'t vivartya (vivarjya in the PaficaviT]'tsatisahasrika) puru~abhavaT]'t pratilabhya.

Here are two more statements close to the expression in the inscription: Samadhiraja 32.l57cd-158ab: vivartayitva strfbhavaT]'t sa bhaved dharmabha1}akaJ:z, na sa puno 'pi strfbhavam itaJ:z pascad grahf~yati. Ratnaketuparivarta (II.27: KURUMIYA p. 50): strfbhavam antardhaya puru~a bhavaJ:z saT]'tvftto. Ratnaketuparivarta speaks also of transformation of marks and organs of women into those of men in the same chapter. 36 Because the major caityas of the Kathmandu valley have been renovated continuously, and since mediaeval times such renovations are done by Tibetan monks or under their guidance, these caityas have taken new components from time to time, reflecting ongoing changes in contemporary traditions. That is why we have to rely on personally donated images or caityas of comparatively small size in order to have an idea of Buddhism in the Licchavi period. No excavation in the vicinity of the major caityas of Kathmandu valley have yet been carried out. It is not easy to excavate a main shrine or stfipa as they are still places of active worship, but it is not impossible to do so in a courtyard. The Buddhist tradition was never discontinued in Nepal. So, such excavations, I must say, would be of great help for the understanding of Buddhism in the middle period and its transformation in later times.


Diwakar Acharya

attention to a:n interesting and exceptional case of the Buddhist donative formula yad atra pUlJ-yaTfl... being blended in a Saiva inscription from Budhanilakantha (Kathmandu) inscribed on the base of a sivaliJiga al):d dated in [Saka] saTflvat 398 (476/477 CE)Y The related portion of the inscription runs this way: .
<srfmlinadeva>nrpati/:t pra1J.ato jaglida tvatsthlipanlijanitam asti yar) atra pU1J.yam tat sarvvaZokasahitasya vivrddhamflZam duf:tkhak~ayliya bhagavan mama sarvathlistu.

The king <Illustrious Manadeva>, bowed to [the god], said: 'What here is the merit produced from this action of founding you, [i.e. the sivaZiriga,] 0 lord, its roots properly grown, may that be for the complete destruction of sorrow of me together with all [my] people.

This indicates that Buddhist ideas were already popular in Nepal by this period and were even adopted by other religious groups. Furthermore, we know from Anuparama's Dvaipayanastotra inscription, installed before 540 CE, that the Buddhists had made good advance by that time, and the orthodox Brahmanical section of society had got alarmed at that development. 38 The two inscriptions presented below are further evidence for their growing influence. There are not many inscriptions until the late fifth and early sixth centuries in India which could even indirectly be related to Mahayana. So, these inscriptions deserve attention and should be added to the list of inscriptions related to Mahayana. First I present aquite damaged inscription from the pedestal of a lost image of Avalokitesvara39 which is dated in [Saka] saY[tvat 479 (558 CE):
(1) saJ?1.vat479 dvitfyli~li4ha ... .. .yajfva ... ... (2) ...... bhagavadliryyli-

37 VAJRACHARYA 1973: 41-42, no. 7. The year of this inscription, first relld 396, has been corrected to 398 in PANT 1986: 275-276. 38 For an elaboration on this, see ACHARYA 2007. 39 At present, this pedestal supports an image of Vi~I)U in a small temple located in Brahma Tol, Kathmandu, but the inscription on it clearly suggests that it once supported a Buddhist image.

Mahayana Buddhism and SukhavatI ;:;ult in India


<valokiteivara> ... .. .miineniirddha ... ... (3) ...... sarvvajiiajiiiiniiviiptaye bhavatu 40

The [Saka] year 479, the second A~a<:lha ...... <Pri>yajlva ...... [an image of] the Blessed One, Arya<valokitesvara> ... . .. half the size of (?) ... . .. may that be for the obtaining of the knowledge of the Omniscient.

Though only a few words of this inscription are preserved, it still contains the last part of a variant of the Mahayana formula, sarvajfiajfiiiniiviiptaye, and parts of the donor's and the deity's names. There is another similar inscription which has almost everything intact except the date in the beginning. The king's name is also damaged partially, but GNOLI reads it Ramadeva (circa 547 CE). VAJRACHARYA reads only -deva, which is clearly visible; nevertheless, he places it before the above inscription of 558 CE in his book on Licchavi inscriptions, obviously following GNOLI'S suggestion. However, as I can read the lower part of the ligature before deva as ga in the rubbing produced by Gnoli, I am of the opinion that it should be Gailgadeva (circa 567 CE). This will make the inscription ten years younger than the one cited above. The place of finding, nature and palaeography suggest that the two inscriptions are somehow related. I present here my reading and translation of the inscription:
(1) Of!l sva<sti sal?1Vat> ++++++++++ bhattiirakama<hiiriijasrfgari>[gii]devasya siigral?1 var~asatar?l samiijiiii<payata!:t> (2) sarvvasattvahitasukhiirtthiiya bhagavata iiryyiivalokiteivaraniitha41 prati~thiipita!:t [SPACE] deyadharmmo 'yal?1 paramopiisakamalJiguptasya (3) bhiiryyayii mahendramatyii saha yad attra pU~lyaf!1 tad bha[va] tu miitiipitrpurvvarigamaJ?l krtvii sarvvasattviiniil?l sarvviikiiravaropeta(4)+++++sarvvajiiajiiiiniiviiptaye42

40 I am unable at present to go and read this inscription on the spot. Therefore I simply reproduce VAJRACHARYA'S reading. See VAJRACHARYA 1973: 185, no. 43. 41 This sentence is grammatically incorrect. It needs to be either bhagaviin iiryyiivalokiteivaraniitha(l or bhagaviin iiryyiivalokiteivaranathasya vigraha(l. 42 VAJRACHARYA 1973: 177, no. 40.


Diwakar Acharya Om, good <luck! In the year... ,> when the lord gr~at <king Illustrious Gari>gadeva is ruling for hundred years and further, [an image of] the lord Aryavalokitesvara, the Blessed One, has been set up. This is a charity of Paramopasaka Malfigupta together with his wife Mahendram at!. Whatever merit [is obtained through this action], may that be for all beings, first and foremost his mother and father, for the obtaining of the ... knowledge of the Omniscient endowed with all excellent forms.

Both of these inscriptions are special as they contain the term sarvajfiajiuina, which is attested in many Mahayana siitras including the Kiisyapaparivarta, A~tasiihasrikii, larger Sukhiivatfvyiiha and also in Asanga's Bodhisattvabhiimi. Compared to sarvajfiajfiiina, anuttarajfiiina is poorly represented in the SLttras, although it seems to be the predominant expression in inscriptions (SCHOPEN 2005: 241, fn. 14; 265). From the second of the two inscriptions we can tell that sarvajfiajfiiina was the last member of a tripartite compound which contained sarviikiiravaropeta as the first and another word of five ak~aras as the second member. Sarviikiiravaropeta appears once qualifying siinyatii in the A~tasiihasrikii (WOGIHARA 1935: 750), and once in the Lalitavistara qualifying suparisodhitajfiiina (VAIDYA 1958: 309). In the DasabhUmika (KONDO 1936: 61), sarviikiiravaropetasarvajiiajfiiina is found as a compound without any intervening element, and in the PaficaviY[lsatisiihasrikii (KIMURA 2006: 166), we find sarviikiiravaropeta compounded with sarviikiirajiiatii. In a seventh century Nepalese inscription, anuttara is combined with sarvajiiajfiiina in a similar donative formula,43 and this combination is also attested in the GalJ-(lavyiiha. However, in our inscription just anuttara is not possible, because we have space for five ak~aras, and -ta- at the end of the first word is intact, which would not have been so if the following ak$ara had begun with a vowel. I therefore guess that the damaged word was sarviinuttara ('supremest') which is attested as an adjective to samyaksaY[lbodhi in the Kiisyapaparivarta. 44

43 SCHOPEN 2005: 256 and fn. 15. 44 Following STAEL-HoLSTEIN (1926: 8),


Mahayana Buddhism and SukhavatI cult in India


The inscription on a caturvyuhacaitya from Tyagal As I mentioned earlier, a Nepalese inscription that mentions the triad of Amitabha and his Buddha world SukMvatT comes second chronologically only to the Govindnagar inscription and is to be placed before the SaiicI slab inscription. This inscription is significant in many respects. It is inscribed on four sides of one of the two stone caityas in a courtyard of Tyagal Tal in Pat an district of Kathmandu valley.45 It is not dated but on palaeographical grounds it is placed about the time of Arp.suvarman, i.e. the late sixth or early seventh century, by VAJRACHARYA. Each side of the square base of the caitya, like the one seen in the photo on the next page, contains a verse, inscribed in two lines, which praises the TatMgata worshipped on that. side together with his two Bodhisattvas. The odd and even piidas of each verse are separated by the niche of each TatMgata lying in the middle of the wall. In the following pages, as I have placed the photos of the two sides on top of each other, the a and c piidas precede band d in these photos. This inscription was published for the first time by the Sarp.sodhana Mat;l<;lala team in the fifth issue of their Nepali journal
et al (2002: 5-6) introduce a wrong word division and read "yils ca satv[iln] paripilcayati tiln sarviln uttarasyil(l!!) samyakSaYflbodh[au]." Since anuttarasyill]1 samyakSaYflbodhau is attested dozens of times in Mahayana sfitras, I suggest to read sarvilnuttarasyilYfl as a compound.

One more expression found in our inscription, sarvasattvahitasukha-, appears in several Mahayana sfitras including the Paficavil!!satisilhasrika and the larger Sukhavatfvyfiha. 45 I am grateful to Nepali historian and writer Devichandra SHRESTHA for his help in locating the caitya. I am also grateful to two researchers, Nirajan KAFLE and Rajit Bahadur SHRESTHA, and photographer Yogesh BUDHATHOKI, all from the Nepal Research Centre, for their help in preparing photographs of the caitya and the inscription. As the inscribed part of the caitya was covered with lime and other substances, the photo quality is not so good. I regret the reSUlting inconvenience involved.


Diwakar Acharya

PurlJima, and it has been included in VAJRACHARYA'S book.46 But the valuable information contained in this inscription has yet to be revealed, so it is necessary to read and interpret it again. It consists of four verses in three metres: the first in Upajiiti, the second in SikhariI).l, and the third and fourth III Vasantatilaka. The first and second verses are in first person singular and the other two are in second person plural. This inscription does not say anything about the donor of the caitya or the context of the donation. Let me now present my reading of the inscription, which includes five improvements as compared to VAJRACHARYA'S edition, and translate it.

East side:
1) [siddham]47 ak~obhyam ak~obhyasitiigramurtin tathiigataf!l staumy abhito bhiratyiim

46 VAJRACHARYA 1973: inscription no. 98, 387- 388. He has misjudged the directions of the Buddhas and placed Ak~obhya in the north, Sakyamuni in the west, Samantakusuma in the south and Amitabha in the east. 47 V AJRACHARYA (1973: 387) reads Of!l.

Mahayana Buddhism and SukhavatI cult in India



samantabhadram bhuvi bhadraclirilJ-an tathiLiva sannirmalakfrtimlilinam bhiratylim 1bhiramyam V; bhadraclirir:tan 1bhadraklirir:tan V


1) 2)


sukhlivatylirrz vande satatam amitlibhan jinaravim saloke1arrzlokodbhavabhayaharam pankajadharam mahlisthlimapr[li}<ptam ~kr>plisnigdhamana[sam}


Diwakar Acharya

West side:
1) 2) saddharmaratnakusumastavakiicitiiligam buddhaJ!l samantakusuman namatiibjavatyiim maiijuriyam paramadharmavidmi kumiiran nityai'i ca susthitamatili karulJaikatiinam

North side:
1) . [yo va]m --m bhaktyadya tan namata akyamunim mu[nfal!l]48
v _vv v_v v_v


maitryiirj.hya --Ii guhyiidhipaJ?l vimalavajradharaJ!l saha[yamJ

_vv v_v v_v

yo va ] ytiva V; maitryticJhya ] maitrytirddha V; sahtiytim ] sahtibjam V

Mahayana Buddhism and SukhavatI cult in India


[East side:] From the front, I praise Ak~obhya Tathagata [residing] in the world of Abhirati, who is the embodiment of the imperturbable and sharp~pointed [nature].49 [I praise] Samantabhra [Bodhisattva], who performs good [deeds] on earth, and in the same way, Sannirc malaklrtimalin [Bodhisattva]. . [South Side:] I always venerate Amitabha, the Sun-like Jina, in the world of SukhavatI, who has destroyed the darkness of the great illusion of existence with the light of great wisdom; Mahasthamaprapta, whose mind is affectionately disposed due to <unlimited> compassion, and Lokesa,so who holds a lotus and wards off the dangers of arising in the world. [West Side:] [0 people,] you must bow to the Buddha Samantakusuma in the world of AbjavatI, whose limbs are covered with bunches of the precious flowers of the True Dharma, to MafijusrI [Bodhisattva], the prince who knows the Dharma best, and to Susthitamati [Bodhisattva], whose mind is fixed on compassion eternally. 51 [North Side:] [0 people,] you must bow now devotedly to Sakyamuni, the lord of ascetics, who ... , to the one who is rich in benevolence (maitryiltj.hya) ... , [and] to the lord of Guhyas who holds the stainless Vajra, [i.e. VajrapaI).i]; [all] in the SaM world. In this caitya, the four Tathilgatas are placed on four sides of the square lower level, and the eight Bodhisattvas at the higher level before the dome begins. Even though the inscribed verses place 49 The original meaning of sita is 'sharpened,' and this meaning fits well here, but VAJRACHARYA (1973: 387) has taken it as 'blue.' Though this wrong interpretation is a result of phonetic confusion of sand s, one can find its roots in Ak~obhya's visualisations from Tantric texts which attribute to him a bluish/blackish complexion. Besides, one could also split a compound like ours where ak~obhya and Sita appear together into ak~obhy and asita, and thus, get closer to 'black' (asita). Something like this could be lying behind the attributed complexion of Ak~obhya. 50 The inscription reads salokeaf!!, which means 'together with Lokesa.' If we translate it faithfully, the next words in the pilda, which in fact describe Lokesa, will be adjectives to Amitabha. So I have translated salokea,!l as 'and LokeSa' following the demand of the context. 51 As an alternative, one can probably take nityaf!! adverbially with the imperative 'namatha.'


Diwakar Acharya

Tathagatas and Bodhisattvas side by side in their respective worlds, the lower level houses four niches and the higher level eight. 52 This clearly suggests that, in this caitya, the Tathagatas are placed in the lower level and the Bodhisattvas in the higher. 53 In the lower level, though bodily shapes are still visible in the images of four Tathagatas, they are damaged beyond recognition; and there is no certainty that these are remains of the original images. The same is true with

52 As one Bodhisattva is exactly above the Tathiigata, the second Bodhisattva is a little bit to the side. Probably this was not the original way that the Bodhisattvas were placed. If the block of Bodhisattvas is rotated just a little, two Bodhisattvas come in the center of each side. It is possible that at a time of renovation people forgot to fix the upper part rightly. 53 This appears a bit odd, but it is also true that in early images the Buddha is depicted in human/ascetic form, while the attending Bodhisattvas are depicted in godly or rather royal forms. Anyway, the fact that sometimes Bodhisattvas seem to supplant the Buddha in importance and stature is not new. To some extent, SCHOPEN (2005: 278- 279) has dealt with this problem while identifying a Mahayana scene painted at AjaD-ta.

Mahayana Buddhism and Sukhavati cult in India


the niches in the higher level, but four of them now contain late .images of the Buddha, MafijusrI, PadmapaI).i, and probably Tara, which are already damaged to some extent, and the other four are either empty or contain pieces of defaced stones (see photo on p. 44).54 So, we do not know how these Tathagatas and Bodhisattvas were originally represented. The original top structure above the dome has been lost, and at present, a rather late and unmatching structure covered with painted metal plates is superimposed on top of the dome (see photo on p. 40).
Four image cults fitted in the caitya

Apart from the evidence it provides for a rather unique form of Mahayana practiced in Nepal in the late sixth century, this last inscription provides evidence for early efforts in fitting various Tathagatas and Bodhisattvas in four directions, and thus producing a cult object acceptable to the followers of specific books, or rather different Mahayana models. The set of four triads found here is not found anywhere else. The cult ofAk.yobhya The beginning of the inscription on the east side of the caitya is indicated by an auspicious symbol, and here is housed Ak~obhya Tathagata together with Samantabhadra and Sannirmalaklrtimalin in the Abhirati world. We know Ak~obhya's Abhirati world in the east from several Mahayana and Vajrayana sources, but the Bodhisattvas associated with him in Vajrayana texts are generally Maitreya and K~itigarbha. This triad is unique in itself and provides evidence of an archaic cult of Ak~obhya or the eastern/earliest 'pure land.' We know from the Ak~obhyavyiiha, one of the early Mahayana texts translated into Chinese, which is also made part of the Mahii54 As ALSOP has argued, the Licchavi stone caityas were originally built with empty niches. It is highly probable that this was the case with our caitya, and whatever we see now under the niches, defaced stones or recognisable images, are unoriginal.


Diwakar Acharya

ratnahita collection, 55 and the portions of theA~rasahasrikaregard ed as additions (but made before the second century CE)56 that the cult of Ak~obhya predated the cult of Amitabha, 57 though we do not have epigraphical evidence for it. Ak~obhya appears in the Mahavastu as one of the irreversible (avaivartika) Bodhisattvas in the ninth bhami. 58 The Ak~obhyavyfiha describes how a Bodhisattva attained Buddhahood to become the Buddha Ak~obhya; however, in added portions of the A~tasahasrika, he is already the Buddha of the east in Abhirati. . The Bodhisattva Samantabhadra is generally associated with Sakyamuni in mediaeval Mahayana sources. However, he is said to be coming from the east, the direction of Ak~obhya, in the Samantabhadrotsahana chapter of the Saddharmapw:ujarlka. 59 This way, there is at least one scriptural indication for Samantabhadra's association with the east, but except for our inscription we do not have any other scriptural or epigraphical evidence for his direct association with Ak~obhya. He is already associated with Vairocana in the Gar:u;lavY~lha, 60 and finally depicted as the primordial Buddha in later Tantric traditions. The name of the other Bodhisattva, Sannirmalaklrtimalin, is not attested anywhere as a Bodhisattva, if we are to take the name as it features in the verse. We could consider that the real name of this Bodhisattva is Vimalaklrti, who narrates Dharma to MaiijusrI,
55 The Bajaur manuscripts in Kharosthi script and Gandhari language also include a large portion of an early Mahayana satra related with the Ab;obhyavyaha, see STRAUCH 2007: 47-60. 56 For identification ofthese portions, see CONZE 1967: 172-173. 57 NATTIER 2000: 101-102. 58 SENART 1882: 139. 59 SaddharmapulJarfka 26: atha khalu sal11antabhadro bodhisattvo mahasattva1:t parvasyalJt dii galJanasamatikrantair bodhisattvair mahasattvai1:t sardhaTJl parivrta1:t ... 60 For example, SUZUKI & IDZUMI 1934: 425: yatha ceha sahayalJ1lokadhatau bhagavato vairocanasya padamalagata1:t sal11antabhadro bodhisattvo dak~ilJalJ1 palJilJ1 prasarya sudhanasya l11L7rdhni prati~thapa yamasa, tatha sarvalokadhatu~u ...

Mahayana Buddhism and Sukhavatl cult in India


Sariputra and others in the Vimalakfrtinirdefa, and that the name in our versei~ a descriptive term, as it is almost parallel in meaning with the original name. 51 We in fact have a secure case of an extension of a Bodhisattva's name with an extra adjective, Vimalavajradhara for Vajradhara, in one of the verses from our inscription. However, the Saraf!lgamasamiidhi hints at a greater possibility of this Bodhisattva's name being a bit longer than Vimalaklrti, something very close to the term in our inscription. This satra mentions the Bodhisattva *Matyabhimukha who visits Sakyamuni from the Buddha Ak~obhya's world of Abhirati, and is predicted to become the Buddha *Vimalaprabhaklrtiraja in a future aeon. 52 As indicated by the application of asterisks, both of these names are reconstructed into Sanskrit from Chinese by LAMOTTE. If we consider chances of error in such reconstructions, we can presume that the original shape of the name reconstructed as *Vimalaprabhaklrtiraja was not far from the name in our inscription. 53 The substitution of vimalawith sannirmala- can be metri causa; the former does not fit anywhere in the verse. I would say, vimalaprabhiikfrti- (,fame of stainless brilliance') of the reconstruction is not so logical or suitable to. Sanskrit word order, but if we correct it to vimalakfrtiprabhii- ('bril51 I do not think that Vimalaklrti's identity as a layman poses problem to his inclusion in the triad of Ak~obhya as an object of devotion. It is not necessary that both of the Bodhisattvas included in a triad are of equal status. In fact in all four triads from our inscription, the first Bodhisattva is superior to the second in the same set (see p. 70). Because in the Vimalakfrtinirdesa Vimalaklrti is made to narrate Dharma even to MaiijusrI, a celestial Bodhisattva with the role of a saviour, one can imagine how much importance is attached to him in certain traditions: he is regarded virtually superior in knowledge and its transmission even though he is a layman. However, it is true that he is dropped off in later traditions (with a few exceptions). 52 LAMOTTE 1998: 78-79. 53 Here I remind the reader that when Khotanese fragments of the Suraf!!gamasamiidhi were discovered, EMMERICK was able to correct three Bodhisattva names reconstructed by LAMOTTE: Merusikharadhara to Merusikharakiitadija, Vimalacandragarbha to Sasivimalagarbha, Sarvaratnaracita to Sarvaratnapratyupta. See, LAMOTTE 1998: xv.


Diwakar Acharya

liance of stainless fame'), it becomes natural a!ld also equivalent to the metaphorical expression sannirmalakrrtimala. As the last component of the reconstructed name, -raja means nothing more than the -in suffix. Thus, this much can be said that VimalakIrti or *VimalaprabhakIrtiraja 64 is the closest match for SannirmalakIrtimalin of our inscription. Vimalaklrti appears also in the first two chapters of the Tantric Mafijusrrmulakalpa (SastrI 1920: 8,40), and in the second occurrence he is made one of sixteen Mahabodhisattvas. It is' noteworthy that the Vimalaklrtinirdeia contains a passage which proves his association with Ak~obhya. When asked by Sariputra, VimalakIrti tells that he comes from Abhirati, the world of Ak~obhya Tathagata, and Sakyamuni confirms his statement. VimalakIrti further clarifies that he has come to an impure world from a pure world for the sake of purification of all beings.65 What is more, upon a request of the assembly, he brings the Abhirati world into the Saha world, i.e. our world. 66 This way, we can prove an earlier association of Vimalaklrti as well as Samantabhadra with the Buddha Ak~obhya on the basis of these hints from the Vimalakrrtinirdeia and Saddharmapw:uj.arrka. However, neither of the sutras can be the source for the triad of Ak~obhya mentioned in our inscription, because both sutras mention only one of the two Bodhisattvas and lack the other. The Ak~obhyavyuha and A~tasahasrika, the earliest sutras which are related to Ak~obhya, do not even mention either of the two Bodhisattvas from our inscription. However, both of these sutras relate the Bodhisattva Gandhahasti with the Buddha Ak~obhya, as the one whose future Buddhahood is predicted at the time of Ak~obhya's departure. If observed properly, it is possible to see that the same motif lies behind the names Gandha64 On the identification of *Vimalaprabhaklrtiraja with Vimalaklrti, see LAMOTTE, 1998: 170, fn. 181. 65 The concept that the land of Ak~obhya is pure lies behind this statement. It seems that by the time of the composition of the Vimalakrrtinirdesa a general concept of 'pure land' was already at work. 66 Vimalakfrtinirdefa 11.2-4.

Mahayana Buddhism and Sukhavatl cult in India


hasti and Samantabhadra. Gandhahasti literally means 'fragrant elephant,' w~ich is a descriptive adjective to an elephant of the best type. Bhadra is the best of elephant types,67 and Samantabhadra's association with elephants is suggested in iconography by placing him on a seat with elephants on all sides (samantabhadra). In this way, both of these names mean almost the same thing. This suggests that Samantabhadra is a metamorphosis of Gandhahasti, which took place after the Ak~obhyavyuha and the 'additions' to the A~!asahasrika, 68 and before the longer Sukhavatlvyuha where Samantabhadra appears. It appears that there existed a tradition that connected Samantabhadra to Ak~obhya slightly posterier to the 'additions' to the A~tasahasrika. As for the triad of Ak~obhya, it must have been formed already along with other triads by the time of composition of the longer Sukhavatlvyuha which mentions Amitabha's triad, and the Paiicavif!lsatisahasrikti which mentions in passing Samantakusuma's triad in a world-system far away (see below).

The cult ofAmitabha

Moving to the south in the path of circumambulation, we find the most famous triad of Amitabha Tathagata and his two Bodhisattvas, Avalokitesvara (here spelt Lokesa possibly for metre's sake) and Mahasthamaprapta. This set is found in two sutras of Pure Land Buddhism: the longer version of the Sukhavatlvyuha and the Contemplation Sutra. 69 The first sutra says that, in the west in 67 The three types of elephants are bhadra, mandra and mrga. Three subtypes, bhadramandra, bhadramrga and mrgamandra are also mentionedin the Ramaya'!a (1.6.22). 68 Apart from these two texts, Gandhahasti appears also in the Vimalakfrtinirdeia and the shorter Sukhavatfvyaha as a member of the assembly of Sakyamuni, when the Samadhiraja makes him visit Sakyamuni from the world of Ak~obhya. Samantabhadra does not appear in these texts. Both of these names are used only in the relatively late Karu'.lapu,!arfka and Manjusrfmalakalpa. 69 The other satra, the shorter version of the Sukhavatfvyaha, spells the name of the Tathilgata Amitayus instead of Amitabha, and does not men-


Diwakar Acharya

SukhavatI Lokadhatu is Amitabha Tathilgata, the Arhat; he has two Bodhisattvas: the first of them is Avalokitesvara, and Mahasthilmaprapta is the other. 70 The second siitra states in the same way, for example, in one place, "when these words were spoken, Amitayus appeared in the air above, attended on his left and right by the two Mabasattvas, Avalokitesvara and Mahastbamaprapta. So brilliant was their radiance that it was impossible to see them in detail" (INAGAKI 1995: 328). This triad is well known and widespread. It arrived in China in the early phase of transmission of Buddhism and is worshipped today in East Asian countries, but in the Indian context our inscription is the first incontrovertible evidence71 for the existence of the SukbavatI cult proper.
The cult of Samantakusuma

Moving to the west, we find Samantakusuma Tatbagata with MafijusrI and Susthitamati. In the Pancavif!1satisiihasrikii at the end of the introductory section, exactly this triad of the Buddha Samantakusuma is mentioned. As the sutra describes, ten Bodhisattvas from the Buddha worlds of the ten directions visit Sakyamuni in Saba, and worship him with jewel lotuses of golden colour as he delivers his sermons. At the end flowers are scattered all around, and he is covered with them and so is his world. It is now comtion Avalokitesvara and Mahasthiimaprapta even among the assembled Bodhisattvas. For the two alternative names Amitabha and Amitayus, see NATTIER 2007.
70 pafcimilyillJl difi ... sukhilvatyillJllokadhiltilv amitilbho nilma tathilgato 'rhan '" (ASIKAGA 1965: 26); ekas tayor ilnanda avalokiteSvaro bodhisattvo mahilsattvaJ:t dvitfyo mahilsthilmaprilpto nilma (ASIKAGA 1965: 49). VAIDYA'S edition of the text has Mahilsthilmaprilpto instead of Sthilmaprilpto. 71 The celebrated Mohammad Nari stele of uncertain date (third or fourth century CE or even later?) could serve as such evidence but it has become quite controversial regarding the date and identification. However, its identification as a depiction of Sukhavati is rejected by many scholars but accepted by some (e.g. HUNTINGTON 1980, QUAGLIOTTI et al 1996).

Mahayana Buddhism and SukhavatI cult in India


posed of jewels and precious stones, and filled with flowers and fruits "just like the world system PadmavatI, the Buddha-field of the Tathagata Samantakusuma, where MafijusrI the Crown Prince resides, and the Bodhisattva Susthitamati, and other very powerful Bodhisattvas."72 Though a Buddha of this name does not appear in the A~?asii hasrikii, an almost synonymous term, Avakln;mkusuma, 'Scattered Flowers,' appears there as the name given to a large group of future Buddhas. In the same sutra, we find yet another Buddha called Suvarl).apu~pa who is named after a similar concept and described in a similar way. The name Samantakusuma means 'Flowers All Around' and Suvarl).apu~pa means 'Golden Flowers.' While the Buddha Samantakusuma is described in our inscription as having his limbs covered with bunches of flowers of True Dharma,73 Suvarl).apu~pa is also described as a future Buddha in a similar fashion in the A~?asiihasrikii: Sakyamuni shines a 'golden' smile when Sister GaIigadeva appears in his assembly. When Ananda asks why he is smiling, he tells that Sister GaIigadeva will become the Buddha Suvarl).apu~pa in the future, and relates the name of the future Buddha with the lady's brahmacarya vow under the Buddha Dlpmp.kara, and her act of covering the latter with golden flowers.7 4
72 CONZE 1975: 44; Sanskrit text (DUTT 1934: 17): tadyathapi niima padmiivatf lokadhiitu/:z samantakusumasya tathiigatasya buddhak~etral'[l yatra mafijusrf/:z kumiirabhutaJ:t prativasati susthitamatis ca bodhisattvaJ:t anye ca mahaujaskii bodhisattviiJ:z.

The world of Samantakusuma is named AbjavatI in our inscription for the sake of metre. The world of PadmavatI is rarely attested. Beyond the Paficavil'[lsatisiihasrikii and AjitasenavyiikaralJa (see fn. 80 below), it appears once in the GalJejavyuha (SUZUKI & IOZUMI 1949: 82) but is spelt PadmavatI and the Buddha there is also different. 73 RatnakusumasaIp.pu~pitagatra, a name almost identical in meaning to this attribute of Samantakusuma, appears as the name of one of the Tathagatas in the Smaller Sukhiivatfvyuha. In both places the key word ratnakusuma is common. Besides, in the GalJejavyuha (VAIDYA p. 66), we find a Bodhisattva situated in the southwest whose long name incorporates the phrase samantakusuma. 74 See A~tasiihasrikii 19 (WOGIHARA 1935: 747).


Diwakar Acharya

TheAk~obhyavyilha, however, mentions the same Buddha under the name 'Golden Lotus' (as rendered into English from Chinese). The topos is basically the same (though it adds the theme of the preceding Buddha's parinirviilJa) but the characters involved are different: on the day of his parinirviir;a, the Buddha Ak~obhya "will predict Bodhisattva Fragrant Elephant's attainment of Buddhahood, saying, 'After my parinirvfir;a, you will become a Buddha, named Tathagata Golden Lotus.' ... At that tit:;Ie, the gods and humans will all scatter over the Buddha garlands of flowers, many kinds of incense, and clothing. The scattered fragrant flowers will pile up around the Buddha to a height of one league" (CHANG 1983: 331). The Buddha figure behind these different names, it appears to me, is the Buddha on the seat of enlightenment (bodhimar;rja).75 Let us read the following two representative7 6 passages from the Mahiivastu, a small portion of a long description of the veneration of the Buddha by the deities:

75 The bodhimalJtj.a was once the most important symbol of Buddhism. It had a status comparable to the caitya; or rather, it was the bodhimalJtj.a which used to make the caitya worthy of veneration. I cite here a passage from the A~tasiihasrikii, as quoted by SCHOPEN in one of his articles, which highlights the importance of the bodhimalJtj.a:

"Just Kausika, as those men and ghosts who have gone to the terrace of enlightenment, or its circumference, or its interior or to the foot of the tree of enlightenment, cannot be hurt by men, or ghosts, or be injured by them, or taken possession of, even with the help of evil animal beings, except as a punishment for former deeds." ... (CONZE'S translation quoted in SCHOPEN 2005: 29) Once Buddha images were introduced, they took the place of the bodhimalJtj.a and also of other aniconic symbols. It appears to me that some of the early Buddha figures were held to be connected with certain aniconic symbols which indicated the Buddha's presence when his images were not yet introduced. So, perhaps, Ak~obhya and Samantakusuma have to be connected with the seat of enlightenment, Sakyamuni with the Bodhi tree, and Amitabha with the wheel of Dharma. 76 Other passages of interest from the Mahiivastu are: SENART 1890:

303,309,352-353,1897: 277-278.

Mahayana Buddhism and SukhavatI cult in India


For seven days while he sat on his solitary seat thousands of kotis of devas paid him honour. Over that seat they scattered powder of sandal-wood tree and flowers of the coral tree. Above it celestial musical instruments struck up and played. Then devas from above scattered down powder of the celestial sandal-wood tree; of the celestial aloe-wood, of the celestial kesara, of celestial tamala. They showered down flowers of the celestial coral tree, of the celestial great coral tree, of the karkarava, of the great karkarava, of the rocamana, of the bhz~ma, of the samantagandha, of the greatsamantagandha, of the mafiju~aka, of the great mafiju~aka, celestial flowers of the parijataka, flowers of gold, of silver, of all precious jewels. There appeared in the sky thirty thousand celestial and bejewelled sunshades shading the Conqueror's body, which was like a rock overlaid with precious stones, like a tope of gold, blessed with the root of virtue acquired in several kotis of kalpas (JONES 1952: 269-270).77 Again, monks, when the Tathagata had awakened to the unsurpassed perfect enlightenment, for a full seven-days he sat alone cross-legged. Then devas of earth, devas of sky, ... and the Akani~tha devas, for a full seven-days honoured, revered, worshipped, and adored the Tathiigata on his noble bodhi throne. And for a full seven-days the whole universe of three thousand worlds became one vision of splendour. On that occasion the Exalted One uttered these verses:
For a full seven-days the perfect Buddha, the monument of the whole world, after awakening to the supreme enlightenment did

77 SENART 1890: 286-287: saptahal?2 ekasane devakotzsahasralJi pujayensuf:z. tasmif!2 asane divyaf!2 candanacarlJaf!2 okirensuf:z pu~pehi ca mandaravehi okirensuf:z divyani taryalJi upari aghattitani pravadyensuf:z tada deva ca divyani candanacurlJani uparito prakirensuf:z divyani ca agurucurlJani divyani ca kealacarlJani divyani tamalapatracarani divyiini mandaravalJi pu~palJi pravar~ensuf:z mahiimiindaraviilJi pu~ pani karkaravani mahakarkaravani rocamanani mahiirocamanani . . bhz~miilJi mahiibhz~malJi samantagandhani mahiisamantagandhiini mafija~akani mahamafija~akiini piirijatakapu~piilJi divyiini suvalJapu~ piilJi rupyapu~piilJi sarvarataniimayani pu~palJi pravar~ensu{2 divyiini trif!2sac-chatasahasralJi divyiini ratnamayani antarzk~asmif!2 pradurbhatiini cchadayensuf:z jinakayaf!2 sailaf!2 ratnamayaf!2 stapaf!2 va suvarlJamayaf!2 naikakalpakofikusalamulasamanviigataf!2.


Diwakar Acharya

not rise from his seat. Thousands of kotis of devas assembled in the sky, and for a full seven-nights poured down a shower of blossoms. Blue lotuses, red lotuses, camp aka, and white lotuses, lovely thousand-petalled and brilliant, did the devas pour down (JONES 1952: 317-318).78

The Buddhacarita briefly describes this episode in the second half of the fourteenth canto, while the Lalitavistara elaborates it in four chapters, 19-23, and even beyond in the twenty-fourth. There, different classes of deities and beings are made to venerate the Buddha with various materials: incenses, lights, flowers, jewels and so on. Both the Buddhacarita (15.5) and Lalitavistara (the last paragraph of the prose opening of the twenty-fourth chapter) mention the name Samantakusuma at the end of this episode, though not as the Buddha but as a god who approaches Sakyamuni after the latter's week-long uninterrupted meditation on the seat of enlightenment. It seems that the name Samantakusuma can be assigned to the entire episode, or one of the significant figures involved there, particularly to the Buddha being worshipped or the deities worshipping him. Indeed, we have two sets of information, one from the Buddhacarita and Lalitavistara, where this name is given to a representative deity, and the other from our inscription and the Paiicavlrrzsatisahasrika (subsequent Prajiiaparamita stUras included), where the name is assigned to the Buddha. In this way, we can see
78 SENART 1890: 348-349: punar aparafTl bhik~u tathiigato anuttariifTl samyaksafTlbodhim abhisafTlbodhitvii saptiihapurafTl ekaparyankena atiniimesi. atha khalu bhumyavacarii devii antarfk~ecarii devii caturmahiiriijikii ca devii ... yiiva akani~thii ca devii saptiihapurafTl tathiigatafTl bodhima1J.g.avaragatafTl satkaronti gurukaronti miinayanti pujayanti sarviivatf ca trisiihasramahiisiihasrii lokadhiituf:t saptiihapurafTl ekiilankiirii abhu~i. atha khalu bhagaviiJ?1 tiiye veliiye imiifTl giithiim abhii~i saptiihapurafTl safTlbuddho bodhifTl buddhitva uttamiifTll iisaniito na utthesi sarvaZ?kasya cetiyo II devakotfsahasrii1J.i gagar;asmifTl samiigatii I pu~pavar~afTl pravar~ensu saptariitram ananakafTlll utpaliifTl padumiiytl campiifTl pur;g.arfkiifTl manoramiifTl I sahasrapatriifTl ruciriifTl tatra devii pravar~i~u I

Mahayana Buddhism and SukhavatI cult in India


how this name was coined, and realise its antecedents. 79 Anyway, it seems that. different names were tried for this Buddha presiding the PadmavatI world covered with flowers/jewel-flowers/jewels. In the Ajitasenavyiikarm:za, PadmavatI is retained as the name of one of the cities of the Tathagatas but the Tathagata there is named Ratnasikhin. 8o MafijusrI is the best-known Bodhisattva. He begins to appear already in earliest Mahayana siitras and his images are widely producedY However, his association with the Buddha SamantakuslIma is not known from any other source than the passage from the Paficavi1'!lsatisiihasrikii cited earlier. Fortunately, his association with Susthitamati, the other Bodhisattva of our triad, is known from one more source. In the Susthitamatidevaputrapariprcchii,82 which makes part of the Mahiiratnakuta collection, MafijusrI teaches Susthitamati the perfection of wisdom. Susthitamati as a devaputra is also present in the Buddha's assembly in the Rii~trapiilapariprcchii. It is also noteworthy that Susthitamati appears as a devaputral-kanyii in Khotanese materials. 83 Susthitamati later disappears from the scene but MafijusrI rises to prominence.


An association of flowers with the attainment of Buddhahood seems strong that the A~tasahasrika makes Sakyamuni worship the Buddha Diparp.kara with five lotuses in his previous life, so that the latter predicts that he will become the Sakyamuni Buddha. See A~tasahasrika 19: WOGIHARA 1935: 747. 80 DUTT 1984: 111-112. 81 See HARRISON 2000.

82 This text surviving in Chinese translation is rendered into English in CHANG 1983, pp. 41-72, under the title 'How to Kill with the Sword of Wisdom.' 83 A reconstruction of the name of this devaputra/-kanya from Tibetan . Blo-rab-brtan into Sanskrit in both THOMAS (1935: 94, 179) and EMMERICK (1967: 9) is Susthiramati, which is very close to the original.


. Diwakar Acharya

The cult of Siikyamuni

Moving now to the north, we find Sakyamuni Tathagata with Vajradhara, the king of the Guhyakas, and possibly Maitreya - the one who is richly endowed with maitrl. 84 It is well-known from early textual sources that Vajradhara/Vajrapani is associated with Sakyamuni. Unfortunately the name of the other Bodhisattva has not survived, but since he is said to be connected with maitr'l, it is logical to identify him as Maitreya. He is known as the Buddha's companion or even as the future Buddha from the Pali sources, and in several Mahayana s~ltras he appears in the assembly of the Buddha asking questions to the Buddha himself or other fellow Bodhisattvas in the assembly. The Mahiivastu mentions Indra as Vajravaradhara, the holder of a choice Vajra, with Sakyamuni, depicting him as the latter's protecter. 8S However, the A~tasiihasrikii states that VajrapaI).i, the great Yak~a, is the constant companion of the irreversible Bodhisattva. 86 Though identified in this way variously as Indra or a Yak~a, there is no doubt that 'the holder of the Vajra' is associated with Sakyamuni as his protector. I am not aware of any text which brings Maitreya and Vajradhara together as the attendants of Sakyamuni or any
84 The names ending in -eya are in principle metronymic, but one should not forget that there are so many words ending in -eya which do not have metronymic connotations (see Wackernagel1987: 505-511). All of them, however, can be interpreted as having some specific, mainly causal, relation with the word they are derived from. But still, 'being rich in X' is not one of the meanings attested and should be taken as an 'interpretation.' 85 SENART 1882: 157: agrato vajravaradharo tridasagurii abaddhamaf}fciio I indro sahasranayano gacchati purato naravarasya I 86 A~tasahasrika 17 (WOGIHARA 1935: 683): punar aparal?l subhate avinivartanfyasya bodhisattvasya mahasattvasya vajrapaf}ir mahayak~o nityanubaddho bhavati I sa durdhar~o bhavati, anatikramm:zryas ca bhavati maml~yair va amanu~yair va, durasadaf:t sarvasattvanal?l. . .. ebhir api subhate akarair ebhir lingair ebhir nimittaif:t samanvagato bodhisattvo mahasattvo 'vinivartanfyo 'nuttarayaf:t samyaksal?lbodher dharayitavyaf:t. The Dasabhilmika echoes the same idea.

Mahayana Buddhism and SukhavatI cult in India


other Buddha, or as one of the interlocutors of Dharma. Since there ~re no relevant textual sources, they are not identified even in the rare cases that they appear in early images. Vajrapal).i is rather the default identification of the first Bodhisattva, but Maitreya always remains unidentified, if he is not misidentified either as Brahma or some other Bodhisattva. 87 As Sakyamuni disappears or is renamed in the scheme of paficajinamaflq,ala, this triad of Sakyamuni, Maitreya and Vajradhara has further special value. The directions of the Buddhas
It is quite striking that Amitabha is placed in the south in this inscription, while the Pure Land siitras, and some other Mahayana satras too, locate him in the west in SukhavatI together with his

two Bodhisattvas. This compels me to investigate further the issue of the assignment of the Buddhas in various directions. TheAk~obhyavyiiha centres on Ak~obhya who presides over the world of Abhirati in the east. The siitra, however, mentions three other Buddhas: Sakyamuni as the narrator of the siitra, *Suvarl).apu~pa/-padma as the successor of Buddha Ak~obhya, and Buddha *Visalanetra88 under whose guidance the would-be Ak~obhya adopted the path of Bodhisattva in the past; but the sLltra does not speak of their directions. The A~tasiihasrikii mentions that there are innumerable Buddha-fields with many Buddhas presiding over them in all ten directions but does not name them. However, in the nineteenth chapter, the siitra implies a set of four Tathagatas in a successive row: Ak~obhya in the world of Abhirati, Dlparp.kara in the city of DIpavatI in the distant past (but it is unclear if it was in the Saha world itself or somewhere else), and Suvafl!apu~pa and Sakyamuni in their

87 The figures of the Ramnagar stele, which dates from the year 32 (equivalentto 110 or 159 CE) and is preserved at the National Museum, New Delhi, can be identified as Sakyamuni with Maitreya and Vajradhara. See the figure numbered 13 in MYER 1986. For representations of Maitreya and his attributes in different periods, see BHATTACHARYA 1980. 88

For the name of this Buddha, see NATTIER 2000: 85, fn. 45.


Diwakar Acharya

Buddha worlds unspecified as regards their naII!e and location in a given direction. Amitabha and his Bodhisattvas, Avalokitesvara and Mahasthamaprapta, do not appear in this sutra. Maitreya appears here as a Bodhisattva, but Vajrapfu:1i is merely a yak~a; and . the lord of Saha is still Brahma, not Sakyamuni. In the Paficavirrzsatisiihasrikii, one Buddha with one Bodhisattva is placed in each of the ten directions, but their names are new and arbitrarily created, following an imaginary scheme. For example, the Buddha in the south is Asokasn, his Buddha world is named Sarvasokapagata, and the Bodhisattva there is named Vigatasoka. Apart from this list, the sutra now and again mentions four Buddhas in their respective worlds, with which its redactor appears to be intimately acquainted: Ak~obhya in Abhirati, Sakyamuni in Saha, Dlprupkara in DlpavatI (though only a city, not a Buddha world),89 and Samantakusuma in PadmavatI with his two Bodhisattva attendants. Amitabha does not appear in this sutra though Avalokitesvara and Mahasthamaprapta show up in the assembly.90 All these sutras look to the east as the direction of Ak~obhya and believe the world of Sakyamuni to be located in the west. Another one of the earliest sutras which could be grouped together with the above sutras is the Surarrzgamasamiidhi, which makes a devaputra named *Matyabhimukha come to the assembly of Sakyamuni from the world of Abhirati in the east; otherwise the Buddhas are not assigned to specific directions in this text. The longer Sukhiivatfvyuha places Amitabha in the west with the . .two Bodhisattvas, and is not concerned with the direction of other Buddhas. There Sakyamuni is also mentioned, located in Saha, but Ak~obhya has been completely ignored; he does not feature even in the long list of arbitrary names of Tathagatas. But, as SCHOPEN has informed us, this sutra "explicitly refers to a samantabhadracaryii," suggesting "some kind of linkage between the Bhadracarfprartidhiina and the cult of Amitabha" (p. 179). Amitabha's location is fixed also in the Pratyutpannabuddhasammukhiivasthitasamiidhi 89 It seems permissible to speculate that the concept of a Buddha city precedes the concept of a Buddha world. 90 See DUTT 1934: 5.

Mahayana Buddhism and SukhavatI cult in India


L a relatively early text that mentions Amitabha even though it is not

concerned with his cult proper. The other sutras of the SukhavatI tult follow suit.-The Bhai$ajyagurusutra, which promotes a new cult of the Buddha Bhai~ajyaguru and shows its affiliation :with the cult of Amitabha, assigns the two in the east and the west respectively. Even after the introduction of the Buddha Amitabha, however, many sutras are reluctant to fix him in the west. The Samiidhiriija 'groups innumerable Buddhas in four directions,91 but does not hame them. More than once the sutra mentions Ak~obhya, Amitabha, Sakyasirp.ha/-muni, and DIparp.kara, but locates only Ak~o bhya in the east.92 If we compare these with the four Buddhas known to the redactor of the Paficaviytzsatisiihasrikii, we can see 'that the Buddha Samantakusuma of the world of PadmavatI has been dropped here in order to accomodate the Buddha Amitabha ~f the world of SukhavatI. Similarly, the Vimalakfrtinirdeia in'i:Iudes Sakyamuni, Amitabha, Ak~obhaya and Prabhfitaratna in a list of thirteen Tathagatas without specifying their directions. ~i"~The Mahayana sutras of the subsequent period present a series Mnew Tathagatas and locate them in six, eight, or ten directions. ;[he shorter Sukhiivatfvyuha, like other Mahayana siitras, first tnentions that innumerable Tathagatas exist in ten directions but ~names only a few of them, and only from six directions. Unlike 'the longer version, it does not drop Ak~obhya but places him in east. The Saddharmapur:ujarfka places sixteen princes of the Buddha Mahabhijfiajfianabhibhfi in pairs in eight directions, where I\lqobhya and Amitabha appear in their usual directions. In the ~uktaka chapter of the Ga1}tjavyuha (SUZUKI & InZUMI 1949:


Samiidhiriija 28.82f: piirvasyiif!L disi aprameyiin asaf!Lkhyeyiin buddhiin bhagavataJ:t pasyati. evaf!L dak~i1}asyiif!L pascimiiyiim uttarasyiif!L diSi aprameyiin asaf!Lkhyeyiin buddhiin bhagavataJ:t pasyati. so 'virahito ,bhavati buddhadarsanena.

,:?iSi 10kavisrutaJ:t I bodhisattvanayutaiJ:t puraskrtaJ:t

~) ,92 Samiidhiriija 14.68-69: gandhahasti purimiidisii gato 'k~obhyak~etra

siikyasif!Lhu dvipa,dendru prcchanii " sukhiivatfya varalokadhiituto mahiisthiimapriipta "avalokiteSvaraJ:t IbodhisattvanayutaiJ:t puraskrtaJ:t siikyasif!Lhu dvipadendru prcchanii "


Diwakar Acharya

81-82), Merchant Muktaka first says that he sees ten Tathagatas in their Buddha worlds (the names of both Tathagatas and their lands sound arbitrary and long), and once again (ibid. 82) says that whenever he wants he can see Amitabha in SukhavatI, Vajrabha in CandanavatI, Ratnabha in GandhavatI, Ratnapadmabha in PadmavatI, Santabha in KanakavatI, Ak~obhya in Abhirati, Sirp.ha in Suprati~!ha, Candrabuddhi in Adarsama1f<;ialanirbhasa, and Vairocana in RatnasrIharp.sacitra. This time the number is nine, the order is unusual, and directions are not specified. 93. Several mediaeval Mahayana sutras composed subsequently mention Amitabha in the world of SukhavatI without specifying the direction. The Ratnaketuparivarta mentions Sakyamuni and Amitayus without assigning them to specific directions. Instead, it states a promise of Amitayus that he would be doing this and that in the future (pascime kiile); a reference to time instead of space. The Rii$trapiilapariprcchii, however, mentions only three Buddhas: Amitayus, Aksobhya, and Siddharthabuddhi (probably an allusion to Sakyamuni) in passing without specifying their directions. The Larikiivatiira mentions Amitabha's SukhavatI as the source of everything including Jinas and Bodhisattvas. Similarly, the Sarvatathiigatiidhi$thiina shows its afiliation with the cult of Amitabha by mentioning him alone and depicting an access to his SukhavatI as the final reward. 94 Now we have a more or less clear picture: the Prajfiiipiiramitii and affiliated sutras invariably assign Ak~obhya in the east; the sutras of the Sukhiivatl cult and those sutras which are under the influence of this cult assign Amitabha in the west (and Ak~obhya
93 It is noteworthy that the list of Tathagatas in this passage of the GalJavyaha begins with Amitabha and it is even possible that he is placed in the east in that scheme. In the same way, Sakyamuni's world, Saba, is positioned in the west in the Larger Prajiiiipiiramitiisatra. 94 The SuvarlJaprabhiisa, which is regarded as comparatively late, gives what appears to be a scheme of a caturvyahacaitya and names Ak~obhya as the Tatbagata of the east, Amitabha of the west, Ratnaketu of the south, and Dundubhisvara of the north. See the satra 1.4: ak~o bhyariijalJ purvasmin dak~ine ratnaketunii Ipascimiiyiim amitabha uttare dundubhisvaralJ II

Mahayana Buddhism and SukhavatI cult in India


in the east if they mention him). However, most of the sittras which combine both of these traditions are reluctant in fixing the directions of Amitabha and other Buddhas, though they generally pinpoint the direction of Ak~obhya. . Implications and outcomes In the light of the above observations, another important issue can be better explained: the process of inclusion of the cult of Amitabha in a unified cult and identification of access to SukhavatI as the ultimate religious goal. Once this process is properly explained, itwill help us to understand the formation of our caturvyiihacaitya ina better way, and it can also shed new light on the chronology of hfew early Mahayana siitras. According to SCHOPEN, who identified SukhavatI as 'a generalised religious goal,' "the fact that rebirth in SukhavatI is promised as a reward in conjuction with the cult of the book, or the cult ;of a specific book ... clearly indicates that SukhavatI hereL in the Samadhiraja and subsequent siitras,] must have been conceived of asa generalised religious goal in no way attached specifically to ~hecult of Amitabha" (p. 166). However, SCHOPEN was unable to determine "the degree to which this process of generalisation and ~isassociation effected a decline and weakening of the specific cult Of Amitabha as a separate entity" (p. 183), and expressed the hope that future studies would shed light on this issue . . . . . He was looking at the issue, I would say, from only one side. 1Iis starting point was SukhavatI's attestation in the Bhai~ajya guru, Samadhiraja and subsequent mediaeval Mahayana siitras asa generalised religious goal. He did not inquire into the pre"~iling situation at the time the cult of Amitabha came into existence. Consequently, he was unable to realise the important point that Amitabha's SukhavatI arose only after Ak~obhya's Abhirati as such a goal,95 There was a stage when Ak~obhya's world of Abhirati

He was, however, aware of the need for defining the relation of This need has by now been served, to certain extent, by Jan NATTIER'S articles on the cult of Ak~obhya.
. 95

Ak~obhya with early siitras.


Diwakar Acharya

was known but Amitabha's world of Sukhavatl was not. The Ak~o bhyavyuha, A~tasahasrika, A~tavi1?lsatisahasrika, Paficavi1?1satisahasrika, and Surarrzgamasamadhi represent this stage. Let me first quote some lines from the Ak~obhyavyuha:
Sariputra, if good men or good women [who follow the Bodhisattva path] after their death in this Buddha-land or another Buddha-land, have been born, are being born, or will be born in the Buddhaland of Tathagata Ak~obhya ... (CHANG 1983: 327). Sariputra, those Bodhisattvas who have received my prophecy and attained nonregressian will be born in Ak~obhya Buddha's land (ibid. 329). Now here are two passages from the ninteenth Chapter of the A~tasii hasrikii with my translation. Both of these fall in the portions identified as additions by CONZE:
seyam iinanda gwigadevii bhaginf strfbhiivaf!l vivartya puru~abhiivaf!l pratilabhya itas cyutvii 'k:jobhyasya tathiigatasyarhataJ:t samyaksa1Jlbuddhasya buddhak:jetre 'bhiratyiif!l lokadhiitiiv upapatsyate. tatra copapannii ak:jobhyasya tathiigatiisyarhataJ:t samyaksaf!lbuddhasyantike brahmacaryaf!l cari:jyati. tatas cyutii satf buddha-k~etriid buddha-k~etraf!l saf!lkrami:jyati avirahitii tathiigata-darsanena. (WoGIHARA 1935: 745)

This goddess of the Ganges, Ananda, when she vanishes from this world, she will sever her existence as a woman, assume manhood, and be born in the Abhirati world, the Buddha-field of the Tathagata Ak~obhya, the Arhat, the fully enlightened. Having reached there she will observe the brahmacarya vow in the presence of Tathagata Ak~obhya, the Arhat, the fully enlightened. When vanished from this world, she will pass from one Buddha-field to another, never deprived of the sight of the Tathagatas.
uttfrIJa-pankiis te bodhisattvii mahiisattviiJ:t, ye ak~obhyasya tathiigatasyarhataJ:t sainyaksa1J1buddhasya buddha-k~etre brahma-carya1J1 caranti. bodhi-parini:jpatty-upagatiis te iinanda bodhisattvii mahiisattvii veditavyiiJ:t. (WOGIHARA 1935: 746)

Those great Bodhisattvas, who conduct the brahmacarya vow in the Buddha-field of Tathagata Ak~obhya,96 the Arhat, the fully enlight-

96 This role of the teacher or guide of the Bodhisattvas born in his world is found attributed also to Amitabha in the Samiidhiriija. See

Mahayana Buddhism and SukhavatI cult in India


ened, should be known as the ones who have got rid of the mire, who have reached near the accomplishment ofenlightenment. The Surwpgamasamiidhi and Vimalaklrtinirdeia, though they do hot refer to access to the world of Abhirati as a religious reward, are engaged with Ak~obhya. They narrate stories of a Bodhisattva who comes to the Sahaworld from the Buddha Ak~obhya's world of Abhirati for the sake of perfection of all beings. In this respect, .these two texts are different from the rest. However, I think this pecularity is intended. These two texts are intended for more advanced and intellectually oriented people; their motive is different from that of the Ak~obhyavyuha and Prajiiiipiiramitii texts, and so the process has been reversed to suggest that they can have the same purity in this world. In the sutras cited or discussed above, except the Vimalakfrtinirdea, Amitabha or his two Bodhisattvas are not attested. The Vimalakfrtinirdea includes Amitabha and his two Bodhisattvas, though only in two separate lists of assembled Tathagatas and Bodhisattvas. 97 Amitabha appears in the same way in the Satasiihasrikii, imd his Bodhisattvas are included in the assembly in the Paiicavirfl'satisiihasrikii also. My guess is that these sCUras stand at the .~eam point of the first and second stages. < At some time in this stage the Amitabha cult, which must have existed as a minority cult in certain secluded regions, rose to promiIlenCe to compete with and finally eclipse the cult of Ak~obhya. :rhe proven existence of Amitabha in the Northwest of the Indian subcontinent earlier than anywhere else in India might support this

2005: 171. 97 As NATTIER (2000: 80, fn. 19) has pointed out, two translations of the Vimalakfrtinirdeia made in the third and fifth centuries present "a particularly intriguing tidbit of evidence" for the rise of the Amitabha cult by a change in the sequence of names in a list of Buddhas. As she . ~6tes, "Ak~obhya appears first after Sakyamuni in the list of Buddhas ~iven in Chih Ch'ien's translation of the Vimalakfrtinirdeia, while by the . time of Kumarajlva's translation Amitabha has now been moved to the first place" from the sixth. This "suggests that the cult of Ak~obhya was ,gradually being eclipsed by that of Amitabha."


Diwakar Acharya

argument. Even though the longer version of tbe Sukhavatfvyuha does not mention Ak~obhya, a reference to Samantabhadra's vows in the stUra is enough to hold that the redactor of this sutra was aware of some earlier cult, in which Samantabhadra had an important role. Unless and until the existence of an independent and archaic cult of Samantabhadra is confirmed, we cannot ignore the evidence of our inscription which makes Samantabhadra subordinate to Ak~obhya.98 There was another stage when both of the two Buddhas were known, and their lands were regarded simultaneously as the ultimate religious goal or reward, as two alternatives. The Samadhiraja and AjitasenavyakarafJa contain some traces of it. In the Samadhiraja, access to the Buddha Ak~obhya and his world Abhirati is described as the final religious goal, side by side with access to the Buddha Amitabha and his world of SukhavatI. 99 Let us look again at the following very exceptional Budhist Sanskrit verse from the Samadhiraja:
tatha punar amitiiyu te~a tatro bhii~ate buddha-aneka-iinusarrt siim sarva imi [sukhiivatfrrt pravi~to abhirati gatva] ak~obhya pasyi buddhalJz.

So also the Buddha Amitayus, to those there declares various kinds of blessings: all these have entered SukhavatI, and having gone to Abhirati, will see the Buddha Ak~obhya.1oo

98 The lexicon Amarakosa, which perhaps belongs to the sixth century (see Vogel 1979: 309-310), lists Samantabhadra as a name of the Buddha, while it does not list any name of celestial Buddhas or Bodhisattvas. This alone cannot be conclusive but can keep open the possibility of Samantabhadra being in the centre of an independent and earlier cult. 99 Samiidhiriija 34.48: susarrtgrhftviin ima buddhabodhirrt dhiiretva
nityarrt ca hi gauravelJa I te arthu krtvii vipulalJ2 prajanii1J2 drak~yanti ak~obhya nariilJam uttamam II 100 SCHOPEN 2005: 163-164; Sanskrit verse as cited by SCHOPEN; he has chosen the reading of the oldest manuscript from Gilgit, sarva imi against the reading of the edition, sarvi imi. However, his translation is problematic. It runs as follows:

"So also the Buddha Amitayus, to those there declares various kinds of

Mahayana Buddhism and Sukhavatl cult in India


that his devotees go and see Ak~obhya, the Buddha spanning all three times, after entering his SukhavatI ","arId. This might point to a stage where the cult of AmiHibha was trying to engulf the cult of Ak~obhya. This goes well with one of the boons promised to all inhabitants of SukhavatI which states that they can fly to other Buddha-fields to make merit by worshipping the other Buddhas in them. In the AjitasenavyakarmJa too, the pure lands of Ak~obhya and Amitabha, Abhirati and SukhavatI, are shown as available goals, but it is less likely that they are "conceived of as being of the same order" as Schopen (p. 158) argued. In the whole of the Ajitasena'Y)lakaralJa, Ak~obhya's Abhirati is mentioned only once. As the passage states, when the Buddha entered the city of SravastI nintynine kotis of niyutas of hundreds of thousands of beings were established in the Buddha world of SukhavatI and eighty-four kotis ofniyutas of hundreds of thousands of beings were established in the Buddha world of Abhirati. 101 Here the number of people established in Abhirati is smaller that the number of beings established inSukhavatI, and Abhirati comes second to SukhavatI in order. This suggests that the cult of Amitabha has not yet entire engulfed the cult of Ak~obhya but was in the process of subduing it. It seems that even this subordinative reconciliation was shortlived. We soon find SukhavatI being unanimously described as the final religious reward in the last phase of this process of identificat~on of such a reward. All the passages SCHOPEN selected and ana~ysed, except those from the Samadhiraja andAjitasenavyakaralJa, represent this stage. In sum, the following observations can be made: a) the presence and influence of the Buddha Ak~obhya is seen in earlier sutras, b) 'so()n Amitabha arrives on stage, and for some time the new cult
blessings: 'You will all go to my Sukhavati.' Having gone to Abhirati, theysee the Buddha Ak~obhya." 101 SCHOPEN (2005: 158) quotes this passage but arrives at the conclusion that ''Abhirati and Sukhavati are here clearly conceived of as being of the same order, and there is no distinction of, or preference for, one over the other."

lIere AmiUibha declares


Diwakar Acharya

. struggles to engulf the older cult, c) gradually AmiUibha becomes so prominent that all other Buddhas including Ak~obhya are subordinated, and d) newer Mahayana siitras do not even mention many of the subordinated Buddhas.

* * *
At this juncture, I am tempted to produce a relative chronology of early Mahayana siitras, using their affinity or affiliation with the cults of Ak~obhya and Amitabha as a criterion.
Original parts of the A!ftasiihasrika: Originally Ak~obhya is foreign to the sutra; cf. CONZE 1967. Ak!fobhyavyuha: This sutra mentions the genesis of the Buddha Ak~obhya, his parinirvalJa is imagined, and his career is modeled after that of Sakyamuni. Ak~obhya additions to the A!ftasahasrikli: Ak~obhya is already a Buddha, his genesis is not discussed. A!ftadasasahasrika and Paficavilflsatisiihasrikii: These do not mention Amitabha or his Bodhisattvas. Longer Sukhavatfvyuha: This sutra mentions the making of the Buddha Amitabha styled after the Ak!fobhyavyuha, and adopts Samantabhadra's vows. Pratyutpannabuddhasammukhiivasthitasamiidhi: The Suralflgamasamadhi refers to this sutra, which mentions Amitabha. Suralflgamasamadhi: This sutra mentions Ak~obhya but is more interested in the conduct of the heroes than the devotional path that provides rebirth in a Buddha world; it refers to the Pratyutpannabuddhasammukhavasthitasamadhi though not to Amitabha. Satasiihasrikli and Vimalakfrtinirdda: These sutras mention Amitabha and Avalokitesvara only in passing. Shorter Sukhiivatfvyuha: The Buddha of SukhavatI becomes Amitayus; the Buddhas of six directions are specified. Samiidhiriija and AjitasenavyakaralJa: The cults of Ak~obhya and AmiUibha overlap, but there are indications that the cult of Amitabha is rising into prominence. Most of the time, he is referred to with his new name.

* * *

Mahayana Buddhism and Sukhavati cult in India


Now let us return to our caitya-inscription. The fact that Amitabha is placed in~e south in our caitya-inscription, suggests that the presentation of a unified cult was the first priority in the choice of scheme followed in this caitya, rather than assigning Amitabha to his original place. In spite of this, his triad has remained intact in the unified cult of our caturvyuhacaitya. Most probably the cults united in the caitya did not lose their .individual identities. Though not separate and independent, they existed embedded in the united cult, as long as the Mahayana perspective prevailed. The reality was that the independent identities of all individual cults involved in the caitya were not highlighted in the 'books' promoting particular cultS.l02 It is not even necessary that what is going on in the realm of lay practices is always reflected in high 'books' of philosophical or mythological nature. Since the triad of the Buddha Amitabha is known from the Sukhiivatfvyuha which was translated already in the second century; since the triad of the Buddha Sakyamuni is depicted in the Ramnagar stele dated year 32 (equivalent to 110 or 159 CE); and since the other triad of the Buddha Samantakusuma is attested in the Paficavil!lsatisiihasrikii which could be placed around the same period (the Chinese translation requires a mid-third century date at the latest), no doubt remains about the fact that triads of the Buddhas were well known by the early second century CEo Let us try to find out when a fusion of these four triads into a caitya would pave occurred. This must have happened before the time assigned ;to our inscription, the late sixth century.l03 We can use the contents of the inscription to find an answer to this question.
102 However, these books sometimes exceptionally allude to some components of these cults. For example, the Saddharmapu1Jt!arfka and Vimalakfrtinirdeia allude to the fact that Samantabhadra and Vimalaklrti are flssociated with Abhirati. 103 The late sixth century is rather the time of late Mahiiyanic development gradually heading towards Tantrism. The caturvyuhacaitya was the working ground for Tantric traditions in the next phase of Buddhism. There we find a set of five Buddhas, four in four directions and one in the centre or pinnacle. In those traditions the set of five Buddhas is completed adding Vairocana at the top of the dome. In this set of five


Diwakar Acharya

As I argued earlier, the triad of Ak~obhya fOl.Jnd in our inscription represents a rare and archaic cult of Ak~obhya, which could have existed even before the available siitras at the level of devotional practice. The triad of Sakyamuni, too, is quite archaic, and cannot help us determine the date of the scheme of our caturvyiihacaitya. The triad of Samantakusuma, too, is archaic, but the evidence of the Pancavif!lsatistihasrikti suggests that it was known to the redactor of this satra but probably not to the redactor of the A~tastihasrikti. Above all, it should be noted that ih our caitya Sakyamuni and Samantakusuma, like Ak~obhya and Amitabha, are placed in their own Buddha worlds. This phenomenon is not highlighted in later periods. We know that the Buddhas flanked by two attendants on four sides of the SaikI stiipa I were added in the fifth century.l04 It is possible that the idea of combining four Buddhas or four triads was in place, at least in a formative state, already in the fourth century. The source of the exact scheme followed in this caitya-inscription remains unknown. lOS Nevertheless, what we can simDhyanI Buddhas, the Buddha Samantakusuma is transformed into Ratnasambhava and Sakyamuni, too, is replaced by Amoghasiddhi. 104 DALLAPICCOLA 2004: 805. lOS I think we cannot expect to find the exact scheme of this caitya in one Mahayana sutra, as it actually draws individual cults related with different books or traditions together, and we do not have access to ritual manuals of the period which might have recorded such schemes. There are only six Licchavi caityas with their Buddha images intact, and two of them do not place the Buddhas in customary directions (GUTS CHOW 1998: 32, and a review: DECLEER 2000). This suggests that more than one scheme was implemented to form caturvyuhacaityas. Things are less clear particularly when standing BuddhalBodhisattva images are involved (These caityas with standing figures probably predate those with triads). Art historians have offered competing theories to identify these images but dispute remains. They have also found 'erroneous cases' like Amitabha appearing twice (see GUTSCHOW 1998: 32). If we keep in mind that there were different schemes at work, it becomes easier to interpret such irregularities, and we do not need to assume any 'errors.' I think the second figure identified by GUTSCHOW or other art historians as Amitabha

Mahayana Buddhism and Sukhavatl cult in India


ply observe here is that it combines the three major streams of Mahayana known to us, the cult of Ak~obhya,106 the Prajfiaparamita cult, and the cult of Amitabha, and perhaps the stream of older Nikaya Buddhism with many cult-branches in which Sakyamuni was worshipped under this or that name. For early Mahayana sutras, as SCHOPEN argues, "the image cult -like the stupa cult - is an already established part of Buddhist cult practice," and they promote "a whole series of already established religious actions undertaken with a specifically defined intention" (p. 118). However, as he states, "early Mahayana was neither involved with nor even interested in the early cult of images" (p. 116). "It was trying, most simply, to send its monks back to their books" (pp. 138-139) by promoting the cult of the book or specific books. On this ground, SCHOPEN concludes, "we are left, it seems, with the apparent fact that, at least in regard to major Buddhist cult forms the stupa and the image cult - the appearance of early Mahayana sutra literature had no effect" (p. 138). Though they sound important, these conclusions are a bit exaggerated. We cannot say that the appearance of early Mahayana sutra literature had no effect on the cult forms. There was some effect - rather mutual effect - and because of that the cult of the stupa, specific cults of images, and the cuIts of specific books eventually coalesced into a caitya. Buddha images were not an essential part of the stupa/caitya earlier, but by the time of mediaeval Mahayana they became so. No doubt, "since each text placed itself at the centre of its own cult, early Mahayana, rather than being an identifiable single group, was in the beginning a loose federation of a number of distinct though related cults, all of the same pattern, but each associated with its specific text" (p. 52). As it is a general tendency, at least of the laity, to reconcile and identify heterogeneous entities, there
could.be some other similar looking Buddha venerated in specific cults. 106 As I speculate, the cult of Ak~obhya, the Imperturbable, probably was originally associated with the heroic path later attached to Samantabhadra, and his name was a constant reminder to a would-be Bodhisattva not to stumble on the path.


Diwakar Acharya

was a clear need of reconcilation and fusion of these specifically heterogeneous but interrelated cults into one unified cult. For this purpose, specific Buddha/Bodhisattva images linked with specific books were made part of the stiipa. This fusion took place at a time when another fusion between the cult of the book and the cult of the stiipa was already at work. 107 If I may take liberty of speculating a bit, I find some scheme in the arragement of the Tathagatas in our caitya. In a sense, these four Tathagatas also represent different aspects of the 'Buddha, or a Bodhisattva. He is Ak~obhya, the Imperturbable, when he cultivates the six perfections and is on the path of universal good. He is Amitabha, the Buddha of unmeasurable light, when he radiates rays of omniscience, compassion and so on. He is Samantakusuma when he is enthroned in the bodhima1Jtj,a and enjoys the bliss of enlightenment, at the time when his achievement is celebrated by all divine and mortal beings. And, he is Sakyamuni, the Sakya sage, when he wanders and imparts the knowledge he has achieved. I want to make one more observation about the arrangement of Bodhisattvas in these sets. In each set, one Bodhisattva is relatively more exalted compared to the other. Samantabhadra has the reputation of a celestial Bodhisattva but Vimalaklrti is a layman living in VaisalI, as the Vimalakfrtinirdesa describes. Maitreya is a celestial Bodhisattva, but Vajradhara is a king of lower beings, the Guhyakas. Again, MafijusrI is depicted as a godly Bodhisattva since early times but Susthitamati is a son of god with no important role. Similarly, the name of Avalokitesvara itself suggests his . innate divine nature and texts describe him as such, but the other attendant Mahasthamaprapta was once in the mundane realm and has attained the Bodhisattva status with his efforts.

107 On the issue of the fusion of the two cults of the stupa and the book, see BENTOR 1995. As BENTOR has stated, early textual evidence for the practice of depositing texts or text portions in stupas is found in the Pratyutpannasutra.

Mahayana Buddhism and SukhavatI cult in India


ACHARYA, Diwakar. 2007. Anuparama's Dvaipiiyanastotra Inscription from the Early 6th Century. In: Journal of Indological Studies, No. 19, pp. 2952. Kyoto University. Ak~obhyavyaha. See CHANG 1983, pp. 315-338. ALSOP, Ian. Licchavi Caityas of Nepal: A Solution to the Empty Niche. Published online: http://www.asianart.com/alsop/licchavi.html (last accessed: 9 December 2009). Astasiihasrikii Prajfi.iipiiramitii. Abhisamayiila'!lkiiriilokii Prajfi.iipiiramitii., vyiikhyii (Commentary on A~tasiihasrikii Prajfi.iipiiramitii) by Haribhadra, together with the text commented on. Ed. Unrai WOGIHARA. Tokyo: Toyo Bunko, 1935. BENTOR, Yael. 1995. On the Indian origins of the Tibetan practice of depositing relics and dharanis in stupas and images. In: Journal of the American Oriental Society. VoL 115, No.2, pp. 248-261. Bhai-?ajyagurusatra. Ed. N. Dutt with the assistance of D.M. Bhattacharya and Shivnath SASTRI. Gilgit Manuscripts, VoL L First edition: Srinagar 1934. Second edition: Delhi: Satguru Publications, pp. 1-32. BHATTACHARYA, Gouriswar. 1980. Stiipa as Maitreya's emblem. In: The SUipa: Its Religious, Historical and Architectural Significance. Ed. A. L. Dallapicc01a with S.2. Lallemant. Wiesbaden: Franz Steiner Verlag, pp. 100-111. Bodhisattvabhami of AsaIiga. See DUTT 1966 and WOGIHARA 1930. Buddhacarita. The Buddhacarita or Acts of the Buddha. Ed. and trans. E.H. Johnston, Punjab University publication 31. First edition: Lahore 1936. Reprint: Delhi: Motilal Banarsidass, 1984. CHANG, Garma C. C. 1983. A Treasury of Mahiiyiina S~7tras. Pennsylvania State University Press. CONZE, Edward. 1967. The Composition of the A~tasahasrika Prajfiaparamita. In: The Bulletin of the London School of Oriental and African Studies, 14, pp. 251-62. Reprinted as a chapter in Thirty Years of Buddhist Studies. Oxford: Bruno Cassirer, 1967, pp. 168-184. --- 1975. trans., The Large Sutra on Perfect Wisdom with the Divisions of the Abhisamayiilmikiira. Berkeley: University of California Press. DALLAPICCOLA, A.L. 2004. Stiipa. In: Encyclopedia of Buddhism. ed R.E. BuswelL New York: 2004, pp. 803-808. Dasabhamikasatra. Dasabhamfsvaro niima Mahiiyiinaslitram. Revised and Edited by Ryilko Kondo. Reprint of the original edition of 1936. Kyoto: Rinsen Book Co., 1983.


. Diwakar Acharya

DECLEER, Hubert. 2000. Chaityas of the Valley: The Grand Art-Historical Inquiry [Review of Gutschow 1997]. In: Buddhist Himalaya, Vol. X, Nos. 1-2 (1999/2000). Available online: http://buddhim.20m.com/1O-6.htm (last accessed: 9 December 2009). DIKSHIT, B.S. 1888. The Twelve-year Cycle of Jupiter. In: Indian Antiquary 17. Bombay: Education Society Press, pp. 1-7,312-317. DUTT, N. 1934. The Pancavi1!lsatisahasrika Prajnaparamita. Calcutta oriental Series, 28. Calcutta Oriental Press. --- 1966. Bodhisattvabhami. Tibetan Sanskrit Works Series, 7. Patna: K.P. Jayaswal Research Institute. . - - 1984. AjitasenavyakaralJa. Second edition. Delhi: Sri Satguru Publications., 1984, pp. 103-146. Based on the edition by N. Dutt with the assistance ofD.M. Bhattacharya and Shivnath Sastri: Gilgit Manuscripts, vol. I, Srinagar 1939. EDGERTON, Franklin. 1953. Buddhist Hybrid Sanskrit Grammar and Dictionary: Vol. II: Dictionary. London: Oxford University Press. EMMERICK, R.E. 1967. Tibetan Texts Concerning Khotan. London Oriental Series, 19. London: Oxford University Press. FLEET, J. F. 1888. The Use of the Twelve-year cycle of Jupiter in Records of the Early Gupta Period. In: Indian Antiquary 17. Bombay: Education Society Press, pp. 331-339. FUSSMAN, Gerard. 1999. La Place des SukhavatI-vyilha dans Ie Bouddhisme Indien. In: Journal Asiatique, Vol. 287, No.2, pp. 523-586. GalJtjavyaha. Ed. D.T. SUZUKI & Hokei IDZUMI. New revised edition. Tokyo: Society for the Publication of Sacred Books of the World, 1949. GNOLI, Raniero. 1956. Nepalese Inscriptions in Gupta Characters. Rome: IsMEO. GUTSCHOW, Niels. 1997. The Nepalese Chaitya: 1500 Years of Buddhist Votive Architecture in the Kathmandu Valley. With drawings by Bijay Basukala et aI, and an essay by David Gellner. Stuttgart & London: Ed. Axel Menges. English translation by Philip Pierce. Lumbini: Lumbini International Research Institute. HARRISON, Paul. 1990. The Samadhi of Direct Encounter with the Buddhas of the Present. Tokyo: The International Institute for Buddhist Studies. --- 2000. MaiijusrI and the Cult of the Celestial Bodhisattvas. In: ChungHwa Buddhist Journal. Vol. 13.2 (May 2000), pp. 157-193. HUNTINGTON, John C. 1980. A Gandharan Image of AmiUiyus' SukhavatI. In: Annali dell' Istituto Orientale di Napoli, Vol. 40, pp. 651-672. INAGAKI, Hisao. 1995. The Three Pure land Sutras. A Study and Translation from Chinese. In collaboration with Harold Stewart. Kyoto: Nagata Bunshodo.

Mahayana Buddhism and SukhavatT cult in India


JONES, J. J. 1952. The Mahiivastu. Vol. II. Translated from Buddhist Sanskrit. London: fali Text Society.

Kiisyapaparivarta: A Mahiiyiinasiitra of the Ratnahtta Class in the Original Sanskrit, in Tibetan and in Chinese. Ed. Baron A. von Stael-Holstein.
Shanghai, 1926.

___ Romanised Text and Facsimiles. Bibliotheca Philologica et Philosophic a Buddhica, V. Ed. M.1. Vorobyova-Desyatovskaya in Collaboration with
S. Karashima and N. Kudo. Tokyo: Soka University, 2002. KETKAR, Venkatesh B. 1923. Indian and Foreign Chronology with Theory, Practice and Tables, B.C. 3102 to 2100 A.D. Bombay: British India Press. KIMURA, Takayasu. See PancaviJ?lsatisiihasrikii. Lalitavistara. Ed. P.L. Vaidya. Buddhist Sanskrit Texts, No. 1. First edition. Darbhanga: The Mithila Institute, 1958.

Lmikiivatiira. Ed. P.L. Vaidya. Buddhist Sanskrit Texts, No.3. Darbhanga:

The Mithila Institute, 1963 . . LAMOTTE, Etienne. 1998. SaraJ?lgamasamiidhisatra. The Concentration of Heroic Progress. An Early Buddhist Mahiiyiina Scripture. English translation from the French original of 1965 by Sara Boin-Webb. London: Curzon Press, 1998. LEVI, Sylvain. 1905-8. Le Nepal.' etude historique d'un royaume Hindou, Vols. I-II: 1905, Vol. III: 1908. Annales du Musee Guimet. BibliotMque d' Etudes, Tome XIX. . LEWIS, Todd T. 2004. From Generalized Goal to Tantric Subordination: SukhavatI in the lndic Buddhist Traditions of Nepal. In: Richard K. Payne and Keneth K. Tanaka, Approaching the Land of Bliss: Religious Praxis in the Cult of Amitiibha. Studies in East Asian Buddhism 17. Honolulu: Kuroda Institute, pp. 236-263 . .,--- 2004a. Nepal. In: Encyclopedia of B!iddhism. Ed. R.E. Buswell. New York, pp. 588-592. LUDERS, Heinrich. 1961. Mathurii Inscriptions. Unpublished Papers edited by Klaus L. Janert. G6ttingen: Vandenhoek & Ruprecht.

Manjusrfmalakalpa. Ed. T. Gal).apati SastrI. Part I. Trivandrum Sanskrit

Series, 70. Trivandrum: Government Press, 1920.

Mahiivastu. See SENART and JONES. MYER, Prudence R. 1986. Bodhisattvas and Buddhas: Early Buddhist Images from Mathura. In: Artibus Asiae, Vol. 47, No.1, pp. 107-142.
NATTIER, Jan. 2000. The Realm of Ak~obhya: A Missing Piece in the History of Pure Land Buddhism. In: Journal of the International Association of Buddhist Studies, Vol. 23, No.1. pp. 71-102.


Diwakar Acharya

--- 2003. The Indian Roots of Pure Land Buddhism: Insights from the Oldest Chinese Versions of the Larger SukhavatIvyilM. In: Pacific World. Third Series, No.5. pp. 179-20l. --- 2007. The Names of Amitabha/Amitayus in Early Chinese Buddhist Translations (2). In: Annual Report ofThe International Research Institute for Advanced Buddhology at Soka University for the Academic year 2006, pp. 359-394. Paficavi'!lsatisiihasrikii Prajfiiipiiramitii. Parivarta I. Ed. N. Dutt. Calcutta Oriental Series, 28. Calcutta Oriental Press, 1934. Parivarta II-VIII. Ed. Takayasu Kimura. Tokyo: Sankibo Busshorin, Vol. I (Parivarta II-III): 1986, Vol. II (parivarta IV): 1990, Vol. III (Parivarta V): 1992, Vol. IV (Parivarta VI-VIII): 2006. PANT, Nayaraj. 1986. Licchavi sa'!lvatko nin:taya [Determination of the Licchavi Saqlvatl Kathmandu: Royal Nepal Academy. The Larger Prajfiiipiiramitii. Ed. Stefano Zacchetti. In Praise of the Light: A Critical Synoptic Edition with an Annotated Translation of Chapters 1-3 ofDharmarak~a's Guang zanjing, Being the Earliest Chinese Translation of the Larger Prajfiiipiiramitii. Bibliotheca Philologica et Philosophica Buddhica, 8. Tokyo: Soka University, 2005. Pratyutpannabuddhasammukhiivasthitasamiidhi. See HARRISON 1990. QUAGLIOTTI, Anna Maria. 1996. Another Look at the Mohammed Nari Stele with the so-called "Miracle of SravastI." In: Annali del Istituto Universitario Orientale, Napoli. Vol. 56, pp. 274-289. Rii~trapiilapariPrccha. Ed. L. Finot. Bibliotheca Buddhica II. First edition, 1901. Reprint: Delhi: Motilal Banarsidass, 1992. Ratnaketuparivarta. Sanskrit Text. Ed. Y. Kurumiya. Kyoto: Heirakuji-Shoten, 1978. REGMI, D.R 1983. Inscriptions ofAncient Nepal. 3 vols. New Delhi: Abhinav Publications. RICCARDI, Theodore, Jr. 1980. Buddhism in Ancient and Early Medieval Nepal. In: A. K. Narain, Studies in History of Buddhism. Delhi: B. R Publishing Corporation, pp. 265-281. SaddharmapulJt;larfka. Ed. H. Kern and B. Nanjio. Bibliotheca Buddhica 10. St Petersburg, 1908-12. Samiidhiriijasatra. Gilgit Manuscripts. Vol. II, part 3. Ed. N. Dutt with the assistance of Shiv Nath Sastri. Calcutta, 1954.

tra. Ed. N. Dutt with the assistance of D.M. Bhattacharya and Shivnath Sastri. Gilgit Manuscripts, Vol. I, pp. 103-146. SCHOPEN, Gregory. 2004. Mahayana. In: RE. Buswell. Ed., Encyclopedia of Buddhism. New York: 2004, pp. 496-498.

Mahayana Buddhism and Sukhavatl cult in India


__ 2005. Figments and Fragments of Mahayana Buddhism in India: More Collectedpapers. Honolulu: University of Hawai'i Press. SCHUSTER, Nancy 1981. Changing the Female Body: Wise Women and the Bodhisattva Career in some Maharatnakutasutras. In: Journal of the International Association of Buddhist Studies. Vol. 4/1, pp. 24.:..69. SLUSSER, Mary S. 1982. Nepal Mandala: A Cultural Study of the Kathmandu Valley. Princeton: Princeton University Press. STRAUCH, Ingo. 2007. The Bajaur collection: A new collection of Kharo:;thf manuscripts. A preliminary catalogue and survey. Online version 1.0 (August 2007). Available online: http://www.geschkultfu-berlin.de/ e/indologie/bajaur/publication/strauch_2007_1_O.pdf (last accessed 25 December 2009). The Larger Sukhavatlvyuha. Ed. Atsuuji Ashikaga. Kyoto: H6z6kan Library, 1965. The Shorter Sukhavatlvyuha. Buddhabhii:;ita-AmitayulJ-Sutra.Translated from the Chinese version of Kumarajlva. Nishu Utsuki. Kyoto: Educational Department of the West Hongwanji, 1924.

Susthitamatidevaputraparipfcchii. See CHANG 1983, pp. 41-72. Suvan:wprabhasasutra. Ed. B. Nanjio and H. Idzumi. Kyoto: Eastern Buddhist Society, 1931. THOMAS, F.W. 1935. Tibetan Literary Texts and Documents concerning Chinese Turkestan. Part 1. London: The Royal Asiatic Society. VAJRACHARYA, Dhanavajra. 1975. Licchavikalaka Abhilekha. [Inscriptions of the Licchavi Period.] Kathmandu: Centre for Nepal and Asian Studies.

Vimalaklrtinirdea. A Sanskrit Edition Based upon the Manuscript Newly Found in the Potala Palace. Study Group on Buddhist Sanskrit Literature, Taisho University. Tokyo: Taisho University Press, 2006.
VOGEL, Claus. 1979. Indian Lexicography. A History of Indian Literature, Vol. V, fasc. 4. Wiesbaden: Otto Harrassowitz. WACKERNAGEL, Jakob & DEBRuNNER, Albert. 1987. Altindische Grammatik. Band rr,2: Die Nominalsuffixe. Unveranderter Nackdruck der ersten Aufiage 1954. G6ttingen: Vandenhoeck & Ruprecht. WOGIHARA, Unrai. 1930-36. Bodhisattvabhumi: A Statement of Whole Course of the Bodhisattva. Tokyo: Toyo Bunko. --- 1935. See A:;tasahasrika Prajiiaparamita.

Early Chinese Buddhist translations

Contributions to the International Symposium "Early Chinese Buddhist Translations," Vienna 18-21 April, 2007

Guest editor

Max Deeg


Max Deeg

It is certainly by chance that this issue of the Journal of the

International Association of Buddhist Studies, containing articles on Chinese translations of Buddhist texts, appears following Prof. Jikido Takasaki's presidential address to the Association's XIVth Conference held in London in 2005, which was published in an earlier issue under the title "Between translation and interpretation - Cases in the Chinese Tripitaka."l The coincidence referred f~ by Prof. Takasaki of recent de~elopments in Buddhist Studies was, however, what led to a conference being held on the topic of Chinese Buddhist translations, of which the papers in this issue are the result.2 In recent years there has clearly been a great increase in the number of scholars doing philological work on the early Chinese
1 Jikido Takasaki, "Between translation and interpretation - Cases in the Chinese Tripitaka - (Presidential address at the XIVth Conference of the International Association of Buddhist Studies, London, August 29 September "3, 2005)," JIABS 29/1, 2006 (2008): 3-20. 2 It is my pleasure, also on behalf of the participants of the conference, to express my gratitude to Prof. Ernst Steinkellner, who encouraged me to organize this event under the auspices of the Austrian Academy of Sciences in Vienna. With support from the team at the Academy's Institute for the Cultural and Intellectual History of Asia under its new director Dr. Helmut Krasser, the conference was held in a warm and relaxed atmosphere. Dr. Birgit Kellner and Dr. Krasser subsequently offered to publish the resulting articles in this special issue of the JIABS. We are grateful for their patience in waiting for the articles and for their diligent editing.

Journal of the International Association of Buddhist Studies Volume 31 Number 1-22008 (2010) pp. 79-82


Max Deeg

translations of Buddhist texts, texts that had been, albeit not entirely ignored, nevertheless neglected by earlier generations of scholars when compared to the Tibetan translations. Although later products, these were thought to represent more accurate versions of their underlying Indian originals. The appearance of new Buddhist manuscripts from Afghanistan, both in Sanskrit (in the Sch0yen Collection) and in GandharI (in the British Library Collection and, more recently, in the Bajaur Collection), has shown the value of Chinese translations, these being closer in time to the original composition of these valuable documents than the equivalent Tibetan renderings. Moreover, the Chinese canon contains, in early Chinese translations, texts that were never translated into Tibetan. Thus, editors of these newly-found manuscripts often find that they must rely on this Chinese material for their work. The traditional focus on the Prajfiaparamita literature has been expanded upon, and investigation of early Mahayana Buddhism and its texts has gained a new impetus. The search for new historical and contextual insights has led to meticulous and detailed studies of individual texts and the problems connected to their transmission and translation into Chinese. Yet when compared to Tibetan Studies, the philological grasp on the Chinese Buddhist translations seems still to be in its infancy. 3 For more than a century, such research was mainly conducted in Japan. The few Western scholars working with these texts were restricted to a few champions of the caliber of Sylvain Levi, Paul Pelliot, Paul Demieville, Etienne Lamotte, Ernst Waldschmidt, Erich Frauwallner and Erich Zurcher. This restriction in manpower certainly was not least due to the linguistic unwieldiness of some of the material. Classical Chinese, let alone the far from normative Buddhist Chinese, was to some extent impenetrable. There has
3 Werner Thomas, in an evaluation of the translations of Buddhist texts into Tocharian, obviously had problems in finding recent literature on Chinese translations when he quotes from an article by Kenneth Ch'en from 1960: Werner Thomas. Probleme der Ubertragung buddhistischer Texte ins Tocharische. Mainz: Franz Steiner Verlag: 7 (Akademie der Wissenschaften und Literatur, Mainz, Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse, Jahrgang 1989, Nr.lO).



been no standard tool for its study; nO grammar or dictionary in a Western or even East-Asian language is available that covers the whole range' of the texts. The widely used "Soothill-Hodous''4was drafted on the basis of the Song dictionary Fanyi-mingyi-ji 1J~~~:g ~~, "Collection of Translated Terms and Meanings," compiled by Fayun )"*~ (1088-1158), and therefore did not cover the early translations. The Soothill-Hodous also has shortcomings in certain terminological areas. 5 Terms and syntactical structures that were not understood either from the standpoint of classical Chinese or the underlying Indic text were easily dismissed as signs of poor comprehension on the part of the translator(s). In contrast, the Tibetan translations were appreciated as being at least as what looked like precise renderings of the Sanskrit originals, down to syntax and terminology.6 Due to political and social circumstances, the terminology of Chinese Buddhist translations was never standardized in the manner done for the Tibetan, where the terminological vyutpatti tradition was established in the period of the "early spread" (snga dar) of Buddhism in the eighth and ninth centuries. Attempts at identifying translation idioms of individual translators or groups, or at analyzing terminological and grammatical-syntactical peculiarities are still confined to a small, albeit growing, corpus of studies. There is still need for comprehensive working tools, a corpus verborum et terminorum, of the early Buddhist translations into Chinese. Habent colloquiafata eorum - the first proposal for a discussion forum on Chinese Buddhist translations was submitted during the Conference of the International Association of Buddhist Studies in
4 William Edward Soothill, Lewis Rodous, A Dictionary of Chinese Buddhist Terms, 1930. 5 See the caveats and remarks by Charles Muller in his "Preface to the Digital Edition" of the dictionary. 6 See David Seyfort Ruegg. "On Translating Tibetan Philosophical Texts." In: Doboom Tulku (ed.). Buddhist Translations - Problems and Perspectives. Delhi: Manohar 1995: 82f. The volume in which this article is published shows the imbalance of perception: despite its general title, it is entirely dedicated to various aspects of translating Buddhist Tibetan texts into Western languages.


Max Deeg

Lausanne in 1999. The urge to diminish my own ignorance and difficulties when dealing with Buddhist Chinese texts prompted me to consider convening a conference on the topic, and I gratefully took up the offer of Prof. Ernst Steinkellner, then director of the Institute for the Cultural and Intellectual History of Asia at the Austrian Academy of Sciences in Vienna, to organize such an event at the Academy with its support. I tried to gather as many of the scholars working in this field as possible, and to cover as many aspects of Chinese Buddhist translations as was feasible, but quite naturally I did not fully succeed: because of the usual restraints of academic life7 as well as the unforeseeable calamities of nature, a some obvi0us contributors could not attend the conference. There is, as far as I am aware, no edited volume or monograph on the subject of Chinese Buddhist translations. Hopefully this issue of JIABS will reflect both individual and general aspects and problems concerning the study of Chinese Buddhist translations and the language they use, and will lead to a broader awareness of this valuable area of Buddhist Studies as well as more engagement therein.

7 I recollect that Seishi Karashima was unable to attend for such reasons. a When already on his way to Vienna, Dan Boucher was caught in a blizzard on the east coast of the US, and after a considerable amount of patient waiting, frustratingly had to give up ever arriving.

Creating religious terminology

A comparative approach to early Chinese Buddhist translations 1 Max Deeg

The only translation "project" in the history of religions which can be said to match the transformation of a huge corpus of Buddhist texts into Chinese2 and later into Tibetan and other Central-Asian languages is the phenomenon of the translation of the Bible into Mediterranean languages as Greek and Latin, into Near Eastern languages as Syriac, Armenian and finally into the evolving nationallanguages of Northern Europe, especially into the different Germanic and Slavonic idioms. 3 The following paper will concentrate on the comparative aspect of Old-High German translations from Latin, the philological categories which were developed in this context by various scholars and the possible contribution of these for the analysis of early Buddhist Chinese terminology. Even if the linguistic and cultural preconditions are, in both cases, as different as will be sketched below, the structuring of the types of terminological creations in
1 It is my pleasure to thank my colleague Dr. Dan King, Cardiff University, for his function as ZokapilZa over my hybrid English, and the editors, Dr. Birgit Kellner and Dr. Helmut Krasser, for their valuable suggestions and comments. 2 See Zurcher 1999: 8: " ... the production of the earliest Buddhist texts in Chinese, around the middle of the second century CE, marks a 'linguistic break-through' in the spread of the dharma ... " . 3 An overview is found in Stegmuller's article in Hunger, Stegmuller, Erbse, etc. 1988: 149ff.

Journal of the International Association of Buddhist Studies

Volume 31 Number 1-2 2008 (2010) pp. 83-118


Max Deeg

Germanic languages is, in my eyes, a fundament~l and crucial one for any translation process. The advantage of such a comparison is that Germanic philology is clearly ahead in its philological analysis.4 The paper will concentrate on the earliest stage of the German language preserved, usually classified as Old High Germim (OHG). This language covers a period - if we start with the earliest literary production of translations in the form of interlinear glosses - of about 600 years between 500 and 1100 and is preserved in a variety of different dialects and mixed forms. Another historical advantage on the Germanic side is that we have a much better insight into the "workshops" of these early medieval translators -like, e.g., Notker (Labeo) of St. Gallen (ca. 950-1022)5 or Otfrid of WeiBenburg (ca. 800-870) - than in the Sino-lndic case where discussions about translation problems are not completely absent - as can be seen in the letter exchange between Shi Daoan ~'m~ (313-385) and KumarajIva/Jiumoluoshi M~.~{t (350-409) - but were held on an abstract level rather than dealing with technical issues of translation. 6 OHG loan vocabulary was developed over a period of half a millennium and a huge bulk of it became an integral part of the modern German language. Many conceptual terms and words for items of material culture in the language stem from a direct or indirect contact with the Romans or with Late Antique / Early Medieval Latinity. It is certainly true that in China the impact of
4 Another interesting example is the translation process of Greek texts into Syriac, that is from an Indo-European to a Semitic language, for the technical side of which see King 2008. 5 The medievalist Curtius has called etymology a specific "form of thinking" (Denkform) in the Middle Ages: see Curtius 1967: 486-490, but it seems to be clear that semantic analysis, whatever its character and quality may be in a specific case, is the basis of translation from any language into another. 6 A list of old versus new translation terms - called Qianhou-chujingyiji iW~CI:I~!J!;*2, "Record of differences in the earlier and later editions of the sutras" - in Sengyou's {~1{] Chu-sanzang-jiji CI:I~~*2~ (T.2145.5a.l3ff.), however, implies that there was a concrete discourse about technical translation issues.

Creating religious terminology


the Indian languages closely connected with Buddhism did not in'fluence the everyday language to the same degree as Latin did with the Middle- and North-European vernaculars, and the vocabulary (and the materialloans)1 was clearly more restricted to a religious sphere. It should be emphasized that this paper will not deal with syntactical issues of the translation process, although it seems clear that loan syntagma - the attempt to (re)construct the syntactical form of the original language in the target language - is an important factor of translation processes. B The composite verb forms in Germanic languages - the periphrastic perfect, plusquamperfect, future or conditional of the type "I have done" ("Ich habe getan"), etc. - are clearly products of original loan syntagma. 9 Loan syntagmata may provide better terminological criteria for identifying and verifying certain translators or translation "schools" but they are a difficult and complex issue in the Chinese translation texts, a discussion of which would go beyond the scope of this study.10 This paper will be mainly concerned with a specific kind of word analysis which must have been underlying the translation processes and is often called etymology but should rather perhaps be called semantic analysis. This analytic process is found in both traditions, the Western-European bible-translation and the Buddhist translations into Chinese and other Central-Asian languages. In the OHG tradition this process of creating terminology in translations is dis-

7 The impact on material culture is aptly discussed in Kieschnick 2003. B On this issue see Zurcher 1977. 9 See OHG. ih habem .gitan. On theoretical analysis of problems of syntax in translation see Notker's work, discussed in Backes 1982. 10 It is fair to state that by the digitization of the Chinese Buddhist corpus (especially the CBETA-edition) there has been a development of philological studies of Chinese Buddhist texts which could be called explosive - did the number of scholars involved not interdict this as an exaggeration. A real goldmine for experts' studies in this field is the Annual Report of the Institute for Advanced Buddhist Studies at S6ka University, as can be seen in the bibliographical references in this paper.


Max Deeg

cussed by one ofthe main protagonists involved, Notker, but on the Indo-Chinese side there is little direct evidence 'of a discussion of the theoretical and practical considerations involved, at least from the side of translators themselvesY The Buddhist translators were nevertheless, as can be shown by the analysis of various examples, trained in the tradition of the Indian semantic-etymological nirvacan a and applied it in their translation work. 12 By concentrating on the terminological aspect and its categorization I hope to contribute to a systematic framework' of analysis for the study of Early Chinese Buddhist terminology which may eventually help to clarify the greater developmental lines in the overall growth of this terminology from the third century until the Tang period and afterwards. While introducing a comparative aspect to the discussion I am, again, fully aware of the linguistic, literary, historical and cultural differences between the two translation processes compared. First of all, in the case of European tribal cultures like the Germanic, Celtic and Slavonic ones there. was no writing system for their languages before the advent of Christianity; the impact of late-Antique Latinity and its scholarship on the translation process applied to the newly arriving scriptures was necessarily an extremely strong one. In contrast to this, Chinese literary culture had already existed for several centuries - from the traditional viewpoint of the Chinese on their own past and antiquity even for millennia - and had a linguistic and expressive cultural apparatus

11 For an example of a theory of translation by the Chinese scholarmonk Yancong ~~ (557-610) see Held 1972. 12 For the development of nirvacana, its techniques and the bearing on Indian exegetic tradition see Deeg 1995, Kahrs 1998, and Bronkhorst 200l. Funayama's article in the present volume impressively presents one of these nirvacana-users, Paramartha/Zhendi Ji:g'lil' (500-569), at work. Similar to the European case where semantic analysis was standing in the Latin tradition of such authors as Marcus Terentius Varro (116-27 B.c.) and the later antique author Isidorus of Seville (ca. 560-636), the Asian translators, even if they do not refer to it verbally, built on the principles developed in the framework of autochthonous Sanskrit philology.

Creating religious terminology


which in quantity at least was a match for the Indian imports arriving with Buddhism. There is the typological difference between the two languages, the original language and the target language, in both translation processes: in the case of Germanic languages two inflective Indo-European languages with a certain flexibility of word formation through prefixes, suffixes and compounding-elements are involved, while in the Buddhist case the translation went from a similarly structured Indo-European language - Prakrt or, later on, Sanskrit - to an isolating language, Chinese, which did not have a Clear prefixal and suffixal word-formative and morphological system. This difference certainly had an impact on the way in which translators mastered their task of creating new terminology in the target language. The awareness of the difference is, though very much "romanticized," visible in Chinese sources, as for instance in Sengyou's {~ttI (445-518) catalogue-descriptor Chu-sanzang-jiji

The plainness and elegance of expressions are [also] tied to the management of the brush. Some excelled in the Hu meanings, but they did not comprehend the Chinese purports; others understood Chinese but they did not know the Hu sense. Even though they may have had a partial understanding, in the end, they were cut off from a complete comprehension .... How could there be obstacles in the satra? There were merely failures [caused by] the translators!13

Another difference between the two translation processes is that in .the medieval period Christian monks translated from one classical language, Latin, the structure of which had already been analyzed by the classical, mainly Stoic grammarians. Problems were rather caused in the target languages, as there were different Germanic ~ialects involved: 14 an originally Old-Saxon scribe and translator,
~Z1!f::X:~D~A~ ~'tfiJj~ , TfiFfT~~ ,~~~
0 0

.R~lii~? ~'Z*j:! Transllltionby Link 1961: 289. 14 This raises the question of Chinese dialects' impact on the early translations, although this is, of course, an almost impenetrable aspect of the translation process, due to the historical sources and the character of

x., TfiFf8~tfiJj~



!mE*i{j~~, ~~r%[J~


Max Deeg

for example, in Fulda could well have been responsible for the writing-down of a translated text and could have infiltrated the OHG with his native dialect's forms.1s Although the Indian grammatical tradition in the para'!lpani of Pa1!ini was not less sophisticated than the Stoic one, the translators of Buddhist texts into Chinese could not completely rely on such a systematic penetration of the language which they translated nor did they have an elaborated linguistic analysis of their own language at hand. They had to deal with the ambiguities and problems of mostly North-Western Middle-Indic Prakrts16 from which they probably mostly translated during the first two or three centuries. Chinese, with her long tradition of written language and literary forms and genres, did not have the same flexibility as the Germanic languages and their oral literature, or, to come up with an example from the Buddhist side, as the Tibetan language which could be remodeled according to the underlying Sanskrit and its grammatical (vyakara1}a) and semantic (nirvacana) hermeneutic tradition. Thus

the largely a-phonetic Chinese writing system. 1S A concrete similar case is the OHG heroic poem "Hildebrandslied" (Song of Hildebrand) where OHG and Anglo-Saxon create a mixed translation idiom. 16 Hypothetically the underlying language of some of these translations has been identified as GandharI. Brough 1975 and von Hiniiber have worked out some examples of "misrenderings" from this language, as did, although being rather careful in his identifications, Karashima 1993, 1994, etc. This hypothesis has gained and is still gaining a steady flow of new textual and linguistic material from the British Library documents worked on by Richard Salomon and his research group at the University of Washington (see Andrew Glass' article in the present volume) and from a group of manuscripts in Pakistan worked on by Harry Falk and his team at the Freie Universitat Berlin (for a first preliminary overview of the new collection see Strauch [2007]). The reservations which Boucher brings forth against this "GandharI-hypothesis" start to become relative if the normal span of dialectic variants and the possibility of blending with a "Hochsprache" - the famous problem of Buddhist Hybrid Sanskrit - is kept in mind which, by the way, is found in a very similar fashion in OHG translations.

Creating religious terminology


the place name Sravasti, so frequently found in Buddhist sutras, could be rendered as Tib. mfian yod, lit.: "hearing-is," because according to the nirvacana tradition of analyzing a word by different lexemes or roots, -{sru (srav-; "to hear") and /"as (-asti; "is"). Despite these differences in terms of cultural development, however, the tasks for the translators in both areas were not so different after all. The outcome was in both cases what Erich Zurcher has called a "scripturallanguage"17 in the sense that it at the same time contained elements of the elite vernacular of the respective period, and foreign elements, mainly in terms of translation terminology. In the following discussion, I will structure my examples on a categorization which was elaborated by the German philologist Werner Betz in the context of Old-High German (OHG) transla.dons of Latin texts. IS I think that this delivers an appropriate heuristic19 tool for structuring early Chinese Buddhist terminology and thus may help to avoid the confusion which can be caused by the use of autochthonous categories such as geyi f11i~, "matching of concepts," which, after all, is rather a theoretical concept used and discussed by scholar-monks who were not necessarily involved in the translation process its elf. 20 I have tried to translate Betz's terminology - which partly corresponds to the systematic English terminology used by the . American-Norwegian linguist Einar Haugen 21 - into English, although I myself have to raise some doubts as to wether I have been able to do this correctly and intelligibly in all cases. For each category I will present one or a few example from the OHG corpus and .will then give examples from the Chinese corpus. These examples

17 Zurcher 1991: 279ff. 1S Betz 1965; Betz 1974. 19 This term is meant to point out that there are, of course, no strict boundaries and that there is overlap between the categories. 20 On geyi see: T'ang 1951, Ito 1990, Lai 1978 . . 21 Haugen 1953. Haugen's and other linguists' systematization were already taken into account in a study by Chen 2004 of translations of Kumarajlva and Xuanzang.


Max Deeg

are mainly drawn from Seishi Karashima's extensive philological and analytical work on the Chinese Lotuss~ltra22 'in general and on Dharmarak~a's/Zhu Fahu's ~5*~i (active 265-313) early translation of the satra, the Zhengjahua-jing jE5*~*Jlf translated in 286, a text which belongs to the older stratum oftranslations. 23 Although having used Karashima's (and others') work heavily I have tried, nevertheless, to use material which has been left open to further analysis. 24 While Kumarajlva's later translation of the Lotus has traditionally been considered to be better in stylistic terms and in terms of translation technique, Dharmarak~a's version definitely has the advantage at giving some insight into the "workshop" of an early translator. In some cases he translates more correctly; for instance, . when Kumarajlvajust usesjo {~or rulai ~D* for the different epithets of the Buddha, Dharmarak~a usually tries to render them semantically correct - at least from his own analytical point of view - as, for instance, in the case of renzhongshang ACPL for puru~ot tama 25 where Kumarajlva has only a simplejo {~.26 After these preliminary remarks I will present Betz's classification system and give examples from his corpus and from the

22 Boucher 1998; Karashima 2001. On examples from another text attributed to Dharmarak~a, the Ra~trapalapariprcchasatra, see Boucher 2001, and, expectably in more detail, Boucher 2008. On an introductory discussion of some examples from Zhi Qian's Jt~ (active 220-252) corpus, some of which may have been taken over by Dharmarak~a and others, see Nattier 2003. 23 On Dharmarak~a as a translator there are two American PhD thesis by Man 2002, and Boucher 1996. On a detailed discussion of the oldest Chinese translations see now Nattier 2008. 24 As the data and their concrete location and bibliographical references are easily accessible through Karashima's three main publications (1993), (1998) and (2001) on the Lotus I have abstained from giving them here and refer the reader interested in these details to these publications. 25 Similarly wushangshi ~...t for anuttarapunt~a in Zhi Qian's translations: Nattier 2003: 227. 26 See Nattier 2003: 234.

Creating religious termino1.ogy


translation corpus of the Lotus s(itra. Betz discerns the following categories and subcategories: 1. Lehnwort / Loanword 1.1. Fremdword / Foreign word 1.2. Assimiliertes Fremdwort / Assimilated foreign word 2. Lehnpragung / Loan moulding 2.1. Lehnbildung / Loan formation 2.1.1. Lehnformung / Loan shaping Lehniibersetzung / Loan translation Lehniibertragung / Loan rendering 2.2. Lehnbedeutung / Loan meaning l.Loanword (Lehnwort)

A loanword is a word which keeps its phonetic form completely

partly; this is what is normally called transliteration and would be called yinyi/onyaku .:g.~ ("translation by sound") in traditional Chinese or Japanese terminologyY For Betz there is a difference between 1.1. foreign word 28 ("Fremdwort") as such and 1.2. assimilated loanword ("assimiliertes Lehnwort"). In Germanic translation languages we also find loanword (transliteration) and loan shaping (see below) side by side: This was, for iI,lstance, definitely the case with the word for cross (crucifix), kriizi, as an assimilated loanword, also wlziboum, "tree of condemnation" (attested once) and galgo, "gallow" (attested 20 times). Thus the .sometimes irritating concurrence of a loanword (transliteration) ~nd a semantic rendering of the same word in early Chinese trans-


Beside the general discussion of early stages of transliterations iil.Chinese in Pulleyblank 1983 there have been studies of transliterations found in texts or collections of texts: see Karashima 1994 on the Dfrghagama, South Coblin 1983 on the early translators in the context of Late Han glosses, and Harrison, South Coblin 1999 on early mantra ~ransliterations from the Drumakimnarariijapariprcchiisutra . ..28 This should be differentiated from the term "borrowed word," as it is not only borrowed but marked by the fact that it is (still) recognizable as Ii foreign word.


Max Deeg

lation texts may be regarded as a normal rather th~n an exceptional case. Otfrid ofWei13enburg shows a clear awareness for such a concurrent situation when he discusses the word "angel" which may be rendered in OHG as engil or as boto:
What we call 'engil' (angel), is called 'baton' (messenger) in the Franconian vernacular for (those) people, (because they) always want to report ... what they have been told (by God).29

There is one clear difference between the OHG and the Chinese translations in this category: while names (personal, geographical) were always transliterated in OHG, from a very early stage onwards the Chinese translations used both transliteration (category 1) and semantic renderings (category 2). Strictly speaking all OHG names would belong to category 1 (although scholars like Betz only discuss appellativa under this heading), and I therefore feel entitled to include Sino-Buddhist transliterations of names under this category.
1.1. Foreign word (Fremdwort)

A foreign word is closely preserved in its original phonetic form in the borrowing language. In OHG these words are rather rare, and this is certainly due to the simultaneous and predominant use of Latin and the quick adaptation of loanwords to the phonetic and morphological system of the target language. The main difference in this category is that in OHG foreign words are mainly found in the area of material culture. Examples are wfn, "wine," from Lat. vinum, or OHG munih, "monk," from Lat. monachus. 30 In the case of real cultural items or institutions being taken over with foreign words, they lost their foreign taint relatively quickly. If there was no material equivalence the words
29 Betz 1965: 71: "thaz wir engil nennen, thaz heizent boton in githiuti frenkisge liuti, thie io thaz irwellent, thaz sie thaz gizellent '" so was so in gibotan." 30 There is, however, also one example of loan translation, einago: cf. Kobler 2006: 138.

Creating religious terminology


tended to be considered as foreign elements for quite some time, as e.g. OHG palma, palmboum, "palm," for Lat. palma. In Chines'e at least the same rate or percentage of foreign words is found in basic religious or doctrinal terminology. In the. Chinese case the oldest words for Buddha, Bodhisattva and sangha, fo {~, pusa if~ and seng {~ - in their Early Middle Chinese pronunciation reconstruct able as *but, *s~1J, and *b:J-sat31 - , were definitely considered to be foreign words when they were first conceived and used. But during the historical development of the language phonetic change and semantic integration - these words certainly shifted more and more to the second subgroup and became assimiiated loanwords, provided that they were kept in use at all. This is the case with the early transliteration Mile 5WJ'-1J/*mjis-lsk for Maitreya, which phonetically makes sense in the light of a form Metrega. 32 This shows that, despite problems concerning the correct reconstruction of the historical phonology of Chinese in the first centuries CE, the aspect of reconstruction is an important tool
31 I here use the reconstructions (indicated by the asterisk * throughout the text) of Pulleyblank for practical rather than "ideological" reasons. (Owing to font limitations I have not been able to represent all IPA characters correctly.) Without being able to go into detail it should be pointed but that there are problems involved in the reconstruction of Early Middle Chinese, that is the linguistic standard of an assumed vernacular of the elite from the end of the Latter Han period (late second century) to the Tang (seventh century and later). Reconstruction of various periods of pronunciation of Chinese in the West started with Karlgren's work and ,has been carried on by scholars such as Pulleyblank and South Coblin, and on the Chinese side by Li Fanggui and others. The hermeneutical cir.cl~ with the Buddhist transcriptional corpus evidently lies in the fact that forms reconstructed from purely Chinese material such as rhyme-tables :and onomastic glosses which are mainly based on a phonological system have to be applied to a corpus of words/forms the correct linguistic judgement of which would require more sophisticated phonetic reconstruction. Taking into account all these points it seems to be fair, however, to use the mentioned reconstructional corpora to think about phonetic implications arising from the underlying Indic terminology of the transcriptions. 32 Another example for le representing r-V-g-V (V = vowel) from Dharmarak~a's corpus is mohoule ~1*1JJ/*ma-y~w-18k for mahoraga.


Max Deeg

at least for the understanding of loanwords and has to be taken into ' account, be it with appropriate caution.
1.2. Assimilated loanword (assimiliertes Lehnwort)

An assimilated loanword is phonetically and morphologically - as the name implies - assimilated to the target language. An example from OHG would be trahton : Lat. tractare, in the sense of "to strive for." Here in OHG the Latin a-stem verb had to be transferred into a Germanic verb-category, into an -on-stem verb. The verb also shows assimilation in terms of phonetic change: it had become integrated in the OHG vocabulary at a relatively early period and already undergone the second - or High-German - sound shift in which the Latin guttural c : /k/ became a fricative h : /X/. Due to the general characteristics of the language it is not possible - and also not necessary - to adopt the word category and the endings of Indic languages in Chinese. Thus the feature which is important in inflective languages in order to identify such an assimilated loanword - morphemes - is of no help in the case of Chinese. One could argue that the way of rendering a word into Chinese characters and the acceptance of such a transliteration may be considered as a marker of assimilation. The assumed early creation of graphs for the loanwords *but {m : Buddha,33 and *S;}I) {~ : saligha,34 and the continuity of their use seem to be such indicators for assimilated loanwords. One could argue that hybrid renderings, in which one part of a name/term is given phonetically while another part is translated semantically, belong to the category of assimilated loanwords from the moment of their creation. In Dharmarak~a's corpus we find examples like Binnouwentuoni(-zi) 5)~m:xW8Jb(T)35/*pin-n;}wh-

33 Semantic (ren) ).., "man" + phonetic fu/*put J:g as a negative particle

("non-," "un-").
34 Semantic (ren) ).., "man" + phonetic ceng (zeng)/*dZ;}1)(*tS;}1)) Wf/, in the sense of "assembly of people." 35 Or: Fennouwentuoni-zi 71JJiYJB'JbT.

Creating religious terminology


::r for -putra.

mun-da-nri-(tsi') for Piin;a MaitrayalJ-'fputra 36 with a semantic -zi

It is clear that a loanword is only considered as a foreign element inside the target language for a certain time. After cultural and linguistic integration - not least caused through sound changes the respective term on a common level will not be felt as a foreign ~lement anymore, although a reflective discourse well may uphold the notion of exoticism or strangeness. Thus the above mentioned examples for Buddha, Bodhisattva and sangha would be considered to be part of the normal vocabulary of the Chinese language, although their foreign origin may cause some dispute. These adopted terms may then be "challenged" or eventually be replaced by new foreign words like fuotuo/*but-da {~W8 or sengjia/*s;}!}-kia {1l!' }till. I would even argue that the word futu/*buw-d:J >-'1lI (and the synonym >-'m) , attested very early for Buddha and considered to be of Central-Asian origin by Prof. Ji Xianlin ~~#,37 is just a concurrent term for the older fo, which was introduced into the language when the Early Middle Chinese pronunciation no longer corresponded exactly to the original Indian pronunciation. 38


There are also complete transliterations such as Binnouwentuofu


fJ~tst~'8~/*pin-n;)wh-mun-da-put, Bintiwentuofu BM1i!:x:Wti~/*pin-dcj

mun-da-put; there is also an attempt to render the name semantically Manyuan-zi )j1jlj)jjr Pilrr:zamaitriiym;fputra, Pa. Pur:zr:za-mantiiniputta (see Karashima 1992: 26), cpo Manzhu-zi :)jijljtJGr in Zhi Qian's Jt~

Da-mingdu-jing *~m:~ (483a.ll) (see Karashima 1992: 277), clearly showing that the name element corresponding to Skt. MaitriiyaIJi- was considered to be mantra-, "word, wish, vow." Ii 1992. For phonetic reasons on the Indian side I would rather propose that ;Chinese had a weak vowel after plosive endings - *but;), *S;)'l.;), *b:J-sat;)in the earliest period of Buddhist transcriptions (first/second cent.) which would well correspond to the proposed weak final vowel < Skt. -a(J:t), etc. :iJy Fussmann 1989 for spoken G1indh1irI. In the case of the obviously very :l.~ar1y standardized transliteration for bodhisattva it seems strange that the 'creator(s) did not choose a character with a dental final for bodhi- instead .,ofpu-/*b:J-. In the light of the G1indh1irI forms of the word, however, as fOr instance bo.isatvo (see e.g. Lenz 2003: 264a.), one could well argue


Max Deeg

Between both groups have to be counted what I would call redundant hybrid loanwords; these are loanwords, transliterations, which are either explained by a semantic synonym or by a generic term. An example would be chatu mU for lokadhiitu in which cha mUl*thait is a transliteration for a Prakrt correspondent ofa synonymous Skt. k:jetra to dhiitu. Another example is Qishejue(-shan)/*gjidzia-gut-O ~liJilIJlliLlJ for Grdhrakuta. From Dharmarak~a's corpus one could refer to chan-ding t~JE: and chan-si t~~, for dhyana where the simplex chan/*dzian t~has certainly already. become an assimilated loanword in connection with the semantic explanatory elements ding JE:, "concentration," or si ~" "concentrated mind." A similar example is sanmei(zhi)-ding/*samh-m;}jh - f17!;:CL)JE: for a simple samadhi. 2. Loan moulding (Lehnpragung) Loan mouldings 39 - what Haugen calls a "loanshift" - with its various subcategories are semantic renderings which in the traditional Chinese/Japanese terminology are called yiyi/giyaku ~ ~ ("translations by meaning"). In the Buddhist context the standardized term for dharma - to complete the triratna - was from a very early period the Chinese semantic correspondent fa 5-*, and transcription was restricted to personal names (tan or tanmo).40 For this group Dharmarak~a's translations are a goldmine because he usually avoids transliterations/loanwords in favor of semantic renderings.

that the Chinese transliteration is based on a sloppy oral (haplologically abridged) form of the GandharI. 39 Or: "loan coinage"? 40 This standardization did not occur in other cases, as Vetter and Zacchetti 2004 have shown for the Dharmarak~a-corpus (example Jingfa ~~)t: for dharma).

Creating religious terminology


This group is divided by Betz into a number of subcategories:

2.1. Loan formation (Lehnbildung)

In this category - called "creation" by Haugen - the translator tries . to copy the structure of the original word by means of semantic elements in the target language. This requires a formal and semantic analysis of the original term and an application of this analytic knowledge to the target language as well. In Latin such analytical know-how was procured, as already mentioned, through the work of the grammarians. In the Indian .context two "schools" of interpretation were at hand when it came t() analyzing a word: the PaI).ini-school of grammar, vyakaral}a, and the nirvacana school of Indian semantic analysis. 41 Now unfortunately no direct discursive reference is preserved, as far as I know, to the use of the nirvacana tradition as a means of semantic analysis for early Chinese translators. I would argue, however, that IIlany examples from the early translation corpus clearly show that this nirvacana was indeed a preferred method of analyzing Indic 'terms and names and translatIng them into Chinese - and Tibetan .as well- instead of using the analysis of the grammarians into prefix~s, roots and derivational elements. A typical nirvacana-analysis breaks a word down into two (or more) verbal elements (roots), as iiithe example of the Tibetan translation of Sravasti. The explanations found in the Tibetan bilingual dictionary sGra-sbyor-bampo-gfiis-pa/Mahavyutpatti (beginning of the 9th cent.) very often indicate a deliberate decision for a nirvacana-analysis although ,they reflect the "correct" vyakara~a-analysis of the same word. I only want to give as one example the discussion of the term arhat:
. The word arhat is on the one hand (explained) as 'puj[a]m ar[a]hatfti arhan,' being called 'worth of veneration' because he is worthy of be.. ing venerated by gods, men and all the others; on the other hand it is said: 'klearihatavan arh[a]n,' meaning 'having defeated the enemy


See note 12.


Max Deeg of suffering.' From these (two ways of explanation) the (one based) on meaning is taken and (the translation) is fixed as 'dgra-beom-pa'."42

Having presented the "traditional" explanation for arhat I cannot resist to think of the Chinese wu(suo)zhuo ~(PJT)~43 which, in my view, reflects another analysis of the word: it is possible that it was parsed as being derived from ..fiabh- - with the interchangeability of r and 1 - for which Prakrt forms laMi and lahar and even Niya-Prakrt (Southern Silkroad) lahafJlti are documented. 44 That such a derivation is possible is shown by Dharmarak~a's translation wuzhuo-guang-sanmei ~~7';~$K, "unattached-light-samadhi"obviously taken as a-nilambh + ..fbha- - for a Skt. anilambhasamadhi (KumarajIva: wuyuan-sanmei ~~-=$K). 2.1.1. Loan shaping (Lehnformung) Loan shapings clearly use semantic equivalents in rendering the original word in the target language. Depending on the level of formal equivalence two subgroups are discernable: Loan translation (Lehniibersetzung)
In loan translations (Haugen: "literal creation") the semantic context and form of the translated word are as close to the original term as possible. OHG gawizzani - prefix gi- and a nominal derivation from wizzan - for Lat. conscientia is a good example of the large number of words belonging to this group.

42 arhan ses bya ba geig tu na I pit ja ma ra ha tf ti a rhan ses bya ste / lha dan mi la sogs pa kun gyis mehod par 'os pas na mehod 'os ses kyan bya / yan geig tit na / kle sa a ri ha da ban arhan ses bya ste / fion mons pa'i dgra beom pa ses kyan bya ste I rnam pa 'di las 'dir ni don bean par bya ste dgra beom pa ses btags. Text quoted after Simons son 1957: 269. 43 See also the discussion in Nattier 2003: 215ff., who thinks that the term was analysed as *a-raga, "devoid of passionate attachments" (translation Nattier). 44 See Turner 1966: 635, s.v. ldbhate.

Creating religious terminology


Most of the lndic compounds translated into Chinese belong to this category and did not cause too many problems. Some examples from Dharmarak~a's Lotus, chosen randomly, are: the dvandva thengying :t~~ for nagaranigama, "city and market place;" edao @m (lit.: "evil path") for durgati; liangzu iA0JE. - as a real bahuvrfhi in Chinese - dvipada, "biped." Dharmarak~a's daci-dabei *1rJi* f.5 mahamaitrf-mahakarU1JJi is giving a word-by-word rendering where KumarajIva has a simple cibei 1rJi1.5. In contrast, in some cases Dharmarak~a, by using this category of rendering, produces a simpler rendering than KumarajIva: lokahita is translated by shimin'ai tltPl.t:R "one who is sympathetic with the world,"45 where KumarajIva has a complex zhu-fo zhi jiushi-zhe mf{~:L.~tlt~, "the Buddhas, savers of the world." Dharmarak~a consequently uses semantic renderings even in cases of generic names (e.g. of plants) where one could argue that a transliteration would have been more appropriate and was actually tised by later translators: shengxiang ~W (lit.: "born-fragrance") forjati(ka)gandha where KumarajIva uses a hybrid sheti-hua-xiang Iilm~W/*dzia-dcj-O; or jietuo-hua fg~g)t* ("liberation-flower") for !'rnukta-kusuma versus KumarajIva's zhong-minghua ~~*, standingfor nanaratnakusuma. There are other examples where it is not obvious in the first place why the translator formed his terms: fangdeng 1JJ!f. certainly stands for vaitulya rather than for (maha)vaipulya, in which fang 1J represents the prefix vi- (in the sense of "apart, spread") and deng J!f. I~nders tulya, "equal, similar;" KumarajIva instead uses the more generalized expression dasheng-jing **~J[, mahayanasutra. But also translations which seem a little bit far-fetched match 'ihis category, as e.g. Dharmarak~a's translation ren ;g, "to bear, to endure," for Saha which takes the lndic word to be derived from ..fsah-, "to suffer, to endure, to tolerate." Xiangyin(-shen) W1-t$ ~or gandharva presupposes an analysis into gandha-, "perfume" + 'rava (..fru-, "to give a sound"); similar is da-yiyin-hua *it1-* for
'.45 -hita here obviously taken in the sense of "sympathetic, friendly"; ",see also the variant reading hitanukaI'J'Lpa, "friendly and compassionate" given in Kern, Nanjio (392, note 1).


Max Deeg

mahiimandiirava where, beside -mandii- being taken to belong to -{man, -rava is taken again to be a derivation from ..fru-. In the Chinese context this and the next group are probably the largest categories due to the typological and linguistic differences with the Indic languages. This is also the group in which a lot of so-called mistranslations 46 are found, which, however, very often as the work of Seishi Karashima and others shows - reveal a clear semantic analysis on the basis of the underlying Pra~ft.47 Due to the ambiguity of the language this even multiplies the possibilities for nirvacana-analysis compared to Sanskrit. One of the wellestablished terms is one of the epithets of the Buddha, shizun tit ~, the "world-honored one," for Skt. bhagavat, which obviously had been broken into the elements bhaga + .fvand-, "to honor, to venerate," and was, at least in the early period of translations, in "competition" with another, structurally similar term, tianzun 7( ~, the "one honored by the gods."4B There are some cases where we may be able to find an explanation for certain renderings as products of a loan translational process. For instance, in Dharmarak~a's Lotus translation we find the strange expression dumen &F~, lit.: "gate of salvation," for paramasukha, "highest bliss," where Kumarajlva has the slightly overstretched jingmiao-diyi zhi Ie 5-'W)%-L~, "pure, delicate and highest bliss." Keeping in mind now that Dharmarak~a renders dharmamukha by famen 5t:F~, which establishes, inter aliud, men F~, lit.: "gate, door," as a translation for mukha, I feel tempted to think that the translator had an original *piira(1'J7.)mukha which he had no other choice than to translate by dumen &F~.

46 On a discussion ofthe applicability ofthe term mistranslation in this context see Deeg 1995b. 47 What level such forms based on a kind ofPrakrtic "proto-philology" can attain can be seen in Karashima's 1999 discussion of the relevant early translations of the name of Avalokitesvara, finally standardized as Chin. Guan(shi)yin W(1!:!:).g-, or Karashima's and Nattier's 2005 observations on the name Sariputra, Chin. Qiuluzi fX.Ar. 4B For a full discussion of these terms see Deeg 2004, and Nattier 2003: 232ff., the latter also discusses other translations for bhagavat.

Creating religious terminology


If we accept the Prakrtic hypothesis for early Buddhist translations we may add a lot of examples to the category under discussion. For this I would like to bring forth some examples from the Lotus sutra translation of Dharmarak~a: Baoyin .if for (the kalpa) Ratniivabhiisa where yin if is obviously taken for a derivation from the root ..fbhii~- (Skt. avabhii~a?) instead of from ..fbhiis-. Sanda -~, lit.: "three achievements" contrasted with KumarajIva's sanming ..::::Ej)j - for traividya, "threefold knowledge," makes sense if we assume a derivation from :fvid-/vindati, "to acquire, to get, to possess." Translations like
ruyi-bao(zhu) 3ZD~.~, ruyi-zhu 3(O~~, ruyi-zhi-zhu 3(O~Z~

for malJiratna may have been caused by an analysis of the first element malJi as belonging to miinin (-(man-) , "thinking to be ... , thinking to have.''49 The rendering fan-renJi-(tianzi) Jit?2,Ifr1Ji(~-r) forbrahmii sahiifJ1pati seems to have an underlying analysis into 1sah- + padi(n) (or: pathi, see above the example of Mahaprajapati), while haozun ~~ for k~itipati/mahipati takes the same, second lnembrum -pati, Pkt. *-vafJ1di as belonging to ..fvand-. 50 Even examples, which, at first view, look very odd such as Dhar~arak~a's muren fB:A, "mother," as a translation for vadhukii, /t",idow" - which KumarajIva "correctly" renders as guanii 'fstart to make sense if we allow a Prakrtic interpretation *vatukii (interpreted as Skt. miitrkii).51 In the same way aihu :R~, "compassion and protection," for kantlJiiyamiina - KumarajIva has datibei *~1,J; - may be taken as derived from kantlJii + fyam- in 'illesense of "to restrict, to protect," or Anyang(-guo) ~.(~), lit.: j~peace-fostering," for Sukhiivatf-lokadhiitu (KumarajIva: Anleih:ijie ~~t!tW), as derived from a *Sukhiivad(h)f, taking -vadhf as ~'Prakrt-form belonging to ..fvrdh-, "to increase, to (let) grow" and tendering it as yang . , "to bring up, to foster, to nourish." Bold as


See GandharI ma':ta for Skt. manas, e.g. Salomon 2000: 232a. /.,. For the interchange of p and v in North-Western Prakrt see e.g. ,~~lomon 2001: 85. I suspect that hao ~ here stands for mahi- rather than :f9rk~iti-, in a slightly redundant meaning of "strong, powerful, leader." ~.rFor the interchange of v and m in North-Western Prakrt see Allan ~,OOI: 86.


Max Deeg

it may appear, I am willing to admit a sound semantic analysis for Shangshi-jiaye ...tH~Jm!~ as a translation for Uruvilva-kiisyapa, in which a Prakrtic *uvvelii could be interpreted and be translated as *ud-velii, "upper time." I would thus argue in general that some cases in which we tend to judge the translations as wrong, are, seen from the standpoint of the translator, products of a clear analytic process: Dharmarak~a's monengsheng ~~MJ, "who cannot be defeated by anybody" for the epithet of Maitreya, ajita, "invincible," is obviously the attempt to render the past participle passive (PPP) of the lndic name in Chinese - a problem which Kumarajlva avoids by using a transliteration. A similar case, where a past participle passive is correctly given by Dharmarak~a, is the translation of the name of the bodhisattva Sadiiparibhata as Changbeiqingman m:t<Bl,~iff, the "one who is constantly disregarded," where Kumarajlva has the opposite and rather nonsensical Changbuqing m/f,~, the "often notdespised," which only makes sense in the chapter to which it gives its title if it is taken in an active sense "always not-despising." Single expressions, due to the tendency of Chinese for twocharacter-words (binoms), often were translated as synonymous dvandva-like binoms: daoxing for caryii. Another example is zongchi $,I~fif for dhiira[tf, in which zong obviously is used - like in another of Dharmarak~a's renderings, zongshui $i~7.K, literally "carrying water," for jaladhara - to semantically reflect chi fif, "to hold," ..fdhr-. To this "redundant" category one may also count examples such as liiluo .ig for gardabha, "ass, donkey," despite the semantically slightly deviating luo ,g, "mule;" or guanzhou ~m ,Ifi (lit.: "connection-shaft") for argala, "bolt." There are also some adverbial-syntactical renderings trying to cope with the Indic word-structure like bubu t!7t!7 - against Kumarajlva's simple jian )$I - for karma-krame[ta. Again, here and elsewhere one can clearly recognize Dharmarak~a's tendency for a word-by-word rendering. See also, as a similar example, changye -&~ for dfrghariitram. As has already been indicated, there is a tendency in Dharmarak~a's translation to give semantic renderings instead of transliterations. He uses buhuan /fji, "non-returner," for aniigiimin, .


Creating religious terminology


and butuizhuan =Fill", "not retrogressing," for avaivartika; Kumarajlva uses the loanword a'nahan roJj3~-a- for the first but in most cases keeps the (semantic) translation for the latter expression. The same stands true for renderings of personal names: Dharmarak~a translates Chimingwen t-if::SIitj, "bearer of fame," for Yasodharii, where Kumarajlva uses the loanword/transliteration Yeshutuoluo lf~~~~~/*jia-hu~-da-ra. The same is seen by Dharmarak~a's Prakrtic nirvacana-rendering of Mahiiprajiipati as Da-jingkui *{;j)(~, "great respected thoroughfare," which - like the alternative translation Da-aidao *~m supports Brough's interpretation52 as going back to an underlying *Mahiipiyapadi (Skt. *Mahiipriyapath1:). Dharmarak~a renders Mafijugho~a as Purouruan-yin "broad (and) soft sound,"53 while Kumarajlva, once more, has the transliteration Wenshushili )zJ;KgijJ 'fU/*mun-dzuj-i-lih. 54 Dharmarak~a renders Siikyamuni as NengTen ijEi= in which neng ijE for Siikya is derived from -fsak- ("to be able to"), and ren stands for muni. 55 The arhat Gaviif(lpati is translated by Dharmarak~a as niushi 4- 00], "cow-ruminating," .which may reflect a North-Western Prakrt *Gaviif(lvadi, "speaker


of COWS."56


Even the word order of the original is sometimes kept by Dharwhere Kumarajlva syntagmatises his rendering and thus changes the order of the single words: 57 e.g. for Mahiisthiimapriipta .52 Brough 1975. 53 I suspect that

Dharmarak~a's rouruan-yin-hua *'X.:g for was influenced by his rendering of Mafijugho~a; Kumiirajlva has manshusha-hua ijkJ*5!yijfU*muanh-dzu;)-~ai . . ,54 For Mafijusri Dharmarak~a has Pushou 5wJ, "broad head," in which (.Mis taken as belonging to sir(as), "head": Karashima 1992: 27. ;'<55 The Tang-period Yiqiejing-yinyi -t)J~.:g~ takes nengren for :.$iikya only (T.2128.46Sb.S), while the Song-period Fanyi-mingyioji lIm~::g;~~ correctly recognizes it to be a rendering for Siikyamuni iJT.2131.10SSa.18f.) . .56 For the interchange of p and v see note SO. 57 Prom Dharmarak~a's corpus the example of huadu(-shu) 1fIJr(tlf) piiracitraka seems to be an example of such an inverted compound, fuafija~aka



Max Deeg

Dharmarak~a's Dashi-zhi *~~ (lit.: "Great-po,wer-attained") versus KumarajIva's De-dashi 1~*~ (lit.: "Achieved-great-power"). The same analytic pattern can be seen at work injieluan M~L (lit.:

"kalpa-confusion") for kalpasalflk~obha, where KumarajIva uses the inverted zhuojie 5JVM (lit.: "chaotic-kalpa"). .
Dharmarak~a's monosyllabic bie 53U - where KumarajIva has the "regular" shouji =R*G - is certainly based on a grammatical analysis of vyiikaralJa as derived from vi-ii-..(kr- "to take apart, to separate, to analyse."58 In contrast in the case of dingyi 5E~ for samiidhi he seems to have added a redundant yi ~ which may be an attempt to render the root ..fdhyii-. 59 Even in cases in which the analysis of a certain word by the translator is not completely clear we might suppose an underlying analysis, as for instance in the hapax legomenon wuke(-yu) ~OJ( 1t\), lit. "(the hell) 'Impossible'" (?), for Av[ci in which a- obviously stands for an a-privativum; the semantic function of ke OJ I am, however, not able to explain. Loan rendering (Lehniibertragung)

In a loan rendering (Haugen: approximate creation) the original word is still clearly recognizable in terms of content and - normally only partly - of form; it does not completely mirror the structure of the underlying original term. An example in OHG is horsam(i) for oboediens, "obedience," a word with a successful history in modern German: Gehorsam. Here the Lat. prefix ob- was not translated and the abstract nominal suffix -sam (i) was chosen instead of the present participle of the verb horen, "to hear."50 Another example .
hua ~ representing citra(ka) and with du fr as a translation of pi'ira, "the other side" (like du as a rendering for pi'iramita). 58 Dharmarak~a is, however, inconsistent in his use, as he also has jue ;;R;, "decide, determine" for vyi'ikara,!a. 59 Inyixin-pingdeng -'L,,3:jZ$ (KumarajIva: renshan if3/g) for sami'ihita or sami'idhi in which sami'i- obviously has been rendered by pingdeng :if $. 50 Betz 1974: 158; an exact loan translation is found in a text from the

Creating religious terminology


is deomuati for humilitas which is composed of dio, "servant," and muat(i), "cast of mind," while the Latin abstract word is derived from humilis, "ordinary, low, humble." OHG tagasprahha, lit. "day's speech, talk," as a translation for Lat. hom ilia is another example for this group. Loan renderings thus involve the greatest degree of creative freedom on the part of the translator or translators as he or they can, to a certain extent, ignore the form and the semantic basis of the original word as long as his or their new creation renders what is meant in the target language. Another example einsidil, literally "who settles alone," for Gr. anachorita or Lat. eremita. 61 What may be called hybrid renderings, because they use one tninsliterational and one semantic element which often clarifies the semantic field, belong to this group as well. Examples from OHG are, for example, salmsang for psalterium, or fimfchust(i) for pente/coste. 62 I categorize in this group Chinese terms which still reflect a .lexical element of the underlying Indic word, even if this is not 'always obvious and only reconstructable by following the semantic analysis which may underlie the rendering. The translation shangzhen 1:.;Et, literally meaning "excellent and rare," for udara, "noble, illustrious," is rather an analytic rendering in which shang 1:. obviously stands for ud-. Other examples are: dade *t~, "one of great virtue" (Kumarajlva:fo {~) for mahamuni, "great sage," or mahayasas, "one of great fame." In jingyu ~:f:~, ~'border region," for vidis(a), "intermediate quarter, region," vi- has Obviously been taken as indicating distance and peripherality. The translation Lingjiu-shan 1I;;;LlJ, lit.: "spiritual vulture mountain," for Grdhrahtta may be counted in this category be.cause it adds an element ling 11 and rather loosely translates kuta, "peak," by the more generic shan LlJ, "mountain."


.. 61

gaganhorenti . A direct loan translation is waldlihher; see Betz 1965: 39.

See Betz 1965: 59, who also quotes the complete Anglo-Saxon translationfiftigdcrg, literally "fifty-days."



Max Deeg

One may also count into this group the rare examples where the translator(s) tried to harmonize between different versions of the text. One example I think I have found in Dharmarak~a's Lotus is ranzhuo ~~, "tainted and attached" - Kumarajlva has the simplex zhuo ~ -, where the Kern edition has sajjati "to adhere, to stick to," but the Kashgar manuscript has rajyati, "to colour, to redden, to affect." It seems that the translator in the binom has tried a harmonization of two versions of the text. Another example in this group is ligou .!), literally meaning "abandoned dust," for Skt. vimala, which is closer to the original than Kumarajlva's simpler jing ), "pure." See also Dharmarak~a's kongji ~;j[, lit.: "void-quiet," for the abstract siinyatii, but on the other hand kongwu ~if*, lit.: "void-nothing," for the adjective siinya which rather belongs to the group of loan translations. One could argue that the translation jingxing ~~1j, which literally may be rendered as "passing and walking," for (anu)cankrama belongs in this group as it seems to attempt, as a semantically redundant binom, to follow the reduplicating structure of the. Indic original. Even loan renderings which at first glance do not look very well chosen sometimes turn out to have a sound interpretational basis: when Dharmarak~a has dian-gui M5Il, literally "jolted, fallen upside down ghost," for apasmiiraka, which is usually taken to mean "ghost of oblivion" (: apa-{Smr-), this does not look very close; but keeping in mind that apasmiira as a medical term also means "epilepsy, falling sickness" the picture changes and Dharmarak~a's translation seems to be an ingenious loan rendering. 63 Dharmarak~a's decision to translate kalyii1'}amitra by shanshi ~gjjj, "good master, teacher" - where Kumarajlva has the regular shanzhishi ~~O~ which should be categorized as a loan translation - is probably based on the fact that a kalyii1'}amitra -like Upagupta for Asoka - usually has the role of a teacher. There are loan creations which became "standard": the word for hell, diyu :f:fu3~, lit. "earthly prison," for naraka or niraya clearly
63 The Song-dictionary J:anyi-mingyi-ji m~.:g~ takes this as a translation for pisaca(T.2131.1086a.26ff.).

Creating religious terminology



;Concentrates on a different semantic aspect than the lndic original terms and was used throughout the history of Buddhist translations ~s well as th'e rendering egui MX)l, lit.: "hungry ghost," for preta . .While long i~, "dragon," for niiga belongs rather to the following group, Dharmarak:~a's longxiang ~~, for hastiniiga, "elephant," based on an already established loan meaning long = niiga, should be grouped under this category. It is not always certain if a term belongs to this group as we can~i16t be sure if it is really a translator's creation or if it was already ~~'part of Chinese vocabulary which happens not to be documented liffi:our extant sources. Dharmaraksa's decision to use kaishi i7fJ bodhisattva - which he has taken over from earlier translators ~hch as An Shigao ~i!t~ and Zhi Qian 3Z:~ - may represent such ... fli\:ase: so long as we do not know wether such a word was already ~rtuse before the earliest Buddhist Chinese documented exam~i~we cannot really say that it is a loan creation of the Buddhist ~trimslators. Later Buddhist explanations consider the word to be of !Biiddhist origin when they explain that it should mean "the gentlewho has an understanding of enlightenment."64 When Dharmarak:~a translates pisiica in a loan-translational way :~rfanzu-Cluocha) .&JECmirtU), "inverted feet-Cdemon),"65 he obvi~risly had a certain traditional description of the pisiica in mind. ~ether this was really how the Tang-period dictionary Yiqiejing~inyi -tv~-=f~ by Huilin 1Iij;f\ describes it, is a different mat""''''66 t


:~ ~;;;~


;;;~ Yiqiejing-yinyi



;1':):1; see also 407a.l3. Il~~ Also found in his translation of the Da-baoji-jing *.f.l:~!lf (T.31O),

ftrujialuoyue-wen-pusaxing-jing fJ~~m~Fo~~i\iff~ (T.323), Mie ang-ming-jing ~+h*~ (T.435), Dajangdeng-dingwang-jing *h ~ (T.477, where we find janzush~u EUE-, "feet and hands inted"), Xiuxing-daodi-jing ~~ff~ftt~ (T.606).

. . . T.2128.376c15ff. EtJE.)l()l;gt!? ft5Jfu~~: *fU~:(-~JIU~A :.@Et, J!JJE.1* J:5'~: ---HtIlEt, ~ElHt!? *h~)jt$~~!lf~: rNJl ~it~ . ;t{:M7CJ([jmg Arm':ffl5.f ~- ~~JE.' MJf ffli. ' Etli!l ' ~j] , f& . ;gS: ~!fo/.l 01ft~~)l~t!? 0) ("Inverted-feet-ghost. A name for a (spe0
0 0


Max Deeg

For another kind of demons called piltana, Dharmarak~a also produces loan creations of different levels of complexity according to the characteristic features of this kind of ghosts: goubian-hungui ~m5~5\1, "dirty ghost on the side of the gulley," hunce-gui ~JJW 5\1, "lavatory ghost," or hunshen ~~$ "dirty ghost," semantically inverting the expected analysis of piltana as derived from -!Pil-, "to cleanse, to purify." A similar case ofloan creation is the continuously used qunmeng f:fa~ for sattviini or priilJin where the Tang dictionary explains that meng a~ is a synonym for meng @, "common people."67 As already mentioned, lndic prefixes could not be and did not always have to be translated into Chinese. 68 They sometimes but
cies of) ghosts: in the 'Geographical Records' (of the historiographies) it is said (in the section) about the kingdom of Rouli that to the east of a certain kingdom there are men with hands and feet inverted, with bent knees and inflexible feet. The above (mentioned) commentary stated that hand and feet are inverted and (people) are crimp. In the Dongfang-shuoshenyijing it is said that in the deserts of the west there are beasts who look like deers, with faces of men, with tusks, hands like monkeys, feet like bears, vertical eyes and horizontal noses, inverted heels, with huge physical strength and malevolent; (they) are called 'monsters' and belong to the species of ghosts.") 67 T.2128.431b.4;mil"! (6:X:'tI%'IOJ'~'; ~m'&'aA5.f'ti:!. Ii'JJim.!l aA~ ti:!. 0~'aA"~~5C' ti:!. 0~'~Eff,M,9;O'ti:!. 0) ("Qunmeng: the old texts use it in the same way as 'men;' spelt as 'mai' + 'geng;' means 'sprout;' in .. ' 'the Guangya it is (called) the beginning of the germination process. Accordingly 'meng' (means) a dump fellow; also for common ignorant people."); see also 443b.l7f. ;mil"! ('aA";~Ujj:'& 1i')ii!f~5.!l B: 'aA'~: 1j[**)]~ti:!. Ii'=~{.t.!l B: $~ti:!. ~'Ji~fLx" ~~fEl~/J\1j[ti:!. 5Z.~=F 1[{,!='tI%'o Ii'=~{.$.!l B: 'tI%'~ti:!.'tI%'~'aA'IOJti:!.) ("Qunmeng: 'meng' spelt as 'mo' + 'geng.' The collected commentary of the Hanshu says, that 'meng' is when grass and trees first sprout. In the Maoshi-zhuan it is explained as the crowd; ignorant common people, also for young grass. The character can also be written as 'meng.' The Maoshi-zhuan says, that 'meng' means 'min' (people), and 'meng' (people) is the same as 'meng' (sprout).") 68 In this respect Chinese differs from Tibetan where the translation of the Indic prefixes was standardized at an early point.
0 0 0

Creating religious terminology


not consistently were rendered by speciallexems which were obvi.ously meant to express the semantic value of the lndic prefixes such as in puman if5iim for paripurlJa, where pu- if ("common, universal") seems to indicate completeness (pari-), while Kumarajlva's chongman 3iS5iim (lit.: "complete-full") translates the same word by a redundant binom. 69 Dharmarak~a, every once in a while, seems to render a certain prefix by one Chinese lexeme, as for instance shen ft ("very, extremeley") for abhi- in shenle ft~ : abhirati, and in shenman :frt~ : abhimiina. Another example would be jie ~5 ("tie, knit") for nir- injiehen ~5t& for ni~kii1ik~a or nirvicikitsa, jieqin ~5~J? - not quite correct as an analysis of a Sanskrit word - for ni~evamiilJa, and jiewang ~5~~ for niftsaYJlsayaf!l. Or, more inconsistent because two prefixes are rendered with one Chinese lexeme (chu tB, "come out, raise") but still semantically correct, .chuxian tBfJ[ and chuxing tBW for utpadyate and chuzai tBtE for ni~kiisayitvii / (manuscriptal) niskrriimayitvii. That the omission of a literary rendering of prefixes was not Seen necessarily as a defect of the translation may be deduced from dIe preface of the Mahiivyutpatti. Although Tibetan translation terminology almost regularly translates lndic prefixes it is stated there that prefixes only have to be translated when the meaning of the basic word - that is the dhiitu, the root - is changed by iUD
69 Another example is chongmanyue ft5Wi'fR:, lit.: "fully happy" (Kumarajlva: xin'anjuzu J~\~~JE., lit.: "piece-of-mind-complete") for salflto~ita where sam- is probably expressed by chong ft, or the binomial chongman 3t:5Wi, "full, complete." 70 See Simonsson 1957: 255: pa ri dan / u pa Zta bu Za sogs te / tshig gi phrad dan rgyan Zta bur 'byun ba rnams bsgyur na don dan mthun tin 'byor ba'i thabs ni / yons su ie 'am / yan dag pa ie 'am / fie ba ies sgra biin du sgyur cig / don Zhag par sfiegs pa med pa rnams ni tshig gi [had gyis bsnan mi dgos kyis don biin du thogs fig. ("In case of the translation of pari, (samyak), upa, and single prefixes (which are used) like ornamentation (of the root) one should, in order (to achieve) accordance and agreement with the meaning, translate 'yons su' (complete = pari), 'yan Jag pa' (real, entirely = samyak), or 'fie ba' (near = upa), according to the :verbal shape (alone). In case that there is no achievement of additional meaning there is no need of adding one element but one should render


Max Deeg

2.2. Loan meaning71 (Lehnbedeutung). Betz distinguishes a category ofloan-meaning (Haugen: extension). A loan-meaning occurs when an original word in the target language adopts a new meaning and a different connotational semantic range by being used as a representative of one (or more) termini of the source language. Thus an originally Germanic word like geist - which had such a strong impact on the German "Geistesgeschichte" from the 17th century onwards - started its "career" as a loan meaning, as it was already existent in Germanic languages before its use in a Christian context, meaning something like "mental movement, inner feeling." In Christian texts it was, however, used for Greek pneuma or Latin spiritus (originally for Hebrew ruach), which had, of course, from a doctrinal perspective, a completely different set of connotations. Another important loan meaning is the word for god, OHG got - already Goth. gup -, which originally in Germanic languages was a neuter pluraletantum - *gooam, *gooo - and denoting rather lower divine beings but became used as a strong masculine noun for the Christian monotheistic God (Germ. *gooaz), often used with attributes like (al)waltant, (al)mahtfg, Lat. omnipotens, in order to show the difference. Dharmarak~a's use of the pre-Buddhist terms huaren 1~A which is already found in theLiezi 5iU.y72 and the Guanyinzi ~m 13" .y73

(lit.: take) it according to the meaning (of the root).") 71 Or: loan-signifier? 72 In the Tang-period this has been recognized and countered by the explanation that the Zhou king Mu ml~x had already been converted (hua 1b) by Mailjusri (Wenshu )(::7*) and Maudgalyayana (Mulian }l). See Daoshi's (second half of the seventh cent.) Fayuan-zhulin $ ~~# (T.2122.394b.20f.), repeated by the Song-scholar Baoyun's .~ (1088-1158) Fanyi-mingyi-ji g~~~~ (T.2131.1166c.29f.). 73 V3<D1bA' ;fi~~::1E,L.\ , ~::1EJL.\ ' J1:.~~Jei:' =f"~~m ("As for example the transformed man: he despises the spirit of birth and death, transcedes the spirit of birth and death, just calling it illusion but not calling it the Dao.")


Creating religious terminology


"-. and huaxiang 1c{* - found in the Huainanzi 5tm-=r74 - for nirmita, "magically produced statue or man" belongs in this category. This category thus comes closest to what is called geyi ;f~~ in Chinese Buddhist texts. Here it is certainly useful to keep in mind the distinction between "formal" and "conceptual loans" ~hich Erich Zurcher made in his discussion of the terminological influence of Buddhism on early Daoist texts. 75 The two categories, although they certainly overlap, refer to the fact that in some cases words from an originally Chinese context are taken over just as, as it were, "empty cartridges" which are filled up with almost completely new Buddhist meaning. I would argue, however, that even in the case of the usually quot~d examples of geyi - dao m, usually being taken as a translation for bodhi, and wuwei fW<~, being a translation for nirva1}a, nirvrta .ornirvrti76 - there may sometimes be some semantic reasoning for choosing them for rendering the lndic terms. In the case of :dao there are enough examples where the term renders an lndic yana77 which semantically can only be derived from rya- (yati), "to lllove, to go." For wuwei one could argue as well that it was some analytic process that prompted the translation: nirvrta or nirvrti, "terminated, emancipated," could be taken, after all, in the meaning of "without action," derived again from nir--fvrt- in a respective sense. 78
In the chapter Yuandao-xun ~mwll: ~::t:LLm' 1:l~?!o/JrmT1'f RX; ("Now the Dao of the Highest Heaven generates the ten thousand things but does not exist; it produces the transformation of appearances but does not regulate.") >.75 See Zurcher 1980.


See the examples in Karashima 1998: 472f., S.v. Karashima 1998: 86f. ';78 From the standpoint of an early Prakrt origin of the equation - see GandharI nivalJa (nirvalJa), nivudu (nirvrta), nivrudi (nirvrti): Brough 1962: 302c - the derivational process was not as clear as in Skt. which, however, shows its own inconsistencies: nir--fva-, nir--fvr(t)-. A GandharI inivalJa could, after all, well be interpreted as *nirvarlJa, being derived trom nir--fvr-. See also Norman 1994: 221ff., and 1997: 13.



Max Deeg

It is in this group that we meet the bidirectional poly-semantic character of the terminology which strikes modern philologists as a feature of inconsistency in Chinese Buddhist translations. By poly-semantic I mean the fact that there is not a one-to-one correlation between the original term and the translation, andbidirectional points out the fact that there may be different renderings for the same Indic words, but also one Chinese translation term which renders more than one Indic original word. We also find such cases in the Germanic languages. I only want to bring up the example of the word "soul" as a loan meaning maybe originally a loan creation - for Lat. anima. It did not cover the whole range of use of anima which could also mean "life" in its physical and mental aspects - which usually is rendered by other words such as lIb (Old-Saxon lif), which also means body, or jerah (spirit) - and OHG sela, already Goth. saiwala, Old-Saxon seola, usually is used mainly for the soul in our modern religious sense, indicating individual transcedency.J9 In the light of Indian grammatical and semantic analysis, however, such a poly-value is easily understood. Already the Nigha1}tulists of synonymic expressions from the Veda, placed in front of Yaska's Nirukta, and Yaska's different explanation of the same word show this clearly. On the Buddhist side it is again the later Mahavyutpatti which may shed some light on the underlying understanding:
In respect to one expression several words (can be) understood. sD

Conclusion The aim of this paper was to discuss more general issues of the translation techniques used in and underlying early Chinese Buddhist translations. A scheme for arranging and analyzing the early Chinese translation vocabulary should raise an awareness of these different categories, and the terminological creativity may warn us not to discard some of the renderings as crude or even "false" be79

Weisweiler, Betz 1974: 112f.

skad gcig la min du mar 'dren pa ni ... ; see Simonsson 1957: 250.


Creating religious terminology


fore we try to understand why they were chosen by the translator iIi the first place. A correct interpretation of these early translation activities will not only throw light on some aspects of the history of Buddhist texts, especially those of early Mahayana, but will lay the foundation of a better understanding for the development and .spread of Buddhism in India and beyond in the first centuries CEo

Backes, Herbert. 1982. Die Hochzeit Merkurs und der Philologie. Studien zu Notkers Martian-Ubersetzung. Sigmaringen: Jan Thorbecke Verlag. Betz, Werner. 1965. Deutsch und Lateinisch. 2Bonn. (First ed. 1949). - 1974. "Lehnworter und Lehnpragungen im Vor- und Frtihdeutschen." In: Friedrich Maurer, Heinz Rupp (eds.), Deutsche Wortgeschichte, Band I, Berlin, New York (Grundriss der Germanischen Philologie 17/1): Walter de Gruyter: 135-163. Boucher, Daniel. 1996. Buddhist Translation Procedures in Third-Century . . China: A Study of Dharmarak~a and His Translation Idiom. University of Pennsylvania (PhD dissertation). '- 1998. "GandharI and the Earliest Buddhist Translations Reconsidered: The Case of the SaddharmapUI;u;larlkasiitra." Journal of the American Oriental Society 118: 471-506. - 2001. "The textual history ofthe Ra~trapazapariprccha: Notes on its third century Chinese translation." Soka-daigaku-kokusai-bukkyogaku-kotokenkyiljo-nenpo Heisei 12 nendo/Annual Report of The International Research Institute for Advanced Buddhology at Soka University for the Academic Year 2000 EU1Jl*~~~~{?t~~~~~ff%pJTif&.l:jZ~ 12 ifJJr, No.4: 93-116 . .;.... 2008. Bodhisattvas of the Forest and the Formation of the Mahayana. A Study and Translation of the Ra~trapalapariprccha-Siltra. Honolulu: University of Hawai'i Press. Bronkhorst, Johannes. 2001. "Etymology and Magic: Yaska's Nirukta, . Plato's Cratylus, and the Riddle of Semantic Etymologies." Numen 48: 147-203. Brough, John. 1962. The Gandharl Dharmapada, London, New York, Toronto. 1975. "Buddhist Chinese Etymological Notes." Bulletin of the School of Oriental and African Studies 38: 581-585 (reprinted in Brough, John. 1996. Collected Papers. Hara Minoru, Wright, lC. (eds.). London: School of Oriental and African Studies, University of London: 433-437.


Max Deeg

Chen Shufen. 2004. Rendition Techniques in the Chinese Translation of three Sanskrit Buddhist Scriptures. Devon: Bardinge Simp ole Publishing (Cambridge Buddhist Research Series 1). Curtius, E.R. 1967. Europiiische Literatur und lateinisches Mittelalter. 6Bern, Miinchen: Francke. Deeg, Max. 1995. Die altindische Etymologie nach dem Verstiindnis yaska's und seiner Vorgiinger. Eine Untersuchung iiber ihre Praktiken, ihre literarische Verbreitung und ihr Verhiiltnis zur dichterischen Gestaltung und Sprachmagie. Dettelbach: Verlag IH. Roll (Wiirzburger Studien zur Sprache & Kultur 2). 1995b. '''Misunderstanding' in Translations of Sacred Texts - The Case of Old-High-German Bible Translations and Sino-Buddhist Sutra Translations." Gokai: sono gengo-bunka-teki-shoso ~M-f0)~~:X1c.~g:M. Nagoya: Nagoya-Daigaku-Gengobunkabu (Nagoya-Daigaku-Gengobunkabu Tokutei-kenkyu-sirlzu 5 ~rS~*~i~ff~%'JEV9 - A <): 165-185. 2004. "Bhagavat in Chinese Buddhist Translation: An Indirect Example of Oral Nirvacana in Buddhist Text Translations?" In: Sh. Hino, T. Wada (eds.). Three Mountains and Seven Rivers. Prof Musashi Tachikawa's Felicitation Volume. Delhi: Motilal Banarsidass: 153-167. Fo Guang Shan Foundation for Buddhist & Culture Education (ed.). 1999. Buddhism Across Boundaries - Chinese Buddhism and the Western Regions. Collection of Essays 1993. Sanchung: Foguang Cultural Enterprise Co., Ltd. Fussman, Gerard. 1989. "GandhlirI ecrite, GandharI parlee." In: Collette Caillat (ed.). Dialectes dans les literatures 1ndo-Aryennes. Paris: College de France: 433-501. Harrison, Paul; South Coblin, W. 1999. "The Oldest Buddhist Incantation in Chinese? A Preliminary Study of the Chinese Transcriptions of the Mantra in the Druma-kinnara-raja-paripfcchli-su.tra." In: Fo Guang Shan Foundation for Buddhist & Culture Education 1999: 138-186. Haugen, Einar. 1953. The Norwegian Language in America, A Study in Bilingual Behavior. Philadelphia. Held, Axel. 1972. Der buddhistische Monch YEN-TS'UNG (557-610) und seine Obersetzungstheorie. Koln (unpublished Inaugural Dissertation/ PhD dissertation). Hiniiber, Oskar von. 1983. "Sanskrit und GandharI in Zentralasien." In: Klaus Rohrborn, Wolfgang Veenker (eds.). Sprachen des Buddhismus in. Zentralasien (Vortriige des Hamburger Symposions vom 2.Juli bis 5.Juli 1981). Wiesbaden: Otto Harrassowitz: 27-34. Hunger, Herbert; Stegmiiller, Otto; Erbse, Hartmut; Imhof, etc. 1988. Die Textiiberlieferung der antiken Literatur und der Bibel. 2Miinchen: Deutscher Taschenbuchverlag GmbH & Co. KG.

Creating religious terminology


;1to Takanobu 19TJiJli%. 1990. "Kakugi-bukkyo-ko - shoki-chiigoku-bukkyo no keisei", ~~{l,~Jt~'WcpOO;lt{l,~O)~RlG (,'A Study of Ke-i ~~ Buddhism: The Formation of Early Chinese Buddhism"). Toyo-gakuho *J[ 71/3, 4: 57-89. jiXianlin~~#.1992. "Zai tan 'futu' yu 'fo"'J~ r):J.mJ~ r{~j. Chong, Hwa Buddhist Journal cp11#**J[ 5: 19-30. Kahrs, Eivind. 1998. Indian semantic analysis - The 'nirvacana' tradi,', tion. Cambridge: Cambridge University Press (University of Cambridge Oriental Publications 55). iKarashima, Seishi *~/1i;D-~. 1993. The textual study of the Chinese ver", sions of the Saddharmapur.u!arfkasutra - in the light of the Sanskrit and 'Tibetan versions. Tokyo: SankibO-busshorin =:m{~.#. ,;~\-,' 1994. Cho-agon-kyo no gengo no kenkyu (Onshago-bunseki wo chushin to-shite) D" ~JloJ-a~~.lI O)i*~O)tiffJ.E(if~~:5tfJT~CPJl,\t ll"). Tokyo: Hirakawa Shuppansha .IJZ)iJJtI:lil&U. 1998. A Glossary of DharmarakCia's Translation of the Lotus Sutra IE ;,$*~~lTI~. Tokyo: The International Research Institute for Advanced Buddhology, Soka University (Bibliotheca Philologica et Philosophica Buddhica I). ~'f'1999. "Hokke-kyo no bunkengaku-teki-kenkyii (2) - Kannon Avalokita;{svara no goi-kaishaku" $~O)X:'*BkJtiffJ.E C=J - _if Avalokitasvara .'. ~~O)m'~. Soka-daigaku-kokusai-bukkyogaku-koto-kenkyujo-nenpo ::; Heisei 10 nendo/Annual Report of The International Research Institute for Advanced Buddhology at Soka University for the Academic Year 1998 'Err1l**~~~{~~*~~tiffJ.EPJT1:FJ[.IJZRlG 101:F)Jt, No.2: 39-66. ~..!." 2001. A Glossary of Kumiirajfva's Translation of the Lotus Sutra ~.Y$ ;, li$!~lTI~. Tokyo: The International Research Institute for Advanced !, Buddhology, Soka University (Bibliotheca Philologica et Philosophica Buddhica IV). ~:;12001b. "Some features of the language of the SaddharmapuI).garIkasiitra." Indo-Iranian Journal 44: 207-230. !i:;;;.2002. "Miscellaneous notes on Middle-Indic words." Soka-daigakuft[,1,:kokusai-bukkyogaku-koto-kenkyujo-nenpo Heise; 13 nendo / Annual . 'Report of The International Research Institute for Advanced Buddhology " at Soka University for the Academic Year 2001 ;U1l**~~~{~~*~~ tiffJ.EpJT1:FIPRlG 131:F)Jt, No.5: 147-152. fashima Seishi; Nattier, Jan. 2005. "Qiuluzi ttiir, An Early Chinese ,;,;c, arne for Sariputra." Soka-daigaku-kokusai-bukkyogaku-koto-kenkyuff:,d0-nenpo Heisei 16 nendo/Annual Report of The International Research !Fnstitute for Advanced Buddhology at Soka University for the Academic ~::Year 2004 Errfl*~~~~{~~~~~tiffJ.EPJT1:FJ[.IJZRlG 161:F)Jt, No.8: 361~f.376.



Max Deeg

Kern, Hendrik, Nanjio Bunyiu. 1908-1912. Saddharmapu7J.tj.arfka. St. Petersburg (Indian reprint Delhi: Motilal Banarsidass 1992). Kieschnick, John. 2003. The Impact of Buddhism on Chinese Material Culture. Princeton, Oxford: Princeton University Press. King, Dan. 2008. The Syriac Versions of the Writings of Cyril ofAlexandria. A Study in Translation Technique. Leuven: Peeters. Kobler, Gerhard. 1994. Althochdeutsches Worterbuch. 4Paderborn: Schoeningh (online-version used); http://homepage.uibk.ac.at/-c30310/ ahdwbhin.html (last accessed: 28 December 2009) Lai, Whalen. 1978. "Limits and Failure of Ko-I (Concept-Matching) Buddhism." History of Religions 18: 238-257. Lenz, Timothy. 2003. A New Version of the Giindhiirf Dharmapada and a Collection ofPrevious-Birth Stories. British Library Kharo~thf Fragments 16 + 25. Seattle, London: University of Washington Press (Gandharan Buddhist Texts 3). Link, Arthur. 1961. "The Earliest Chinese Account of the Compilation of the 'Tripitaka' (II)." Journal of the American Oriental Society 81/3: 281-299. Man Mai Lai (Mai Man Lan). 2002. Dharmarak~a and His Work: The Impact of Central Asian Buddhist Thought in Translating Buddhist Texts in the Third to Fourth Century China. Ann Arbor (UMI dissertation print; originally University of Wisconsin Madison 1994). Nattier, Jan. 1990. "Church Language and Vernacular Language in Central Asian Buddhism." Numen 37/2: 195-219. - 2003. "The Ten Epithets of the Buddha in the Translations of Zhi Qian 3Z:~." Soka-daigaku-kokusai-bukkyogaku-kOto-kenkyujo-nenpo Heisei 14 nendo / Annual Report of The International Research Institute for Advanced Buddhology at Soka University for the Academic Year 2002 EU 1I*~~~{~fjt~~~:oJf~pff1:f.IPjZpJG 141:f.Jjt, No.6: 207-250. - 2003b. A few good men: The Bodhisattva Path according to The Inquiry of Ugra (Ugrapariprccha). Honolulu: Hawai'i University Press. - 2008. A Guide to the Earliest Chinese Translations: Texts from the Eastern Han *~ and Three Kingdoms ':::'!giQ Periods. Tokyo: The International Research Institute for Advanced Buddhology, Soka University (Bibliotheca Philologica et Philosophica Buddhica X). Norman, K.R. 1994. "Mistaken ideas about nibbiina." In: Tadeusz Skorupski (ed.). The Buddhist Forum Ill. London: School of Oriental and African Studies: 211-225. (Reprinted in: Norman, K.R. 1996. Collected Papers, VOl.VI. Oxford: The Pali Text Society: 9-30). - 1997. A Philological Approach to Buddhism. The Bukkyo Dendo Kyokai Lectures 1994. London: School of Oriental and African Studies (The Buddhist Forum 5).

Creating religious terminology


Pulleyblank, Edwin G. 1983. "Stages of Transcription ofIndian Words in Chinese from Han to Tang." In: Klaus Rohrborn, Wolfgang Veenker (eds.). Sprachen des Buddhismus in Zentralasien (Vortriige des Hamburger Symposions vom 2.Juli bis S.Juli 1981). Wiesbaden: Otto Harrassowitz: 73-102. _ 1991. Lexicon of Reconstructed Pronunciation in Early Middle Chinese, Late Middle Chinese, and Early Mandarin. Vancouver: University of British Columbia Press. Salomon, Richard. A Gilndhilrf Version of the Rhinoceros Satra. British Library Kharo~thf Fragment SB. Seattle, London: University of Washington Press (Gandharan Buddhist Texts 1). South Coblin, W. 1983. A Handbook of Eastern Han Sound Glosses. Hongkong: The Chinese University Press. Simons son, Nils. 1957. Indo-Tibetische Studien. Die Methoden der tibetischen Ubersetzer, untersucht im Hinblick auf die Bedeutung ihrer Ubersetzungen fur die Sanskritphilologie. Uppsala: Almqvist & Wiksells Boktryckeri. Strauch, Ingo. 2007. "The Bajaur collection: A new collection of Kharo~thI manuscripts - A preliminary catalogue and survey" (http://www.de/indologie/bajaur/publication/strauch_2007_I_O.pdf; accessed 28/01/2008). Tang Yung-Tung. 1951. "On 'Ko-Yi', the Earliest Method by which Indian Buddhism and Chinese Thought were synthesized." In: W. Ringe, et.al. (eds.). Radhakrishnan - Comparative Studies in Philosophy Presented in Honour of his Sixtieth Birthday. London: Allen & Unwin: 276-286 . .Turner, Ralph L. 1966. A Comparative Dictionary of the Indo-Aryan Languages. London: Oxford University Press. Vetter, Tillman; Zacchetti, Stefano. 2004. "On lingfa ~Jlf5ii; in Early Chinese Buddhist Translations." Soka-daigaku-kokusai-bukkyogaku-koto-ken. kyajo-nenpo Heisei IS nendo / Annual Report of The International Research Institute for Advanced Buddhology at Soka University for the Academic Year 2003 EU1l*~~~H~~r'aJ~liJf5'CpfT~&'~ 15 ~.oc, No.7: 159-166. Weisweiler, Josef; Betz, Werner. 1974. "Deutsche Friihzeit." In: Friedrich t . Maurer, Heinz Rupp (eds.), Deutsche Wortgeschichte, Band I, Berlin, New York: Walter de Gruyter (Grundriss der Germanischen Philologie 17/1). Zacchetti, Stefano. 2005. In Praise of the Light. A Critical Synoptic Edition with an Annotated Translation of Chapters 1-3 of Dharmarak~a's Guang zan jing 7\:;W~, Being the Earliest Translation of the Larger Prajfiilpilramitilsatra. Tokyo: The International Research Institute for Advanced Buddhology, Soka University (Bibliotheca Philologica et Philosophic a Buddhica VIII).


Max Deeg

Zurcher, Erich. 1977. "Late Han Vernacular Elements in the Earliest Buddhist Translations." Journal of the Chinese Language Teacher's Association 12/2: 177-203. 1980. "Buddhist Influence on Early Taoism - A Survey of Scriptural Evidence." T'oung Pao 66/1-3: 84-147. - 1991. "A New Look at the Earliest Chinese Buddhist Texts." In: Koichi Shinohara, Gregory Schopen (eds.), From Benares to Beijing. Essays on Buddhism and Chinese Religion in Honour of Prof Jan Yiin-Hua. Oakville, New York, London: 277-304. - 1999. "Buddhism Accross Boundaries: The Foreign Input." In: Fo Guang Shan Foundation for Buddhist & Culture Education 1999: 2-59.

Early Chinese Buddhist translations Quotations from the early translations in anthologies of the sixth century Hubert Durt

Anthology, an old Greek term used for collections of poetical flowers, selected as the best and preserved for the enjoyment of future generations, corresponds more or less to the two Chinese encyclopedic collections of the early sixth century which will be introduced here. The two compilations offer a selection of texts which is not based on their literary value, or on their antiquity. Sengyou {~tb (445-518) wished to compare complementary and sometimes contradictory tales about the life of the Buddha. We have kept what seems to be his first attempt (thirty-four chapters of very different length, twice edited): the Shijia pu ~~~* (T. 2040). An important characteristic of the Shijia pu is the importance attached to the family relations. As indicated by its title "Genealogy of Sakya (~~*)," the Shijia pu is obviously concerned with the family relations of the Buddha: that is to say, his lineage and his relatives. Among the thirty-four chapters, fourteen chapters may be considered, already from their titles, as mainly family related: three chapters (I, II, III) deal with the ascendents of the Buddha; three chapters (VI, VII, XVIII) deal with his relatives in a broad Sense; five chapter~ are named after his immediate relatives: father (XV), mother (XVI), aunt and stepmother (XIV and XVII), and son (XIII); three chapters are labeled after his cousins, either followers such as Anuruddha and Bhadrika (XI), and Sundarananda (XII), or dissidents, like Devadatta (X).

Journal of the International Association of Buddhist Studies Volume 31' Number 1-2 2008 (2010) pp. 119-139



In a more ambitious work, a kind of encyclopedia of "practical" Buddhism (exempla, ritual, devotion), the lost Fayuan zayuan yuanzhi ji 5t:~i~J1~~J]l:5Ei~ or Fayuan ji, Sengyou had devoted a section to the "Treasure (consisting of the) Buddha" (jobao {~.). In this section, there are twenty-two chapters whose content seems to have been close to at least seven of the last chapters of the Shijia pu. This supposition is based on the table of contents of the Fayuan ji which has been kept in the Sengyou's Canonical Inventory, the Chu sanzang jiji tB - ~~G~.l In the linglii yixiang ~!1.fJ1if (T. 2121, dated 516), the intention of Baochang .[l~ (ca. 466 - ca. 518) who was a collaborator of Sengyou was different. He collected 782 extraordinary exempla (yu D.Jl) or "strange tales" (yixiang)2 which must have come from Indian Buddhist Scriptures (jinglU). He grouped them under a system of more or less "anthropological" categories consisting of twelve main sections (bu frG) devoted to living beings. 3

T. 2145, j. 12, pp. 87a - 94c. 2 In the text itself of the Jinglii yixiang (see T. 2121, j. 30, p.l60a 4, quoting the Chinese version of the Samantapasadika, T. 1462, j.lO, p. 743b20), the use of the term yixiang with the meaning of "strange story" confirms us in the choice of the translation adopted here for the title of the Jinglii yixiang instead of the more diffuse "Various aspects of sfltras and Vinayas." 3 Apparently under Chinese influence, these sections mix, on the one hand, the fourfold Mahayanic distinction (Buddha, Bodhisattva, pratyekabuddha and sravaka), fourfold sangha distinction (bhik.~u, bhik$ulJf, upasaka, upasika), the six classical Buddhist destinies (gati, i.e. deva, human, aSLtra, animal, preta and hell), and, on the other hand, social or secular categories where gender distinction is strictly observed. The titles of the twelve sections (bu %G), added presumably by later editors, are: 1) heaven and earth (i.e. devas), 2) Buddhas, without mention of pratyekabuddhas, 3) Bodhisattvas, 4) sravakas, 5) kings, 6) eminent people (changzhe ~:t3), 7) upasakas and upasikas, 8) heretics and hermits/immortals (waidao xianren :9~m1llJA, including brahmacarins and brahmins), 9) landlords and ordinary people (jushi shumin deng '@ffll~';~), 10) ghostly spirits (guishen 5Ib1$) including on a very limited scale asuras and pretas, 11) animals, 12) hells.

Early Chinese Buddhist translations


By far the largest and most important part of the compilation is devoted to the human destiny U. 4-45), starting with a section 'on the Buddha (fo bu {~:g~, j. 4-7), centered on events in his life. Baochang often used the same material as Sengyou. As, observed by Ouchi Fumio, three of the four subsections of the linglU yixiang's section on the Buddha are based on certain chapters of the ;Shijia pu. 4 As the scope of Baochang is much more extended than the genealogy and biography of the Buddha dealt with by Sengyou, yve find episodes of the life of the Buddha in almost all the twelve bu of his encyclopedic anthology. There is even a difference of material between Baochang and Sengyou in the section on the Buddha of the linglii yixiang. Among the relatives of the Sakya (the Buddha's father, mother, etc.), Baochang introduced a tale about the antecedents of the famous Chandaka and Kal).thaka, the esquire and the horse of the Buddha. I will revert to this tale in the last part of this paper when dealing with the hbrmalisation of archaisms in the anthologies.

fI. The archaic translations used in T. 2040 and T. 2121

:What is the interest of these two extant (and extended) works for ~!p.e textual study of archaic translations? ~tWe have to take into account, first, that both compilers, Sengyou ~~nd Baochang, were first interested principally in narrative litera;tiue (and not in doctrinal texts)5 and that, second, they were not ~jarchaeologists" in the sense that they were not primarily inter"~sted in the most antique translations, but they made use of them ~\Vhen their narrations seemed valuable. The name of a translator is ':hever mentioned. It seems that Sengyou and Baochang liked to be ' ~~up-to-date."6 We find long quotations of texts which had just been translated at the time of the compilation of the two anthologies. ,~uch is the case for the two extant translations (?) made by the un:t~-"-----------

Ouchi 1977. Especially in the linglU yixiang, doctrinal developments have been ~:xcised from the quoted texts. ,1 ,6 Durt 2006: 51-86.




known Danjing II:~ at the end of the fifth century. These translations are the Mahiimiiyiisutra/Mohemoye jing ~gay~IfG~ (T. 383), on the mother of the Buddha, and the "Sutra of the extraordinary conditionment" (Weicengyou yinyuan jing *~1lZSI~~I, T. 754), which is one of the most interesting documents on the legend of Rahula, the son of the Buddha? The two translations (?) of Danjing became quite authoritative. Among the archaic translations used by Sengyou in his research on the biography of the Buddha are translations' attributed to three early translators: Kang Mengxiang .@:~~, Zhi Qian 3Z: ~, and Dharmarak~a/Zhu Fahu ~5t;gi. They were also used by Baochang, who for his extended narrative researches, made large use of a fourth early translator, Kang Senghui .@:{~~ who translated the Liudu jijing 7\)jt~~ (T. 152), an archaic collection of tales connected with the six perfections (du )jt, piiramitii).
1. The translator Kang Mengxiang (active 194-210) is mostly known for the "Sutra of the practice and original rise" (Xiuxing benqi jing MHT::Z[S:iEe~J., T. 184 in 7 chapters) and the "Middle sutra of the original rise" (Zhong benqi jing CP::Z[S:iEe~J., T. 196 in 15 chapters). He was for the first text working with Zhu Dali ~*1J, and for the second text working with Tanguo 1I:5It. These two benqi jing form one biography of the Buddha cut in two parts for unknown reasons and divided into monographical chapters. T. 184, on the birth and the youth of the Buddha, is used only for short references, . as Sengyou disposed already of two main biographical sources: the "General radiance s~ltra" (Puyao jing ~81~I, T. 186), a translation attributed to Dharmarak~a which was used for the first edition (in around 502) of the fourth chapter of the Shijia pu, and the "Sutra about causes and effects" (the abbreviated title for Guoqu xianzai yinguo jing ~~;JJjl.1:ElZSI5It~, T. 189), a translation by GUI).abhadra (of the fifth century), used for its second edition (of around 515).

7 I have two articles in preparation on the originality of these Maya and Rahula narratives. This original character must have been influential in the attraction of Sengyou and Baochang to the two translations (?) of Danjing.

Early Chinese Buddhist translations


2. For Zhi Qian (active 220-252), we are mostly confronted with his "Satra of the auspicious original rise of the prince" [i.e. the Buddha to be] (Taizi ruiying benqi jing *T:fMJJ!*re~~, T. 185, without chapter division) which can already be considered an anthology, or at least a compilation made of elements of diverse origins. 3. Among the abundant translations attributed to Dharmarak~a (active 265-313), several works (among them the [Upiiyakausalya] lfiiinottarabodhisattvaparipfcchii/Huishang pusa wen dashanquan jing ~L:gil7iFo5*;gfl~~, T. 345) are quoted for short references, but it is the influential Puyao jing which is the most extensively quoted, sometime erroneously.8 I say "influential" because Sengyou considers the antecedents of the Buddha from the standpoint of the Lalitavistara of which the Puyao jing represents an archaic and in some places more extended version. This alignment of Sengyou with a Sarvastivadin tradition made him almost suppress the popular Sumedha tale from the Shijia pu account on the origins of the Buddha. Quoting a satra of different origin (the already referred to Weicengyou yinyuan jing *~~IZSJ~~~, T. 754), Sengyou has been obliged to introduce a story suppressed in his first chapters (the marriage of the Buddha's predecessor Sumedha) in his chapter XIII on Rahula. 4. Kang Senghui (active around 247-280) is only known for two translations, both collections of exempla, the "Compendium siltra ()f the six perfections" (Liudu jijing 7\N~~, T. 152)9 and the "Siitra of various ancient example stories" (Jiu zapiyu jing tHfE':" nR~~, T. 206).10 There are still other archaic translations quoted here and there in the Shijia pu and the JinglU yixiang. There are also many trans8 The "error" alluded to is an attribution of the source of chapter XII on Saundarananda and Bhadrika to the Puyao jingo Examining Sengyou's Catalogue (T. 2145, j. 2, p. 87c23), one can get the Correct reading Chuyao jing 1:B8I~J[ (T. 212). 9 There are 25 extracts from T. 152 in T. 2121. 10 There are 8 extracts from T. 206 in T. 2121.



lations which are at the frontier of the archaic translation such as the works of Zhu Fonian ~{~~ (active 365 - early fifth century). Zhu Fonian is well represented in the abundant extracts from the Agamas and the Vinayas but also in his more original works, as the "&Ura of the embryonic stage of the Bodhisattva" (Pusa sitOtaijing :g~J1JTij~~!, T. 384),11 also called "Womb sutra."12 As could be expected, the previous collections of tales were greatly exploited in the two compilations of the early sixth century. We have already mentioned the Liudu jijing 7\Jjt:*~! (T. 152) and the liu zapiyu jing !i~t~DJfu~ (T. 206). As we will see, the case of the anonymous translation entitled "Sutra of great salvific artifice being the indebtedness toward the Buddha" (Da fangbian fobaoen jing *}]1J!{~$R)t~!, T. 156) deserves a special study.13 Among the "classical" translations, extracts from the "Great treatise on the perfection of wisdom" (fJa zhidu lun *~jjt~Hfi, T. 1509), the "Sutra of the wise and the fool" (Xianyu jing Jl~~!, T. 202) and the "Storehouse of sundry valuables" (Zabaozang jing ~t. ~~~, T. 203)14 appear frequently, as well as collected tales from the Dharmapada tradition such as the Chuyao jing tf::\8m~~ (T. 212). There is even, both in the Shijia pu and in the linglit yixiang, a quotation from an unknown source, probably lost, the lijing chao *~:ty.15 As expected, there are many more references to unknown works in the linglii yixiang than in the comparatively short Shijiapu.

11 See the episodes from T. 384 artificially bound in the Nirva1fa Chapter (XXVII) of the Shijia pu T. 2040, j. 3, 73b. 12 See the unpublished PhD thesis of Elsa Legittimo, International College for Postgraduate Buddhist Studies, Tokyo 2006. 13 See the section hereunder about shortenings (chapter XIV of the Shijia pu). 14 Translated under that title by Willemen 1994. 15 See the chapter XXII of the Shijia pu.

Early Chinese Buddhist translations


II Characteristics of quotations from early translations

How to consider quotations from archaic texts? I see five approaches: 1) the way of shortening the quotation, 2) the kind of Buddhist terminology used, 3) the differences in transcription and translation of proper names. 4) In which measure are the quotations close to the original texts as we know them in modern editions? 5) Is it possible to detect sectarian influence? Before investigating a few exemplary cases, I have to make a general observation on the way the quotations are introduced. Inihe Shijia pu, under the title given to each of the 34 chapters, there is a mention of one source. Some short chapters are based ()nly on this mentioned source. In most of the cases, this source is a reference to the first quotation. Thereafter come two or three supplementary quotations from other texts. In the longest chapters, dozens of texts are quoted. At the end (and exceptionally in the middle) of a chapter, Sengyou gives his own short comment, which alludes often to the difference of content among the quotations. 16 In the linglU yixiang, the quotations are generally short. Their origin, and exceptionally a variant source, are quoted, without comment, at the end of the quotation. 17

I Shortening of quotations
Abbreviated quotations are very frequent. Looking on these reductions, I must say that on the whole I admire how accurately the meaning of a passage has been respected and how the vocabulary remained faithful to the original. Generally, the accumulated epi!hets and the fastidious formulas of greeting are suppressed. Often, the gathas are also suppressed. These elements belong probably to the oral recitation of texts. With the anthologies, we move from
16 For the identification of the sources of the Shijia pu, we are helped by the annotated Japanese translation of that work by Hasuzawa 1936. See also Kanno 1996. 17 The source in the TaishO Canon of every quotation of the linglii yixiang is now very clearly listed in the book of Sakamoto 2005.



orality to silent reading. When several sources ,are used for the same tale, an effort has been made to avoid repetitions. I will take as an example chapter XIV of the Shijia pu on the ordination of MahaprajapatI Gautaml. The mentioned source is the ninth chapter of T. 196, translation attributed to Kang Mongxiang. Sengyou quotes it at length but avoids mentioning the eight gurudharmas of the bhik~u1Jls. He shifts shortly to the "Vinaya of the Mahlsasakas"/Wufen lil n:51-j$ (T. 1421) probably because these rules are a Vinaya matter and he concludes with a quotation from a collection of tales on gratitude (T. 156) in order to insist on the gratitude owed to Ananda who has been so influential in the opening of the Sangha to women. In his conclusion, Sengyou points out that the fourfold community was already a tradition of the past Buddhas. For this episode, Baochang refers directly to the Shijia pU. 18 Another example of shortening an archaic translation could be taken from chapter XIII on Rahula. There, the quotation from the Puyao jing figures between a lengthy quotation from the already referred to "modern" source on Rahula, the Weicengyou yinyuan jing *~1[ZSJ~~~Jlf (T. 754) translated by Danjing, and a quotation from the "Vinaya of the Mahlsasakas" (T. 1421) about the rule that the parents' agreement is compulsory for the ordination of children as novices. The story deals with the famous theme of Rahula recognizing his father whom he had never seen. A particularity of the tale of the Puyao jing is that the mother of Rahula, called Yasodhara in the previous quotation, is here called Gopi Such shift of name is not unusual in the anthology of Sengyou. Telling shortly the same Rahula story, but in much shorter terms, the linglii yixiang does not quote the Puyao jing but makes again use of T. 754. Other versions of the Rahula story are taken from the Da zhidu lun (T. 1509) and from one of the DharmapadaUdanavarga related collections of tales (T. 211).19

18 There are 14 quotations made directly from the Shijia pu in the Section on the Buddha of the linglii yixiang. 19 Translated by Willemen 1999.

Early Chinese Buddhist translations


2. Evolution of the Buddhist terminology

As with the shift of proper names, the changes of vocabulary for terms as important as nirvCilJa (nihuan 5Jb5.: and niepan 5.!E~), arhat (yingzhen B!~ and aluohan IloTR)~), etc. seem to be generally accepted. An archaic translation such as the ninth chapter of 'T. 196 completely integrated in the chapter XIV of the Shijia pu on the ordination of MahaprajapatI GautamI is suffused with freshness (the description of the bride)20 and of vigor of expression (the Buddha to Ananda: dan ru-suo-zhi bu ru wo-zhi {g)it:PfT~ 7F:f<oiR; ~ "What you know is not yet to be compared with what I knoW").21 Some references to proverbs,22 or to gestures later forgotten (such llsdisplaying one's hair on the ground in order to let a sramalJa 'step on it),23 bring us back to primitive flavors reminiscent of the Sumedha tale. The quotation of this text in the Shijia pu is some'times deprived of details which were too well known, such as th~ seven days that the future Buddha lived with his natural mother before her death. 24 , ", Nevertheless, an influence of the "classical" language of the "Chinese translations of the fifth century is conspicuous in our two ;:lllthologies of the early sixth centuries. Both anthologies owe much .to the Da zhidu lun (T. 1509), translated by J(umarajlva. Although ibeing first of all a Prajiiaparamita commentary, the Da zhidu lun ,has also been a kind of Mahayana encyclopedia, especially rich in exempla, and a standard of the Chinese Buddhist terminology until the middle of the seventh century. It was in the Tang period that the coining of new translations and new transcriptions was attempted for certain terms by Xuanzang ~~ and Yijing ~). The linglii yixiang presents a particularity that we do not find in the Shijia pu. I have previously mentioned the use made by Baochang of the liudu jijing (T. 152). In nine of the twenty five cases, the quotation 20 T. 159, 159all-15. 21 T. 159, 159a25.
T. 159, 158b29.

23 T. 159, 159bl-2. " 24 T. 159, 158c7.



is not referred to one of the multiple titles of the Liudu jijing, but to unknown siltras with the name of the main figure of the tale . . These siltras faithfully duplicate the original tale, but make use of a "normalised" terminology for some technical terms and for some proper names. They probably had a more "modern" independent existence parallel to their presence in the collection of the Liudu jijing.25 It is thus their normalized version of the tale which figures in the Jinglii yixiang. . A typical example of such a normalization may be studied in the text called Chandakasiltra which is tale no. 14 of the seventh juan (on the relatives of the Buddha) in the Jinglii yixiang. 26 It is inspired by tale no. 83 (on the perfection of wisdom) of the Liudu jijingY Modifications have been made in the terminology ( e.g. aluohan ~OJi5~ for yingzhen H!Jt,28 cf. supra) or in proper names: Maya, the mother of the Buddha, is ordinarily called Shemiao ~frrj> in the Liudu jijing but is called Moye f,J1fG in the Chandakasiltra reproduced in the Jinglii yixiang.29
3. Differences in transcriptions and translations ofproper names

As already pointed out, the changes in the proper names did not seem to have mattered much. We read in the chapter XXVII (on the ParinirvaIfa) of the Shijia pu the name of Cunda, the last amphytrion of the Buddha, quoted under four different transcriptions. Three texts are quoted: the Mahayanic Mahiiparinirvii1}asiltra, with Chuntuo ~~W8; the Mahiiparinirvii1}a account of the Dfrghiigama, with Zhouna JBJ11G; an archaic Nihuan jing 5Jb5~, with only Chun ):; and the comment of Sengyou himself who coins Chuntuo ): W8 as a kind of compromise between the texts of the Great and the Small Vehicle. 3D There are differences which derive from the

See on this topic the observations by Pelliot 1920: 340.


28 29

T. 2121, j. 7, 37a29-39a3. T. 156, j.8, 44b-46c. Translated by Chavannes 1910: 293-304. Compare T. 152, 45c29 and T. 2121, 38c12. Compare T. 152, 46a29 and T. 2121, 39a3. See T. 2040, j. 4, 70a15, bI7, c3 and c16.

Early Chinese Buddhist translations


lndic language in which the translated text was originally written, but there were also - it seems to me - a few differences between the official (or pedantic) way to write a name as, e.g. Suddhodana, and the more familiar way to write it: Jinghan )w ("Pure Rice"). I do not know how long in China a familiar expression such as Shijiawen ~'Jm!X for Sakyamuni has been used, but it is clear that it appears in archaic translations that found its way to the Japanese polemical writers of the eighteenth century. We sometimes meet difficult place names figuring in the original text but suppressed in the anthology. Such is the case for the Nasi xian jj~fl-~ appearing in T. 196 and omitted in T. 2040. 31 The tale started in Kapilavastu but there is a question of changing the place for the rainy season to a place close to a river. Could Nasi jj~ fl- be an unorthodox transcription for Vara:r:tasi or for KaSi?

4. Differences in the content of a quotation

A fundamental question is the closeness of the quotations to the original texts as we know them in modern editions. We may answer .that, taking into account the shortenings (our section one), there is generally a striking similarity between the text of the TaishO and its quotation in anthologies which are also edited in the TaishO. The references given in the Shijia pu and the linglii yixiang generally include mentions of the juan. They correspond generally to the present situation. It is exceptional that the text of a quotation cannot be identified. Such cases may be worth of further research. Nevertheless, it may be the result of uniformisation of the original text with its anthologica1 quotation which may have taken place since the time of the first compilations of the Chinese Buddhist Canon. On the other hand, a few differences between the original text and its quotation in an anthology may be the result of a normalisation or a wish to correct some incongruities in the content of a tale, as we will see in the case of the famous K~antivadin tale. This is the point that I wish to pursue here.


Compare T. 196, j. 2, 158bll, and T. 2040, j. 2, 52b11.


. Hubert Durt

We have said previously that in the archaic translations belonging to the narrative genre, we have fewer problems of doctrinal terminology than in the more speculative texts. We have seen that Sengyou ,and Baochang generally respected the archaic transcriptions or translations of proper names (our section three) but that some technical terms (like nirval}a, arhat, cakravartin) had often been "normalised" (our section two). Archaic elements in tales: the eclipses of the transformation of blood into milk One problem in which I feel particularly interested in the reading of archaic Buddhist narratives is the way of handling some fantastic hagiographic details which seem to be un-Chinese or even repellent for a Chinese mind. One such case is the transformation of blood into milk which is a miraculous phenomenon illustrating particularly the virtue of patience during the extreme suffering due to self-sacrifice. Those tales belong to the Jiitaka literature, therefore their hero is a bodhisattva who will become in a distant future the Buddha Siikyamuni. Several tales are known for this blood/milk transubstantiation. Although not popular, this theme is not ignored, as we will see, in the Indian Buddhist literature where the miracle had to be ex:' plained: milk is the symbol of motherly love. Such a love is necessary in order to condone and to wish the happiness of one's persecutor. The most famous example of patience cum compassion is the story of the dismemberment of the body of the ascetic K~iintiviida or K~iintiviidin or K~iinti ("Patience") the hermit, as he is often named in Chinese: Renru xianren ;>g~{ULA. He will simply be called K~iinti hereafter. A ferocious king chopped one after the other the limbs of K~iinti as a revenge for the attention devoted to his sermons by the ladies of his court. On this gory tale there exists an abundant bibliography.32 The other tales with which I will deal 32 Lamotte 1944-1980.: I: 264-265, II: 889-890, III: 1670; Panglung'
1981: 92-93.

Early Chinese Buddhist translations


here are less known. The miracle appears in one version of two stories of self-sacrifice. 1.) The tale of the boy Sujati who offered his body, scrap after scrap, to feed his parents and 2.) the tale of the king Kaficanasara whose body was pierced by one thousand holes filled with oil and put to fire as a worship offered to a brahmin in exchange of the second part of a dharmic gatha. An interesting point is that in both the K~anti !radition and the Sujati tradition, the persecutor is connected with Ajfiata Kaul).<;linya, the first disciple ()f the Buddha. The conversion from the extreme of ugliness to the extreme of holiness is typical of the Buddhist dialectics. As shown by the story of Angulimala, he who is able of the worst may be able ()f the best! About the content of the tales and their hagiographic elements, We face delicate questions. As said before, many tales belonging to the Jataka literature and kept in archaic collections have been reproduced later in anthologies under a shortened form. Is it because details had to be eliminated in order to abbreviate the tale? Or did soine events disappear because they had been considered as inc()ngruous? Incongruity meaning unpalatability with the decorum prevalent at certain ages? A comparison between archaic tales and their treatment in old anthologies (since the sixth century) or more recent anthologies (as the Japanese Konjaku Monogatari 4--~?D~ ?fthe early twelfth century) is often rewarding.
{a. The dijing

legend in the Zhong benqi jing and the the Liudu

i'rhe most ancient Chinese version of the tale appears in a "Life of tpe Buddha"33 where references to Jataka are not numerous: the 'Zhong benqi jing cp*m~34 ("Middle sCitra of the original rise," T. 196). As said earlier, with its first part, the Xiuxing benqi jing ~~1=r ,*Jm~If. ("Satra of the practice and original rise," T. 184), this translation is attributed to Kang Mengxiang ~~f active at the end of the second century. 33 See ZUrcher 1978. 34 T. 196, shang, 148c15-149a5.



The Buddha tells his own story of the "Man of patience" (here presented as daojin mA) to remind Kam;l.(;linya that he had been in a previous existence the king Esheng who dismembered K~anti. At the last moment the King asked his victim to be condoned. The transubstantiation blood/milk is produced as the result of a "claim to truth" (satyavacana) made by K~anti and is accompanied by the customary miracle of the restitution of the body of the hero in its former state. The same final scenario figures in the tale of the boy Sujati, but the converted man who will become Kaul).~inya is not the persecutor as the ferocious king, but the benevolent god Sakra of that age who had first derided the suffering boy who refused his boon to become god or king but wished to progress on the way to bodhi. Seeing the transubstantiation miracle, Sakra implored the boy to be the first saved by him when he will become a Buddha. A second ancient Chinese version of the miracle is the tale no. 44 of the Liudu jijing 7\Jjt~~, the "Compendium of the six perfections," translated by Kang Senghui of the third century. It is the version of T. 152 of the K~anti story which will be used in the anthology linglit yixiang (T. 2141). We already know that this anthology "Strange tales from Sutra and Vinaya" makes extensive use of the archaic collection of tales examplifying the six perfections



The tales of the Liudu jijing present evidences of archaic character. Such is the case of the tale of the martyrdom of the hermit (r~i) whose name is almost universally based on the virtue of forbearance (k~Cinti), as is the case for "K~antivada" in the MahCivastu. 35 In all the traditions, his body is chopped in numerous parts by an irate king, but in the Mahavastu, the link between blood and milk is made explicit. Milk is bound metaphorically to the maternal feeling of the victim toward his persecutor. 36 Its effect is that this miracle became a hagiographical mark of the "uncorrupted" (apradu~ta)Y
Senart 1897: 356-36l. Senart 1897: 358.3 Mtitaye putras premena stane~u ksfralp pravaheya evalp - Jones 1956: 356: " ... just as a mother's milk flows from her breasts out of love of her offspring." 37 Ibid. p. 359. 2 and 4.
35 36

Early Chinese Buddhist translations


In the archaic tale of the Liudu jijing, the gushing of the milk is compared to the love of a benevolent mother toward her baby (you cimu zhi ai qichizi ye j@~aL:a;!'t*'Tfu).38 Quoted in the JinglU yixiang, we read again the mention of the love of compassion (min J&;) compared to that of a mother thinking of her baby (you mu zhi nian chizi j@aL~*,T), but there is no further mention of the milk gushing from the wounds. 39 May we consider this absence of the milk reference as the result of a normalisation through erasement? It could be the case, and, in the context of the K~anti the hermit legend, Chavannes has pointed out that the theme of the blood transformed into milk had not been universally accepted. 40 Actually, when we compare the versions of the K~anti tales, which are more numerous than the material used py Chavannes, we may observe that milk is mentioned not only .in influential texts of the fifth century such as the popular tales collection entitled Xianyu jing Ji-~~ ("Sutra of the wise and the fool"),41 or the venerable and universally used Da zhidu lun *~Jjt ~ ("Great treatise on the perfection of wisdom")42 again, although the Paficavif(lsatisiihasrikii Prajfiiiparamitii commented on by this text does not refer, in its numerous versions, to that miracle . . In another passage, commenting on the "Sadaprarudita romance" which closes several Prajfiaparamita texts, the author of the Da zhidu fun explains that the washing of the Dharmodgata sanctuary with blood could be effected by the transformation of blood into scented water: a miracle analogous to the transformation of blood into milk that had been the case with K~anti the hermit. 43

. 38 T.

152, j. 5, 25b2L

T. 2121, j. 8, 40c14.
Chavannes 1934: 113.

41 T. 202, j. 2, 360a17. 42 T. 1509, j. 14, 166c.


Ibid., j. 99, 749b13-24, See Durt 2000: 21, n. 52.



b. The Dafangbianfobaoenjing and other legend~

Although the transformation of blood into milk seems to be a characteristic element of the legend of K~anti the hermit, there is a "sutra," actually a collection of tales, the Dafangbianjobaoen jing *:1J~{iJI;~)JBI,~ ("Sutra of great salvific artifice being the indebtedness toward the Buddha," T. 156) whose unknown translation has been attributed to the late Han period, where, in two tales, the same transformation is mentioned. 44 As mentioned above, these two tales are the dismemberment of the boy Sujati and the worship of king Kaficanasara through an embrasement of his body. This last tale is representative of a hagiographic motive showing an exchange between a devotee's body and the last part of a giithii possessed by a malevolent figure. These two tales which have in the T. 156 narrative a mention of the transubstantiation blood/milk are known without milk mention in several other versions. In these two tales, we have to take into account that the change of blood into milk became a kind of curse, used in satyavacana. In both stories of self-sacrifice,45 the victim claims that the blood of his wounds should become milk or white as milk if his purity of intention should be fallacious. The result is that his tortured body recovers its pristine state. The Dafangbian fobaoen jing, which collects these two tales with a . mention of the transformation miracle, has been abundantly quoted in the linglii yixiang.46 It is thus earlier than the sixth century. The Dafangbianfobaoenjing seems to be a rather eclectic collection of tales which, strangely, have not been included in the compendium of Chavannes 47 nor in the comparative list of Jataka tales made by Higata. 48 In this eclectic collection, we do not find the well-known . 44 Milk is mentioned in T. 156, j. 1, 129c26 and T. 156, j. 2, 135a13. 45 See Durt 2000: 13.
46 Correspondences: T. 156, j. 1, 129c26 = T. 2141, j. 31, 164b26, and T. 156, j. 2, 135a13 = T. 2141, 24, 132b20. 47 Chavannes 1910: I-VIII. See De Jong in the reprint of Chavannes 1962, and De Jong 1965: 240-242. . 48 Higata 1954.

Early Chinese Buddhist translations


legend of K!?anti the hermit, but a different legend of the self-sacrifice of a prince named K~anti.49 About the possible archaic character of a part of the Dafangbian fobaoen jing, I may now only say that some of these tales present archaic translations (e.g. nihuan )Jb).) and other tales more modern translations (e.g. niepan )!E~). The archaic elements may have led to the attribution of the whole sutra to the late Han period. As this text deserves a minute analysis, we cannot deduce from the two passages referred to above any conclusion about the antiquity (or eventually the geographical origin) of the miracle of the transformation of blood into milk.
c. Later occurrences of the theme of blood and milk in the legend

We find again mention of milk in the Chuyao jing tBBI~, a collection of tales translated by Zhu Fonian ~{JI3~ of the late fourth century50 and in the Candragarbhasutra (Yuezang wen fj~x)51 translated by the translator Narendrayasas of the late sixth century. We still find a mention, in relation with Kam;u;1inya, of the miraculous transubstantiation in a famous commentary on the Lotus sutra by Zhiyi ~~~ (538-596).52 In Sanskrit, besides the already quoted Mahiivastu's explicit passage, it seems that we may refer only to the implicit mention that the blood of K!?antivadin was transformed: tasya tad rudhirarrz pariivrtam, without indication of milk, that is found in the Miilasarvastivadavinaya. 53 The detail of the transubstantiation blood/milk seems to be absent from the late tradition. Xuanzang !::~ had the occasion of dealing with the story of K~anti in his translation of the Abhidhar-

49 T. 50 T. 51 T. 52 T.

156, j. 3, 137c-138c. 212, j. 23, 730b13. 397 (15), j. 50, 330b8. 1718, Miaofa lianhua jing wenju *,Y*}j*~![:x1l], j. 5 shang,

Gnoli, Venkatacharya 1978: 10.11.



and in his travelogue, the Xiy~ ji j:t~C,55 but does not mention the miracle. The Vibha~a tale of K~anti became the source quoted in a new encyclopedia, more doctrinal and more ambitious than the lirigliiyixiang, the late seventh century's Fayuan zhulin ~~~B3)Kif*. 56 It is in the Korean hagiographicalliterature that we find a revival of the miracle. According to the Chronicles of Silla quoted in the "Memoirs. of the three [Korean] kingdoms" (Samguk yusa =OOm~, T. 2039), compiled by Ilyon -f~ (12061289), the martyrdom of the hero Ichadon ~"*rt!J[, with milk gushing from his severed head, helped the introduction of Buddhism in the Kingdom of Silla. 57 As I centered this paper on the ancient data collected in the encyclopedic anthologies of the early sixth century, I may conclude this short inquiry on the vicissitudes of what seems to have been an archaic hagiographic element in considering that at least the faithfulness of the linglii yixiang to a more ancient model was conditional. Baochang admitted the maternal love of the tortured K~anti, but did not go as far as reproducing the tale of the miraculous transformation of blood into milk. More generally, it has to be said that this kind of transubstantiation is not limited to the Buddhist tradition. It has probably Indian origins but it can also be found in the hagiography of Islamic or Christian martyrs, as shown by the legend of Saint Catharina of Alexandria in the Legenda Aurea.

5. Sectarian influences
Sectarian differences are rather inconspicuous in narrative texts. We know that Sengyou belonged to the Mahayana, the denomination prevalent in China since the first centuries of the introduction of Buddhism in China. We know through his works, of which much is lost, that he belonged to the Sarvastivadin erudite tradition. He was a Master of Vinaya. We may observe that when there is a con54

T. 1545, j. 182, 915 abo

55 T. 2087, j. 3, 882b25-26. 56 T. 2122, j. 83, 896ab.


T. 2039, j. 3, 987c9.

Early Chinese Buddhist translations


cern for Vinaya rules in a sutra, Sengyou prefers to refer directly to .a Vinaya text. His choice of Vinaya texts is eclectic: Sarvastivadin/ Shisong lii +rnff (T. 1435), MahIsasaka/Wufen Iii li?t:jf (T. 1421), and Dharmaguptaka/Sifen Iii lZ:9?tff (T. 1428). The Vinaya of the Mahasanghika (T. 1425) seems less conspicuous. All these Vinaya translations are not very ancient. They were only one century old when our anthologies were compiled. Both Sengyou and Baochang made use of the recently translated (488-489) commentary of the Pali Vinaya, the Samantapiisiidikii, but Baochang's extensive use of the Vinaya of the Mahasanghika (21 quotations) has to be pointed out. His turbulent life58 does not seem to have been particularly centered on Vinayic orientations. In conclusion, anthologies are a mirror of the objects of interest of their age, especially in our case for religious matters. An anthology like the Shijia pu made Buddhism familiar to a number of readers. It has to be considered in the encyclopedic current of the early sixth century, but we may add that the Shijia pu generated many successors, from the linglii yixiang to the later illustrated lives of the Buddha accompanied with a reduced explanatory text always including a mention of its canonical origin. Reading anthologies has the merit of showing us a Chinese tradition both highly conservative and very "lively." To many texts the anthologies have given a diffusion, and sometimes a protection, that they would not have enjoyed if they had been kept only inside the bulky books where they had originally been written down.

Chavannes, Edouard. 1910. Cinq cents contes et apologues extra its du Tripitaka chinois. vol.l. Paris: Leroux. (reprint 1962, Paris: AdrienMaisonneuve) Chavannes, Edouard. 1934. Cinq cents contes et apologues extra its du Tripitaka chino is. voUY. Paris: Leroux.


See De Rauw 2005.


Hubert Durt

De Jong, Jan Willem 1965. Review of Edouard Chavannes, Cinq cents contes et apologues etraits tu Tripitaka chinois, Paris, 1962, 4 tomes. IndoIranian Journal 8: 240-242. De Rauw, Tom. 2005. "Baochang: Sixth-Century Biographer of Buddhist Monks ... and Nuns?" Journal of the American Oriental Society 125/2: 203-218. Durt, Hubert. 2000. "Du lambeau de chair au demembrement," Bulletin de l'Ecole Franraise d'Extreme-Orient 87/1: 7-22. Durt, Hubert. 2006. "The Shijiapu of Sengyou: The first Chinese attempt t6 produce a critical biography of the Buddha." Journal of the International College for Postgraduate Buddhist Studies 10: 51-86. Gnoli, Raniero; Venkatacharya, T. (eds.). 1978. The Gilgit Manuscript of the Smighabhedavastu. Part. II. Roma: Istituto per il Medio ed Estremo Oriente. Hasuzawa JOjun li5R.RlG~. 1936. Shakafu ~}!J[~. Tokyo: Daito shuppansha (Kokuyaku Issaikyo ~aR-i;)]~, Shidenbu 6 ~:(~fi~7\). Higata Ryusho -=f~~t$. 1954. Honsho kyorui no shisoshiteki kenkya ~~o)J~,ma~rliJf~ [A Historical Study of the Thoughts in Iatakas and the Similar Stories]. Tokyo: Toyo Bunko (Oriental Library Series A, vol.35). Jones, John James 1956. The Mahlivastu (Translated from the Buddhist Sanskrit), voLIII. London: Luzac & Company, Ltd. Kanno Ryusho ;g!ff~~. 1996. "Shakafu inyokyoten ni kansuru ichi kosatsu" r~}!J[~J 5[m~$t'::'~T~-~~. Osaki Gakuh5152: 123-129. Lamotte, Etienne. 1944-1980. Traite de la Grande Vertu de Sagesse, 5 vols. Louvain: Institut Orientaliste/Institut voor Orientalistiek (Bibliotheque de Museon).


Ouchi (Ohuchi) Fumio :kpgX:ME. 1977. "Ryodai Bukkyo ruishusho to Kyoritsu iso" ~{i;{<tfx:!I&i~~t~lt'-<ffi ["Buddhist Encyclopedias of the Liang Dynasty and JinglU yixiang"]. Toh5 Shukyo *7J*fx: 50: 55-82. Panglung, Jampa Losang. 1981. Die ErZiihlstojfe des MalasarvlistivlidaVinaya. Tokyo: Reiyukai Library (Studia Philologica Buddhica, Monograph Series III). Pelliot, Paul. 1920. "Meou-tseu ou Les doutes leves." T'oung Pao 19: 255433. Sakamoto Kobaku :l&*JZ:If. 2005. Kyoritsu iso no kenkyu: Ryodai no bukkyo bunka ~1f~~0)1iJt~: ~{i;0){<tfx:. (privately published). Senart, Emile (ed.). 1897. Le Mahtivastu. Texte sanscrit, vol.III. Paris: Imprimerie Nationale. Willemen, Charles. 1994. Storehouse of Sundry Valuables. Tokyo: Numata (BDK English Tripitaka 10-1).

Early Chinese Buddhist translations


Willemen, Charles. 1999. The Scriptural Text: Verses of the Doctrine, with Parables .. Tokyo: Numata (BDK English Tripitaka 1O-Il) . . Zurcher, Erik. 1978. HetLeven van de Boeddha: Xiuxing ben qijing & Zhong benqi jingo Vertaald uit de vroegste Chinese overlevering. Amsterdam: Meulenhoff.

The work of Paramartha

An example of Sino-Indian crosscultural exchange* Funayama TOfu

This paper is a preliminary investigation of the life and work of paramartha (Ch. Zhendi .~*; 499-569 CE), an Indian commentator active during the late Liang and early Chen ~t periods of the Six Dynasties. Paramartha is sometimes counted among the four great translators in the history of Chinese Buddhism. His oral commentaries on the texts he translated were written down by his disCiples, which distinguishes him from other translators. These commentaries were often far more voluminous than the translations

This is a revised version of Funayama 2005a. The paper, originally in Japanese, was translated by Mr. Benjamin Brose, subsequently reworked by the author. I am grateful for his patient work translating an article with intricate problems. I also want to thank Dr. Michael Radich, Prof. Jonathan A. Silk, Dr. Max Deeg and Mr. Ching Keng for their invaluable suggestions. Since I wrote the original Japanese article, I have organized a seminar called "Shintai sanzo to sono jidai" ~~*=it~..:cO) ai\'{i; ("Paramartha and His Times"), a five-year group study with the collaboration of multiple scholars in different fields of research. This seminar was begun in April 2005 and will end in March 2011, and is being held at the Kyoto daigaku jinbun kagaku kenkyusho J?-:il*~A)Cf-j.~:(jJf JEPff (Institute for Research in Humanities, Kyoto University). However, the result of this group research is not reflected in the present article. Any errors in this paper remain my own responsibility. In this essay T refers to the Taisho shinshii daizokyo *1E*MIf:f*itff; Z refers to the Dainippon zokuzokyo * 13 **,iitff. For example, Z1.34.4, 351d refers to Zokuzokyo first volume, case 34, book 4, folio 351, verso, lower register.
Journal of the International Association of Buddhist Studies Volume 31 Number 1-2 2008 (2010) pp. 141-183


. Funayama Toru

themselves. Unfortunately, almost all of Paramart~a's commentaries have now been lost. However, it is possible to gather some of them from the commentaries of Tang Dynasty scholar-monks such as the disciples of Xuanzang 2:~ and Daoxuan m'". These commentarial fragments, rather than Paramartha's translations, are the primary concern of this paper.l I will not be focusing on aspects of Paramartha's doctrinal or theoretical positions but rather on the basic circumstances under which he expressed those ideas. 2 In particular, I would like to consider the blend' of Indian and Chinese cultures that is evident in the works of Indian scholar monks who immigrated to China. This blend is especially apparent in the works of Paramartha. A better understanding of the specific features of Paramartha's commentaries may further expand our understanding of his translations and as well as his thought. The current knowledge of Paramartha's translation activities was long ago enriched by Ui Hakuju's *.iHs!i': detailed study entitled "Shintai sanzo den no kenkyu" J{:~*=-it{f0)1i3f5i:.3 The present paper generally follows this article. However, as will be discussed below, some amendments and supplements can be made to Ui's study. Before examining the special characteristics of Paramartha's work we should first consider some basic biographical facts.4
1 Aspects of Paramartha's translations have been addressed in Takasaki 1979 and Okada 2002. 2 The characteristics of Paramartha's thought and doctrinal studies have been discussed in a number of studies. Among the most important are the articles entitled "Jilhachi kuron no kenkyu" +}\.@~WHO)1iJf~, "San mu sM ron no kenkyu" ~j!\H~HWHO)1iJf~, "Kenjiki ron no kenkyu" ruiJ~ ~WHO)1iJf~, "Tenjiki ron no kenkyu" ~~~0)1iJf~, and "Ketsujo ron no kenkyu" ~fE~~WHO)1iJf~ in Ui 1965: 131-497; Takasaki 1981; Katsumata 1961, vol. 2, chapter 3, section 2, "Shintai sanzo no shikisetsu" ~~*~~ O)~~Jt; and section 3, "Shintai sanzo no yakusho to Muso ron" ~~*~~ 0)~1f,~2::1!W1~~WH; and Iwata 2004. 3 Vi 1965: 1-130. Incidentally, nearly at the same time, Paul Demieville published an important article for the study of Paramartha, "Sur l'authenticite du Ta tch'eng k'i sin louen": Demieville 1929. 4 Su 1978 and Yoshizu 2003 are two major comprehensive studies which concur with Vi Hakuju's "Shintai sanzo den no kenkyu."

The work of Paramartha


Paramartha's biography in the first fascicle of the Further Biographies of Eminent Monks (Xu gaoseng zhuan is well known, but earlier and later accounts are also relevant. The most important are those found in the following sources:


Huikai :I:'~ (Chen dynasty):5 Preface to the Mahiiyanasaf!lgraha (She dasheng lun ;J1i**~illl; T31, 112b-113b == 152c-153b). Id.: Preface to the Abhidharmakosa (Apidamo jushe shi fun ~iiJm:ilM 1~*'f~~illl; T29, 161ab). Id.: Postface to the *Mahayana-Vijfiaptimatratasiddhi (Dash eng weishi lun houji **Ill~~illlt&gc; T31, 73c). Faqian ~:t;: Postface to the Vajracchedika Prajfiaparamita (Jingang bore boluomijing :31t:JlUJIJ~~t~MHi*; T8, 766bc).6 Fei Changfang Jt~m (Sui ~ dynasty): Lidai sanbao ji ~{i;-=.*c, fascicle 9 (T49, 87c-88b), and fascicle 11 (98c-99a). Yancong ~~ (Sui dynasty): Preface to the Hebu jinguangming jing -a-$:31t:76~* (T16, 359bc). Lidai sanbao ji, fascicle 12 (T49, 105c106a).

5 Huikai (518-568), also known as Zhikai ~'~, was one of Paramar'tha's eminent disciples. His biography is found in the Xu gaoseng zhuan, fascicle 1 (T50, 431b). He was often involved in the completion of Paramartha's commentary in the role of bishou $:5t "scribing" (literally "taking down with the brush"): a person who was ordered to transcribe oral instructions. 6 The author of this postface is not explicitly recorded, but judging from the contents I suspect that the author is Faqian ~:t;. In the postface it states that in the renwu :=Elf year (562), ninth month, twenty-fifth day a translation in one fascicle with a commentary in ten fascicles was completed, and at that time, Faqian had made one hundred copies and had them circulated. The vow appended to the end of the text also reads: "So that all beings may, due to these true words, quickly reach nirvii:fJa and always teach in accord with conditions." Therefore, Faqian was probably also the author of the colophon. Cf. Ui 1965: 26-27. Other studies which also conclude that the author of the postface was Faqian include Yabuki 1933: 78; and Xu 2002: 172. In both studies the authors do not state why they concluded that Faqian was the author of the postface.


. Funayama Toru

Author unknown: Preface to the Suvan:taprabhasa-sutra (Jinguangming jing ~7'tlj,EU') in the Shogozo canon M~]!}iU ' Author unknown: Preface to the Sheng tianwang bore boluomi jing ~ x::E~)t*r&'*i~*' in the Shogozo canons and the Fangshan shi jing

Author unknown: Postscript to the Guangyi famen jing ~~:M~*' (TI,922a).

According to his biographies in the Xu gaaseng zhuan and other sources, Paramartha first arrived in the Nanhai ffi~ 'district of Guangzhou in the twelfth year of the Datong :::kIP] era (546), at the age of forty-eight. This being the case, the new information that he brought to China concerning Indian Buddhism dates from this year at the latest. However, Paramartha had also stopped at Funan :J1:;;ffi before arriving in Guangzhou. lO Therefore, the Indian texts brought by Paramartha probably originated some years earlier than 546. With regard to his name, the eleventh fascicle of the Lidai sanbaa ji states that "the Tripi!aka Master11 Boluomotuo tBlffitt*WE of
7 See Gno 1929; Niryo gakujin 1930; and Niryo sei 1934. All three works are by the same author under different names. I have not yet been able to see the Shogozo M~JJi preface. The above three articles contain photos and transcriptions of the preface, but these are incomplete and must be used with caution. With the exception of Su Gongwang's recapitulation of Ui Hakuju's work (Su 1978), subsequent research on Paramartha and the Jinguangming sutra has not touched on this important information. Especially valuable is the record of the expansion of the Shouliang chapter found in the "Daizo bunko koitsu zenpon mokuroku, I." S See Gno 1988. 9 For this see Fangshang shi jing (Sui-Tang ke jing) mf.lI1'l*'(ll13m~V *,) 2000, Vol. 2: 209. The same text is recorded in the Zhonghua dazang jing r:p~*JJi*' Vol. 8: 109, but note that it is copied by a contemporary person (an anonymous editor?) and contains some errors. 10 For a discussion of Paramartha's place within the history of Buddhism in Funan, see Shizutani 1942: esp. 24. According to Shizutani, during Paramartha's time the king of Funan, Rudravarman (Liutuobamo -w~tJl#:~, ca. 514-550), favored Buddhism. 11 "The Tripitaka Master" is called tripita or trepitaka in (Buddhist Hybrid) Sanskrit. Forte 1990: 247f. n. 7.

The work of Paramartha


Youchanni {ifm~Ft:. (Skt. UjjayinI; Pali UjjenI; Pkt. UjenI) in Western .India was known as Paramartha during the Liang" (T49, 99a).12 It is certain that Boluomotuo corresponds to the Sanskrit name of Paramartha, or Paramattha in Pali. Also, in Huikai's Preface to the MahiiyiinaSaf!lgraha, it says that "The Tripitaka Master was from a brahmin family in the kingdom of Youchanni. His gatra-name was Poluoduo Brn:~~ (Bharadvaja or Bharadvaja)13 and his personal name was Juluonata :J'RJ~JJ~{m (Kulanatha), which is translated as 'Qinyi' m~ in this land" (T31, 112c=52c).14

12 UjjayinI was the capital of the ancient country called Avant! (present Ujjain). Incidentally, in historical records dating from the time of Paramartha, Youchanni/UjjayinI was sometimes classified as western India (as in Paramartha's biography), and other times as central India (as in the biography of Yueposhouna Jl ~1tjj~in the Xu gaoseng zhuan TSO, 430b as well as in the Preface to the Shengtian bore jing in the Sh6goz6 canon). Hence such designations were not always consistent. 13 I prefer the form Bharadvaja which signifies 'Bharadvaja's descendant.' But in Sanskrit texts, Bh~radvaja is sometimes written as Bh.!!radvaja in the same sense. Demieville (1929: 16) also indicates the form Bharadvaja. Ui (1965: 9) assumes the Sanskrit equivalent ofPoluoduo ~JUi!i! to be "either Bharata or Bharata," but this is incorrect. A typical example of Poluoduo as the transcription of Bharadvaja is Bintoulu Poluoduo film t!~JHi!i! (Pil).<;lola-Bharadvaja). . 14 The Xu gaoseng zhuan and later sources often wrongly use the form "lunaluotuo" #Jjj~~~t. This version of the name comes from a metathesis of the sounds la (~) and na (jj~). Huikai's usage of luluonata #J ~UMtl1 (he also uses tuo ~t) is certainly correct. Since Huikai was a direct disciple of Paramartha and participated in his translation activities, his record is the most reliable. Huikai's own biography, under the name of Zhikai ~'i:~, is attached to Fatai's r:t~ biography in the first fascicle of the Gaoseng zhuan (TSO, 43lb). His family name was Cao . The biography of Zhikai is followed by another biography, which states wrote a lost biography of Paramartha that a layman named Cao Pi called Sanzang lizhuan =Hi~1J (also called "Cao Pi bieli" .m;5iJ~ [A Separate Biography of Paramartha compiled by eao Pi] which is mentioned in the biography of Paramartha [TSO, 430b]) on which the biography for Paramartha Sanzang in the Lidai sanbao ji was based (T49, 88a; cf. 99a). The author is called a (lay) disciple who received bodhisattva



. Funayama Toru

Regarding Paramartha's school-affiliation, it has already been pointed out that he probably belonged to the Sarpmitlya (Zhengliang IE;;;:) school. 15 This assumption is based on the fact that Paramartha translated the Lit ershier mingliao lun f=!t - -1jFj T~ilil, the author of which was Buddhatrata (Fotuoduoluo 1~~t~Hi) of the SaipmitIya school (T24, 665b).16 Moreover, it has also been noted that the terminology found in the Mingliao tun, such as ren II, (acceptance; Skt. k~anti), ming 15 (name), xiang f (characteristic), and shi diyi fa i:lt~-1~ (the highest worldly elements; Skt. laukikagradharma) are also used in other branches of the Vatslputrlya (Duzi school to which the SarpmitIya belongedY The terms ren, ming, xiang and shi diyi fa correspond to the Sarvastivada's nuan ~ (the heated; Skt. u~ma- / u~magata), ding TJ: (summit; Skt. murdhan), ren and shi diyi fa which are also called shun jueze fen JI~bI<!j"::S (aids to penetration; Skt. nirvedhabhagfya). In Paramartha's case, the usage of terms such as ren, ming, xiang and shi diyi fa was not limited to the Mingliao lun but also occurred in the Bu zhi yi lun :tf~ )L~~ffa and the Xianshi lun ~j~~ffa.18 However, Paramartha was not exclusively connected with the SarpmitIya school. If we consider his theoretical views, the Sarvastivada and Yogacara schools also played an important role. The well-known fact that throughout his life he devoted


precepts (pusajie dizi i'ilil1X:m-r-) under Paramartha's supervision and . listed as a son of Huikai's uncle. That is, Cao Pi and Huikai were paternal cousins. 15 In Paramartha's case, the expression "Sa~mitlya" ("rna) is preferable to "Sa~matIya" -(man) because in the Eu zhi yi fun :g:~~~~ffii, the school is called "Zhengliang dizi bu" IE~m-r-:g:~ (T49, 20b13) and "Sanmeidiyu bu" = JlJt~$ (T49, 22c14). 16 Concerning Buddhatrata, the colophon to this text further states that he was a saint who had attained the third stage (i.e., aniigilmin) of the Sravakayana practice (T24, 672c). 17 See Ui 1965: 395; and Namikawa 2000, especially from page 189. See also Namikawa 1995. 18 Further, the definition of arm;ya as translated below in Section 2 reveals that Paramartha took the Sa~mitIya view as his own.

The work of Paramartha


himself to the translation and explication of Vasubandhu's Abhi'dharmakosa(bhii:jya) and AsaIiga's Mahiiyiinasaf(lgraha as well as Vasubandhu's commentary thereon should not be overlooked. These works demonstrate that Paramartha was closely related to Vasubandhu and also possibly to other commentators of his era such as Dignaga (Chenna ~-*JJG, ca. 480-540). Paramartha translated two of Dignaga's works: the Wuxiang si chen lun ~t}l!;,~~illI in one fascicle (T1619, AZambanaparlk:jii) and the lie juan lun f~m ~,also in one fascicle (T1620, *Hastaviilaprakaraf,la).19 Although Dignaga's theoretical position can probably be said to be that of the syncretic faction of the Yogacara and Sautrantika schools, he also wrote an outline (i.e, the Marmapradfpa) to the Abhidharmakosa which makes it clear that he valued that text. It is noteworthy that Dignaga had a close connection with the Abhidharmakosa of the Sarvastivi'ida, although he was probably ordained by a master of the Vatslputrlya school and did not belong to the Sarvastivada. 20 Dignaga's school-affiliation is instructive when reflecting on the same issues in Paramartha's life. We .should consider the possibility that in India during the fifth and sixth centuries several commentators belonged to schools other than the Sarvastivi'ida school and nevertheless were skilled in the Abhidharmakosa. We can say that Dignaga and Paramartha shared similar positions in that both of them made much of the Abhidharmakosa and Yogacara thought as masters from schools other than Sarvastivada. The Sarp.mitIya is generally considered to be one of the four branch schools stemming from the Vatslputrlya.

19 For the Hastaviilaprakaral}a and the Tibetan translations, see Frauwallner 1959: 127-129, and 152-156; and Nagasawa 1978a and 1978b. The treatises Paramartha brought to China include texts by Dignaga (ca. 480-540). Most probably it was thus Paramartha who first introduced this most recent Indian literature to China. On the chronological relationship bet:'leen Paramartha and Dignaga see Hattori 1961: esp. 84-85. 20 Obermiller 1932: 149. Chimpa/Chattopadhyaya 1990: 181. Frauwallner 1969: 390.


Funayama Toru

1. Paramartha's compositions

This paper is primarily concerned with Paramartha's original compositions. I would like to begin with the following information recorded in the seventh fascicle of the Kaiyuan Shijiao lu I*J::JC*~

[In addition to the texts translated by Paramartha] the Changfang [lu] *m[~], Neidian [Iu] P'lA[~] and so on also refer to a list of [Paramartha's] texts which begins with the Zhenglun shi'yi lUfll~~, amounting to thirteen texts in 108 fascicles. [In the present catalogue] these treatises on sutras as well as commentaries are not listed [among the list of translations] because they are Paramartha's compositions, and not translations from Sanskrit. 21

When we compare this passage's reference to "thirteen texts" with the information given in the Lidai sanbao ji (T49, 88a) and the Datang neidian lu *mP'lA~ (T55, 273c), we can see that it probably refers to the following thirteen texts:
(1) Explication of the Authentic Treatises (Zhenglun shiyi JE~ifll~~), in five fascicles. (2) On Buddha Nature (Foxing yi (~'rt~), in three fascicles. (3) On Meditation (Chanding yi ~,-JE~), in one fascicle. (4) Commentary on the Abhidharmakosa[bhii~ya] (Jushe lun shu {Jl;.'j~Hifll iE,itt), in sixty fascicles (or fifty-three fascicles according to Huikai's Preface to the Abhidharmakosabhii~ya). (5) Commentary on the Vajracchedika Prajfitiptiramitti (Jingang bore shu ~Jllll'M5tlfiE,itt), in eleven fascicles (ten fascicles of commentary and one of the sutra). (6) Commentary on [the Distinction between] the Eighteen Nikiiyas (Shiba bu lun shu + )\. 'ifB~iflliE,itt), in ten fascicles. (7) Commentary on the SaJ?1dhinirmocana-sutra (Jiejie jing shu ~~l!M~ iE,itt), in four fascicles. (8) Commentary on the *Anuttariisraya-sutra (Wushangyijing shu ~l: fj\;J~iE,itt), in four fascicles.

21 x*mP'lA~~, w~JE~ifllf~~~-+='ifB-B)\..{f;, 4.\;I-frft1H~~ ~iE,itt, }[~*BTfJ, ?F1t*l!m, ~f!Ij!J;;::~. (T55, 546c).

The work of Paramiirtha


(9) Commentary on the *TarkasZistra (Rushi lun shu FO'Jrmleiffit), in three fascicles. 22 (10) Commentary on the Catur[ -Zirya-]satya-sastra (Si di lun shu IZ.9 ~*mle iffit), in three fascicles. (11) Commentary on the Refutation of the Atmavada [in the Abhidharmakosa IX; viz., Po wo lun shu 1iErR~~iffit], in one fascicle. (12) Commentary on the Theory of Sixteen Truths found in the Suixiang lun (Sui xiang lun zhong shiliu di shu illIt~iffir:p+A~*iffit), in one fascicle (extant as the Sui xiang lun illIt~iffi, T1641). (13) The Opening Set Passages common to all Siltras (Zhong jing tong xu ~*~@Ff:), in two fascicles.

Most of these texts are no longer extant but a few details are . known. First, fragments of the (4) lushe lun shu exist in Puguang's ~7t lushe lunji 1~*~ila~c. That is to say, Puguang's text contains anumber of Paramartha's statements and it can be assumed that these were quoted from Paramartha's original work. Similarly, (5) the lingang bore shu, (6) Shiba bu lun shu (also known as the 13u zhi [lun) shu 'ff~~[~ila]ifJIE or the Buzhi [lun} ji 'ff~.A[~ila]~E), and (7) the liejie jing shu are no longer extant, but can be partially reconstructed from the fragments cited in the works of Sui and Tang Dynasty commentators. 23 The high probability that (12) the Suixiang lun zhong shiliu di shu is the same as the Suixiang lun of the Taish6 canon has already been established in a previous study. 24 It is likely that some other texts can also be attributed to Paramartha. For example:
(14) Translation of Foreign Words (Fan waiguo yu JJl;9}~1l!), in seven fascicles (also known as Za shi or the lushe lun yinyuan shi {~ ~mlelEJ~*, T49, 88a). - This treatise is listed in the Lidai sanbao ji and the Neidian lu as the last one in the list of Paramartha's works,


22 I tentatively follow Giuseppe Tucci's reconstruction .of the title as "Tarkasastra" in Tucci 1929. . 23 An earlier important study of the Bu zhi lun shu is Demieville 1931. 24 Ui 1965: 96-97; Aohara 1993 and 2003. For a study which does not hold that the Suixiang lun contains Paramartha's commentary, see Yoshizu 2003: 241.' However, I am not fully convinced by Yoshizu's claim.


Funayama Toru

immediately following the above-mentioned (13) Zhong jing tong xu. We know from its name that it was probably not a translation but a composition by Paramiirtha. (15) Commentary on Lucid Explanations Concerning the Vinaya (Mingliao lun shu ~ Tfnj;i9fE) or Commentary on Twenty-two' Lucid Explanations on the Vinaya (Lii ershier ming liao lun shu 1$:=+ =Il Tfnj;i9fE), in five fascicles. - The translation of the Lii ershier ming liao lun has a postscript (T24, 673c) in which it is stated that the translation was completed in the second year of the G,uangda ** era of the Chen (568), and that at the same time "a five fascicle commentary was made." The same postscript further states that Huikai who belonged to Ayuwang Temple ~iiJ1f::E~ in Jiankang at that time was in charge of bishou.~ in the translation. 25 It is possible to recover a large portion of this commentary from the quotations found in later texts such as Dingbin's J'tfi Sifen Iii shu shizong yi ji 1Z9~1$:~!fIi*~~ and Dajue's *Jl: Sifen Iii chaopi 1Z9~1$:~) ttl: (both of which were composed in the beginning of the eighth century). (16) Commentary on the Mahliyiinasa7!lgraha (She dasheng lun yi shu:fi **fnj;~i9fE), in eight fascicles. - This text isknownfrom Huikai's Preface to the Mahiiyiinasa7!lgraha. The lost fragments have been collected by Ui.26
(17) Commentary on the Suvan:taprabhlisa-sutra (Jinguangming Uing] shu 1ik*ijJ:H*~]i9fE), in thirteen fascicles. - There is a seven fascicle version of the Jinguangming translated by Paramartha and also a thirteen fascicle commentary. A portion of the lost text can be reconstructed. (18) Commentary on the Renwang borejing (Renwang bore Uing] shu 1= ::E~~~[~]i9fE), in six fascicles. - Paramartha's lost Renwang bore shu can be reconstructed from the works of Jizangi5~t Zhiyi ~jj, and Yuance (Wonchuk) IIIn!U. It is clear that Paramartha's commentary was based on the Renwang bore jing translated by Kumlirajlva. It is significant that the Indian monk Paramartha would (perhaps at ' the request of a Chinese monk) write a commentary to an apocryphal text composed in China. There is a range of opinions regarding this point which will be discussed below.


26 Ui

For bishou see n. 5 above. 1935.

The work of Paramartha


(19) Exposition on the Ninefold Cognition (Jiu shi lun yiji :fL~~~~), in two .fascicles, or alternatively, Thesis on the Ninefold Cognition (Jiu shi zhang :fL~1i!:), in three fascicles. - These texts contained Paramartha's advocacy of the ninefold consciousness as opposed to the more common eightfold consciousness of the Yogaciita school. Unfortunately, only a few fragments remain. 27 (20) Exposition on the Turning of the Dharma Wheel (Zhuanfalun yiji. ~~~~), in one fascicle. Details unknown. (21) Commentary on the Madhyantavibhaga (Zhong bian lfenbie lun] shu q:r3i[~5JIj~]~). Three fascicles. Details unknown.

:22) Commentary on the Vi1']1iatika (Dasheng weishi lun zhuji **JJt~ ~r.~i:l). Details unknown. :23) Biography of [the Buddhist Master of the LaW] Vasubandhu (Posoupandoufashi zhuan ~fi~R~~iiJ1i TI049), in one fascicle. - This is traditionally taken to be a translation, but an examination of its contents reveals elements which deviate from pure translation and suggest that it represents Paramartha's commentary or a mixture of commentary and translation. 28 This will be discussed further below. '(24) Xianshi lun U~~, in one fascicle (T1618). - This text is also tra'. ditionally believed to have been translated by Parami'irtha but its contents make it doubtful that it was a work of pure translation. The text has features of a commentary on the MahayanasaTJ'lgraha. Therefore, there is a high probability that it is a record of one of Paramiirtha's lecture series on the MahayanasaTJ'lgraha. (25) Treatise on Buddha Nature (Foxing lun #1l't1:~), in four fascicles. .. - As has already been pointed out in a previous stuqy, the Foxing lun is closely related to the Ratonagotravibhaga (Baoxing lun .'11: ~).29 Although the contents of the two texts are partially the same, there are also a number of significant differences. For example, the Foxing lun has repeated expressions such as "The commentary sta-

~i:27 Recently, there has been some doubt regarding Paramartha's author'fship of the text; see Yoshimura 2002 and 2003. It seems to me, however, ~thatthe authorship of this text is still an open question. l;' 28 Takakusu 1904: 293 n. 110, Takakusu 1905: 38, and Frauwallner ~1951: 17-18. { 29 Tsukinowa 1971 and Hattori 1955.


Funayama Toru

tes ... " (shi yue *~ El) and "The record (i.e., commentary) states ... " (ji yue ~[,El) in various contexts. According to Sakamoto Yukio l:Ei: *$515, these are Paramartha's commentaries. 30 (26) Notes on the Seven Items (Qi shi ji t*~[,). - This text is not mentioned in the records of works attributed to Paramarfua but it is cited in Yuance's commentary on the Sa,!1dhinirmocana-siitra. The "seven items" refers to the standard seven words stated at the beginning of Buddhist siitras, namely "thus" - "I" - "have heard" - "at one time" - "the Buddha, World-honored one" - "was staying at such-and-such place" - "together with great bhik~us." It seems that the Qi shiji gave a detailed explanation of these opening lines. 31

Thus we can list at least twenty-six works of Paramartha. It is likely that there are some more texts which have yet to be identified as Paramartha's compositions. 32 2. Characteristics of Paramartha's commentarial method Paramartha was one of India's eminent scholar-monks and many of his compositions naturally reflect an orthodox translation style.
30 In the Foxing lun, "The commentary states" (shi yue ~El) and "The record (i.e., commentary) states" (~[,El ji yue) appear seventeen times. See Sakamoto 1935: 264-267. For further discussions of this issue see also Takasaki 2005: 61-63. 31 Judging from the quotations, this text appears to be closely related to the Jingang bore shu. It is possible that this was simply another name for the beginning section of the Jingang bore shu. At the same time, from various citations of the name Qi shi ji, it could be that the original first portion of the Jingang bore shu was later circulated independently as an extended commentary on the beginning section of siitras. On the Qi ski ji, see Vi 1965: 85; and Funayama 2002: 28 n. 4l. 32 The Zhonglun shu ~~iiBi9iE (Commentary on the Madhyamakasiistra) referred to in the Lidai sanbaoji 11 (T49, 99a) might have been the work of Paramartha. Generally speaking it can be assumed that in the list of Paramartha's works the word shu i9'lt for a commentatorial work (in contrast to the word shi *~) suggests that it is not a translation. lowe this suggestion to Dr. Otake Susumu in personal communication. See also Imazu 1925: esp. 79. I am indebted also to Dr. Otake for this reference.

The work of Paramartha


However, Paramartha also had his own unique style, compared "to other commentators. In what follows, I would like to point out some examples of Paramartha's commentarial method. Revealing the multiple meanings within a single phrase One of the identifiable characteristics of Paramartha's commentarial style is his frequent listing and explanation of the various meanings present within a single phrase. For example:
In Paramartha's commentary it says: The term aralJya (alianruo ~iiJ**

*) has ~Pf~~. m~.~~ings.. T1:t~ .:!l:J;~.t. [Ip.~.~~i~gJ is 'a place far from noise
(ralJa).' That is to say, a place where the sounds of large cities do not reach. T1:t~.~.~~.9p.~.[m~.~~ingl is 'a place far from deforested areas.' That is to say, a place where people do not go to collect firewood. And ~p~.H*~.rm~.a~ingJ is 'a place far from conflict (ralJa).' By 'a place of conflict' is meant a place where defilements disrupt good actions; those who live in such places will fall prey to defilements. For this reason, an aralJya is called 'a place far from conflict.' Places that are from one krosa up to a hundred or a thousand yojanas away [from noise, deforested areas, or conflict] can be called an aralJya. According to the Sarvastivada school's interpretation, one krosa is five-hundred dhanu (gong ~).33 According to the Sammitlya's interpretation, on the other hand, one krosa is equal to one thousand dhanu. Since one dhanu is equal to eight chi R, altogether it is a place eight-hundred zhang Jt34 . distant. Based on the measurements qf.thi~)?l:1!.4.n:~:).~.1:t~p.~], it would be a little over four li35 [from areas of disturbance].36
33 This idea is defined in the Abhidharmakosa III 87cd and the Bhii~ya ,thereon. 34 1 zhang = 10 chi; therefore, 800 zhang is 8,000 chi = 1,000 dhanu= 1 krosa. 35 More accurately, it is 4.444... li.

~nf~, ~1ll~EJf~BT::f~MI: ..~.*-liltr6JT{::Z~, ~1llf*irBT*~$::-:~i~:~it

l;~*hi, ~1ll--!w~Jjj1U!~J~llbjtJL~1*, 1l~lm~*. *{1jt~, ~1{ftJJH~, ~~, -#JI'@?:n:s~. {t(lE;li~~A~, -#JI'@?fL-=f~ill. -~)\R, fLAs Jt!. *{1t.J!:~.~F!" ff!P.l<:II9!E!.:Pllt. (Yuance's lie shenmi jing shu, third fascicle, Z.l.1.34.4, 351b)

36 The text runs as follows: JJ;~* ((~)) :13;-, ~iiJ***1lf, 1'f'::'~.




7t~n~s=fIi11U, ~1l~iiJ***~. *ili~$-~~



'Funayama Toru

Here Paramartha analyzes the Sanskrit term aralJya (P. arafifia), a quiet, forested place of practice, as a-ralJaY 11' is a negative prefix and 'ra1Ja' is defined in the three ways given above. This demonstrates Paramartha's commentarial style of revealing the ~ultiple meanings inherent within a single word. There are also other examples. According to Paramartha, "sons of the Buddha" has five meanings,38 "at that time" has eleven meanings,39 "spiritual powers" has three meanings,40 and "great" also has three meanings. 41 Paramartha frequently employed formulas for listing explanations: "such and such has '""" meanings," "such and such has '""" types," or "such and such itself has '""" meanings." This is not to say that this technique was unique to Paramartha. Rather, it was a general characteristic widely employed by Indian commentators, but it is nonethelesss strongly represented especially in Paramartha's work. It is noteworthy that Paramartha's conversion from Indian to Chinese metrology was based on the view of the SarpmitIya school, and not of the Sarvastivada. This would be possible only if his main standpoint was the SarpmitIya. Furthermore, in the passage cited above, Paramartha points out the diverging interpretations of the length of one krosa within the Sarvastivada school and the SarpmitIya school. At the end of some passages, when comparing Indian and Chinese meanings, Paramartha frequently uses the term "here [in China]" or "in this place" (cijian JltF"m to explain the equivalent Chinese measures.
37 For a-ralJ-a see Edgerton 1953: 64, "a-ralJ-a," q.v., where it is explained that ralJ-a can signify klda (defilement). 38 Paramartha's commentary is quoted in the first fascicle of Yuance's lie shenmijing shu ~~mgW~ifit, (Z1.34.4, 317c; cf: 324a). It may also have been quoted from Paramartha's liejie jing shu. Furthermore, an analogous commentary which is not listed as the original work of Paramartha, but rather as one of his translations can be found in the She dasheng [un yi 1jlld\:*~~ (Vasubandhu), eighth fascicle (T31, 306b). 39 Yuance's lie shenmi jing shu, third fascicle (Z1.34.4, 349a). 40 Ibid., second fascicle (Z1.34.4, 334a). 41 Ibid., first fascicle (Z1.34.4, 317a); Jizang's ~it Fahua yishu t~ ~ifit, first fascicle (T34, 457c).

The work of Paramartha


Among the Indian monks who came to China, this way of explication was unique to those who were skilled in the different views of various schools, and Paramartha was a typical example of such monks.

Interpretation ojthe meanings ojproper nouns

Another unique characteristic of Paramartha's work is his explanation of proper nouns. Two passages exemplify this approach. The first discusses the origin of the name Mahakasyapa: In the Shiba bu [un shu +Ag:~~illijlfrt42 it says: Correctly speaking, for Jiashe ~~ we should say Jiashebo ~~l'BZ (*Kasapa / Kasapa?; P. Kassapa; Skt. Kasyapa). R~.r:~ [in China], jiashe means 'light' (kasa) and bo means 'to drink' (-{po) Taken together, they mean 'drinker of light.' 'Drinker [of Light]' is a surname. There was an ancient ascetic (*r0'i) called 'Drinker of Light.' He had a luminous body and was able to drink various types of light and make them invisible. The present Jiashe belongs to a clan of this light drinking ascetic and therefore has the surname 'Drinker of Light.' His name was derived from his surname and so he was called 'Drinker of Light.'43 This explanation is the same as Paramartha's free translation of the Kasyaplya (Jiashewei imI~*1t) school as Yinguang bu ~Yt'if~ (literally "drinking light school"), but in the above passage his explanation is more detailed. A second example is found in an explanation of the origin of the name "Mulian" IJf: The Tripitaka Master Paramartha said: Correctly speaking, for [Mulian] we should say Wujialuo :?v~~ (*Mudgala?; ct. P. Moggallana, Skt. Maudgalyayana). R~.r:~ [in China], wujia is called

42 Also known as the Buzhi [un shu :~mW\~iffrt.

43 The text runs as follows:





~t Jl:tj3;-7t. tEl, Jl:tj3;-~X. -g- i Z, 1ifz:j3;-~7t. ~ C7tJ ~Aft!E. J::r!i{!-1J;\' ~~7t. YJl:t{!-1J;\., ~1f7tIj!j, fj~~~7t, ~::f1;im. 4J1:t~~, ~~7t{!-1J

A1I, P!PY~7t~ft!E, qfl':ft!E3'z:, ~~7tili. fascicle, T34, 459b).

(Jizang's Fahua yi shu, first


Funayama TOfU

'western bean' (hudou iiJlli), which is a green-colored bean44 and luo means 'to receive' (-{la) here. Taken together, they have the meaning of 'receiving western beans.' Probably this surname comes from an ancient ascetic (*r~i) called Wujialuo who only ate these beans and no other food. Therefore, he was named 'Receiver of Western Beans.' [Mulian] belongs to his clan and hence he has this name. 45

These explanations of people's names are not generally found among the explications of other Indian commentators, but they do exist in the form of fragments of Paramartha's works which are still preserved in the Buddhist canon. This suggests that this style of explanation was unique to Paramartha. It is quite possible that Paramartha's explanation here is based on his knowledge of the nirvacana tradition. 46

44 Maudga > mudga (Pali mugga). These green colored beans probably correspond to modern "mung" beans. 45 The text runs as follows: ~~'i'if=ii:iS;, ~~{bo~. ~{bo1!f, J1:t

1'f{UJA~~{bO~, /Git:.:.:.-wJ~, PfUtJl:tli, ~q;5tii)jli. A~{UJA*i, ~

Y~~mo (Jizang's Fahua yishu, first fascicle, T34, 459c). Cf. Kuiji's Amituo jing shu JliiJ5i~t*~fiE: ~~~*iS;, ~~~{:l:m~.J1:tiS;~(5t)t)jR.~ (5t)ii)jliE!p*lim.J:tl1'f{UJArIl1tJl:tli,~1&:{UJ~,lZ9fr,i~~ (T37, 315c). Furthermore, in the original text of the same commentary "~" is erroneously given for "5t." Also, Paramartha's interpretation of 'Mulian' is found in the sixth fascicle of the Sanlun xuanyi jianyou ji =~3Z~~1iliI :* by Chiigan ChOzen 9=JilMU~ (T70, 465bc). It is clear from the context that the passage is quoted from the 'Chaopi' fy:jJI:, namely the Sifen Iii chaopi 1m:S1*i'):j:ll: by commentator Dajue:k:Jt (fl. ca. the beginning of the eighth century) in the Tang. However, the passage is not found in the extant version of the Sifen iii chaopi. Therefore it is highly probable that the extant version is different from the Chaopi consulted by Dajue. 46 See, for example, the explanations of "kasyapa," q.v. in Deeg 1995: 328 (also 425), "ara1Jya," q.v. in ibid.: 362 (also 422) and "kacchapa," q.v. (cf. ktiSyapa) in Kahrs 1998: 142. Note, however, that these explanations are not the same as Paramartha's.

+815" J,>.1.J;f~-e:.R. ff$, JJ..-LJS.x A=-'h.-=- )}{!l:.p:JZPJ]li . ..tttL}E/,j(:t, Bnv.~h. 'ifIi' "l-- - 2'f;\i[-=-2-+8- *,Fl_1t-hfl!=I/J..lL-, t=J IllJ~i=i,


The work of Paramartha


Comparing India and China

Paramartha, was an Indian who had gone to China, and some of his comments comparing India and China have been preserved. The following statement about seasons is one example:
Dharma Master Paramartha declared that there are three seasons, as follows: The four hot months [in India] span from the sixteenth day of the first month to the fifteenth day of the fifth month ~n., th~~ )::tn.~H~.,Y" ~l:I.i,I,l~]. The four rainy months span from the sixteenth day of the fifth month to the fifteenth day of the ninth month. The four cold months span from the sixteenth day of the ninth month to the fifteenth day of the first month. From the ninth day of the later half of the second month of the rainy season the nights gradually grow longer. JI,l..t.4~:s. p.~~<?~,. this [begins on] the ninth day of the seventh month. From the ninth day of the later haIf of the fourth month of the cold season the nights gradually become shorter. J.J.l..~l:I.~~, p.~~<?~, this [begins on] the ninth day of the first month. 47

,Descriptions of the seasonal divisions of the year in other texts ',' such as the Datang Xiyu ji differ from those given by Paramartha. 48 'However, the passage cited above provides a concrete description of the months and days which mark the three seasonal divisions of the year in China (referred to above as "this place"), In the "Yiyi" -"" chapter of the liejie jing, i.e., SaY(ldhinirmocana-siltra, there is a reference to a musical instrument pina m~.

47 The text runs as follows: x~~*:tjji:lz:':::'~;z3;, 11fJl:tFFl,lEJj +1\ S, ~ :liJj+3iS, ~w'\If~/2:9Jj.11f3iJj+1\S, ~1LJj+3iS:'j'ffi~/2:9Jj.11f1L Jl+1\S, ~lEJj+3iS, ~~If~/2:9Jj. ~~~=jH&"-~1LS~l$i~, -&Jl:t f!'!',tJj 11 S. ~~~/2:9Jj 1&"~1L S ~l$i~, -&ft.:~.I!-:',lEJj 1L S. (Puguang',; :'1*71:; lushe lunji {~*~iliI~c, eleventh fascicle, T41, 188a).

48 For an introduction to the six yearly divisions, see the seviii~~mmm~?,j> enth fascicle of the Sapoduo pini piposha (*Sarviistiviidavinayavibhii~ii T23, 547c; translator unknown). For reference to the theory of three divisions, see the second fascicle of Daoshi's iit!!: Pini taoyao mmit~ (Z1.70.2, 134b).


. Funayama Toru
vl~a. 49

It is a translation of Skt.

Paramartha explains the word in

the following way:

In Paramartha's note it says: A pina is a musical instrument. It approximately resembles the piba me:. ):).~r~ ..50

Historically speaking, the creation of the Chinese piba (or pipa) was partially influenced by the Western Regions. This sort of information is rarely found in other Buddhist texts.

Comparing the theories of various schools

We have already seen how Paramartha's commentaries include elucidations of the various meanings inherent in individual phrases. In a similar way, Paramartha sometimes explained a given point from the perspective of different schools. His comment on the robe colors of Indian monks is one example. A monk's robes in India, called ka:jaya or ka:jaya 'deteriorated clothes,' had to be neither . new nor of a pure color. Paramartha described how monks' robes. were altered to meet this requirement. The following is the Sui master Jizang's 51 statement:
Tripitaka Master Paramartha said: "The ka~iiya of ~<?r~igI.1. ~~n9..~ H..~" ~.I.14ifll are of blood-red color (crimson). Although [the robes] of the five schools51 are different, they are all red." Question: It is often said that robes are of three deteriorated colors. Why do you say that they are all red? Answer: It is usually explained that new robes are first stained blue, then they are soaked in mud, and next they are soaked in the sap of the Mulan (magnolia) tree. Therefore they can be called either blue, mud[-colored], or Mulan[-colored]. Tripitaka Master' [Paramartha] said: "This method is not used in the M.i.44~Y.~ingq.~m

49 The word pina is found in T16, 713b25-26. Its Tibetan equivalent is pi bang. For the Tibetan translation of the Sa1?'ldhinirmocana-siitra see. Lamotte 1935: 46, chap. 3, 6, 11. 4-7.

The text runs as follows: ~~ ~e. ii;, m~*~lf~~. J!:~.I!!,me:., (Yuance's lie shenmi jing shu, fascicle two, interlinear note, Z1.34.4, 347b). 51 For the notion of "the five schools" (wubu 1l.tfi3), see Funayama 2007: esp. 86-89.


The work of Paramartha

[i:~ . ,Jr:t.4~ilJ.. The


three types of deterioration means that [monks' robes should] be stained by using one of three colors. They are stained blue if blue dye is available in the place. If no blue is available there, then the robes are stained with mud. If mud is not available there, one can grind iron to make a liquid [so that the robes] are stained. If one of these colors can be obtained it is sufficient. The colors will, however, vary according to differences of time and place. Because there is concern that bhik~us will have doubts and regrets, it is said that one [of these] colors should be used. Although the doctrines of the eighteen schools are different, the color of their robes is the same. Therefore the Great (NirvQ1}a) Sutra says: '[Those who] see my disciples wearing crimson robes say that [the robes] are [the color of] blood.'52 But since the method of staining is not the same, there are differences among the various schools. [For example], the Sarvastivada school stains the visible areas [of their robes]; the Sthaviravada (Theraviida) school stains the seams [of their robes]; and the Sammitiya school stains the four corners [of their robes]."53

Different versions of the above explanation can be found in the lost fragments preserved in the fourteenth fascicle of the Xuan ying yin yi ~~if~ (also known as Hui lin yin yi ~$if~, fascicle fiftyrune; T54, 699a), and Daoxuan's Jiemo shu (in Sifen lujiemo shuji yuanji 1Z9:5:H~mJliEJlEf'A'~~c., fascicle eighteen; Z1.64.5, 459b) and elsewhere. These fragments are similar in that they all preserve Paramartha's explanations to a Chinese audience regarding the color of monks' robes in India. According to Paramartha, although
52 Cf. TI2, 457b; 699b. Although the reference to the MahaparinirvQ1}asatra is included in Paramartha's statement, we cannot exclude the possibility that it was Jizang's addition. 53 The text runs as follows: ~~*.=.~~, ~Ifn.i3~, Jj-~~~, ifHi3i

l'fIPFIP], IP]~~B. F",~. ~~'='flti3. ~{PJ~JLJ1:~B. ~."1ii~~~, fJT~iW 1&.~, j'j;:J!IjAr~, j'j;:fiJH~z, ~~*)ji, ~~~.~r~~*lIl. .=.~~. rJH~:.!f.~A, il~lJUt. ~'='fltB1f, =i3zi:f:1, ~ffl-i3, YI.!i~FIJ;z. ~ 1f.~, J!ljffl.I.!i~. ~~1f.~, fflr~~I.!i~. ~r~~, RJM~i'tl.!i~z, Jill:1EL R!n&-13i!jE, {EL~aif~1-~, -B::f'r~, i5~!tli::i:;/J~~'I19f, ~ ~;/J~ =fl
.~1&-B. +}\Tm~',Ut~, ~B~-. ~ :;kJ ~, JVfJt3f1T~~B-t<, ~Ill 1lIfn..1ELI.!i~::fIPl~, 1f~Tm~~. ~iil~~l'f~, !ti*Jf~~. J::~l'f~J!IjWW ~!~. ~lE.ml'f~, {ELl.!iti[9ftJ-t!1. (Jizang's lingang bore jing yi shu, second

fascicle, T33, 97bc).


. Funayama TOfU

the staining methods of various schools were different, their robes were all considered to be red. 54 The next section discusses the positions of different schools regarding the number of teachings, 80,000 or 84,000.
Master Paramartha said: Question: What does it mean to say that among the eight thousand teachings there is a single position regarding things such as the five skandhas? According to the Sthaviravada (Theravada) school, there are 84,000 teachings, while according to the Sammitlya school, there are only 80,000. Answer: In terms of the six types of dharmas, all teachings interpret the meaning in the same way.... 55

We should be careful to note that the style here is roughly the same as a few other of Paramartha's works which have been handed down as "translations." For example, in the Xianshi tun, the following commentary comparing schools is given:
[Regarding the iilayavijniina of the Yogacara school,] among the teachings of the Lesser Vehicle, the SammitIya school calls it *avipranasa [wushi "without expiry"], using the analogy of a 'written contract.' '" The Mahiisamghika school calls it sheshi m~56 The Sarvastivada school calls it *samanviigcltapriipti 57 and the Sthaviravada (Theraviida) school calls it *bhaviirigavijniina . ... 58

... .

Furthermore, a discussion of various schools can also be seen in the Sui xiang tun. 59

54 55

On staining, see Hirakawa 1994: 606-616; and SatD 1963: 683-690. The text runs as follows: .~*Bijl:iS.", F,,~. Jlt3ig~)\"~1ti:f~1~-~~,


jti':iS."1iiJ. tf{JtJ:.~(~)lm, fo(lj~)\"~[)]TM~. 4{JtJE.m:g:~, f..~)\..~. ffj:Afirti:i', ~-~~ ... , (Yuance's lie shenmi jing shu, third fasci-

cle, Z1.34.4, 352c) 56 The Skt. equivalent is unknown. 57 The original Sanskrit term is not clear; it could also be samanviigatii priipti/:!. 58 The text runs as follows: 1f!H~~, JE:I;~~~~~~, :tzo#t*1 ....
~~{~'I'tiiJlm4S~ti~ ... , ii1~?t~~~ 1RJ1IlI1~.... {illm~lm~~)J-~~ .. ,.

(T31, 880c-881a). 59 Cf. tft&ii1~?t~, ~1RJ1IlI1~~ZM~, ~!i!fHMIRJ1IlI1~~, jt{tE~

The work of Paramiirtha


;pse of Chinese rather than Indian generic names in examples

~iri Indian Buddhist texts, there are cases where it is necessary in ithe course of an argument to provide a proof by means of an ex'imple that distinguishes between two different people. In such in$tances, we frequently see the use of the names 'Devadatta' and ,;'Yajiiadatta,' just as in English, we might use generic names like '~j6hn' and 'Tom.' Paramartha also uses this rhetorical device. For ;6Xample, in the Po wo 1i&13t chapter of the Abhidharmakosa, an Wustration is used to distinguish between the minds of two peo:~ie, referring to the minds of Devadatta (devadatta-cetas) and !Yajfiadatta (yajfiadatta-cetas), which Paramartha translated direct,ly'as "Tianyu's *~ (i.e., Devadatta's) mind and Ciyu's ffrliJ~ (i.e., ;yajiiadatta's) mind" (T29, 308blO). However, in other, similar cases 'Yore find examples in which Paramartha used the Chinese names :Zhang ~ and Wang ::E, rather than Indian names. An example is ~found in the sixth fascicle of Dingbin's Sifenlu shi zong yi ji where ,lhe Ming liao [lun] shu is quoted: "The three families of Zhang, ,Wang, and Li::$ each in turn provided food for bhik~us" (Z.l.66.2, t'173ab). Other examples are not restricted to Paramartha's composi~ions, but also appear in his translations. For instance, this usage is 11so seen in the first fascicle of the Foxing lun: ~ J,,>
. .. First, the distinction between self and others is established with reference to [mental] continuums of different bodies. For example, when two people face each other, there are the concepts of self and other just as when Zhang faces Wang, Zhang is self and Wang is other; when Wang faces Zhang, Wang is self and Zhang is other. This logic also applies in the case of [non-human] objects .... 60

liJ:'he same sort of example is also found in the fourth fascicle of the :rSi di lun:


Funayama Toru

You ask: [Inasmuch as] all conditioned things are momentary without abiding, how can memory be possible? Why? Because it is incoherent to suppose that one person sees, and a different person remembers. Answer: If the cognizer changes, then memory would be impossible, just as if Zhang saw and Wang remembered. If the continuum of cognition changes, then in that case, too, memory would be impossible, just as one cannot remember a horse when what one saw was a cow. If cognition is unitary, then memory is still impossible, since no subsequent state of cognition could come into existence. That which is different from these three cases is called memory.61

It seems that the reference to Zhang and Wang was provided by Paramartha or a member of his translation group in view of those in his Chinese audience who would not be familiar with Indian names such as Devadatta and Yajfi.adatta.

Commenting on sutras composed in China

That Paramartha and his translation group were conscious of their Chinese audience is also revealed in other ways. For example, we know that he made commentaries for sutras composed in China. These sutras include the Renwang bore jing 1=::.Ef,gt~*~ (Scripture of Benevolent Kings) which was transmitted as one of Kumarajiva's . translations. Since the Chinese provenance of this text has already been discussed by Mochizuki Shink:6 ~A f~ 7, Gno H6d6 *!frt and other scholars, there is no need to re-examine it here. 62 It is certain that Paramartha composed a commentary on this apocry. phal sutra in which he advocated some of his own views. As mentioned in the first section of this paper, according to the list of siitras recorded in the eleventh fascicle of the Lidai sanbao


61 The text runs as follows: &FI=I~. ~~~7~, ~Ij~~/F{, :@;;;i5;{iiJf.iX;. {liT

ftg~ftg't,g1!Ht;~$:1lf. ~. ~~o1lfA, :@;;Jl,Ij/Ff.iX;. frlJ*~::E'I:i. ~ ~;fIH.A, :@;;!1J'/Ff.iX;. frlJ~4=-/F't,g,~~. ~~-, :@;;!1J'/Ff.iX;. 1!W~~$:.


&:*j/:t;.=.~, Jl,ljq<;~:@;;. (T32, 397b) * The Song, Yuan, and Ming editions of the Canon read fan &: while the Korean edition reads ji lJ... 62 For a discussion of past research on the Renwang bore jing as an apocryphon, as well as my own thoughts on the matter, see Funayama



The work of Paramartha


ji (T49, 99a2; alO), Paramartha translated the Renwang bore jing in one fascicle and composed a commentary, the Renwang bore shu, in six fascicles. Other citations found in the works of Jizang, Zhiyi, and Yuance, confirm that Paramartha's own commentary on the stitra (which is sometimes called Benji *~c "[Paramartha's] . original/root record (i.e., commentary)" in later references) did exist. However, the existence of Paramartha's translation of the sutra was denied by scholars such as Mochizuki and Ono who maintained the Renwang jing was produced in China. There is an interesting fact which supports their view: In many, if not all cases, the words from the sutra contained in Paramartha's Renwang bore shu as quoted by Yuance are exactly the same as the apocryphal text said to have been translated by Kumarajlva, the Renwang bore boluomi jing 1=::EfiJ~;E'1~~!P:*~ (T245). As I indicated in a previous paper,63 a close examination of the relationship between the apocryphal Renwang jing and the fragmentary quotations of Paramartha's "Original Note" (Renwang bore shu)64 reveals the following three points: (1) Paramartha never translated the Renwang jing; (2) Paramartha certainly composed some kind of commentary to the Renwang jing; and (3) Paramartha based himself on the apocryphal text of Chinese origin, whose translation was traditionally attributed to Kumarajlva. This so-called 'translation' by Kumarajlva, which formed the basis of Paramartha's commentary, also reveals Paramartha's

63 Funayama 2006: 53-54. A close examination ofYuance's quotations of the benji clearly reveals that Paramartha's text was a commentary on a text that had already existed as Kumarajlva's translation, inasmuch as we presume that Yuance's quotations are correct and trustworthy. It is certain that he sometimes did not quote the passages of other texts verbatim, l:mt I assume that his quotations of the benji are largely trustworthy. At least it is evident that he knew Paramartha's commentary on the Renwang jing and that what is called "Paramartha's translation of the Renwang jing" did not exist in Yuance's times (see T33, 361c).
64 The fact that Paramartha's Original Note quoted by Yuance is the same as Paramartha's Renwang bore shu is discussed in Vi 1965: 53. For a summary of Paramartha's lost text cited in Yuance's Renwang jing shu see Kimura 1982.


. Funayama Toru

system of panjiao *U~ (critical systematization .of the Buddha's teachings). Although I am not able to give a complete account of Paramartha's system of panjiao here, a few brief points can be introduced. First, there is a section of Paramartha's commentary contained in the eighteenth fascicle of Puguang's 1f:7\:; Jushe lun ji ~*mlli~c. which reads: "Furthermore, Paramartha said: 1265 years have now passed since the Buddha's nirviilJa" (T41, 282a).65 This reveals Paramartha's historical perspective as a Buddhist and, at the same time, it shows that Paramartha assumed the development of the Buddha's teachings even within the Buddha's fortyfive year teaching career. This citation was probably drawn from Paramartha's lost texts, the Jiejie jing shu and the Bu zhi lun ji. From these two texts we see that in the Jiejie jing (also known as the Jie shen mi jing) there were three types of teachings, or turnings of the wheel of the dharma. It seems, however, that Paramartha advocated a panjiao, which, while analogous to this, also differed somewhat in form. He developed this panjiao in his commentary to the Renwang jingo That is to say, the forty-five years of the Tathagata's preaching career can be divided into three "wheels of teaching" (falun ~.): zhuarifalun ~r~. ("turning the dharma-wheel"), zhaofalun ~r~. ("illuminating the dharma-wheel"), and chifalun fi'fr~. ("upholding the dharma-wheel"). This can be seen in the' following passage:
Paramartha said: The Tathagata preached three types of dharmawheel during his forty-five years in this world. These were the zhuanfalun, zhaofalun, and chifalun. Among these three dharma-wheels , there are the revealed and the secret. The secret [teachings] are found i among all three turnings of the dharma-wheels, from the night he . attained emancipation to the night he entered nirvii1Ja. The revealed [teachings were given] during the first seven years after he had attained emancipation. In the thirty-one years after the first seven, he turned the zhuanfalun. 66 During the seven years after the thirty-eighth'

65 Frauwallner 1951: 7-8. 66 The expression zhuan zhuanfalun *f*fr*lIiiii would be unusual in
Indic language, because the first zhuan is certainly a verb, whereas the compound zhuanfalun, which includes the second zhuan, should be a noun as the object of the first zhuan. .

The work of Paramartha


year, he turned the chifalun. [Thus we know that] after he turned the dharma-wheel he preached the Wisdom S[Uras up to the twenty-ninth year, that is {one year] before the thirtieth year, and only when it came to the eighth day of the first month of the thirtieth year, he preached the Renwang. Therefore, [the satra] states, "The eighth day of the [first] month of the [first] year [after the twenty-nine years]." Namely, he preached this satra in the thirty-seventh year after he attained emancipation and he was seventy-two years old. 57

In the history of Chinese Buddhism, there are two traditions regarding the chronology of Sakyamuni's teaching. In the first, Sakyamuni left home at nineteen, attained the way at thirty, preached for fortynine years, and died at the age of seventy-nine. In the second, he left home at twenty-nine, attained the way at thirty-five, preached for forty-five years, and died at the age of eighty. 58 The passage

57 The text runs as follows: JnmiS;, ftD31n-tttIZ9+Ji~, m =r:tl\iml. ~~~. ~ . ff. ?tJl:t-=.l\iml, ~ml~W. W]l,Jjfi1~Ji11Z, ~~~11Z, f:!Ut-=.r:t i/\l[. ml]l,ljf)JJJX:~-t;~, {.iftiftl:tl\iml. -t;~1~ =+-~>=p, ift~l:tl\iml. =+A 1f1~-t;~>=P, iftffl:tl\iml. fiiftiftl:tl\iml*, ~-=. +~u~ - +:tL~Bm~R~ $, 4~-='+~1'0JjAI3, jJ~J(. {=::E. ~i "f)J~JjAI3", Jl:tJlIjPX:-f~~

=+-t;~mJl:t*~, 7J~-t;+-~illjS;jS;. (Spoken by Zhiyi, recorded by . Guanding, Renwanghu guo bore jing shu {=::Eii~R~;g:*~itJ1t, second fascicle, T33, 263b). Note that the same content is also given in a different quote in the following way: ~iS;. Ji.:im-='iI:~, ftD*1:E-tttIZ9+3i~, iJ(. = *(forf!?)r:tl\iml. -iftiftl:tl\iml, mIJ'*~. ~ift~*jW. WJlJjftMi1~~11Z, .~~~11Z, 1E!.Jt.ift=r:tl\iml. *j~P1if)JJJX:~-t;~, 1E!.iftiftl:tl\iml. ~-t;~1& =+ 7f,>=P, 3l1tift ~1:tl\iml. 1i-='+)~~1~, M'-t;~>=p, iftfI(forff?)r:tl\iml. fi1'0 fffi~Tiftfi~(for ff?)*, -r~=+-~. JlrJ-+:tL~Bm~R~;g:, 4~=+ @'.1'0Jj)~I3, jJiJ(. {=::ER~;g:, ~jS; "f)J~JjAI3" . ~4 ;<js:ic jS;, i " ;f)Jf,JjAI3"~, ~PJEJjAI3. ftD*px:~-t;~m R~;g:. ~JlU~:)(, B -+ :tLf" ~Jl:taif, H!~PX:~1& =+1\~. Jl:t ;<js:ic ~,~ftDl::.ic. (Yuance's Renwang jing shu, end of the first fascicle, T33, 376bc. Cf. alsQ Jizang's Renwang bore jing shu, first fascicle, T33, 321a). 58 The belief that the Buddha preached for forty-nine years is found in Bai Fazu arMll (Western Jin), trans., Fo bannihuanjing 1~RnJt)fj)~, last fascicle (T7I, I71bc, I72a); (translator unknown), Bannihuan jing, last fascicle (Tl, I87a); Daoan ~* (Former Qin / Eastern Jin), Binaiye xu "~lf~ff;; KumarajIva (Later Qin), trans., Chan miyaofa jing *'fM ~rM~, middle fascicle (Tl5, 256a); Xiao Zilang .-r~ (Southern Qi),


Funayama Tom

quoted above in which Paramartha speaks of the Tathagata's "forty-five years in this world" conforms to the latter tradition. The same point is explained in the Renwang bore jing, where it says: 1i'a~+~=~*~~*~/ffiJIJ~~~-$.fB{~,fJJ~J3)\.. S ,jj~+iH. (T8, 825b). The expression "chu nian yue ba ri" (fJJ~J3)\" S "on the eighth day of the month [sic!] of the first year") does not seem to occur in other siitras and is one of the unique characteristics of the Renwang jingo Therefore, the above passage can be interpreted as Paramartha's development of an original classificatory system for siitras for the purpose of explaining the Renwang bore jingo Incidentally, the locus classicus for the meaning of the three types of dharma wheel (zhuan, zhao, and chi) is not the Renwang bore jing or the lie shen mi jing, but rather a passage in Paramartha's translation of the chapter called Ye zhang mie pin *~!~Jb of the lin guangming (diwang)jing ~7t~(*::EWf.69 One thing that is made clear from the above passage is that our previous assumption that, since he was Indian, Paramartha would not comment on apocryphal texts is incorrect. Moreover,
Jingzhuzi ~-fT(T52, 318c), and so forth. On the belief that the Buddha preached for forty-five years, see Dasheng beifen tuoli jing 7c~?,~)j wtflj~~ (translator unknown), sixth fascicle (T3, 276b); Tanwuchen ., ~~ (Beiliang), trans., Bei hua jing ?,~~jH~, eighth fascicle (T3, 219c); SaIp.ghabhadra f~1JJoJ/&~t~ (Sengjia Batuoluo, Southern Qi), trans., Shan jian iii biposha ~J!~f'iMttb, first fascicle (T24, 675b); Bodhiruci ~fJi': 1J1E:SZ (Northern Wei), Jingang xian lun 1it:1ilJ'HwmfB, third fascicle (T25, 818b), and so forth. 69 I.e., iwi(JJ~IiiI-tlJ~f~iit., ~f:E+jjiitJ1l, B1~~iiJ~$~-=-~-=-~

fR:1lr, ~tt~, f!f!tt~, fiftt;;~, ffi7ctMffl, ~7ctt;;t_t r:J;:7ctMl, tlW:wJ; ~, ~7ctt;;$I, *7ctMg (translation omitted). This passage is contained
in the second fascicle of the lin guangming jing (T16, 368b). On the lin guangming jing as the basis for Paramartha's theory, see the first fascicle of Chengguan's ~fI Da fangguang huayan jing shu: .~*-=-jjifft 1it:7t~)) lL~' f!f!. fif-=-~z~, ZlJ'7clPJtt, ffilS~ifri+~. ~~, t:-trm [g~, q;~tt;;~. t:-1~m fl~;j')) , ~~. f!f!=~, ~~f!f!:ffi;j:. =+:1~~~' f!f! . fif, ~~f!f!~:fffiftr=~ (T35, S08c; translation omitted). Incidentally, in Yijing's translation it reads: ~~tt;;~, fif f!f!ttiliilli (T16, 414a). In this translation, the three categories of turning, illuminating, and upholding are not clear.

The work of Paramartha


as has already been established by Mochizuki, it would be a mistake to conclude on the basis of Paramartha's commentary that the Renwang bore jing is an authentic sutra. In other words, we cannot deny the possibility that Paramartha made commentaries even on sLltras which he knew to be apocryphal (such as the Renwang jing). Having been trained in the orthodox Indian method of sutra commentary this should have been unacceptable to him. Why then would Paramartha do this? The reasons for this cannot be discussed in detail here, but, for one thing, Paramartha was invited to preach to a Chinese audience so it is possible that he made use of sLltras that were already established and well-known in China in order to spread the Buddhist teachings.
Approval of the characteristic doctrines of Chinese Buddhism The use of the theory of san shixin

In his commentaries on Buddhist sutras, Paramartha was conscious of the technical terminology currently used in China, and made use of terms such as shixin +{" ("ten faiths"), shijie +~~ ("ten comprehensions"), shixing +1T ("ten practices") and shihuixiang +~rPJ ("ten dedications") when referring to the grounds of bodhisattva practice. Shijie, shixing and shihuxiang are called san shixin -+'L' which means "three sets of ten minds." As has already been demonstrated, these terms related to the theory of bodhisattva practice were unique to Chinese Buddhist doctrine, and are not found in Indian texts. It has also already been pointed out that Paramartha used the term shijie to refer to the traditional "ten abodes" (shizhu +{i:).70

70 Mizuno 1984. Further, the following examples offer evidence of Paramartha's use of terminology coined in China: 1. ~~$:::.it :fL~ ~)) ;13;, FfO!~. *;;t)) (i.e., 1IE~*~);13; "*~Jl;+=ft}itU" (cf. T12, 491c), itU


~fPJ*. ~.



1@'L'***, A+1", 1"{t

ftoftO. 11E*O}1U$?J)\m+1"f.%.FJTitU;4(Yuance's lie shenmi jing shu, fourth fascicle, Z1.34.4, 391bc; translation omitted). 2.13X ;;tile)) ;13;, tH -+1"3~+~~, 7F/E. gill~=*. +D~+;I:-!H, /E. $: i "1T11**" (Yuance's Renwang jing shu, fascicle "shang ben" ~J:*, T33,369a; translation omitted). 3. - ;;tile)) ;13;, +1"~~fI'r~E. +~~~tE




Funayama Tom

3. Interpolated elements within the translated

Another issue regarding terms unique to the Chinese theory of stages mentioned at the end of the last section is the fact that Paramartha used technical terms such as shixin, shijie, shixing and shihuixiang in both his commentaries and in what have been identified as his translations. This is one of the reasons why it is problematic to take some of Paramartha's works as pure translations of Indic texts. As many previous studies have already shown, there is a passage in the third fascicle of Paramartha's translation of the She dasheng lun shi which reads:
Bodhisattvas are of two kinds: (1) one who abides at a worldly stage; and (2) one who abides at a holy stage. The stages from the initial arousing of the mind (of enlightenment) to the Ten Faiths are all worldly stages. The stages of the Ten Comprehensions or higher all belong to the holy stages.7 1

There is also an explanation given in the fourth fascicle of the same text:
Bodhisattvas are of two kinds: namely, worldlings and saints. Those who are in the stages up to the Ten Faiths are worldlings, and those who are in the stages of the Ten Comprehensions or higher are saints.72

From these passages we can see that some of Paramartha's theory clearly diverged from the Indian terminology and doctrine of practice and was derived instead from Chinese Buddhist doctrines.

fl'~1. +1t~Jifjj1. +jjg[rPJBJ::~P}l;J!Ji. *~~3Hi~~;ltt1~. JZ-n~j3;.

+{i+ll::.+~lH,'. i!ij:~+{i~~fl'r1(Yuance's Renwang jing shu, fascicle zhong ben ~I:j:l* T33,386c; translation omitted). These passages are enough to clarify Paramartha's use of technical terms such as shixin and shihuixiang in his explanations of the theory of practice. I have already discussed this issue in Funayama 2002: esp. 22; and Funayama 2003: esp. 126. 71 ~ilI~=fl, -1:EfL{lr, =1:f~{lr. 1tf1)]~~,[," ~t:+1i),~, :Mztl::fL{lr. 1tf+~~YJ::, ~}I;~{lr. (T31,174c) 72 ~ilI~=fl. ~~fL~ ~A. +fiY~tl::fL~, +~~YJ::tl::~A. (T31,177c)

The work of Paramartha


Briefly stated, in the history of Chinese Buddhism from the Six Dynasties through the Sui and Tang, the standard theory for the stages of the bodhisattva path contained the following fifty-two stages after chufaxin f)J~~JL' (generation of bodhicitta): ten elementary stages called ten faiths (shixin +f~) [stages 1-10] --.. ten abodes (shizhu +{; called shijie +f1~ in Paramartha's texts) [stages 11-20] --.. ten practices (shixing +17) [stages 21-30] --.. ten dedications (shihuixiang +@~) [stages 31-40] --.. ten grounds (or ten stages, shidi +:f:-!!l) [stages 41-50] --.. final two grounds (hou erdi 1~=:f:-!!l) [stages 51-52]73 The stages from chu faxin to the end of the shixin were known as the "stages of outer (bahya) worldlings (prthagjana)." Next, what is called san shixin "three sets of ten minds" were known as the "stages of inner worldlings" and the chudi f)J{g ("first ground" of the ten holy grounds) and up were regarded as the "stages of saints." In contrast with this system, we know from the two passages cited above that Paramartha's theory held that the stages from chu faxin to the end of the shixin were known as the "stages of worldlings" (jan wei }L{ft or fanfu wei }L~{ft) and the stages from the beginning of ten abodes on were known as the "stages of saints" (or holy stages; sheng wei ~{ft or shengren wei ~A{1L). This way of establishing the boundary between worldlings and saints (or holy beings) was a significant divergence from contemporary Chinese doctrines. From the perspective of his Chinese audience, this way of explaining the theory of the bodhisattva path had the value of being easy to understand. On the other hand, it is problematic that the texts which are transmitted as "Paramartha's translations" contain those non-translational elements. Which part of the translation was literal and which part was added by Paramartha or his translation group? These issues have not yet been completely resolved.

73 For this point, see also Funayama 2005b: 388-J92.


Funayama TofU

Translating one Sanskrit word with two Chinese giving different explanations for each


Closely related to the preceding discussion is the fact that Paramartha often used two Chinese characters to translate a single Indic word and provided different explanations for each of those characters. Of course, the use of two similar Chinese characters to express the meaning of a single Indic word is not unusual, but to give different explanations for those two characters is quite rare. A straightforward example, already discussed by Nagao Gadjin, is the way in which the word "huanxi" ikl} (joy/joyous) is explained by its components "huan" and "Xi."74 Huanxi is a simple word which corresponds to the original adjective pramudita- (to be delighted, to be happy), namely the first stage of the earliest ten stages of the bodhisattva path, also known as the "joyous ground" (pramuditti bhilmiJ:t). In the eighth fascicle of Paramartha's translation ofVasubandhu's commentary on the MahtiytinasaY[lgraha, it is explained in the following way:
To abandon affection for oneself is called huan, and to produce affection for others is called xUs

This explanation is completely based on the Chinese language and is not possible in Sanskrit. From his investigation of the context in which the word occurred, Nagao points out that this is not just limited to the explanation of the word huanxi but can be extended throughout the entire section in question and that those elements cannot be taken as translations. Furthermore, Nagao also indicates that in addition to huanxi in Paramartha's translation of Vasubandhu's commentary on the MahtiytinasaY[lgraha, there is also the example of the differentiation of the characters yi and yang in the compound yiyong ltffl (tisaya). Moreover, in the ninth fascicle of the same text, the phrase xinyaoyi 1~~lt (adhytisaya) is separated into xin, yao, and yi. The explanation given for the difference between xin and yao is as follows:


Nagao 1987: 60.

M=~~iX, ~ftI1~~*. (T31,206a).


The work of ParamarthR.


Because the mind is settled and without doubt about the orthodox teachings of the six piiramitiis, it is called xin (faith). And because one wishes to practice in accord with the object towards which one holds faith, it is called yao (desire)?6 Xinyaoyi corresponds to the Sanskrit term adhyasaya. The same Sanskrit word is translated as shenxin mgJt,\ by Buddhasanta and Gupta, and as zengshang yiyao ~l:~~ by Xuanzang. Therefore, it seems that Paramartha's distinction between xin and yao does not make sense in Sanskrit. The explanation of the term ritnhua ~~ in second fascicle of the Foxing lun is yet another example. There, runhua is divided into run ~ and hua ~ as two separate notions?7 The corresponding portion of the section of the Sanskrit text of the Ratnagotravibhiiga leads us to believe that the original Sanskrit for runhua was the single word snigdha "moist / lubricating" and that distinguishing between run and hua does not make sense in the context of lndic languages. Again, as has already been discussed in a previous study, in the Sui xiang lun 1li;J;mlll, the term aiyu ~W\ (chanda) used in the Abhidharmakosa is analyzed in terms of ai and YU. 78

~:t\~lE~"P, IL\~~~, Mc~~{~. :tmm{~~, *~{~1T, Mc~~

(T31, 213b). 77 In the passage which gives a detailed explanation of this term, run and hua are distinguished: ~m-1!f, t~Ymt!,I;;~m~, m-1!fB~'W~rPJ1~ ~ (T31, 797a12-13). The term originated as an explanation of the phrase "san runhua xing zhe" =t~m-,t:1!f (T31, 796c17~18) and originallyappeared in the following passage: .\j1J:tIHr=fl. {iiJ1!f~~. -1!f:tlD~;;b1~'~, -1!f~.'~, =1!fWilm-'~ (T31, 796b5-6). Fortunately, these three terms were translated in the thirty-first verse of the original Sanskrit text of the Baoxing [un .t:~* and its prose commentary as ruyi gongdexing :tlD~ ;;b1~'~, wuyixing ~.,~, and runhuaxing ~l'1H: and correspond to the Sanskrit words prabhiiva, ananyathiibhiiva, snigdha (or snigdhabhiiva). In spite of the fact that the passage following "san runhuaxingzhe" (T31, 796c17-18) does not exactly correspond to the Baoxing [un, we can safely infer that the term runhua is equivalent to Skt. snigdha. *For this passage see Takasaki 1989: 47-48. 78 Namely, ;ft&~~JE,i:m1iBl, ~~f~i:m1iBl (T32, 165c4-5; transla-



funayama TofU

We have seen examples of how, in texts such as the Sh;e dasheng fun shi and the Foxing fun, one Sanskrit word was translated using two semantically similar Chinese characters which were interpreted as having different meanings. With regard to these examples, previous research has generally held that they resulted from the scribal errors of Paramartha's disciples, since an Indian scholarmonk such as Paramartha was not expected to give such explanations. 79 It has been concluded, therefore, that those aspects of Paramartha's translations which are inexplicable or inconvenient are attributable to his disciples' misunderstandings, but I doubt that such explanations are correct. As shown above,Paramartha's commentary on the apocryphal Renwang bore jing and the Chinese Buddhist doctrinal terminology used to discuss the beginning stages of practice (ten faiths, ten practices, and ten dedications of mind) cannot be explained in terms of disciples' errors. Setting aside the question of whether or not Paramartha fully endorsed this approach, we can say that using elements unique to China in commentaries was in some form approved by hini or by the consensus of his siitra translation group.
Pure translation should contain small-print interlinear notes

In Paramartha's "translations" we sometimes observe that those words which should have been written as small interlinear notes if the text was a pure translation are included in the body of the text.
tion omitted). For this see Aohara 2003: 845. 79 Nagao Gadjin writes:. "It is difficult to believe that this sort of Chinese commentary could have corne from Paramartha. Therefore it was probably the commentary of his disciples which was mixed in with [the original translation]." (Nagao 1987: 60.) Furthermore, when proposing the theory that there existed two different Vasubandhus who were later mistakenly identified as a single person, Frauwallner also noted the possibility that Paramartha correctly understood the difference between the two persons but that his disciples wrongly confused them as a single person and therefore compiled the Posoupandoufashi zhuan ft.~~Rl't i[i1t which took Vasubandhu as a single person. See Frauwallner 1951: 18.

The work of Paramartha


There are a number of such passages. A typical example is found in the following passage of the Xianshi lun ~~~lffl:
Third, yang shi JtJiWl is of six types such as [cognition] in the realm of the eye; these are the six cognitions. IJ::t~'p'q!~t~.~~J.~~.tht:!mxh~.l:zg0h~~~
;>.f:t.~ .:V;.~.~.80

In Chinese Buddhism, the Dalun frequently refers to the Da zhidu lun, but this is not the case; here Dalun refers to the MahiiyiinasaY(lgraha. It is easily surmised from the general context that the Xianshi lun is a kind of commentary on a certain section of the MahiiyiinasaY(lgraha. It should be noted that yong shi and zhengshou shi both derive from the same Sanskrit word aupabhogikaY(l vijfiiinam (or upabhogavijfiiinam).81 In other words, although the different translations of yong shi and zheng shoushi are meaningful in Chinese, they create a tautology in lndic languages and are meaningless in the given context. Therefore, the underlined words ofthe above passage probably did not exist in the original lndic text.

80 ~= JtJiWl1lf, hf1~iWlc:W-~, Plp;l'thiWl.j<?';W6i!t~.l1.~~~~ (T31, 879a). 81 The translation of yang shi is found in a stanza in the first fascicle of Paramartha's translation of the Madhyiintavibhiiga r:p~)j-}jiJ@jj (T31, 451c28) and in the prose commentary thereon (452al-2). The corresponding Sanskrit for the term in the verse (1, 9b) is aupabhogikam (vijfiiinam); in the verse commentary it is upabhaga(-vijfiiinam). Furthermore, a stanza of the Zhongbian fenbie fun is quoted in the Mahiiyiinasaf!1.graha, where the same term is translated as shou shi stiWl (T31, 1I5c19). This shou shi corresponds to what was translated in the preceding prose commentary as shouyong shi stJtJiWl (T31, 1I8el8). In short, in comparing Paramartha's translations of the Zhongbian fenbie fun and the She dasheng fun, he translates the same Sanskrit word alternately as yong shi, shoLt shi, and shouyong shi. Moreover, in the first fascicle of Paramartha's translation of the She dasheng [un there is the term zhengshou shi lEst iWl. This corresponds to what Xuanzang translated as bi neng shou shi ;JEt. ~~stiWl and probably refers to the Sanskrit upabhoga. On this point see Nagao 1987: 275-277. Taking all this into account, it is evident that there is no essential difference between yong shi and Zhengyong shi. Using both of them forms a tautology in Sanskrit.


Funayama Tom

As is well-known, the Biography of the Buddhist Mqster of the Law Vasubandhu has also been identified as one of Paramartha's translations. However, unless it is assumed that there are elements added to the translation, passages like the following cannot be . properly understood:
In this land there was a Brahman, who was the state master, with the surname Kausika .. He had three sons who were all named Vasubandhu. y (!.~.l!.. . tr(!.n~~~~~.q. ..(!.~.. ~~~~Y~n'.. (tJt;ll!} .. i.~ l?(!.n4hl!..~~. ~.I:(!.p.~:: J.<~t~.q.. !'t.~. '.f~rp.ni. (q~Y!):)n )ng.i;:~, .~~~r~ .~~ .~~~):)..!'t. ~~W~rp.. in .I:1.(!.rntng. ~J:l.~~-:
4r~n. AIHl,9~g~ .t.l?~Y. ):).!l:Y~ .~J:l.~ .~!l:rp..~ .n(!.m~.,. J:h~y. !l::r:~. ~!'t.C:~ g~~~n. ~1!9th~r

Iq.iff~l:"~ml w~m~..~~. nl:~t..t.J:l~Y.. ~.~~. ~~. .9..i.~tt~g.l!~.~J:l.~q.

The third child named Vasubandhu became a monk of the Sarvastivada school and attained the fruit of arhatship. He was also named Bilinchibasuo J::!: ; ~f Ji!k~ (Virificivatsa?). ;I?~~~I,l.cht .!,(!l:~. hts. rn9.t.l:t.~r:~. p.!l:rp..~. !'tP4. ~ ~~~~ (v.t;l.~s.t;l) m~~R-~. )9R-: .9.1: .'.<;:hM/. P~9P'~~ .~~~.<? ~~~..t.l?~ .I!~m~. ~ .~~~~ .t.<? r~f~.I: ~q .t~~. ~if.~p.riR-g .<?f ~~~*" .~M. in .t~~~ .p.~~s:.~. [i:~: 1. Ghi~~1th~. qff~pripK 9.f ~!'t~.t.l~ N~. S:.~n~.g. t!:~~ JJ.82

In this passage, an explanation of the name Vasubandhu is given. It is evident that the name Vasubandhu (Tianqin ~m, Shiqin itt m) was not the name he received when he was ordained as a monk (i.e., what is called "dharma name") but rather his original birth name conferred on him by his parents. 83 It is possible to assume

Jl:t;ff~gijJ~~F~~~'tJiF3m!. ;ff=+[FlJ~~IH\~R ..~ __ ~.~7,;,J~

..l:~J'~AA....?'!-.~.~~.4S.~~~B... ~J.'.~4S.,..1jl;;:lW4S.,.. 10.~.;;<:. m-=+~~

~:.:~*?).~~:T.l.RfF.!m: ..~~~~.A~!.fr.~.1::T.~.4S.~*..:{iii.~:#'~ (1) The original text gives po ft. but, given the context, it should be read suo ~. (2) Basha Ji!k~ - The Korean edition has Jupo *0tft.; The Song, Yuan, and Ming editions all have bapo Ji!kft., but it should read basha in accord with its meaning. (3)
T~.~(3). (T50, 188b). (1) Suo ~ -

~R, h~ili~~B'~tJj*, 1~~~{~*. 5JiJ~J::!::;~~fJi!k~(l). J::!::;~f:&;ttS:

In the Taish6 canon (as well as the Shukuz6 ~JIi) this section is punctuated as " ... ~!/JHfr. ft.~~R:&1filitRt~A." However, it should read


~4t. *rft.~~R:&1filitR'r:tA."

Likewise, according to the first fascicle of Shentai's *$,* commentary on the Abhidharmakosa (Jushe lunshu 1:l'!.*~iiilit;t first fascicle, Z1.83.3, 277cd), the "Vasu" of Vasubandhu means Vasudeva and his parents were given the child because they worshipped at Posou tian miao ft. ~:::RJWi (that is, Vasudeva shrine).

The work of Paramartha


that the underlined words were supplemented when the text was translated into Chinese. The next passage is also from the Biography of Dharma Master Vasubandhu:
" Within five hundred years after the Buddha's nirviilJa, there was an arhat named Jiazhanyanzi ~:h'fij:JJfT (Katyayanlputra). kH~.!w,m~ :o/.~~ g~t:ty~4 Xmm .ht~. m~t~wr) .~~:m~x:r:tXI. wh~~l:)... W~~. h.~~.l:)..!H:mm. He first became a monk of the Sarvastivada school. He was Indian and later went to the kingdom of Jibin i'ij2.~.84 l.~Q~~j~j~.th~..I;Wr.~l:).:o/.Y~.U~( JI.l4~i1: .. In collaboration with the five hundred arhats and the five hundred bodhisattvas, he compiled the Sarvastivada school's Abhidharma text, the Bajialanta j\..{iJDlJHt!! (*A~tagrantha). H~J:yjq~.1w.9:o/.~.~~Jh~
!!ftj~f:l.rHff:l..!\"J1... . ... 85

It is hard to believe that the underlined words are translations from the original Indic text. Although the above-cited passages are traditionally regarded as translations, it is noteworthy that the phrases "this place" (citu J!t) and "here" (cijian JIUi'I') which refer to China are used in the above two passages. The Ba jiandu signifies the Apitan ba jiandu lun [5iiJm~J\. .8t~ (T1543) translated by Sarp.ghavarman and Zhu Fonian. It is a version of the Jiianaprasthana, the text later translated by Xuanzang as the Fazhi

The possibility that the Biography of Dharma Master Vasubandhu is not a pure translation, and that it contains elements of Paramartha's oral commentaries was first pointed out by Takakusu Junjir6 ~1mI'~{jz~~, and I concur with his conclusion. However, I would like to correct an error Takakusu made regarding the following passage:

84 Jibin if. signifies the northwest region which includes Gandhara and Kashmir. 85 {ijlH)~J3[1&]is:9=J, 1f~ilJ~~~~:h'fij:JJfT. fif.~$~M~,.. 1~:Ii~.~.. )t; ~roi~~WB/l*. *~;R~A, 1&1ifJl~. It~~~~.~.W;j~ _]is~ilJ

... (T50, 189a). * "Jian" . , following the Song, Yuan, and Ming editions. The Korean edition has "qian" ~. .

~t~&]is~roi, #.~roi~~WB~mJm, ~~ (U~{iJD.Ht!! , fi.P.J!tr..Ij~


Funayama Tom

Up to this part, the text records the [three] brothers of Vasubandhu and so forth. Hereafter, it records that Sanzang Sheli .::.~1tJ~86 went east from the imperial palace of Taicheng ~ [and later] arrived at Guangzhou where he again translated all the Mahayana treatises. It also records the affairs which occurred after his death so that these things would be passed on to later generationsY

In this passage, Takakusu's translation of Taicheng as "the capital of Tai-chou," that is to say, the capital city of Tai prefecture (present Zhejiang province) is incorrect. 88 Moreover, Takakusu also mistakenly disregards the distinction between the characters tai If (Taicheng If~) and tai ~ (Taizhou~1'1'1).89 Furthermore, the identity of the author of this passage is quite problematic. I would suggest that this postscript was not a later addition but was present from the beginning. One reason for this is the unique expression used to refer to Paramartha, "Sanzang Sheli." First, at the end of the Niepan jing benyou jinwu ji lun ~1E ~*~*if4-~1"~iil, translated by Paramartha, it notes that it was "Sanzang Sheli's oral exposition" =-IJIlM~jlj~\' (T26, 282c). This passage reveals that it was Paramartha's own oral commentary transcribed as an appendix to the translation. And second, at the end of Paramartha's translation of the Guangyi famen jing m~{:ft F~*~, there is a passage which reads:
This sutra issued from a chapter of the Middle Agama [MadhyamagamaJ. Paramartha Sanzang Sheli was requested to translate it on the tenth day of the eleventh month of the fourth year (563) of the

86 Sanzang is "the Tripitaka Master" and sheli (for asheli ~PJItJ~) signifies "master" or "mentor"; that is, Paramartha. 87 M*~t:Jj:t, ~cx~~5L$. tt1~~c.::.~ItJ~f{{:~tlJ<*~[l1'1'I, .I; ~~*31Hif~i1Ii*~1~1&*, {tjjN~1-t. (T50, 191a). 88 Takakusu 1904: 293 89 Takakusu's error is not explicitly mentioned by Frauwallner 1951: 18, but his translation shows the referent for the place name Taicheng as "the city of Tai (Nanjing)." Taicheng should rather be understood as indicating the inner city of Jiankang, namely the imperial palace where the emperor resided. This was correctly indicated in Demieville 1931: 18. For a classic study on Taicheng, see Zhu 1936: 108-116.

The work of Paramartha Tianjia era of the Chen, guiwei year? at the Zhizhi Temple Guangzhou. 90




Considering these two examples, it is possible that Paramartha was reverentially called Sanzang Sheli by his direct disciples .. In any event, if the above-mentioned affairs of Paramartha after his arrival in China were recounted in the now lost second half of the present Biography of Dharma Master Vasubandhu, it goes without saying that such records cannot be called translation in the strict sense. Conclusion This essay has examined some of the unique aspects ofParamartha's compositions (or rather, his oral teachings as recorded by his disciples) through quoted fragments. These fragments provide clear evidence of the proactive techniques utilized by Paramartha when commenting on siltras, which include: revealing the multiple meanings within a single phrase; interpreting the meanings of proper nouns; comparing India and China; comparing the theories of various Indian schools; the use of Chinese rather than Indian names; and commenting even on apocryphal siltras. Further, we have discussed elements within his "translations" such as his method of translating one Sanskrit word with two Chinese characters and giving different explanations for each, and the presence within the body of the text of passages which, if the texts were translations in the strict sense, should have been given as small-print interlinear notes. Generally speaking, this essay has shown that one of the primary characteristics of Paramartha's compositions was his consciousness of the culture of his Chinese disciples and audience. In this we see a concrete example of the intersection of Indian and Chinese cultures. Paramartha actively made use of elements unique to Chinese culture. This may have been one of his unique traits or it may have been a common pattern among Indian scholar-monks of
90 J1U~II9='~~-. IljtR~Jm1~t'X~*+~J3+I3, ~1l1\"'fIltl\,"*, mn~~;r.ff=i;J~~~. (T1; 922a).


Funayama Toru

that period. 91 The latter possibility is also suggested in sections of the Da zhidu [un *~l~JilH. 92 Either way, it is a tangible 'example of the Buddhist monk Paramartha's "preaching the dharma in accord with circumstances" and his practice of "skillful means."

For confirmation of the Northern Wei monk Bodhiruci's 'fr~yJfE use of Chinese doctrine and sutra exegesis, see Otake 2001: 65-68. Moreover, Yuance's lie shenmi jingshu, Renwang jingshu, as well as other Tang commentaries, quote the teachings of the Indian monk called Chang'er sanzang ftJ}=jt (i.e, "the Tripitaka Master 'Long Ears"'). Some of those passages state that Chang'er Sanzang explained compounds such as rushi .t1oti':: (evam) of the phrase rushi wo wen .t1Dti':::fZ11fl by dividing it into ru .t1o and shiti':: as two separate notions. The identity of Chang'er Sanzang is uncertain, but in the second fascicle of Zhanran's Weimo jing lueshu *1t~*ll!:~i9IE (T38, 583b5) there is mention of a "Shang tongshi" ~i@'CgiP, that is to say Fashang r:tJ:: (495-580 cf. "Gaoqi Shang tongshi" ~"i!f~i@'CgiP T85, 514b4-5), who once spoke with with Chang'er Sanzang. Therefore, it may be the case that Chang'er Sanzang was the monk from the Northern Qi who worked as a zhao xuantong 8j1~i@'C (governmental monk-administrator; for this see also the Lidai sanbao ji, fascicle 12 in T49, 102c20-'-21), that is Narendrayasas jj~~~l[J1 '@tJ490-589) who was explicitly characterized by his long ears (see T50, 433a17-20; T55, 365bll-13). For Narendrayasas as one often members of the xuantong under Fashang, see Yamazaki 1942: 521 and 545-556.



;fl1l ((r:t~~1S:~~ w~: ~ftJ}~;g-, ~i3;jj~~~~J1~'@t, ~(read~)i3

. , ~tFPBt~ftAllio (Zl. 53,4, 326c). Further, according to Yamaguchi Hiroe, an eighteenth century Japanese Tendai monk Shutoku Honjun "'f ~*i#!i, also identifies Chang'er sanzang with Narendrayasas; for this, see Yamaguchi 2004: 116f. 92 I.e., ~JJ:;g-, ~;!t:9J1~~~, ~z::fYE:, Xmmzi!ij[i3Jl (T25, 277a). Lamotte translates the passage as follows: "Tsan-t'an ~~ (varlJana louange ). - Lauer leurs qualites, c'est tsan; les vanter sans cesse et les exalter, c'est t'an." And on this runs his brief comment: "Ces explications semantiques sont evidemment des gloses chinoises a l'usage des Chinois." For this see Lamotte 1976: 1934.

The work of Paramartha


Aohara Norisato wJJlf4t%l 1993: "Tokue no Zuiso ron" 1!~0) Wf\jt~ilii~ , Indogaku bukkyogaku kenkyu FP~~{5!l~~1iJf~ 41/2, pp. 978(185)974(189). 2003. "Tokue Zuiso ron no jittai gy6s6 kaishaku" 1!~ Wf\jt~~ O)~ ~11't~~~, Indogaku bukkyogaku kenkyu FP~~{5!l~~1iJf~ 51/2, pp. 847(186)-842(191).

Chimpa, Lama and Chattopadhyaya, Alaka (tr.), Chattopadhyaya, Debiprasad (ed.), 1990: Tiiraniitha's History of Buddhism in India, Delhi: Motilal Banarsidass (originally Simla, 1970). Deeg, Max, 1995: Die altindische Etymologie nach dem Verstiindnis Yiiska's und seiner Vorgiinger: eine Untersuchung iiber ihre Praktiken, ihre literarische Verbreitung und ihr Verhiiltnis zur dichterischen Gestaltung und Sprachmagie, Dettelbach: J. H. Roll. Demieville, Paul, 1929: "Sur l'authenticite du Ta tch'eng k'i sin louen," Bulletin de la Maison Franco-Japonaise II, 2, Tokyo, pp. 1-79 (reprint: Paul Demieville, Choix d'hudes bouddhiques (1929 -1970). Leiden: E. J. Brill 1973, pp. 1-79). 1931: "L'origine des sectes bouddhiques d'apres Paramartha," Melanges chinois et bouddhiques 1, 1931, pp. 15-64 (reprint, Paul Demieville, Choix d'hudes bouddhiques (1929 -1970), Leiden: E. J. Brill 1973, pp. 80-130).

Edgerton, Franklin, 1953: Buddhist Hybrid Sanskrit Grammar and Dictionary, Vol. II, New Haven: Yale University Press (Reprint Delhi: Motilal Banarsidass 1970). Fangshang shi jing (Sui-Tang ke jing) mU1~*~(W1fm$(ti*~), 30 Vols, Beijing: Huaxia chubanshe 2000. Forte, Antonino, 1990: "The Relativity ofthe Concept of Orthodoxy in Chinese Buddhism: Chih-sheng's Indictment of Shih-Ii and the Proscription of the Dharma Mirror Sutra," In: Robert E. Buswell, Jr. (ed.), Chinese Buddhist Apocrypha, Honolulu: University of Hawaii Press, pp. 239-249. Frauwallner, Erich, 1951: On the Date of the Buddhist Master of the Law Vasubandhu, Roma: IsMEO. 1959: "Dignaga, sein Werk und seine Entwicklung," Wiener Zeitschrift fiir die Kunde Siid- und Ostasiens, 3, pp. 83-164.

-1969: Die Philosophie des Buddhismus, 3. durchgesehene Aufiage, Berlin: Akademie-Verlag.


Funayama Toru

Funayama Toru iYci-0J1~ 1996: "Gikyo Bonmo kyo seiritsu no sho,mondai" ~ *~ W'J-t*il1*~~ px:.lz:GI)MfF,,~~, Bukkyo shigaku kenkya {L.~5t:,.JiJffE 39/1, pp.54-78. 2002: '''Kan'yaku' to 'Chugoku senjutsu' no aida - kanbun Butten ni tokuyu na keitai 0 megutte" r~~~J ~ r9=I~mJzttJ Gl)F",' - ~)({~JltiJ;:!f,f lff.t7[;~~IiO('0-C, Bukkyo shigaku kenkya {~~5t:~.JiJffE 45/1, pp. 1-28.

- 2003: "Ryuju, Mujaku, Seshin no totatsushita kaii ni kansuru shodensho" llMM,~w;tltmGl)IJ~L-t=-jll'~{ft~;:::ImT0Mf1t;iJ:, TohOgaku J!f:1:7~ 105, pp. 134-121. 2005a: "Shintai sanzo no chosaku no tokucho: Chuin bunka koshO no rei to shite" ~~-=-itGl)Wf'FGI)!f,f1~ - 9=IF~)({l:zli:1g<GI){7iJ ~ L--C, Kansai daigaku, Tozai gakujutsu kenkyfisho kiyo ImW*~*W~*i.JiJffEffl*C.~ 38, pp.97-122 2005b: "Seijakan no ni keitO - RikuchO Zui To bukkyoshi chOkan no isshiron" ~1!fWl.GI)=*iIJC - :AlM~~m1~~5t:,iifo~GI)-~~, in: Mugitani Kunio ~~*~~ (ed.), Sangyo koshO ronso -=-~~~tAli~, Kyoto: Kyoto daigaku jinbun kagaku kenkyusho, pp. 373-408. 2006: "Masquerading as Translation: Examples of Chinese Lectures by Indian Scholar-Monks in the Six Dynasties Period," Asia Major, Third Series, Vol. XIX, Pts. 1-2, pp. 39-55.

- 2007: "Kyoten no gisaku to henshu: Yuigyo zanmai kyo to Sharihotsu men kyo ," *~JI!l,GI){$1'FUi'l1' - WJi~=IVI::*~~ ~ W*;fIj*F,,~*~~, in: Kyoto daigaku Iinbun kagaku kenkyusho (ed.), Chfigoku shfikyo bunken kenkya 9=I~*~)(Jfk.JiJffE, Kyoto: Rinsen shoten, pp. 83-107. Hattori Masaaki ~1X~~JE@.Ii 1955: "BusshO ron no ichi kosatsu" W1~'t1tilll~ GI)~~, Bukkyo shigaku 1~~5t:~ 4/3&4, pp. 160-174. -i961: "Dignaga oyobi sono shuhen no nendai" 7''-1;7''-j--jJ,&V-cGl)$jil Gl)ip{-t, in: Tsukamoto hakase shOja kinen bukkyo shigaku ronsha ~* t.f0~i.f~c.;%{~~5t:~~~, Kyoto: Tsukamoto hakase shoju kinen kai, pp.79-96. Hirakawa Akira If)II~ 1994: Hirakawa Akira chosaku sha, dai 16 kan nihyakugojikkai no kenkya III If) II~ ~If) II~w1'F~~ 16 5ff =l3:Ii:+1l3GGI) .JiJffE III, Tokyo: Shunjusha. Imazu Kogaku 4'$r#~ 1925: "Ragorabadara no churon chu ni tsuite" m ~~!uj;(:~tmGl)9=I~m;:::&t-c, Gendai Bukkyo m{-t{~~, 2-16, pp. 72-84. Iwata Taijo *B3~1f 2004: Shintai no yuishikisetsu no kenkya ~a*GI)pt~m GI).JiJffE, Tokyo: Sankibo busshorin.

The work of Paramartha


Kahrs, Eivind, 1998: Indian Semantic Analysis: The 'Nirvacana' Tradition, , Cambridge; University of Cambridge (Oriental Publications 55). Katsumata Shuukyo .Il91JZ.~tJ& 1961: Bukkyo ni okeru shinshikiki setsu no kenkyu fflltJ&f::;j3~t.QJL'~~O)1i3f~, Tokyo: Sankibo busshorin. Kimura Kunikazu**t;F~fo 1982: "Saimyoji Enjiki ni okeru Shintai sanzo shoden no gakusetsu ni taisuru hyoka, 2" [,~~OO?llUI::;j3~t.Q~~=iGm {iO)~~f::ttT.Q~fJfC-), Kenkyu kiyo (Nagaoka tanki daigaku) 1i3f~*[i, ~(*IMJ~lUlJl*~) 6, pp. 82-67. Lamotte, Etienne, 1935: Sa1'[ldhinirmocana sutra: ['explication des mysteres, Louvain: Bureaux du recueil, Bibliotheque de l'Universife. 1976: Le traite de la grande vertu de sagesse de Nliglirjuna (MahliprajiilipliramitliSlistra), Tome TV, Louvain: Institut Orientaliste.

Mizuno Kogen 7.kJfijJ,.5I; 1984: "Gojiinii 15 no bosatsu kaii setsu" 3i O)l1fiii~{ll:~, Bukkyogaku f~tJ&~ 18, pp. 1-28.

+- {ll:~

Nagao Gadjin *~:jltA 1987: ShO daijo ron - wayaku to chakai (ge) mfB - flJ~t:7'i:~~ Tokyo: Kodansha.



Nagasawa Jitsudo *rRJfJ: 1978: Yugagyo shiso to mikkyo no kenkyu 3:1iffJo 1TJIl!!,Mt:!lf;tJ&O)1iJf~, Tokyo: Dait5 shuppansha. 1978a: "Kan'yaku nihon taisho Chibetto yaku Shuryo ron cha wayaku" ~ ~=*f.tJffiT"V r~ W~:l::mfBtt~ fp~, in: Nagasawa 1978, pp. 291-301. 1978b. "Muso shijin ron no keitaironteki kent5" W~rl\JIl!!,~mfB~ 0)~1IElmfB ~~1;J, in: Nagasawa 1978, pp. 301-307.

Namikawa Takayoshi 21ft) II:;:{~ 1995: "ShOryobu no shi zenkon i setsu" lE:I:: ffIlO)Im~m{ll:~. Indogaku bukkyogaku kenkya FPit~f~tJ&~1i3f~ 44/1, pp. 399(96)-393(102). 2000: "Chibetto yaku Vi mui kecchaku no Shoryobu setsu to Ritsu nijuni myoryo ron" fT"V r~ W;ff~~~#C~~ o)lE:l::ffIl~t: n~-+-~7 mfB~ , in: Kato JunshO hakase kanreki kinen ronsha, Abidaruma bukkyo to Indo shiso .tmRi~.t.~Mic,~mfB~, 7l:::'7'/V7{~tJ&b1~P',Il!!,M, Tokyo: Shunjiisha, pp. 181-194.

Niryo gakujin =:m~.A 1930: "Hobo renpeki, I" r~.J.l~(-). Gendai bukkyo m{Jd~tJ& 3, January, special edition, pp. 175-179. Niryo sei

-:-:m1:. 1934: "Daizo bunko koitsu zenpon mokuroku, I" *iGJtJil[

Pitaka l:::0~jJ 5, pp. 15-46.

Obermiller, E., 1932: History of Buddhism (Chos-l;.byung) by BLI-ston. II. Part. The History of Buddhism in India and Tibet, Heidelberg: O. Harrassowitz.


Funayama Toru

Okada Yukihiro!lilJl33iisl 2002: "Kan'yaku butten kenkyu josetsu - Shintai yaku Rogyo osha ron 0 megutte" ~~{qjlA1iJfJi:ff:m - ~~~ nlHi::E lE~ililJJ ~oQ('0-C, in: Kimura Kiyotaka hakase kanreki kinen ronsha, higashiajia bukkyo -sono seiritsu to tenkai *Hjij':t:W)i:M'1lc~~ilil*, 7-/7{qjl~ - -t:-O)p;I(;3:'z:(:~OO, Tokyo: Shunjusha, pp. 47-63. Ono Genmyo +Jf3'z,:W} 1929: "Ryo Shintai yaku Konkomyo kyo jobun" ~~ ~*~~:&jt~*~ff:Y::, Butten kenkyu 1qjlA1iJfJi: 1/2, p. 5.

Ono Katsutoshi +JfJlJ;<if. 1988: "Sh6gozo no Shatenno hannya haramitsu kyo no kyojo ni tsuite" ~~(sic!)~O) 1JlJ;<.:R::E*5!:~~~IHt*~J O)*~ff:(;:"':). \; '-C, Nanto bukkyo m'l:fj){qjl~ 9, pp. 48-67. Otake Susumu
~JJ O)p;I(;3:'z:F,,~~,


2001: "Kongosen ron no seiritsu mondai" W:&1lliJ1J1ilJ Bukkyo shigaku kenkyu 1b~5t:~1iJfJi: 44/1, pp. 49-70.

Sakamoto Yukio j:jZ*~~ 1935: "Bussha ron kaidai" 1qjl'r~Hilil~~~, Kokuyaku issai kyo, yugabu, jaichi ~~-W*~:lIl11JDff~+-, Tokyo: Daito shupp ansha, pp. 255-268. SaW Mitsuo{6::$iWd$ 1963: Genshi bukkyo kyodan no kenkyu )Jj'l:fril1qjl~~[I;! 0)1iJfJi:, Tokyo: Sankib6 busshorin. Shizutani Masao lW:lEd$ 1942: "Funan bukkyo ko" :jJ(m1qjl~~, Shina bukkyo shigaku ~Jj~{~~5t:~ 6/2, pp. 12-37. Su Gongwang i.Wf~~ 1978: "Zhendi sanzang yishu kao" ~~=.~~~, in: Zhang Mantao ~~{if; et al. (eds.), Fodianfanyi shi lun 1qjlAiTI~5t:~ilil (Xiandai Fojiao xueshu congkan :~.I1J\";{qjlWl~#T~flj 38), Taibei: Dasheng wenhua chubanshe 1978, pp. 67-108 (first published in Weimiao sheng 1%9: fr!};fjj 2, pt. 2-6, 1937, and again as Zhendi sanzang nianpu, fu yishu kao ~~*=.if.~ , {t~~~, Beijing: Beijing foxue shuju chub an 1940). Takakusu, Junjiro, 1904: "The Life of Vasu-bandhu by Paramartha (A.D. 499-569)," Toung pao, Serie II, Vol. V, pp. 269-296. -1905: "Study of Paramartha's Life ofVasu-bandhu; and the Date ofVasubandhu," Journal of the Royal Asiatic Society 1905, pp. 33-53. Takasaki Jikido ~PIiiJ:Ji 1979: "Shintai sanzo no yakukyo" ~~=.O)~ *~, in: Mori Mikisaburo hakase shaju kinen: Toyogaku ronsha **=W= Jj!~t:W0Jr~'1lc~ J!~j'$~~ilil*, Kyoto: Hoyu shoten, pp. 1109-1125. 1981: "Shintai sanzo no shiso" ~~*=~0),1j!;,1~, in: Daijo bukkyo kara mikkyo e: Katsumata Shunkyo hakase koki kinen ronshU**{qjl~iJ'GW ~.r-.. - JlJ;<y.{&~t:w~m-'1lc~~*, Tokyo: Shunjusha, pp. 701-716. 1989: RoskO ron.tUilil, Tokyo: Kodansha.

The work of Paramartha


- 2005: "BusshO ron kaidai" {5t'ti~~~M, in: Takasaki Jikido and Kashiwagi Hiroo fs*iJU$, BusshO ron. Daijo kishin ron (kyu shin ni yaku) {5t'ti~iiil j;JI~~{~~iiil. (jifJT=~i), Tokyo: Daizo shuppan, pp. 15-64. Tsukinowa Kenryu Jl ifml~~.i 1971: "Kugyo ichijo hosM ron ni tsuite" J'E ~-*.ti~f;:gtT, in: Butten no hihanteki kenkya {5t~O):ltt*IjiJ1iJfJ'E, Kyoto: Hyakkaen, pp. 364-381 (first published in Nippon bukkyo gakkai nenpo r:pjs:{5t~~ff<a 7, 1934). Tucci, Giuseppe, 1929: Pre-Diriniiga Buddhist Texts on Logic from Chinese Sources, Gaekwad Oriental Series 49, Baroda. Ui Hakuju "f:#{sii 1935: "Sho daijo ron giso no dampen" t.fi**~~i91E0) IIYTJl, in: Ui Hakuju. Sho daijo ron kenkyu t.fi**~iiil1iJfJ'E, Tokyo: Iwanami shoten, pp. 117-130. 1965: Indo tetsugaku kenkya, dairoku FPm:fg~1iJfj'Emj;;", Tokyo: Iwanami shoten (first published in Tokyo: Koshisha shobo, 1930).

Xu Ming ~'F~ (ed.) 2002: Zhongguo Fojiao jinglun xu bajiji, 1 ~ff:lI!~[.~(-), Shanghai: Shanghai cishu chubanshe.

Yabuki Keiki 9;;:~ 1933: Meisa yoin kaisetsu B!1H:J>~lt~~~~, Tokyo: Iwanami shoten. Yamaguchi Hiroe 0J D iJMI 2004: "Yuima gyo bunso shoin no Fushu kyo ni tsuite" W*I~*,lEJt~U fflIO) r1iff~*,lEJ f;:"":)v \T, Indogaku bukkyogaku kenkyrt FPm:~{5t~~1iJfj'E 53/1, pp. 116-123. Yamazaki Hiroshi 0JJlJi'tr5.t: 1942: Shina chrtsei bukkyo no tenkai jztf~.pi:!t{5t ~O),@OO, Tokyo: Shimizu shoin. Yoshimura Makoto 'Ei"tIDiX; 2002: "Shoron gakuha no shinshiki setsu ni tsuite" :tlmJjj~*O)JL.'~~~f;:"":)v \T, Indogaku bukkyogaku kenkyrt FPm:~ {5t~~1iJfj'E 51/1, pp. 61-65. 2003: "ShOron gakuha no shinshiki setsu ni tsuite" :tlmJjj~*O)JL.'~~f;: "":)v \T, Komazawa daigaku Bukkyogakubu ronsha foJ?~*~{5t~~*G~ ~ 34, pp. 223-242.

Yoshizu Yoshihide 'Ei1It1t~ 2003: "Shintai sanzo yakushutsu kyo ritsu ron kenkyu shi" ~mli=ik:~tf:l*,lE~~iiil1iJfj'E~, Komazawa daigaku bukkyogakubu kenkyrt kiyo foJ?~*~{5t~~*G1iJfj'E*[.~ 61, pp. 225-285. Zhu Xie *{~ 1936: Jinling guji tukao ~~i'~~.llll~, Shanghai: Shangwu yinshuguan.

GUJ;labhadra, Baoyun, and the Sarpyuktagama1 Andrew Glass

Introduction The only complete version of the Sa:rp.yuktagama available in Chinese is the Za ahan jlng 5Mf~~~~ in 50 rolls (juan off, T 2 no. 99). The main facts regarding this translation are not in dispute, namely, that the Indian monk GUIfabhadra / QiunaMtu6lu6 :5:f(J3~~j(:~B~ (394-468) recited the text for the Chinese monk Shi Baoyun ~ (376-449) to translate during the period 435 to 443 in Nanjlng then Jiankang 9t~, the Capital of the ne established LiuSong ~U* Dynasty (420-479). This version of the Sa:rp.yuktagama is considered to be a Sarvastivada recension based on similarities between the translation and suriving Sanskrit fragments of this s~ltra collection and quotations and commentaries in other extant sources (Mayeda 1985-7). Other details regarding this translation are less clear. 2 One problem is the specific location of the translation activity, whether it was done at Qihuan temple 1[b][~ or at Waguan temple n:g~. The available sources differ on this point. A second, and more interesting problem, is the source used for



1 I would like to thank the organizers and participants of the symposium on Early Chinese Buddhist Translations held at the Institut fUr Kultur- und Geistesgeschichte Asiens, Osterreichische Akademie der Wissenschaften. I am particularly grateful to Zhangcan Cheng, Max Deeg, TofU Funayama, Zequn Ma, and Stefano Zacchetti for their assistance with this paper. 2 One problem with this translation that has been largely solved is the disorder in the sequence of the rolls. For a summary of the scholarship on this see Glass 2007a: 39-42.

Journal of the International Association of Buddhist Studies

Volume 31 Number 1-2 2008 (2010) pp. 185-203


Andrew Glass

the translation of this text. Did GUI,labhadra read out the text from the manuscript which Hixian 5~r.~ obtained in Sri Lanka around 410/411 or did he use another manuscript, or did he recite it from memory? Inconsistencies in the accounts of the extant catalogues have caused confusion over this point and fertilized academic debate for the past eighty years. In this paper, I offer an explanation that attempts to reconcile the differences between the sources regarding which temple hosted the translation work. I also consider the problem of the translation source, and hope to convince the reader that in the absence of concrete evidence which could put an end to the debate, the weight of circumstantial evidence falls heavily in favour of the source being Fiixiiin's manuscript. I will also show that the main arguments that have been used to dismiss Faxian's manuscript from being the source do not stand up to scrutiny. The location of the translation of the Sa1flyuktagama Prabodh Chandra Bagchi was the first scholar to identify the discrepancy in the location of the translation work. 3 He quoted two opposing reports concerning GUI,labhadra's translation of the Smpyuktagama but did not pursue the problem. These reports state:
In the 12th year ofYuanjili (= 435) he [Gulfabhadra] reached GuangzhOu ... at first he lived at Qihuan temple ... At Qihuan temple he gathered many scholar monks and translated the Za liMn jIng.4 Za liMn jIng, 50 rolls: translated at Wagulin temple. 5

The source for the first account is the Chfi sanzang jlji t:IJ =~[.~ which was compiled by Sengybu 1~tti (445-518) in about 515. This source is widely regarded as the most reliable extant catalogue of the early translations. It would, therefore, be easy to dismiss this problem since the source of the contradicting report is the Udai
Bagchi 1927: 382.
0 0


1)]{H[:bg~ o




CSJ 105c6-14. LSJ 91a24; DNL 258c12.


Gur:tabhadra, Baoyun, and the Sarpyuktagama


sanbao j'i lrHt - .~C, a catalogue prepared by Fei Zhiingfang Ji:-t!f in 597, and which is held in rather less esteem by modern scholars. However, there is some additional support for the facts given in Fei's report:

Gur:tabhadra arrived in Jia.nkang in the 12th year of Ywinjia (435) and was ordered by the emperor to live at Qihuan temple; until the 20 th year of Yuanjia (443) he worked on translations at Waguan temple in Jiankang. 6

This account comes from the GujIn yijIng tuji c!J~~'~~IIl~c, compiled by Shi J'ingmai *'ft~~ in 664-665. Qfhuan temple and Waguan temple were both located in the Sanjlng -it district of Jiankang, and were probably at most about two kilometres distant from each other.? For Gm;J.abhadra, a man in his early forties who had travelled from India to China by way of Sri Lanka, this must have been within easy commuting distance. Therefore, the details in this account are at least plausible.

The source of the translation of the


The source of GU:Q.abhadra's translation of the Sarp.yuktagama is not specified in the account given in the CM sanzang jijf (see above). This omission has led to considerable debate; since, if Faxian's manuscript of this text had been used, some modern scholars feel that Sengybu would have mentioned it. 8 On the other hand, if GU:Q.abhadra had provided the source, or recited it from memory, this might equally have been mentioned.

tz*::x:*~~+ :q::~1(~i~JJ'b~ *)~mZ*iJ{tEh:~ ~*~~

0 0

:q::~::x#* D<'fJ'b11fG1Lg~~ GYT 362b4-6.

0 0


? Today there is a new temple next to the site of the old Waguan temple, which burned down at the beginning of the Ming dynasty ({~~*i'U #!!" S.v. 1Lg~, accessed from the China Buddhism website http://www. cnbuddhism.com/cidian/ShowArtic1e.asp? Artic1eID=42478, 31 October 2007). The exact location of Qihuan temple is not known, but was in the same district (Lu 2002: 251). 8 Mizuno (1988: 8), Enomoto (2002: 37), and Nagasaki and Kaji (2004: 46).


Andrew Glass

It has long been known, however, that the Ud~ii sanbao ji specifies that the source of the Sarp.yuktagama translation was'Paxian's manuscript:

Za ahanjIng, 50 rolls: translated at Waguan temple, Fiixian brought it

back. Seen in Daohui's Songqf catalog. 9

of the Za ahlin jlng in fifty rolls, as such a translation would have had to have passed otherwise undetected into obscurity; and further, the fact that GUIJabhadra worked on the same text very dose by (as mentioned previously) must preclude such a hypothesis. The crux of this debate therefore, amounts to whom to believe; does the eM sanz~mg jiji's silence imply Phian's manuscript was not the source; or does the Udai sanbao ji actually contain some facts not reported by the earlier source? Perhaps more important than these two reports is the subtext of the debate: how could a copy of a Sarvastivadin Sarp.yuktagama have been made in Sri Lanka at the beginning of the fifth century? I suspect this problem has determined the shape of much ofthe debate more than the matter of whom to believe. io Several new arguments have been put forward in order to advance the view that GUIJabhadra provided the source. Most of these try to read between the lines of the sources cited above with the aim of detecting new evidence. Those who accept Paxian's manuscript as the source have largely been content to accept the Udai sanbao ji and have not gone into further detail. In order to move the debate forward we must consider other details that relate to the

It does not seem possible that this could refer to another translation

9 *lfliJJ'i3-~1i+;ffiD1:ECg~~ $f.~li* ~lliJf~~*~$;R LSJ 91a24; DNL 258c12. 10 Scholars who reject theLSJ version, and therefore claim the source to be other than Fiixian's manuscript include YinsbUn (1983: 1, 3), Mizuno (1988: 8), Enomoto (2002: 37), and Nagasaki and Kaji (2004: 46). Those who accept that the LSJ may be correct include de Jong (1981: 108), Tsukamoto (1985: 439), and Tseng (2000: xxviii-xxx). Akanuma (1939: 51 n. 8) and Demieville (1953: 418) were aware of the issue but did not commit to either side.
0 0 0

GUIJ.abhadra, Biioyun, and the SarpyukHigama


problem.!1 In this respect, I would like to pursue two questions: was there an opportunity and motive to use Faxian's manuscript? And conversely, was there an opportunity and motive for GUJ).abhadra to have provided the source? Faxian's manuscript To discover if there was a motive and opportunity for Filxian's manuscript to have been used we must consider the people involved, and their histories. As is well known, the person identified as the translator (yz g~) is often not whom we would regard as the translator in the usual modern sense of the term. This is true in the case of GUJ).abhadra's Za aMn j"ing, as we learn from Sengyou: The Indian Mahayana Master GU.Q.abhadra ... recited the texts. The monk Shl Baoyun f~Jf~ (376-449) and the disciples Putf iftlt (Bodhi) and Fayong 5~~ (Dharmodgata) interpreted them. 12 According to the same source GUJ).abhadra had not long been in China when he began work on this text, so we can easily accept that Baoyun, Bodhi, and Dharmavlra were responsible for the actual work of translation. The most important of these interpreters is Baoyun, whose biography is recorded in the CM sanzang jiji (113a5-b2). This biography also appears in the Gaoseng zhuan ~{~{$: (339c18-340a14) by HUljiao ~BX (519) with minor differences.13

11 Nagasaki and Kaji did investigate the relationship between the traps.. lation team that worked on the Za ahcin jlng and Fiixian, and identified the connections, but they did not as~ert their findings in their final conclusion (2004: 38-45). 12 *':!::~~DJ*5~gffl*f:j~JWW8H ...' Il~~!f )j>r~f~Jf~.&5I3Tiftlt5~ ~{t~~ (CSJ 13a6-8). Sanghavarman is said to have had an "eminent disciple" called Bodhi (1$)E5I3Tiftlt), who may be identified with Baoyun's assistant on this translation. This fact is recorded in the fragmentary citations of a lost work of Sengyou, Sap6duo shlzI zhuan i\i~~gfflji{t (Funayama 2000: 349; forthcoming). 13 One difference worth noting is that Huljiao seems to have known that Baoyun died at the age of 74 (-f::+:fiI21J, GSZ 340a13) whereas Sengyou


Andrew Glass

That fact that Baoyun was the primary translator of OUIfabhadra's Za aMn jlng is central to this investigation, because Baoyun travelled with Hixian through Central Asia, as far as Puru~apura (modern Peshawar). Details of their journey are provided both in the Gaoseng Faxian zhuan iWi{~5~r.~{$ (T 51 no. 2085) and in Baoyun's biography. Unfortunately Baoyun's own account of his travels has not survived. An outline of their journey based on these sources follows: Baoyun was probably born in 37614 in Liangzh6u 5~Hr15 (present-day western Gansu, to the north west of Lanzhou). This means he would have been about 24 when he met Faxian in Zhangye ~ t-& (a city in LiangzhOu) in 400. 16 Fiixian would have been some years older - perhaps as little as one year or as much as 15 or so, but almost certainly Faxian was not 63 at the time (i.e., the age traditionally ascribed to him)P Faxian tells us that he and his four companions met Baoyun and four of his friends in Zhangye. The ten of them had in mind to travel to the West, and so they spent a happy summer together anticipating their journey.1S Of this journey, Baoyun's biography tells

was not so exact, giving his age as "70 something" (t:+e~, CSJ 113bl) unless he made this up! 14 Calculated from his age at his death in 449, provided in GSZ .r~)j:c~ - +7\lf~b~rlJ~ ~f-*t:+~J2] (340a13). 15 CSJ ~j'ltAm (113a06); GSZ ~j'ltA (339c18). 16 GFZ 857alO-12. This date is based on Legge, who determined the year of Fiixian's departure based on the GFZ, and the biography of Faxian in the GSZ (Legge 1886: 9; also Deeg 2005: 23-4). In Baoyun's biography the date is given as l-~i~L1J] (GSZ 339c22) which refers to the beginning of the period 397-402. 17 Faxian's dates are uncertain and problematic. Traditional dates for him are 337-422, but this means he would have been 63 when he crossed the desert to Khotan and the Karakoram to Skardu, which seems quite unlikely. Legge suggested he may have been 25 when he went to India (Legge 1886: 3); Deeg suggests he may have been a little older, perhaps thirty or forty (Deeg 2005: 29). 18 GFZ 857alO-12; Legge 1886: 11; Deeg 2005: 496-7.
0 0

Gur;tabhadra, Baoyun, and the Sarp.yuktagama


us simply that they "walked across the Taklamakan and climbed over snowy mountains, [they] struggled with sufferings and dangers without thinking it difficult, and reached Khotan."19 We get rather more detail from Fiixian, according to whom ten of them went as a group as far as Dfmhwlng 9:J~, whereupon Fiixian and his friends went on ahead via Shanshan NB~ to Yanyi ~~, also known as Sorcuq, where they rested for two months. During this time they were rejoined by Baoyun and his companions. 20 From there, seven of the travellers, including Faxian and Baoyun set out for Khotan across the Taklamakan Desert. The journey took one month and five days, concerning which Faxii:'in tells us: "The sufferings they endured were unparalleled in human experience."21 The distance from Yanyi to Khotan is about 600 miles (1,000 km). To have walked that distanceinjust over a month would mean they must have been walking about 18 miles (30 km) per day. After more than three months in Khotan Fiixian, Baoyun and two other companions continued on their journey, crossing the Karakoram mountains to reach Skardu, where they met up with Huijing and two others who had gone on ahead from Khotan. The seven travelled together as far as Udyana (Wilchang }~r~), where Huijing and his two companions went on ahead again. After spending the summer in Udyana, Faxian, Baoyun and two others continued South visiting Suhata? (Slihedu6 1'ITEJ7) , GandhavatI (Jiantu6wei ~IW81~J), Tak~asila (ZhlichashIlu6 ~~tUP~), and finally Puru~apura (Ful6usha JtfJ5j')Y It was perhaps the autumn of 401 by the time they arrived. The purpose of this summary is to point out that Baoyun travelled with Faxian for about one year through extremely dangerous and testing terrain. After their two groups reunited in Yanyf,

19 ;&;J1i;;'1Uj>:<:~trJ~~JVlitJEiNfct/FJ;)~il ~J1lTIil (CSJ 113all-12; GSZ 339c23-4). 20 GFZ 857a12-28; Legge 1886: 11-5; Deeg 2005: 497-500. 21 pJT~~EAf1l[~tc GFZ 857b3; Legge 1886: 16; Deeg 2005: 500. 22 GFZ 857bl-858b12; Legge 1886: 16-33; Deeg 2005: 501-12.


Andrew Glass

FaxHin and Baoyun stayed together while their other companions came and went. To have undertaken such a testing journey together would surely have made them either close friends or bitter enemies. The fact that they subsequently worked together in China, suggests it was the former. Faxian's biography tells us that Baoyun and Sengjing returned to China while Fiixian went alone to HaMa to see the skull-bone relic. Baoypn's biography tells us that while in India, he studied the local language before returning to China.
[Bao]yun, while in the foreign lands, studied the foreign books extensively. He became thoroughly accomplished (J[~,\f) in all the sounds, scripts, and exegesis of the countries of India. Afterwards he went back to Chang'an.23

We do not know exactly how long Baoyun stayed in Gandhara, but it must have been long enough to give him a good start in Sanskrit. That he returned to CMng'an is also interesting since he was not from there. Perhaps this was Fiixian's suggestion, maybe they planned to meet there, or maybe it was just the obvious place to go for a monk interested in translation at that time. While in CMng'an Baoyun met and worked with Buddhabhadra. 24 When Buddhabhadra was expelled from CMng'an by Kumarajlva's followers, Baoyun and his friend Hulguan ~fi went with him. First they travelled to Mount Lu mtLlJ and then, toward the end of 412, they continued on to Jiankang and took up residence
23 ~l:9~J~' ji~t!iFJif 7(~~~-N*~M/lI' ~~J[*,\f i!t}@-&'ft:0 (CSJ 113a13-4); Tsukamoto 1985: 439. The Chil sanzang jiji reads husha t!iFJif, where the Gaoseng zhuan has fanshu Jitif. It is tempting to follow Dan Boucher's suggestion regarding Msha t!iFJif (Boucher 2000), and understand that Baoyun studied Kharo~thI, however, Baoyun was in Gandhara about 100 years after Kharo~thI fell out of use in that area (Salomon forthcoming; Glass 2007b: 72), so this most likely refers to Sanskrit or Hybrid Sanskrit books written in BrahmI. 24 CSJ 113a15; GSZ 339c27. Buddhabhadra had travelled from Kashmir to Chang'an with Zhiyan, who was one of Baoyun's companions on the journey to Turfan. Zhiyan and two others left the main group there and later reached Kashmir.
0 0

Gm;tabhadra, Baoyun, and the Sarp.yuktagama


at Daochang temple m:f;~~. 25 At about the same time, FhHin returned to China, and having heard of the problems in Chang'an, went directly to Jiankang. There, Faxian and Baoyun were reunited, ten years and almost 3,000 miles from where they had parted ways. Faxian also took up residence at Daochang temple and together with Buddhabhadra, Baoyun and Hulguan, they produced numerous translations. The working relationships are documented in the catalogues, for example, "II [Faxian] demanda au maitre de Dhyana du pays etranger Buddhabhadra, de traduire et de publier, dans Ie Tao-tch'ang sseu, Ie Mo ho seng k'i liu ~~DJ{~tBJRW;"26 and "The Dhyana master Buddhabhadra held the foreign book (MahfiparinirvfilJasutra], Baoyun translated."27 Further support for the connection between Baoyun and Faxian during this period can also be found in the texts themselves. Max Deeg has recently reported that GUl).abhadra's translation of the Smpyuktagama contains some terms which follow Fhian's transliterations; he gives as an example Pali ghosito gahapati > qushlluo' zhlingzhe .gffimft~. 28 The first occurrence of this name and title comes in Fhian's translation of the Mahayana MahfiparinirvfilJasutra. The same transliteration appears fifteen times in the Za ahan jIng.29 The reason for this connection must be Baoyun who, as men25 Tsukamoto 1985: 453, 884. 26 Shih 1968: 114. 27 t!jlgffl{~*~.ftW8' -=F-AtIi,* jf~{$~' CSJ 60blO. The Udai sanb1lo
0 0 0

ji has a slightly different report of the translation of this text: "an old catalogue says Buddhabhadra recited this text, and Baoyun held the brush" $.,-RE li::i':ili jf~~~ LSJ 71b7. 28 Deeg 2005: 485-6. 29 E.g., T 2 no. 99, e.g., p. 117c24. This phrase also occurs in three other works of this period: Dharmak~ema's version of the Mahfiparinirvfil!asL7tra (T no. 374) in 421; Buddhajlva's translation of the MahIsasaka Vinaya (T no. 1421); and Hulyan @~, Huiguan @WL and Xie Llngyun's UJm~ re-edition of Dharmak~ema and Faxian's versions of the Mahtiparinirvfil!asL7tra (T no. 375), prepared in Jiankang (NanjIng) and dated broadly to the Yuanjia era (424-52). It is interesting to note that Huiguan had served as scribe for GUl].abhadra's Za ah:in jIng, while his friend Huiyan had done the same for BuddhajIva's MahIsasaka Vinaya,
0 0 0


Andrew Glass

tioned above, was involved in the production of Fiixian's translation of the Mahaparinirvar:zasutra and also translated Gu~abhadra's recitation of the Salpyuktagama into Chinese. Since Fiixian wrote the Gaoseng Faxian zhuan while living at Daochiing temple, we may assume that his manuscript of the SaIp- . yUktagama had not been lost on his journey from Sri Lanka to Jiankang, as he would probably have mentioned such an important detail. The fact that Baoyun and Fiixian lived and worked at the same temple from 413 to about 422 shows that Baoyun would, in all likelihood, have had access to Faxian's manuscript of the SaIpyuktagama. Therefore, we can deduce that GUl:,labhadra's translation team, which included Baoyun, would have had the opportunity to make use of Fiixian's manuscript. The next thing I wish to show is that there was a concerted effort to translate those manuscripts which Faxian had brought back with him. This effort began soon after Fiixian's return and extended into the period following his retirement from translation work. 3D It seems to have continued as long as his colleagues, especially Baoyun, were active. According to his own account, Faxian obtained the following manuscripts during his journey to the West.
In Pa!aliputra (GFZ 864b19-28; CSJ 112a20-1): The Mahasanghika Vinaya / M6M sengqi zhOng ~~tiJ{~1~,w :f$: (T 22 no. 1425) The Sarvastivada Vinaya / Sap6duo zhong ~~~,w1$: . The *Sarrtyuktabhidharmahrdaya / Zci ap(tcin xfn ~1ltiJ~ftJ~\ A satra, Yan jlng ~Jjf~J[ The first chapter of the VaipulyaparinirvalJasatra / Fangdeng bannfhucin jfng ]J~~~)Jb)g~J[ The Mahasanghika Abhidharma / M6M sengqi apitan ~~tiJ{~




and that the two collaborated in the re-edition of the MahaparinirvalJasatra. 3D See Deeg 2005: 27-8.

GUIfabhadra, Baoyun, and the Sarp.yuktagama In Campa (GFZ 864c8): Unspecified siitras


In Sri Lanka (GFZ 865c24-5; CSJ 112a26): The Mahfsasaka Vinaya / Mfshasai III jj5Y~f=l: (T 22 no. 1421) The DIrghagama / Chang ahan jlng -&1loJ~m, The Sarpyuktagama / Za ahan jlng *lE1loJ~m, The *K~udrakapitaka31 / Zazang jlng *lEiJ'~U~ (T 17 no. 745)

In 416, Faxian and Buddhabhadra translated the manuscript of the MahasaIighika Vinaya / M6hesengqf lu *~a;1~1t!;~ (T 22 no. 1425).32 In 417 they began work on the Mahiiparinirviil}asutra / Da biinniyuan jlng *~~5Jb5g~ (T 12 no. 376). Faxian tells us he obtained the first chapter of this text in Pataliputra, but this translation mayor may not be connected with that manuscript. We learn from Sengyou and the Gaoseng zhuan 33 that Fiixian had Buddhabhadra read out (~~tD this text. They also translated the *K~udraka-pitaka (T 17 no. 745); the Yan jlng ~,i~; and the *Sa1Jlyuktiibhidharmahrdaya *lEJloJ~E(Jt)G,,34 the last two of which had been lost by the time of the Kaiyuan shijiao III Ifflj[;f~~~i (730) and probably much earlier.3s The first of Faxian's manuscripts to be translated after Fiixian's "retirement"36 was the MahIsasaka Vinaya translated between 423 and 434 by Buddhajlva and ZM Daoshi'Sng ~.ill1: at L6ngguang

31 ]~)~1~jj5Y~f=l:~2fs:' 1~-&IloJ~*lEIloJ~' ~1~-gB*IE~ (T 51 no. 2085 p. 865c24-5). See also Tsukamoto 1985: 436-7, and Deeg 2005: 572. 32 KSL 505b27; Lancaster 1979, K 889; Deeg 2005: 561 n. 2455. 33 GFZ 864b27; CSJ 60b2-1O; Bagchi 1927: 348. 34 CSJ 112b20; Another part of Sengyou's work has the comment "Yan jlng (Sanskrit, not translated)" ~Jjfm, (JiI::X*~ID (CSJ 12a3); however, at least one, and possibly two more texts from this section of the CSJ are similarly labelled but are known to have been translated. 35 Bagchi 1927: 348. According to Pelliot "cette traduction etait deja perdue vers l'an 500" 1930: 272. 36 See Deeg 2005: 27-8.


Andrew Glass

temple R~:YC~ in Jiankang. 37 The fate of the remaining m;muscripts is not spelled out in the catalogues. We do know, however, that Baoyun worked on a translation of the SarJ1yuktiibhidharmahrdaya in 433 or 434 with Sanghavarman. 38 The same Sanghavarman is credited with the translation of the Sarviistiviidavinayamiitrka / Sapoduobu pin! modeleqie ili?J:~gG~Jb"1~~JJWD (T 23 no. 1441) done in the following year, 435. Sanghavarman's connection with Baoyun provides the opportunity to have had access to Filxian's Sarvastivada Vinaya manuscript (ili?J:~~1')' The fact that Faxian tells us his manuscript was 7,000 verses long, and Sanghavarman's translation is also 7,000 verses long, adds weight to this idea that the latter may be a translation of the former. 39 Therefore, if we ignore the unspecified s(itra obtained from Campa, only three of Faxian's manuscripts were left untranslated when GUJ?-abhadra arrived in 435: the Mahasanghika Abhidharma, the Dlrghagama, and the Sa:q1yuktagama. The case of the Dlrghagama is easily explained as this text was translated from another source by Buddhayasas and ZhU Fonian ~{~~ in CMng'an around the time of Faxian's return. Even though this translation was done in another city, knowledge of that translation would have spread to liankang as there was frequent contact between the translation centres. 40 The case of the Mahasanghika Abhidharma is different, as no other version was available, and this text cannot be connected with any translation done since. We must conclude in this case that

37 CSJ 12b3; GSZ 339a6-8; Lancaster 1979, K 895; Bagchi 1927: 364; Kamata 1998: 383. 38 The circumstances of the translation of this text are confused, and it is uncertain whether this was a new translation of the same text that FiixHln obtained in Pataliputra (Le., the Za'apftan xln) , and which was translated by Fiixian and Buddhabhadra, probably in association with Baoyun; or whether this was a separate text entirely. The details of this situation are described in Dessein 1999: lxxvii-lxxxii. 39 On the length of Faxian's manuscript see GFZ 864b23-4 == Deeg 2005: 561; CSJ 21a18. For the length of T 23 no. 1441 see Kamata 1998: 389. 40 See Tsukamoto 1985: 440.

GUJ;tabhadra, Baoyun, and the Sarpyuktagama


it was not translated but it is impossible to guess exactly why this was so. When GUJjabhadra arrived in Jianhing, the Sa:qlyuktagama would have been the most important work in the colle<;;tion of Faxian's manuscripts that had not yet been translated. As shown above, Baoyun would have had access to this manuscript, and further, he may well have had an interest in seeing this manuscript translated out of a sense of loyalty to his former travelling companion and colleague. It is easy to imagine that Baoyun could have persuaded GUJ!abhadra, a man eighteen years his junior, to recite the Sa:qlyuktagama for him to translate when the latter had only just arrived from India.

GU1}.abhadra's source
According to the biography given by Sengyou,41 GUJ?abhadra was born into a Brahman family in North Central India = Madhyadesa). He is said to have converted to Buddhism after encountering the *SaY(lyuktiibhidharmahrdaya (JraJ~tt*l)l,\), then, not satisfied with mainstream Buddhism (lj\*), he went on to study under a Mahiiyana master. Like Faxian and others before him, he went to Sri Lanka, and onward by boat to China. After arriving in GuangzhOu .1'f'f the monks Huiyan and Huiguan (an associate of Filxian and Baoyun) were ordered to go to meet him and take him to Qfhuan temple 1E\;5g~. The first text he is said to have worked on after arriving in Jiankang is the Sa:qlyuktagama. His biography does say that he had mastered the Tripitaka (t~ :W -~, CSJ 105b23), but this does not mean that he was capable of reciting the entire canon from memory. Certainly memorization is a well known feature of Indian learning, and such learning might well have been part of his training, but we do not know if this included memorizing the Sa:qlyuktagama. As mentioned above, Sengyou reports that GUl!abhadra was interested in the SaY(lyuktiibhidharmahrdaya and the Mahiiyana, and he is known to have


41 CSJ 105b17-106b21 and GSZ 344a5-345a23.


Andrew Glass

worked on translations of several important Mahayana te~ts. 42 If he had memorized the whole of the Sarpyuktagama in particular - a text almost equal in length to all of his other translations combined - this might well have been mentioned. Therefore, while it is perhaps conceivable that GUl;mbhadra could have provided the source of the Za aMn jlng from memory, there is reasonable doubt that this was so. We also learn from GUl).abhadra's biography that he was familiar with writing and using written texts, for example "the Mahayana master tested [Gul).abhadra], ordering him to take out [a text] from the sutra bOX''43 (i.e., a box containing written texts); "then [Gul).abhadra] read out the commentaries."44 Therefore, he could have brought a manuscript of the Sarpyuktagama himself. But such a position seems doubtful. In Fiixian's case, he went to India with a plan to gather manuscripts. He must have known at the time he left China that there was no complete translation of the Sarpyuktagama in Chinese, therefore we can see a clear reason for him to have obtained a copy of this text during his journey, and his own travel account and other biographies make it clear that he did obtain a manuscript of this very text. The same is not true for GUl).abhadra. GUl).abhadra would not have known the Sarpyuktagama was needed in China and he not did he have an obvious interest in this text. Furthermore, his biography does not mention that he brought any manuscripts with him. Therefore, it is unlikely that he would have brought a manuscript of the Sarpyuktagama himself. As seen above, we know that GUl).abhadra was literate, therefore, he would have been able to read Faxian's manuscript of the SarpyUktagama and explain the details for Baoyun to translate. Even if GUl).abhadra had been a specialist in this text, it is also quite likely that he would have made use of Faxian's manuscript; just as Buddhajlva, a Mahisasaka monk and specialist in the Vinaya, did when
42 These include, among others, the Srfmala(devf)siY!lhanadasL7tra (T 12 no. 353), the LaJikavatarasL7tra (T 16 no. 670), and the SandhinirmocanasL7tra (T 16 no. 678). 43 44

**gffl~~J;inr9:~!f1!! 0 CSJ 105b25.


CSJ 105b27.

GUI).abhadra, Baoyun, and the Sal1lyuktligama


he was asked by the monks of JUmkang to translate the MahIsasaka Vinaya manuscript that Hixian had brought back from Sri Lanka. 45 Arguments against Fixiin's manuscript The primary argument against Flixian's manuscript providing the source for the translation of the Za aMn jlng has been that is it not explicitly identified as such in the Chii sanzang jiji. The problem with this argument is that the CM sanzang jijf does not specify a different source either. We must accept that, for whatever reason, Sengybu did not have this information. Therefore, his silence regarding the source should not be taken to support either side of this argument. In an earlier portion of the same catalogue, Sengybu records a list of Faxian's manuscripts specifying that some of them, including the Za aMn jlng were not translated. 46 However the MahIsasaka Vinaya is similarly recorded and is known to have been translated, and so is the Yan jlng.47 Since the details given concerning these two texts are inconsistent with reports later in the very same catalogue, the information given for the Za aMn jlng is not reliable. As suggested earlier, one of the main perceived problems seems to have been the fact that Flixian obtained his manuscript of the Salpyuktagama in Sri Lanka. Since the translation of the Za aMn jlng is widely regarded as belonging to the (Miila)sarvastiviida tradition, some scholars have been uncomfortable with identifying this with Flixilin's manuscript since Sri Lanka is a long way from the homeland of that school. 48 However, prior to the 12th century Theraviida Buddhism did not enjoy a monopoly position in Sri Lanka. Bechert has argued that the Jetavanarama Sanskrit Inscription and other evidence suggest the presence of other schools (nikiiya). He tentatively identifies these schools as the Miilasarvastivadins, the
GSZ 339a3-13; Shih 1968: 118-9. 46 "Za liMn jing (Sanskrit, not translated)" ~jliiJ~Mf(jt)(*~) CSJ 12a5. 47 ~~.1>~1f(jt)(*~) CSJ 12a6, see also n. 35 above. 48 See for example Ylnshun 1983: 3; Nagasaki and Kaji: 46.


Andrew Glass

MahasaIighikas, the SaI!lmitlyas and the Sthaviras (Ther~vadins).49 This conclusion is supported by the fact that FiixHin brought back a copy of the MahIsasaka Vinaya from Sri Lanka. Therefore it is also quite possible that he obtained a Sarvastivada manuscript of the SaI!lyuktagama there. . Mizuno has argued that because an audience of many monks was invited to hear GU:Q.abhadra's reading of the text,50 this indicates a new version of the text was being used rather than one that had been available for twenty years - as FlixHin's manuscripf had been by that time. 51 However, the fact that the manuscript had been in Jiankang for twenty years is no reason to suppose that its translation was any less important - after all, since the manuscript was in Sanskrit, the contents would not have been accessible to the many monks who were invited to listen to it.

The above survey of the circumstances surrounding the translation of the Za ahan jlng has shown that while there are problems connecting the translation done by GU:Q.abhadra to the manuscript brought back by Flixilin, there is ample circumstantial evidence to support this claim. Furthermore, the alternate hypothesis, that GU:Q.abhadra himself provided the manuscript, either in manuscript or oral form, is more problematic with the current evidence.

CSJ DNL GFZ GYT GSZ CM sanzang jlji tl=it~[.$ (T 55 no. 2145) m.tang neidian lil *~1*1~~ (T 55 no. 2149) Gaoseng fi1xian zhuan 11\'r{~$~{.\J (T 51 no. 2085) Gujln yljlng t6jl ~~~~!I~[' (T 55 no. 2151) Gaoseng zhuan ~{~{.\J (T 50 no. 2059)

49 Bechert 1998: 3; see also Bechert 2005: 48-9. 50 CSJ 105c13; see above n. 4. 51 Mizuno 1988: 8.

GUI}.abhadra, Baoyun, and the SaI1lyuktagama KSL LSI Kaiyuan shljiao lil j,fj51;~~~ (T. 55 no. 2154) Lldai.sanbao ji fff{-t=.~G (T 49 no. 2034)


Akanuma Chizen ~m~~. 1939. Bukkya kyoten-shi ron f~~g~5!:~. Nagoya: SanbOshoin. Bagchi, Pramod Chandra. 1927. Le canon bouddhique en Chine: les traducteurs et les traductions. Sino-Indica; publications de l'Universite de Calcutta, tome Ier. Paris: Librairie orientaliste Paul Geuthner. Berchert, Heinz. 1998. "Remarks on Buddhist Sanskrit literature in Sri Lanka from the 9th century until the end of the DambadeI}.iya period." In Paul Harrison and Gregory Schopen, eds., Suryacandraya: Essays in Honour of Akira Yuyama on the Occasion of his 65th Birthday. Indica et Tibetica: Monographien zu den Sprachen und Literaturen des indotibetischen Kulturraumes, vol. 35: 1-8. Swisttal-Odendorf: Indica et Tibetica Verlag. - - - 2005. Eine regionale hochsprachliche Tradition in Sudasien: Sanskrit-Literatur bei den buddhistischen Singhalesen. Veroffentlichungen zu den Sprachen und Kulturen Siidasiens, Heft 37. Wien: Verlag der Osterreichischen Akademie der Wissenschaften. Boucher, Daniel. 2000. "On hu and fan again: the transmission of barbarian manuscripts to China." fournal of the International Association of Buddhist Studies 23: 7-28. - - - 2006. "Dharmarak~a and the Transmission of Buddhism to China." Asia Major 19: 13-37. Deeg, Max. 2005. Das Gaoseng-Faxian-Zhuan als religionsgeschichtliche

QueUe: der iilteste Bericht eines chinesischen buddhistischen Pilgermonchs iiber seine Reise nach Indien mit Ubersetzung des Textes. Wiesbaden: Harrassowitz. Demieville, Paul. 1953. "Le bouddhisme: les sources chinoises." In Renou and Filliozat, eds., L'Inde classique: manuel des etudes indiennes. Vol. 2: 398-463. Paris: Ecole fran9aise d'Extreme-Orient. Enomoto Fumio fl::zfs:X:H!. 2002. '''Zoagongyo' no yakushutsu to genten no yurai" lI'*lEjliiJ~K!I O)~tf:ltJffi:~O)Ell* (On the translation of the SaI1lyuktagama into Chinese and its Indic origin). In Bukkya bunka no kicM
*G~ ~x: ~:{b~x:O)IDl\i t!j,fj~friiU (Basis and Evolution of Buddhist Culture: Essays in Honour of Professor Zen'o Ishigami on his Seventieth Birthday): 31-41. Tokyo: SankibO Busshorin.

to tenkai: Ishigami Zen 'a Kyaju koki kinen ronbunshil E...t~~ ~~ ~1*

Funayama Toru WtitlrOOi. 2000. "Ryo no SoyU sen Satsubatashi shiden to Todai bukkyo" ~O)f!rf.b~ 1I'~~~W~1t.dJ tFt'{-t~~ (Tang Bud-


Andrew Glass

dhism and the Sapoduo shizi zhuan compiled by Sengybu of the Liang Dynasty). In Yoshikawa Tadao 5") I [,'&7:, ed., Ti5dai no shaky6 )l!ffi;0* ~: 325-354. Kyoto: H5yti shoten. - - - Forthcoming. "GuI}avarinan and Some of the Earliest Examples of Ordination Platforms Uietan) in China." In James Benn, Iinhua Chen, and James Robson, eds., Images, Relics and Legends: Formation and Transformation of Buddhist Sacred Sites. Glass, Andrew. 2007a. Four Gandharf SalJzyuktagama Satras: Senior Kharo~thf Fragment 5. Gandharan Buddhist Texts 4. Seattle: University of Washington Press. - - - 2007b. "The Chronology of the Kharo~thI Inscriptions: A reassessment in light of recent discoveries." Gandharan Studies 1: 61-76. de Jong, J. W. 1981. "Fa-hsien and Buddhist texts in Ceylon." Journal of the Pali Text Society 9: 105-116. Kamata Shigeo ~EEa,fz;1t 1998. Daizi5kyi5 zenkaisetsu daijiten :k~~:Jl!~ R::k~$. Tokyo: Ytizankaku Shuppan. Lancaster, Lewis. 1979. The Korean Buddhist Canon: A Descriptive Catalogue. Berkeley, CA: University of California Press. Legge, James. 1886. A record of Buddhistic kingdoms being an account by the Chinese monk of his travels in India and Ceylon (A.D. 399-414) in search of the Buddhist books of discipline. Oxford: Clarendon Press. LV. Hiiiming. 2002. LUi chao da cheng 7\jjWf~. Nanjing: Nanjlng Chtibanshe. Mayeda Egaku. 1985-7. "Japanese studies on the schools of the Chinese agamas." In Heinz Bechert, ed., Zur Schulzugehorigkeit von Werken der Hfnayana-Literatur. Abhandlungen der Akademie der Wissenschaften in Gottingen, philologisch-historische Klasse, dritte Folge, Nr. 149 & N~. 154/ Symposien zm Buddhismusforschung, III: 94-103. Gottingen: Vandenhoeck & Ruprecht. Mizuno K5gen 7.l<.!lffs/;7'C. 1988. "'Z5agongy5' no kenkyu to shuppan" ~*i 1lof~~IdJ 01iJU'U::IlIW (Research and Publication ofthe Chinese Version of the SalJ2yuttanikaya). Bukkyi5 kenkya {~~1iJfY'E (Buddhist Studies) 17: 1-45. Nagasaki H5jun -&ilJiif5!)J'o\! and Kaji Y5ichi 1Jo5)-. 2004. Z5agongy5 *illof ~~I. Shin kokuyaku daiz5ky5 ifJiOO~*~~: Agonbu 1lof~'iiG4. Tokyo: Daiz5shuppan. Pelliot, P. 1930. "Les stances d'introduction de l'Abhidharmahrdayasastra de Dharmatrata." Journal asiatique 217: 267-73. Salomon Richard. Forthcoming. "Whatever happened to Kharo~thI? The Fate of a Forgotten Indian Script." In John Baines, John Bennett and

GUJtabhadra, Biioyun, and the Sarp.yuktagama


Stephen Houston, eds., The Disappearance of Writing Systems. London: Equinox Publishing. Shih, Robert. 1968. Biographies des moines eminents de Houei-Kiao: Kao seng tchoua. Bibliotheque du Museon 54. Louvain: Institut orientaliste, Bibliotheque de l'Universite. Tseng, H. Vinita. 2000. The Nidiinavagga of Siiratthappakiisinf. PhD diss., Faculty of Oriental Studies, Oxford University. Tsukamoto Zenryu ~::if';:;g;~i. 1985. Leon Hurvitz, tr., A History of Early Chinese Buddhism: from its Introduction to the Death ofHui-yiian. Tokyo and New York: Kodansha International Ltd. YlnshUn EPJII~. 1983. Za iihdnjfng lun huibiiin *IE/loJ--a-~~Rll~~i 3 vols. Taipei: Zhengwen ChUbanshe.

Experimental core samples of Chinese translations of two Buddhist siitras analysed in the light of recent Sanskrit manuscript discoveries 1
Paul Harrison

Introduction and survey of the texts Reports of the death of Buddhist philology have been greatly exaggerated, to borrow Mark Twain's famous words. Evidence that it is neither dead nor even dying can be found in the healthy audience numbers in recent years at conference panels dealing with Buddhist manuscripts. This shows that lively curiosity - one might even say excitement - has been aroused in our field by the emergence of new textual material and the philological enterprise devoted to it. Manuscripts in the British Library, Senior, Schoyen, Berlin (Bajaur) and Hirayama Collections (the list is not exhaustive), most of them coming from Afghanistan and Pakistan, have not only attracted the dedicated attention of small groups of scholars, but have aroused keen interest in a wider scholarly public, and continue to do so. Afghanistan and Pakistan - which together encompass the region now referred to as "Greater Gandhara" - are of course not the only source of these new finds: the People's Republic of China has also produced many significant discoveries, significant not
1 This paper is a re-edited version of a presentation made at the International Symposium on Early Chinese Buddhist Translations held in Vienna 18-21 April 2007. My thanks go to my fellow participants at this event, and especially to its organizer, Prof. Max Deeg, for their critical comments. Any mistakes remain my responsibility.

Journal of the International Association of Buddhist Studies Volume 31 Number 1-2 2008 (2010) pp. 205-249


Paul Harrison

least because the manuscripts emerging from that quarter (most of which come from collections in Tibet) are often complete. Here, of course, it is not entirely appropriate to speak of discoveries, since the manuscripts are not being unearthed, but merely retrieved from storage, from the shelves to which they have been consigned for centuries in the Potala and various monastic foundations. In what ways these new additions contribute to our knowledge of Buddhism is a story still in the process of being written. I do not intend to survey the whole field here, but merely pick out two texts and assess the way in which recent advances in their study might impact on our approach to Chinese translations of Buddhist texts, especially those produced during the early period. The two texts are the Vajracchedika Prajiiaparamita (hereafter abbreviated Vaj) and the Vimalakfrtinirdda (VKN). Both works are undoubtedly important Mahayana siitras, and, although they are very different animals in many ways, the history of their transmission is potentially illustrative of many aspects of the passage of Buddhist sutras from an Indic medium into Chinese. In both cases multiple Chinese translations are preserved, and in both cases the emergence of new Sanskrit copies has recently amplified and enhanced our understanding of the Indic texts. In the case of the Vaj, we can now reconstruct the entire work as it circulated in Greater Gandhara in the 6th-7th centuries. For the VKN, we have for the first time access to a complete Sanskrit text, whereas previously our knowledge of this work was based almost entirely on the Chinese and Tibetan translations. It is timely, therefore, to ask what difference these new finds might make to our approach to the Chinese translations, and what light they might throw on them. But before we set about answering these questions, let us first describe the new finds and how they relate to the previously known textual tradition. In the case of the Vaj our knowledge of the Sanskrit manuscript tradition was already quite rich. The editio princeps of the Sanskrit text was published in 1881 by F. Max Mi.iller working from four copies of the text, two from Japan (which he designated together as J) and two from China (his Ch & T). Here I quote from the description given in Harrison & Watanabe 2006:

Experimental core samples


Muller used four witnesses to establish his text: two handwritten copies of .an old manuscript preserved in the Kokiji f.ijJt~ temple in Osaka,. Japan, and two blockprints from China. Since the two manuscripts from Japan are copies deriving ultimately from the same original, they can be regarded as a single witness. That original is apparently a Sanskrit text of the Vaj discovered after the death of the eminent priest Jiun Onko ~~ax:Y6 (1718-1804) by his disciple Chido ~~I (1776-1854). This text was reproduced in fascicle 320 of the Bongaku shinryo ~*)$W!, compiled by Jiun and his disciples. In this compendium it appears that the Sanskrit text was written vertically, with Chinese equivalents for the Sanskrit words in the column to the right and a Chinese phonetic transcription to the left, followed by the Chiriese translations of KumiirajIva and Dharmagupta in the next two columns. One of the copies acquired by Muller, made by the priest Kanematsu Kuken ~r~~Ji in September 1880, contained all of this material, while the second, made by the priest Kurehito Kaishin {3Z: .A.W(>L, Qf Kokiji (presumably around. the same time), contained only the Sanskrit text, written horizontally. Together they constitute what Muller refers to in his apparatus as J. As for the two woodblock prints from China, one is a woodblock edition printed in Beijing in 1760, probably at the Songzhusi ~15t~. In this print, the Sanskrit text appears both in Lafitsha script and in Tibetan transliteration, to which has been added a Tibetan translation made at the Chos 'khor rab rgyas glin temple in Beijing by the lha bris (painter) Dam pa, working under the auspices of ICait skya II Rol pa'i rdo rje (1717-86), state preceptor during the reign of the Qing emperor Qianlong (this is M's T). The other woodblock print of the Vaj was included in a book of Sanskrit texts acquired by the British collector Alexander Wylie in Beijing, in which the Sanskrit text was engraved in the Lafitsha script and printed in red ink (this is M's Ch).

The first of the two Japanese copies, with all the material in it, MUller (1881: 16) records receiving on 15 February 1881. He refers to it as Cat. Bodl. Japan. No. 54 (it is now MS. Sansk. d. 28 in the Bodleian Library). One page is reproduced in MUller 1881, Plate 1. The second, with Sanskrit text only (one page reproduced in Plate 2), is his Cat. Bodl. Japan. No. 55 (now MS. Sansk. d. 29). Since the above description was written, I have been able to inspect these two


Paul Barrison

items in the Bodleian,2 and compare them with print-outs of digital scans ofthe original Bongaku shinryo materials still kept at K6kiji, among which there appear to be at least 8 copies of the Sanskrit text of the Vaj, complete or incomplete, with or without other materials in Chinese and so on. 3 As a result of these investigations, some of the details given in the above description now need to be corrected or refined. First, the Bodleian copies. MS. Sansk. d. 28 (Muller'S No. 54) consists of three stitched booklets, covered in blue, bearing the title Kongo hannya gyo shoyaku gosM ~!lflJUm~~~J[~ii~~1Lm and, in devaniigarf, Vajracchedikiisutra1J1 Part I-III. They are enclosed in a brown case, bearing on the outside the title MS. SANSK. d. 28: VAGRAKKHEDIKApRAGNApARAMITA-sUTRA, WITH 3 CHINESE TRANSLATIONS & CHINESE TRANSLITERATION 1880. The pages of the booklets have vertical lines printed in blue, within which the text is written in a very fine and regular hand. At the end of the text, on p. 81 of Part III, after a lengthy postscript in kanbun taking up two pages, appear some shorter notes in kanbun in black ink:
Copied at K6kiji in the middle ten-day period of September, Meiji 13 [= 1880].4 Head of Survey, India School (?),5 Kanematsu Kuken :&to
2 I thank Dr Gillian Evison, Indian Institute Librarian & Head of Research Support (Special Collections), for helping me gain access to the two IVanuscripts in question. 3 For assistance in locating these materials and securing copies of some of them I am indebted to the kindness of Prof. Sh6ryu Katsura and Prof. Motohiro Yoritomi, President of Shuchiin University. I would also like to acknowledge the help of the staff of the Shuchiin Library, where I had my first sight of these treasures and was able to examine them. 4 Someone has written 1879 in arabic numerals at the top of the page above this line of text. 5 This is a guess at Kanematsu's title. The six characters are written in a cursive hand, with the fourth especially hard to make out, but indogakk6 shirabegakari fpJj'~~IDlilm seems to be the most likely reading. I thank those colleagues who offered suggestions, even if the puzzle is not yet solved.

Experimental core samples


Acquired this manuscript copy in threy volumes on 15 February, Meiji 14 [= 188116 and finished inspecting it on the 17th. Student abroad in England, Nanj6 Bunyu.

Then, in red ink:

Finished inspecting it for a second time at 3.17 p.m. on 27 February, Meiji 14 [= l88lJ, Bunyu.

This note is followed by a poem in Chinese, written on the inside of the back cover, also in red ink, in 4 lines with 14 characters to the line (i.e. 8 lines of verse, 7 characters to the line), with a final inscription in Chinese, signed at the end by Sekka Nanj6 Bunyii of Japan. MS. Sansk. d. 28, then, is full of historical interest. By contrast, MS. Sansk. d. 29 lacks any such embellishments. It is a single booklet, with pages of thin, translucent buff-coloured paper doubled back and stitched, bearing on the title page the words "Bodleian MS. Sansk. d. 29," and vertically in siddha7[l: Vajracchedakaprajfiii-piiramitasutra7[l [sic]. The booklet has a total of 97 pages (with two sides each), but from p. 60 onwards they are blank: the text finishes on p. 59 verso at the end of the sutra, with the words vajracchedikii prajfiiipiiramitiitra7[l [sic] : samiipta7[l. There is no further text or annotation. Each page carries 6 lines of siddha7[l characters, written horizontally. On the basis of my inspection of these items, carried out during a visit to Oxford during the week of 19-23 February 2007, I was able to establish with a reasonable degree of certainty that Muller's Cat. Bod!. Japan. No. 55 (now MS. Sansk. d. 29 iIi the Bodleian Library) is a faithful hand-copy of K6kiji Text 0162, while Cat. Bod!. Japan. No. 54 (now MS. Sansk. d. 28) is an equally faithful hand-copy of K6kiji Text 0165-0167.7 The Bodleian copies are in both cases extremely accurate, and in a trial collation of many pag6 Again, 1880 is written at the top of the page. The date is that on which Miiller records receiving this copy. 7 The certainty is greater in the latter case because of the careful reproduction of the handwritten postscript at the end of 0165-0167, including its erasures and corrections. The text nos. used here are those assigned to the relevant copies in the digital scan collection held in the Shuchiin University Library.



es I noted very few errors. Knowing now where they come from, we can therefore focus our attention on their Kokiji exemplars. Kokiji Text 0162 (hereafter KI) consists of a single stitched booklet with an indigo blue cover and unruled pages of the same kind of buff-coloured and shiny translucent paper that MS. Sansk. d. 29 is written on. The cover bears the title (in black siddharJ1 in a long vertical strip which has been left unpainted) VajracchedakaprajiiapiiramitasutrarJ1 [sic], to the right of which the tiny Chinese characters Kongokyo -ili:IlQjU~~ are written. Nearer the spine, in red ink, is the inscription Yuinyo bhik~u haisha Hoju no zo, i.e. ~}(D bhik~u [in siddharJ1] fl~ )*fML~ (Respectfully copied by the bhik~u Yuinyo; the collection of Hoju). The pages are numbered 1-48 in Part I (Jo -1:) and 1-52 in Part II (Ge ~) (both within the one booklet), each part being preceded by a page bearing the title, again in siddharJ1, VajracchedikiisutrarJ1 [sic).B The pages contain the Sanskrit text, written horizontally - in lines running parallel with the spine - in black ink, three lines to the page, accompanied by pronunciation in Chinese characters (also in black) and a word-for-word Chinese translation or, more accurately, gloss on the Sanskrit (in red).9 In the Sanskrit text word division is marked by red dots, with horizontal red strokes linking the ak~aras of each word. Occasionally ak~aras have been corrected or cancelled, again in red; in some cases the correction is written above, in oth8 The break between Parts I & II falls in the middle of a word in 14b:~avakaZpaya - mi. Since there is almost half a page left blank after

avakaZpaya at the end of Part I, this can best be explained as the result of copying from another copy in two booklets in which the scribe ran out of space at the end of the first booklet. This is indeed the situation with K6kiji Text 0075-0076, where the text is broken at exactly the same point but the last page of Part I is full to the last character on the last line (only then does it make sense to break in the middle of a word). On this evidence 0162 appears to be a later copy, but it is not yet clear whether 0075-0076 was the original from which it was made. Further detailed text-critical work on all the K6kiji copies of the text to determine their stemmatic relationships is a desideratum. 9 The Chinese equivalents given are often similar to those used in Dharmagupta's translation (see below), but not identical with them.

Experimental core samples


ers the original ak~ara has been modified (as when -a is corrected to -a). The section divisions of Kumarajlva's translation are also inserted in red. The text covers approx. 100 pages, and finishes on 52 recto in Part II. On 52 verso there is a colophon in red ink, and on the verso of the next page (not numbered), there is a considerable amount of additional colophon text in black. 1o This copy was apparently based on one made in 1838 by Senkai {}F,j (1786-1860). Kokiji Text 0165-0167 (hereafter K2) comprises three volumes of Kongo hannya gyo shoyaku gosho Jt)(:~JlilJU~[:;fi~~!E:-m, i.e. Fascicule 320 of the Bongaku shinryo, the 12th item of the second part of the Matsusen *i: section. This is a manuscript written in such a regular and even hand, and with so few corrections, that it appears at first sight to be printed; indeed, it almost certainly represents the final redaction, made in preparation for printing, of several other versions in the collection, which are explicitly designated as drafts. The compiler and editor is named on p. 1 as Hoju ~t~j, a sramal}a of Kokiji Temple in KashU (now the eastern Osaka area); this is the same person as the abovementioned Chido. This edition comprises the following elements, all written vertically, arranged from right to left: (1) Chinese glosses, aligned word-for-word to the right of the Sanskrit text; (2) Sanskrit text, immediately to the right of which has also been written in small letters the pronunciation in the Japanese katakana syllabary; (3) Chinese phonetic transcription; (4) Kumarajlva's translation; and (5) Dharmagupta's translation. These have all been reproduced in the Bodleian copy, with the sole exception of the katakana pronunciation guide. The copying is so exact that the position of the text on each line and each page has been scrupulously observed. Above and outside the page frame the section divisions of Kumarajlva's version are given. The Bodleian copy omits 1, but replicates all the rest. The date of this manuscript is not clear at the time of writing, but on the last two pages of Vol. III of K2 appears a long handwritten note by Senkai apparently added in the year 1847

10 For an analysis of these colophons, see Okukaze 2008. I am indebted to Mr Okukaze for kindly clarifying certain aspects of the K6kiji materials in recent personal communications.


Paul Harrison

(Koka 4). This is the postscript reproduced, with some v:ariations in the wording, in the Bodleian copy MS. Sansk. d. 28. The ultimate source of all the Kokiji copies of the Vaj appears to be a single copy sent back to Japan by Ennin ~1= (792-862) who was in China 838-847, where he learned Sanskrit and collected Buddhist texts to take back to his homeland. In the colophon notes to several of the Kokiji copies it is described as a folding book in two fascicules containing the Sanskrit and Chinese texts of the Vaj written horizontally.11 This was stored in the Zentoin ffiJj;~HJG on Mt. Riei tti~LiJ with other texts and ritual paraphernalia brought back from China by the master. It was recopied in the Eikyu & Roan Periods (1113-1118, 1120-1124) by Yakken ~JI and proofread by Yakunin ~?2,,12 and long kept in the Saito "EJ:g: area on Mt. Riei. Later, in the Tempo Period (1830-1844), Yakken and Yakunin's copy was rediscovered by the monk SMen *)JHl (1786-1859), also known as Shin'a Shonin llJpJ~A or Shin'amidabutsu ~JloJ~~B 1Jj5. Shuen made it available to Senkai, who recopied it in Tempo 8 (1837), and then sent it to Roju (aka Chido) at KokijiP

11 Some of the material in these colophons is quoted from T 2166,

likaku daishi zaito soshinroku ~:I:*gjjjiE'w~iijfu1 (see esp. 55.l078b8-

12 The colophons in red ink given at the end of the three Kokiji copies 0073-74, 0075-76 and 0162 all mention Eikyu 4 (1116) and then copying in midwinter of Hoan 1 (1120) by Yakken and a proofreading on the 6th day of the 12th month in the same year by Yakunin. The precise relationships of these copies to each other remain to be worked out, but it is to be noted that in formal respects they resemble the manuscript brought (or sent) back by Ennin, i.e. they are in two books (in the case of 0162 copied into one volume), the Sanskrit text is written horizontally, and it is accompanied by Chinese glosses. We can infer from this that Yakken's copy (from which the Kokiji copies are descended) may have mimicked the original rather closely. It also appears that 0073-74 and 0075-76 both attempted to represent this copy exactly, since the first and last siddhafTL characters of each line are the same in both of them. 13 According to the colophon a further copy was made and despatched in Tempo 9 (1838). Okukaze (2008) is sceptical about the notion that the manuscripts were kept on Hieizan right up to the 19th century, given

Experimental core samples


With these K6kiji copies now accessible, therefore, we are in a position to check Muller's edition against two of its witnesses, or rather, against their sources. 14 The same is true of T (see Muller 1881, Plate 3), for which I have secured a digital copy of the print of the same work kept in the library of the School of Oriental & African Studies, London. is I have not been able at this time to locate a copy of Ch. However, the results of the comparison of Muller's edition with the three witnesses which have come into my hands turn out to be rather surprising (see below). Muller's edition was subsequently supplemented by the discovery of two incomplete but nevertheless sizable manuscript copies bearing an older recension of the text, one being the Stein manuscript from Central Asia (late 5th or early 6th century), published by Pargiter in 1916, the other being the Gilgit manuscript (6th or 7th century), first published by Chakravarti in 1956, and later, in a much more reliable edition, by Schopen in 1989. Edward Conze had all these versions available to him (Schopen's work excepted, of course) when he reedited the text in 1957. His edition has become the standard point of reference, even though it largely reproduces the text as established by Muller, while adding to it information about the Stein and Gilgit manuscripts (not always complete or correct) and a fair number of mistakes. Other editions published during the 20th century take a similar approach, and it has to be said that they generally do not make a positive contribution to our knowledge. At the same time a small number of manuscript fragments from Central Asia have been published, mostly in out-ofthe wholesale destruction of the mountain's temple complexes by Oda Nobunaga in 1571. He considers the possibility that the manuscripts were preserved at ShOju Raigoji ~~*}ffi~ at the foot of the mountain. 14 It will be another matter to determine which, if any, of the surviving Kokiji copies is the source of all the others. The preliminary collation of the sections of the Vaj dealt with in this paper suggests that K2 cannot be a direct copy of K1. 15 Dr Ulrich Pagel kindly assisted me in obtaining this, for which I thank him. The copy in question was also consulted by Conze (1957[1974]: 1). Here it is referred to as T2, to distinguish it from the copy which Max Muller used.


Paul Harrison

the-way places where they have escaped notice. 16 There are also several Nepalese mss of the text, which, as far as I know, nobody has yet taken the trouble to consult. Although it is possible that others survive, two are known to me, both of them microfilmed by the Nepal German Manuscript Preservation Project. The first is Ca 267; NGMPP A913/9; Acc. No. 4/267, a paper ms dated 1701 containing a complete copy of the text on 52 folios (not 62, as given in the NGMPP database). I will refer to it as Ne2. The second is NAK (National Archive, Kathmandu) Acc. No. 5/186 (NGMPP B 90/16), another paper manuscript of uncertain date, hereafter Nel. Collation of both these manuscripts indicates that Ne2 is a direct descendant of Ne1, with no strong evidence of contamination from any other source. Therefore in this paper only the readings of the latter will be considered. 17 We can see, then, that the number of copies or parts of copies of the Vaj is quite numerous, and since the oldest of them dates from around the 5th century, we can track the Sanskrit tradition for this text back quite a long way. The Sch0yen ms (which we presume to come from the Bamiyan area) is an especially significant addition to our knowledge since it covers the first 60% of the text in a continuous run, and, like the Gilgit manuscript, dates from the 6th or 7th century. Since the Gilgit manuscript covers the second half of the text, but not continuously (one folio is missing), putting the two manuscripts together gives us our first look at the whole text as it must have circulated in the "Greater Gandhara" region. 1s We should note in this regard that the Stein
" ,,~,

16 Details of 11 of these can be found in Harrison & Watanabe 2006: 93-94. Recently further fragments of the Vaj have come to light in the British Library's collections, and have been published in the series Buddhist Manuscripts from Central Asia: The British Library Sanskrit Fragments, edited by Seishi Karashima and Klaus Wille (see Harrison 2009). 17 For help with securing copies of these manuscripts I am indebted to the staff of the National Archives, Kathmandu, and to the generosity of Dr Dragomir Dimitrov of the Nepal-German Manuscript Cataloguing Project and Dr Christoph Ciippers of the Lumbini Research Institute, Nepal. 1S For an English translation of this composite text, see Harrison 2006.

Experimental core samples


manuscript, while somewhat older, is full of gaps, since not only are many folios missing, but those that have survived are in very poor shape. . The Chinese translations of the Vaj present an equally ~ich picture. Six have survived, as follows:
1. T 235: lin'gang boruo boluomijing :lIz:~u~;s:~.&m~~, by Kuma-

rajIva, 402 A.D. (hereafter referred to as K). 2. T 236: lin'gang boruo boluomijing :lIz:~U~~.&m~~, by Bodhiruei, 509 (= B).19 3. T 237: lin'gang boruo boluomi jing :lIz:~um!:;s:~.&m~~, by Paramartha, 562 (= Z, for Zhendi 1i~). 4. T 238: lin'gang neng duan boruo boluomijing :lIz:~U~6ITm!:;s:~.&m ~~, by Dharmagupta, 605 20 (= Dh). 5. T 220(9): Da boruojing dijiuhui nengduanjin'gangfen *m!:;S:~rlf~ :tL~~6IT:lIz:~U5t, by Xuanzang, 648 (= X). 6. T 239: Fo shuo nengduanjin'gang boruoboluomiduo jing {~m~1T :lIz:~um!:;s:~.&m~~~rlf, by Yijing, 703 (= Y). Given the sizable number of Sanskrit manuscript remains and the survival of six Chinese translations made over a period of 300 years, the comparative study of the Indic and Chinese versi9ns of the Vaj has no shortage of material with which to work. 21 The situation is quite different with the VKN, the Sanskrit text of which was until recently thought to be lost, and known only

19 Under this TaishO number there are actually two translations attributed to Bodhiruci, the second of which (757a20-761c29), however, is a copy of Paramartha's translation (T 237) with occasional variant readings. It is not taken into account in this study. 20 On the "unfinished" nature of T 238, which distinguishes it from all the other Chinese versions, see Zacchetti 1996. Yuyarna 1967: 73 gives the date as ca. 592, and notes the existence of a second version by Dharrnagupta dated ca. 613 embedded in the lin'gang boruo lun :lIz:~um!: ;g:~ (T 151Ob). However, although this translation of the commentary ascribed to AsaIiga is by Dharmagupta, the complete translation of the Vaj which it contains appears throughout to be that of Bodhiruei. 21 Here we take no account of the versions in other languages, such as Tibetan, Khotanese, etc.


. Paul Harrison

through a few scattered quotations in other works. 22 CO)1sequently the emergence of a complete Sanskrit manuscript of this text, discovered in the Potala Palace in 1999, caused a considerable stir. In 2004 a team of scholars working at TaishiS University published a transliteration of the manuscript (collated with Tibetan and Chinese translations), and in 2006 they followed this with an edition. The transliteration attempted to be absolutely faithful to the manuscript, while the edition, making no claims to be critical, also changed the readings of the manuscript as little as possible (see the comments on pp. xi-xii), since the editors held that it "bore no substantial damage, was copied by a fairly good scribe with good scripts, and its reading was, in general, reliable." While acknowledging that in a number of places the manuscript failed to agree with the Tibetan and/or the Chinese translations, the editors found the relationships between the various witnesses unclear, and expressed the hope that more detailed study of the individual cases would clarify the situation, and enable their own rather minimalist emendations to be improved upon. The Chinese translations of the VKN are only half the number of those of the Vaj, but stretch over a somewhat longer period. Three complete translations survive, as follows: 23
1. T 474: Weimojie jing ~t.~B~, by Zhi Qian JZ~*, in the pe22 For a complete list of such citations see Lamotte 1962: 91-95, and for the text of those in Sanskrit, with Tibetan and Chinese versions as appropriate, see VKN Intro., pp. 23-41. The most important are the eight citations in Santideva's Sik-!asamuccaya. Lamotte lists nine, but the third (in Bendall's edition at 153.20-22) is a reference to the text, not a citation as such (cf. VKN Intro., pp. 24-25). Lamotte's list of citations in the Satrasamuccaya attributed to Nagarjuna (but certainly not by him) is incomplete: there are 7, not 3, as follows: 73-76 (U6-12), 163 (IV.l4), 163 (VIII.l7), 182-183 (IV.l7), 183-185 (IV.20), 187-188 (XII.17-19), 188-'190 ev.20). References are to Bhikkhu Pasadika's edition of the Tibetan text. It is unfortunate that the Satrasamuccaya is extant only in Tibetan and Chinese, since its citations of the VKN are quite extensive. 23 For full details, including other attributions of allegedly lost translations, see Lamotte 1962: 2-14, from which the following details are taken.

Experimental core samples


riod 222-229 (= Q). 2. T 475: Weimojie suoshuo jing ~t.5pJTn~J[, by KumarajIva, 406 (= K). 3. T 476: Shuo wugoucheng jing nM,:i:Jf.jij.~~, by Xuanzang, 650

Here again we are faced with the handiwork of KumarajIva and Xuanzang (and their respective teams), but this time we also have to deal with a product of a much earlier period, from one of the pioneers of Chinese Buddhist translations, Zhi Qian. 24 What are the implications of the Sanskrit versions of these two texts, especially the new finds, for the study of their Chinese translations? The appearance of the VKN, in particular, is likely to be welcomed by those who still adhere to older ideas about the inadequacy and undependability of Chinese versions for the study of Indian material. Gregory Schopen (2005:4) provides a recent and carefully worded statement of such scepticism.
Chinese translations have also been used - less successfully, I think - to try to track what have been seen as developments within a given Indian text. The nature and number of assumptions and methodological problems involved in such a use have not, however, always or ever been fully faced, and it is not impossible that some - if not a great deal - of what has been said on the basis of Chinese translations about the history of an Indian text has more to do with the history of Chinese translation techniques and Chinese religious or cultural predilections than with the history of the Indian text itself.

There are indeed some serious difficulties to be faced, among them the challenge of understanding the language of these translations, especially the older ones, which is often obscure, or downright chaotic, or the challenge of working out in full the modus operandi of the individual translators and their collaborators, and then there are the vagaries of an uncertain and still largely unexplored Chinese manuscript tradition to contend with. But sometimes scepticism and caution can be taken too far. Edward Conze, for example, was

24 On the work of Zhi Qian 3t~, see now Nattier 2008. On his version of the VKN, see pp. 139-141.


Paul- Harrison

inclined to dismiss Kumarajlva's version of the Vaj, despite its antiquity, as throwing little light on the problems of textual criticism "partly because it was not made directly from a sanskrit original,25 and partly because it is less concerned with literal accuracy than the later Tibetan translations were" (1957: 1-2). While forced to concede that some of its testimony was borne out by other early sources, he suggested (p. 2) that "[m]any of the verbal differences, abbreviations and omissions may very well be accounted for by Kumarajlva's methods of translating rather than by a divergent sanskrit original." Conze's assessment in this regard surely resulted from taking Muller's 1881 edition as the standard by which to judge all the Chinese translations, but it should have been clear to him even from the subsequently discovered Stein and Gilgit manuscripts that this would not do, and that Kumarajlva's text may indeed represent more accurately an earlier - and considerably shorter - recension of the Vaj, even allowing for the distortions arising from his "methods." Scepticism, then, is all very well, but it is no better than blind faith when it chooses not to look at the evidence. Scholars working in this field need no further persuasion in respect of the value of Chinese sources, even though they are for the most part well aware of the problems attached to their use. Nevertheless, I still think it useful to submit the assumption that we can track changes in Indic texts through their Chinese translations to some kind of test. To do this, I shall in this paper take a small number of short passages from our two sutras and carry ou(:what is in effect a kind of "core sample" experiment, arranging the various Chinese versions in chronological order to see how they change over time, and assessing the results against a similarly stratified arrangement of the Indic manuscripts. This is .of course virtually impossible to do with the VKN, but even a comparison of the unique 11th-13th-century manuscript 26 with the Chinese ver-

25 Conze provides no evidence for this blunt claim. 26 This is the conclusion about the date reached by the Taish6 team.
See VK.N Intro., pp. 74-75. However, if the script suggests such a date, then the King GopaIa mentioned in the colophon is very probably the third and last to bear that name, whose regnal years are variously given

Experimental core samples


sions furnishes us with some useful results. We start with the Vaj, picking out three sections (7, 12 and 26) for closer scrutiny, before turning to the VKN, from which four samples are taken. For the Vaj the Sanskrit text of each passage is given. first, in two forms, according to the Shorter (in the case of the Vaj presumably earlier) and Longer (presumably later) VersionsY The Shorter Versions are represented by the Gilgit (G), Sch0yen (S) and Stein (P) manuscripts, as well as by various Central Asian fragments. 28 The Longer Versions reflect the text established by F. Max Muller (MM), which Conze (Cz) reproduces with minor changes, as noted. The presumed exemplars of MM's Japanese witnesses (K1, K2) have also been collated, as well as Ne1 (Nepalese ills, NAK Acc. No. 5/186 [NMPP B 90/16]) and T2 (7: fols. 13a-15a; 12: 25b26b; 26a-b: 63a-64b). In order not to encumber the apparatus with excessive detail, the use of class nasals instead of anusviira and gemination of consonants after r are not noted in K1 and K2, while minor spelling mistakes in Ne1 & T are also ignored. Bold in the text of the Longer Versions indicates where the wording differs from the Shorter Versions, bold with underlining indicates an amplification of the wording. A double asterisk after a footnote
as 1128-1143 or 1143-1158 (see ibid., p. 18). This would narrow down the date of the manuscript to around the middle of the 12th century. 27 In an earlier version of this paper I used the terms "Shorter Recension" and "Longer Recension" in the singular, but now realize that this could be misleading, not simply because "recension" may imply deliberate editorial revision, but chiefly because it might give rise to the idea that all these older, shorter versions of the text are somehow copies of a single form of the work (the Shorter Recension), and so too with the later and longer copies. In other words, one would dispose of the fantasy of the Vaj as a single text only to replace it with the illusion of the Vaj as two texts. Nor is it always and necessarily true that the shorter versions - or, we might better say, performances - of a work are older than the longer ones. However, in the case of the Vaj the weight of the manuscript evidence is certainly in that direction. These and other related issues will be discussed at greater length in the introduction to my forthcoming edition of the text. 28 See Harrison & Watanabe 2006 for full bibliographical details.


Paul Harrison

indicates that the reading of Kl and K2 is in substantia~ agreement with the Shorter Versions. 29 The Chinese translations of each passage are then given in chronological order. 30 Each translator is assigned his own colour (in the case of the Vaj, KumarajIva maraschino red, Bodhiruci maroon, Paramartha clover green, Dharmagupta black, Xuanzang blueberry blue, and Yijing magenta; for the VKN Zhi Qian black, KumarajIva maraschino red, Xuanzang blueberry blue), but this colour is applied only to the wording which is original to that translator's text. In this way one can easily see, even without reading Chinese, how much of the wording each translator has taken from his predecessors. An exception is made for Dharmagupta, since his version of the Vaj is not a translation like the others, and it would be very difficult to ascertain the extent to which he borrowed anyone else's wording. Instead his Chinese rendition of all three passages is given again at the end of the Vaj section, set against the Sanskrit, which it follows verbatim in Sanskrit word order. In the case of the VKN the Sanskrit text has come down to us in a single version, so it is presented once only, with minimal or no apparatus (the Tibetan version is also supplied, as given in VKN Text). After the presentation of the VKN passages, I present some remarks about the significance of the general patterns which emerge from this exercise. Within the confines of this paper it is unfortunately not possible to discuss all the points of interest and problems of interpretation which can be found in each passage, so notes on specific variants have been kept to a minimum.

29 In this paper the readings of Kl and K2 are given in the apparatus to the Longer Versions since copies derived from them were collated by Max Muller when he established his edition, and since it is important to show the extent to which he suppressed their testimony in his own apparatus (on this point see below). However, it is evident that this is not their proper place, since they tend very often to agree with the Shorter Versions, i.e. G, S, P and the Central Asian fragments, as will be abundantly clear in the apparatus to my new edition of the complete Sanskrit text. 30 The text is taken from the CBETA version, checked against the printed copy in the Taish6, and repunctuated. Variant readings are generally not noted.

Experimental core samples


Text Passages Vaj 7 Sanskrit (Shorter Versions)a punar aparrup bhagavan ayu~marp.tarp. subhiitim .etad avocatb / tat ki<rp.> man ase subhute kacit tathagatenanuttara{rp.}C samyaksarp.bodhir abhisrupbuddha kascid va dharmas tatha atena desital:.J." subhiitir aha / yathaharp. bhagavand bhagavato bha~itasyartham ajanami nasti sa kascid dha(r)m(o ya)s tathagatenanuttara{rp.} samyaksambodhir abhisarp.buddha / nasti sa kascid dharmo (as tatha)g(at)ena desital:.J. < > tat kasya hetol:.J. < > 0 'sau tathagatena dharmo desit(al:.J.e agrahyal:.J. s)o 'nabhilapyal:.J.f <I> na sa dharmo nadharmal:.J. < > tat kasya hetol:.J. < > asarp.skrtaprabhavitag hy aryapudga1lil:.J. < >

This section missing in P and G. Text here that of Sand Frag d (see Harrison & Watanabe 2006). Underlining indicates where there are gaps in Frag d, so that for these sections of the text we have only the testimony of S. b It can be inferred from the number of missing ak~aras in Frag d that it did not contain this sentence in this form. There appears to be enough room for liha alone, or, more likely, nothing at all, in which case Frag d would have read withK. C Frag dreads: [lin](u)[t](ta)[r]li. d It can be inferred from the number of missing ak~aras in Frag d that it did not contain bhagavan. e Frag d (deSfta agrlihya sau anabhilli) supports our reconstruction of S, in which what can be seen of the missing ak~aras renders desito impossible. f Frag d: agrlihya sau anabhilli. g S reads: aSaJ'!1,skrtathlibhlivitli. This scribal error is accounted for by the close resemblance of the ak~aras pra and thli in this script.

Vaj 7 Sanskrit (Longer Versions) punar apararp.a bhagavan ayu~marp.tarp. subhiitim etad avo cat , tat kirp. manyase subhiite asti sa kascid dharmo yasb tathagatenanuttara samyaksarp.bodhir ityrl abhisarp.buddha1;te kascid vaf dharmas g tathagatena desita1;t , evam ukta liyu~mlin subhiitir bhagavarptam etad avocath , yathaharp. bhagavani bhagavato bha~itasyartham ajanamii nasti sa kascid dharmo yask tathagatenanuttara1 samyaksarp.bodhir itym abhisarp.-


Paul Harrison

buddha~n nasti dharmoo yas tathagatena desital]p I tat kasya hetol]q I yo 'saur tathagatena dharmo 'bhisa:qtbuddho' desito va agrahya~t so 'nabhilapyal]u I na sa dharmo nadharmal]v I tat kasya hetol]w I asarp.skrtaprabhavitaX hy aryapudgalal] I

apararrz MM, Nel, T2: aparan* Kl, .K2 (not noted by MM) asti sa kascid dharmo yas MM, Nel, asti sa kaccid dharmo yas T2: kacit KI,

K2 (not noted by MM).** . samyaksal[!bodhir MM, Nel, K2, T2: samyasar[lbodhir KI, corrected to samyaksarrzbodhir KI (not noted by MM). d So MM, but he notes that his Ch, J & T all read samyaksarrzbodhir abhi- (thus KI, K2, Nel, T2). lty is thus his emendation. e MM notes that his J reads abhisarrzbuddha (thus KI, K2). abhisarrzbuddhaf:z Nel, T2.** f va MM, Nel, T2: om. KI, K2 (not noted by MM). g dharmas MM, Nel, T2: dharmmaf:z KI, K2 (not noted by MM). h evam ukta ayu.'jman subhiltir bhagaval'{ltam etad avocat MM, Nel, T (T2: subhilter): subhiitir iiha KI, K2 (not noted by MM).** bhagavan MM, Nel: bhagavarrzn T2, bhagaval'{l KI, K2 (not noted by MM). iijiiniimi MM, T2: ajaniimi KI, K2 (not noted by MM). k MM notes that Ch & Tread dharmas for dharmo yas (thus T2, so too Nel). dharmmo yas KI, K2.** I tathiigateniinuttara MM: tathiigateniinuttara T2, tathiigateniinuttaral[! KI, K2 (not noted by MM).** m MM notes that his J reads samyaksarrzbodhir abhi- (thus KI, K2). ity Nel, T2.** n abhisarrzbuddhaf:z MM, KI: abhisarrzbuddha K2, abhisar[lbuddha Nel, T2. o niisti dharmo MM, Nel, T2: nasti sa kascid dharmmo KI, niisti sa kascid dharmmo K2 (not noted by MM).** P MM notes that Ch & Tread bha.'jitaf:z and bha.'jitaf:z respectively (thus T2). bha.'jitaf:z Nel. q hetof:z MM: heto KI, K2, hetor T2 (not noted by MM). heto Nel. r 'sau MM, Nel: 'so T2, Kl, K2. , 'bhisarrzbuddho MM, Nel, T2: am Kl, K2 (not noted by MM).** t deito va agriihyaf:z MM, T2 (agriihyas): deito 'griihyaf:z KI, K2 (not noted by MM). yasito [I] va agrahyas Nel.** U 'nabhilapyaf:z MM, KI, K2: 'nabhiliipyo T2. v niidharmaf:z MM, T2: nadharmmO(1 KI, nadharmmaf:z K2 (not noted by MM). w. hetof:z MM: hetor KI, K2, T2 (not noted by MM). x asarrzskrtaprabhavita MM, KI, K2: asarrzskrtaprabhavita T2, na sarrzskataprabhiivitii Nel.

Experimental core samples


Vaj 7 Chinese


Kumarajlva (T 235, 8:749b12- 18)

~}'[:g:m ! ~~EfOJ ? YD*1~/loJ~~Hi 5~IfG?J ~}'[:g:fJt : 'YDiZMf1tpJT5l~ ' ~1:5E a

0 0

vJ\~1DE5~YD*PJ5l fOJ.l~i!iJ: ? YD*pJT5l5~'::fPJ1fx ' ::fPJ5l ' ~F

5~ ' ~F~F5~ PfTJ;J~fOJ ? -t)JJl~'.l~~~5~fffi1:~53U

m- :g:f]f)fG ? YO * 1:pJT5l 5~i;flilJ~~*I- m -:g:m '


2. Bodhiruci (T 236, 8:753b17- 23)

1~-'KI7P p ~ OfJ~.R
"f=j-f.-7j;iPL::fiA-/. . ::ti:tg.

I / ... ::ti:-I:.. , tA"""- - lnT ? .f.t 7:b:/.8 [(nr~::9: ~ -;u; 1JE . ~.R Cl 1JE . JJ ~ ,'G, -Z:;;-IOJ . )'.D~ 1"ifIJ []J'Wf 3"&;lE'='ea5L

- :g:mIfG ? YD*1:pJT5l5~IfG ?J
~}'[:g:m : 'YDiZM{1tpJT5l~ ' ~~5E5~YD*i~ flilJ~~*I-:g: m ' vJ\~1:5E5~YD*PJ5l fOJJ;JiW ? YD*pJT5l5~'::fPJ1fx ' ::fPJ5l '

~F5~ ' ~F~F5~

fOJJ;JiW ? -t)J~A '.l~~~5~i~i; J


3. Paramartha (T 237, 8:762c16- 22)

1~*{1t (variant: {1t1![)*5*$~}'[:g:m : I ~}'[:g:m ! 5Y: ~EfOJ ? YD*1~ /loJw~*I:g:mIfG ? YD*~pJT5l5~IfG ? J


~}'[:g:m : 'YDiZM{1t5l~ ' ~pJT~ b 5~YD*pJT1~i;flilJ~~*I-:g:m ' vJ\~~5~YD* pfT 5l fOJ J;Ji!iJ: ? ~;~ YD*pJT5l::fPJ1fx ' ::fPJ

, ~ F5~ , ~F~F5~ fOJ J;JiW ? -t)J~A 'tz~~~YDpJTfJJn~i5J: J

0 0

4. Dharmagupta (T 238, 8:767c3- 1O)

1![=Xt!t~$~~'Jf}lYD~: 'iElfOJ~~' ~'Jf? ~YD*B!lE~9ill ~..t.lE~~O~i:? ~1![5~YD*5l? J
~'Jf: 'YDiZt!t~t!t~5l~M:f\(;, ~~-5~~YD*~..t.lE~~O~

i: ' ~~-5~~YD*m iElfOJpJTlB? ~iElYD*5~5l::fPJ1fxtel '::fPJ 5l' ::ffEl$ , ~F::f5~ iElfOJlB? ~~5~ c ffJ[8jj~A J
0 0 0


Paul Harrison

m~.~H-~-~m ' ~.~~$ ~~* ~~W. ffi~ o W~

i& ? i!t~ ! 3tD* ~~W.pJT~ e pJTftPJT,IGI,'rtE jt ' \"/FP}!f:z ' /FP}E'f

;''' ,


-;I r/"ZA

::Ib~+ , ::Ib::lb~+


r-;l r/"ZA

J;:;fhl-f+r ? hl=~ g;:SOrD -2;f;If:j-h = ~b EI ~"Po -'7 hi=::~;::; -f+r llJj YAIJy" . YAii-a ~.=E TFfl'f'1' I/JO&;1E E'l:XE~m~ m "~oy"

6. Yijing (T 239, 8:772b22- 27)


*y~ ! ~jY:~~1iJf ? 3tD*D'2'.L~m~pJT~/F ? 1~~~jt;~PJTft

/F ?J
-Wy~ : '3tDfZM~{~pJTft~ , 3tD* ~'~L~mJf f 1*pJTii~ ' ~. PJT

ft fiJftJi& ? {~PJTftjt;/FP}!f:Z ' /FP}ft ' l&: ~Fjt; , ~F~Fjt; 1iJftJi!i)(: ?

0 0

h I ~ !l'D # ~b EI 1ffi: "Po hi=::f!,ifl=l-f+r 0 YA,m .=E"B s }E,,,,mnl ,,"R ;n:,oy" J

" Here and in the next clause K inserts the word ding JE ("settled," "definite," "fixed"), for which there is no support in any Sanskrit text. B replicates this. b Here and in the next clause Z drops the word ding JE inserted by K (and B). c The appearance of wuweifa ~,g,;:t in Dh increases the likelihood of K reflecting an actual Sanskrit reading asa'!lskrtadharmaprabhiivitii, rather than being a commentarial amplification made in China. d X is the first and only Chinese translation to reflect the asti sa kascid dharma yas of the Longer Versions. e X is the only translation to reflect the (dharma) 'bhisa'!lbuddha .. . vii of the Longer Versions. It also adds a third term to the series, sua-siwei pfi}~:r!E, "pondered," "meditated on." f Here Yijing resorts to a device favoured by KumarajIva, albeit not in this section . See below.

Vaj 12 Sanskrit (Shorter Versions)"

api tu khalu b subhilte yasmin Prthivlpradese ito dharmaparyayad arptasas d catu~padikame api gatharpf bha~yetag va desyeta yah sai prthivlpradeaSi caityabhUto bhavetk <I sadevamanu~asurasyal > lokasya ka!:l punar vada!:l subhUte m ya imarp dharmaparyayarpn dharayi~yarptiO parameI;la te P ascaryeI;la samanvagata bhavi~yarpti tasmirps caq prthivlpradese sasta viharaty anyataranyataro va gurusthanlya!:lr

" Text that of S, with notes on readings of P and Frag e (see Harrison & Watanabe 2006). Underlining indicates where the Stein ms (P) has gaps. G not extant for this section. b P, Frag e omit: khalu .

Experimental core samples





k I


P q

Pargiter's reconstruction omits ito, but there is enough room for it in the gap. Lacuna in Frag e. P apparently omits: aY[!tasas. Frag e has anta ... P: clitu-rplidlim (originallyclitu-rPlidlipi, with ma inserted above dli). Lacuna in Frag e. Frag e: glithli. P: bhli-ryate; Frag e: bhli~i-ryate. P, Frag e: tena for vli deiyeta vli. P: saf:!. P: prthivfpradeiaf:!. P: bhavi~yati for bhavet. Frag e reads ~yati, therefore must have read with P. Frag e sadevamlin[u]~yli(surasya). Lacuna in P. Both S (I indicates the use of virlima in the ms) & P punctuate before this term, but punctuation seems unnecessary. Frag e does not have it. Frag e omits: subhute. Lacuna in P. Lacuna in P, but P probably read with S. Frag e reads with S. Frag e: udgrhr;.f~yan(t)i for dhlirayi-ryaY[!ti. Lacuna in P, which may have read either udgrhr;.f~yanti or dhlirayi~yanti, but has in any case only enough space for one verb. Frag e breaks off at this point. P adds: satva. P: (tas)[m](iY[!) (without ca, since it is followed by (P)[r]-). Lacuna in P would apparently permit vijiiaguruO,

Vaj 12 Sanskrit (Longer Versions) api tu khalu puna1;t subhute yasmina prthivlpradese ito dharmaparyayadb rup.tasas catu~padikamc api gathamd udgrhyae bha~ye ta va sarpprakasyetaf va sa prthivlpradesasg caityabhfttoh bhavet sadevamanu~asurasyai lokasya ka1;t punar vado ya imrupi dharmaparyayarp sakalasamaptarpk dharayi~yarpti vacayi~yaq:ttiI ~ avapsyaq:tti parebhyas ca vistareJ,la saq:tprakasayi~yaq:ttim I parameJ,la te subhiita iiScaryeJ,la samanvagataO bhavi~yrup.ti I tasmiq:ts ca subhiiteP prthivlpradese sastaq viharaty anyataranyataror va viD



e f

yasmin MM, Nel, T2: yesmiY[! Kl, K2 (not noted by MM). dharmaparyliylid MM, T2: dharmmaparyliyad Kl, dharmmaparyliyad K2 catu~plidikam MM, Nel, T2: catu-rpadiklim Kl, K2 (not noted by MM).** gathlim MM, Nel: glithli Kl, K2 (not noted by MM). udgrhya MM, Nel, T2 (in Nel & T2 followed by vli): om Kl, K2 (not noted byMM).** saY[!prakliSyeta MM, Nel, T2: deieta Kl, K2 (not noted by MM).**


Paul Harrison

pradeas MM, T2 (T2 regularly PrthiviO): pradea/:t KI , K2 (not noted by MM). h caityabhuto MM, Nel, KI: cetyabhuto K2 (not noted by MM). i sadevamanu$asurasya MM, Nel, T2: sadevamanu$yasurasya KI, K2 (not noted by MM). j immr! MM, KI, T2: imalJ1 K2 (not noted by MM). k sakalasamaptalJ1 MM, Nel, T2: om KI, K2 (not noted by MM).** 1 dharayi$yalJ1ti vacayi$yalJ1ti MM, Nel, T2: udgrhlsyanti dharayi$yanti vacayi$yanti KI, K2 (not noted by MM). m paryavapsyalJ1 ti parebhyas ca vistarena samprakasayi$yalJ1ti MM, T2: om. KI , K2 (not noted by MM); parebhyas ca vistarena samprakasayi$yanti Nel (omits paryavapsyanti).** n paramena te subhuta ascaryeyta MM, Nel, T2 (subhute): parameytascaryeyta KI, paramenascaryeyta K2 (not noted by MM).** o samanvagata MM, Nel, KI, T2: samanvagata K2 (not noted by MM). P subhute MM, Nel, T2: om. KI, K2 (not noted by MM).** q sasta MM, Nel: sasta KI, sasta K2, ccasta T2 (not noted by MM). r anyataranyataro MM, T2: anyataro KI , K2 (not noted by MM). anyataro nanyataro Nel. S vijiiagurusthanlya/:t MM, T2: gurusthanlya/:t KI , K2 (not noted by MM). vijiiasubhasthanlya Nel.**

Vaj 12 Chinese

1. Kumarajlva (T 235, 8:750a6- 1O)

' 1~:::X ' ~&:g:m ! r.iiiR a ~~~ , 7J:[2]{IJ1~~ , ~ b llt~ , -tJJtltFs, :;R: 'A' PDJ~~~,JJ!1~. ' 3<01?ttg:$j f6r5C~A~ c ~E'5'tj:~ ' iil"m d , ~&:g:m ! ~~Am)ji,;t~L~-;ffi~Z)~ , ;fi~~~~pfT1Z~ , fW ~~{?t ' ;fi#.m%r e J

2. Bodhiruci (T 236, 8:754a19- 24)

'1~:::X ' ~&:g:m ! P.iPfT~~ , iiR~)~F' ' 7J~[2]{IJ1~~ , ~llt~ , tJJtltFs,:;R: 'A' PDJ{I~~'~{~. ' 3<o{?ttg:$j f6J)5C~A~~E5j:~ ' iii "m llt~~ ~&:g:m ! ~o~Am)ji,;t~L~-;ffi~Z)~ , ;fi~~~~pfT1 Z~ , flrj~~{?t ' ;fi#.m1~1?t J

3. Paramartha (T 237, 8:763b19- 25)

'1~:::X ' ~&:g:m ! P.i pfT1 ~ ' ;fi~A~E1jt~~J! ~ ' 7J:[2]{IJ1~~ ' iii "miifiiiR ~llt~ , tKtltFs,cp~Pm5z:m ' -tJJA' :;R: ' PDJn~~ ~ '~

Experimental core samples



:mW9: fof5J:1 AJliiijg~t~ , ~Jffti 3ZD ltt~~ ~ , &,W~}J:WW~L;ffi-1

L)~fffj~f~ , ~ttt~ ,

::kgffjtP:P ' ~IlI1-PJ ~~ A

4. Dharmagupta (T 238, 8:768b17- 22)

:g:Jf ! tltcpttt5t ' ltt):t*7J~I2]{IJ~~ '~fi~5t ;fi' 5t55U;fi, JJ[5t;fi' fEZJtt5t3tw17C 'A' ~MHi@ {DJii" , :g: 'Rl,.:tI.iJ.- /Z::>;/+':T'1"83 O~83 offi 83 l'C,~JX-/J JJ~I5"!oJL83 1&1Yf:t1lX.,ttrA~, k&'1 lttcp , :g:Jf !ttt5t~gffj}Qt1'J ' 55U~#~~ffij, ~jt1'J J


-++-r~t;+++#~, ~ffij~, ~~~, fjh,,"k-p.L'>.t;lrlffi:~"~' r=T ~/,t~-t=;-H

0 0

5. Xuanzang (T 220(9), 7:981c27- 982a5)

' ii=x:g:JJ[ !;fitttnrJT~ltb:tr~ , 7J~~{i''5tIffl7G I2]{IJ{1]o{i ' Ittttt

nffi~~@OO~ 7C& A ~~~ ~Lffim.~m. D o W~1~ ~
.iJ.- /Z::>;I

J~ ,At:::tL.% s ~ ' 5<..1'1" ' "Jllpffi ' %.%m1!f' JX-I5"!h>0 l'C,J3.pJL ' I7fJ/G '





';;:f.UP. ffi:-'P.fjhr.=,.y,


3ZDf'ft~ , 3ZD~1'~m<WX:tNH9J ;ffi-1 J}]1,~ , ltttttnrJT::kgffjrJTf ' ~1lI --~~~rJT ' ;fiti1~roJjt1'J=t3 J

6. Yijing (T 239, 8:773a17- 22)

~J>~ ! ;fi~cp~ltb:tr~ ' ~fi M5t , 7J~I2]{IJ f {1]ofili ' &'wlttttt , ~~~113U11! ' -t)]7C 'A' ~ fhHI~ , \"~ ;:b*fffj~1ff9J~ fDJ)5Lijg~ t~ , Jffti ' &,W~A ' ~IJ Py:tNL%-;ffi-1 ' 5Z.lttnrJT ' ~~~1fm , &~ ~~T oJ


With only one verb here, K reads with the Shorter Versions as represented by P and Frag e. Here and at the start of the last sentence, K inserts the words dang zhi ~%l, "one should know that ..." presumably to improve the balance and clarity of the translation. Unsupported by any Sanskrit version of the text, this stylistic amplification is repeated by Band Z, and survives even in Y. K is either translating a shorter version with sakalasamaptaJ?1, or, more likely, adding the word jin iii to bring out the undoubted implication of the text. Only X provides a clear equivalent for sakalasamaptaJ?1, which itself looks like a commentarial amplification. All Sanskrit mss of the Shorter Version group lack it, including KI & K2. K suggests dharayi:jyanti (or udgrhf:jyanti) vacayisyanti, and lacks the further amplifications of the Longer Versions, first evident in Dh. Reflects the Shorter Versions (no equivalent for vijiia-, which is not attested in the Chinese until X). Emend text from 1oJ.


Paul Harrison

Vaj 26a-b Sanskrit (Shorter Versions)"

[26a] tat kirpb manyase subhUte lak~aI!asarppadac tathagato dra~tavyaDI . aha I evamd etade bhagaval f lak~a:r;tasarppadag tathagato dni~tavyaD I -h I sacet punaD subhUte lak~a:r;tasarppadai tathagato bhagavan aha dra~tavyo 'bhavi~yad rajapi cakravartI tathagato 'bhavi~yat Ij ahak I yathaharpl bhagavato bha~itasyartham ajanami I na 1ak~a :r;tasarppadam tathagato dra~tavyaD II atha khalu n bhagavarpso tasyarp velayam ima gatha abha~atap : II ye marp rupe:r;ta adrak~urq ye marp gho~e:r;ta anvayuD I mithyapraha:r;taprasrta na marp" drak~yarpti te janaDS 111 II [26b] dra~tavyo dharmato buddho dharmakayas tathagataD I dharmata capy avijileya na sa sakyarp vijaniturpt 11211

b c d


Base text is G. Underlining indicates where the Stein ms (P) has gaps. Frags h & i (see Harrison & Watanabe 2006: 94) contain material from this section, but it was not possible to collate them when this paper was prepared. See now Harrison 2009 for their testimony (No.1; cf. also No.6). ki'!l G: kin P. lak~alJaSmFpada G: lak~alJaSmFpadayas P. aha I evam P: ahaivam G. etad G: eva P. bhagaval G: bhagaValJl P.

G: lak~m:tasa'!lpaday(as) P. bhagavan aha G: aha P. lak~alJasa'!lpada G: lak~alJasa'!lpadayas P. j P adds here: tasmad alak~alJasa'!lpadayas tathagato dra~tavyal:t. k aha G: ayu~ma'!l subhatir aha P. I P has space here for bhagavan (cf. Dharmagupta). m lak~alJaSmFpada G: lak~alJasa'!lpadayas P. n atha khalu G: atha P. o bhagava'!ls G: bhagava'!l P. p abha~ata G: ... ~ft P. Pargiter conjectures abha~i~ft, but the form adhyabha~ft (often written adhvabha~ft) occurs frequently enough to be more likely. q adrak~ur G: adrak~Ci) P. " ma'!l G: me P. S janal:t G: na(raJ:t) P. t P omits the second verse in its entirety (cf. K).

g lak~alJasa'!lpada

Experimental core samples


Vaj 26a-b Sanskrit (Longer Versions)

[26a] tat kirp. manyase subhUte lak~aIfasarp.padaa tathagato dra~ta vyal).1 subhfitir aha I no hlda:rp. bhagavan I yathiiha:rp. bhagavato bhii~itasyarthamb ajanami na lak~aIfasarp.pada tathagato dra~tavyal).d I bhagavan aha I sadhu sadhu subhfite evam etat subhfite evam etad yathii vadasi I na Iak~a1}.asa:rp.pada tathiigato dra~tavyal).1 tat kasya hetol).e Isacet punal). subhute lak~aIfasarp.pada tathagato dra~tavyo 'bhavi~yadf rajapig cakravartI tathagato 'bhavi~yath I tasman i na Iak~a1}.asa:rp.pada tathiigato dra~tavyal).1 ayu~manj subhutir bhagava:rp.tam etad avocat I yathaharp. bhagavato bha~itasyartham ajanami k na lak~aIfasa:rp.pada tathagato dra~tavyal).11 atha khalu bhagavarp.sl tasyarp.m velayam ime gathen abha~ata I

ye marp. rfipeIfa cadrak~ur" ye marp. gho~eIfa canvagul).P I mithyapraha1}.aprasrtaQ na marp. drak~yarp.ti te janal).r 111 I [26b] dharmato buddho s dra~tavyot dharmakaya hi nayakal).v dharmata ca naW vijfieya na saX sakyaY vijaniturp.z 11211


saf[!padii MM, K2, Nel, T2: saf[!pada KI (not noted by MM).

bha~itasyiirtham MM, Nel: bhii~itasyartham T2.

h i

iijaniimi MM, Nel: ajanami T2. no hfdaf[! bhagavan I yathahaf[! bhagavato bhii~itasyartham ajiiniimi na lak~m:tasaf[!pada tathagato dra~ravya/:! MM, Nel, T2: evam etad bhagavaf[! na lak~m:zasanpadii (K2: saf[!npada) tathiigato dra:jtavya/:! KI, K2 (not noted by MM). [KI, K2 read with Shorter Versions except for insertion of the negative!]** siidhu sadhu subhute evam etat subhute evam etad yathii vadasi I na lak~a I}asaf[!pada tathagato dra~!avya/:! Itat kasya heto/:! (hetos T2)1 MM, Nel, T2: om KI, K2 (not noted by MM).** 'bhavi~yadMM: 'bhavi~yen KI, 'bhavi~yet K2 (not noted by MM). 'bhavi~yat Nel, dra~!avya/:! bhavi~yat T2. riijiipi MM, K2: riijapi KI (not noted by MM). tad riijiipi Nel, T2 (not noted byMM). 'bhavi~yat MM, T2, Nel: 'bhavi~yet Kl, K2 (not noted by MM). tasmiin MM, Nel, T2: tasman KI, K2 (not noted by MM). ayu~miin MM, Nel, T2: athiiyu:jmiit KI, K2 (not noted by MM). iijiiniimi MM, T2: iijaniima KI, iijaniimi K2 (not noted by MM). bhagavaf[!s MM, T2: bhagavan Kl, K2 (not noted by MM).

m n

Paul Harrison

tasyam MM, Kl : tasya K2 (not noted by MM). ime gathe MM, Nel, T2: ima gatha Kl, K2 (not noted by MM).** cadrak:jur MM, T2: adrak:ju1:z Kl, adrak:ju1:z K2 (not noted by MM). madrak:ju

Nel (Nel is very corrupt in this verse).**


canvaguh MM: MM notes that J reads anvayu1:z (thus Kl, K2), Ch reads canvayo, Treads canvayot (in fact T2 reads candhayo, with the pada end marker being misread by MM as at). Conze changes to canvayu1:z. yandhayo:

prasrta MM, Nel, Kl, K2: prasrta T2 . jana1:z MM, Nel, Kl, K2 : jana T2. S buddho MM: buddha Kl, K2, T2 (not noted by MM). buddha Nel. Conze changes to buddha. t dra:jtavyo MM: dra:jtavya T2, KI , dra:jtavya K2 (not noted by MM). drastavya Nel. Conze changes to dra:jtavya. U dharmakaya MM, Kl , Nel , T2: dharmakaya K2 (not noted by MM). v nayaka1:z MM, Nel: nayaka Kl, K2, nayaka T2 (not noted by MM). W ca na : MM notes Ch & Tread casya (thus T2), J reads ca na (thus Kl , K2). marrya Nel (corruption of casya?) x sa: MM notes that Ch & T read sa (thus T2), J reads sa (thus Kl, K2). sa Nel. Y sakya: MM notes that J reads sakyaf!! (thus K2; Kl sakyaf!!, corrected in red to sakyam), Ch & Tread sakya (in fact T2 reads sakyti). sakya Nel.** Z vijanituf!! MM, Nel, T2. MM notes that J & T readjanituf!! (i.e. vijanituf!!?). Thus Kl, K2: vijanituf!!.

Vaj 26a-b Chinese

1. KumarajIva (T 235, 8:752all- 18)

r~~HHft ! m-~"fr{DJ ? PJJ:J -+=fElW'l~D31Cf ?J ~~:g:m" : r~D~ ~D~ J:J -+=fElW'l~D* J {iJt" : r~~:g:m ! ;g:J:J -+=fElW'l~D* :g- , ~~niifl~.JVJ~~D* J ~~:g:mB{iJt" : rtit~ ! ~Dfl(:M{iJtpJT5t~ ,

/FI!!J:J -+-fElW'l~D* a J ~8~tit~ffij5t{~" :


2 Bodhiruci (T 236, 8:756b14- 23)

-== . D

r~~:g:m ! :a'2'~"frfDJ ? PJ t)-tElmGWly%~ ~D*/F ? J ~~:g:m" : r~Dfl(: M3<D*pJTm~ ' /FJ:JfElmGWlY%5e,3<D* J {iJt" : r3<D~3<D~ ' ~~:g: m !/F.l~MElmGgtY%~3<D* J {iJt" : r~~:g:m ! ;g:J:Ji'ElmGgtW'l~D*:g- , ~'~~~I!!~3<D* ~i&~Ft~fElpxgtY%~~D* J ~8~tit~ffij5t{~
0 C
0 0 0

Experimental core samples


3. Paramartha (T 237, 8:765c26- 766a7)

r~&{H:ft ! )Y: ~~fDJ ? DJtJ ~;Et ~~D*~ ?J ~&{HJt~ : r~DIZ~lt {~PJT5t~ ' ~tJ ~;E;fh!tm ~D* J {~~ : r~D~ , ~~:gm ! ~D~ ~ t,Z ~;E;fh!~ ~D* fDJtJi!& ? ~tJ ~;E;f ~~D*=tf ' "~~' h! ~ ~D* ~i!&~tJ ~;E;fh!il~D* J ~ 8~i!t~ffiJ5tf~~ :
0 0

~tJE~IZ E8)th!~{~

tJ1-~*IZ ~f&P)t~!t

~A1'J;f~m Il:bt~F~~

~h!l~~IZ )t:$tO)#~~


Dharmagupta (T 238, 8:771a15- 28)


rlElfDJ~~~.?;f~;E~D*~h!?J ~.~: r~~Dllti!t~! ~DIZ i!t~5t~MIZ' ~;f~;E~D*~h! J i!t~~: r~~~.! ~D~' ~D

~ , ~. ! ~D ' :$tD~)Y: 0 ~;f~;E:$tD*~h! 0

lElfDJpJT[3J? lEz.W ' ~. !


;f~;E~D*~h!~' 1El'''~~D*~ 1Ez.i!&~;f~;E:$tD*~h! llt ffi~Fffii!i:x:' ~D*~h! J m8~$=tf~.i!t~}I~D~~: r:$tDIZi!t~! i!t~5t~MIZ '~;f~;E:$tD*~h! J m8~i!t~1Ez.8~1l:tfj]o~'85t :

~IZE~ 5tJil{JI;~Ji!

~IZ~* 5t!t1Ez.~D*

;f~Mg5t1'J 5tJilR~~

~IZ~1Ez.A 1Ez.~(variants: ~ 1Ez. & i!&1Ez.~)gE:m

5. Xuanzang (T 220(9), 7:985a15- 26)

{~-5~:f~ : r~5Y:~~fDJ ? DJt,Zt;f~;Etm~D*~ ? J ~:fR~~ : r~D IZM{~pJT5t~=tf ' ~h! tJt;f~;E~ ~~D* J {JI;" : r~:fR ! ~~ !

! :$tD)y:pJT5t ~Ji! tJ t;f~;E tm ~ ~D* ~:fR ! ~ tJtffi~;Etm~D*=tf ' "'ifff~'h!~~D* ~i!i:x:~h!tJ~;f~;E~ ~~D* :$tD~Ji!tJt;f~F;f~~~D* J m8~i!t~ffiJ5t~J[EI : ttJEtmIZ tJ1-~~IZ 1El~JTi;f~ ~ ~gg &,~IZ
~~ ! :$tD~ , :$tD~
0 0 0





6. Yijing (T 239, 8:775a9- 17)

r -Ittl'lf~YJ:..
' t0 ~-ft- =fIi~f[jT
. IJ:::" I>'--. J~."t-::...


1!~ YA ~-....

rm; h/ g-kR o);8.ftD::w:-r Tt:l fEJl.)(' /1'/1'

r -rw , ++l-~ ,. /I"' J)&' YA g . . . -r rm; h/ / JI'F'(IXJ L::..~


;f ~~:$tD*

i!&~JJ! tJ ~ffi ~ ~ :$tD* JJ!tJt;f~Fffitm~~D*


J r-WJ>~ ! ~tJ~t tm~D*=tf ' "~~'h!~~D* ~ J m8~i!t~ffiJ5t



Paul Harrison


.L-~--~*fJi: b!P~ffi5~~

~A illi ~[) ii 5~ti;j~pfT~

1'"~5 ~Je,fJi:
t)z:fEZ1'"~5 7

K is the only Chinese translation to accord with the Shorter Versions in the structure of this passage, and to have Subhuti come up with the wrong answer for the first time in the text. See Harrison 2006: 156, n. 112 for a brief note on this important recensional variant. K alone lacks the second verse, in line with the Shorter Versions as represented by P. Although coded maroon, this sentence shares much of the wording with its counterpart in K, but it occupies a different position.

Selected sections of the Vaj according to Dharmagupta

Sanskrit underlined and in bold indicates amplifications not found in the Shorter Versions. Word order has not been changed in either language. Punctuation has been removed.a
Vaj 7 (T 238, 8:767c3- 1O) punar [+ apararp?] t!:!:~ bhagavan $:t:l ayu~marptarp ;gJf avocat fEZ tat {OJ kirp 1E:~ manyase ;g :t<D* tathagatena ff! arhata lE samyak ~ 9ill sarp.buddhena ~L anuttara lE~9ill samyaksarpbodhir ~~ abhisarpbuddha ~ kascid 1 va ~~ dharmas :t<D* tathagatena ~)t deSitah
~ subhutim :t<D~ etad Jf subhUte ~ kacit


;gJf subhutir aha :t<D yatha fJi: aharp t!:!:~ bhagavan t!:!:~ bhagaartham MfJi: ajanami ~ na ~ asti - sa kascid )~ dharmo ::g: yas :t<D* tathagatena ~L anuttara lE samyak ~ 9ill sarpbodhir ~~ abhisarpbuddha ~ na ~ asti - sa kascid ~~ dharmo ::g: yas :t<D* tathagatena ~)t desitaJ:!fEt tat {OJpfT kasya IZSI hetoJ:! ::g: yo fEZ 'sau :t<D* tathagatena )~ dharmo ~)t deSitaJ:! 1'"PJIf;Z agrahyaJ:!fEZ so 1'"PJ~)t 'nabhilapyaJ:! 1'" na fEz. sa )~ dharmo ;j~ na 1'"~~ adharmaJ:!fEZ tat { kasya IZSI hetoJ:! ~~)~ asarpskrtadharma OJ ~ftEl prabhavita [hi unrepresented] ~A aryapudgalaJ:!
vato~)t bha~itasya

Experimental core salllple'>


Vaj 12 (T 238, 8:768b17-22)

m?~ api tu 1!!"*~ khalu puna1;l ~. subhiite lftcp yasmin [?] tit prthivI5t pradeselft ito dharmaparyaylid 7J!i. arp.tasas 1ZY-1.D~ catu~padikam [api unrepresented?] ~ gatharp. 1.i!,1~~ parebhya1;l [?] bha~yeta ;fi va 5tJjU desyeta [?] ;fi va sa1p.prakasyeta [?] ;fi va fEi sa tit prthivI 5t pradesas 3tW caitya[bhiito unrepresented?] fiF bhavet 7:. [sa?]deva A manu~a JlnJ{I$f~ asurasya "tit lokasya foJ kal). 1!! punar ~ vlidal). ~. subhiite ;fi ya lft imarp. $ dharmaparyayarp. :f,!f" udgrh.t;lI~yanti ~" dharayi~yarp.ti [?] ~iti " vacayi~ya1p.ti ftE~ parebhyas & ca 5tJjU vistare.t;la [?] " srup.prakiiSayi~ya1p.ti miH9j parame.t;la fEi te lfsfiF ascaryet;la ~JE. samanvagata "fiF bhavi~yarp.ti lftcp tasmin rca not represented, cf. P] ~. subhfite tit prthivI 5t pradese ~gffl sasta mth viharaty JjU ~ anyataro [va not represented] #m guru }$E~{J;l. sthanlyal). ~~ iT sabrahmacari [?]




Vaj 26a-b (T 238, 8:771a15-28)

fEi tat foJ kirp. 1!~ manyase ~. subhiite ffi lak~at;ta Jt.JE. sarp.pada ~O* tathagato ~Jf! dra~tavyal). ~. subhfitir ~ aha ::f no ~Olf:t hlda1p. "tit# bhagavan ~O yatha fl(; aha1p. "tit# bhagavato bha~a ~ artham ~fl(; arana-mi ::f na ffi lak~at;ta Jt. JE. sarp.pada ~* tathagato ~ff! dra~tavyal). "tit# bhagavan ~ aha ~ sadhu ~ sadhu ~. subhfite ~~ evam ~O~ etat ~. subhfite ~O evam [etad unrepresented?] ~O yatha ~)3t: vadasi ::f na ffi lak~at;la ~JE. sa1p.pada ~* tathagato ~ff! dra~faVYaJ:i fEi tat foJpfT kasya IZSI heto1;lfEi sa[cet unrepresented?] 1!! punal). ~ . subhilte ffi lak~at;la ~JE. sarp.pada ~O* tathagato ~ff! dra~tavyo fiF 'bhavi~yat 1~[= tad] ::E raja". cakravartI [api unrepresented] ~O* tathagato fiF 'bhavi~yat fEiitl: tasman ::f na ~ lak~at;la ~ JE. sa1p.pada ~* tathagato ~ff! dra~tavya1;llf:Etal[or evarp.?] ffi 3Fffiitl: lak~at;lalak~at;latas ~O* tathagato ~Jf! dra~tavya1;l 8~ atha $:::[f ayu~man ~. subhutir "tit#JI bhagava1p.tam ~O ~ etad ~ avo cat ~O yatha fl(; aharp. "tit# bhagavan "tit# bhagavato bha~itasya ~ artham ~fl(; ajanami ::f na ~ lak~at;la ~JE. sarp.pada ~* tathagato ~Jf! dra~tavyal). m8~ atha [+ khalu?]"tit# bhagavarp.s 1~ tasyarp. 8~ veIayam lft ima 'ffia~'8 gatha abha~ata :


Paul Harrison

ye fl(; marp.

B rupel).a

~ ye fl(; marp.

fJi;, gho!?el).a



;r~ mithya mg)l prahal).a prasrta



5!:lm dharmato {~ buddho ~B! 5!:lm dharmata ca [capi?] =f" na [or a-?] ~~ vijfieya

=f" na fl(; marp. ~ drak!?yarp.ti t~ te A janalf )i;;JBr dharmakayas :fEz. [=?] 3<0
;$ tathagatalf ::fna :fEz. sab ~5 sakyarp. ~D vijaniturp.

The punctuation of T 238 is more than usually unreliable, since it tends to construe the text as Chinese, rather than Chinese characters arranged in Sanskrit word order. It should be ignored entirely. Here we accept the variant (see p. 771, n. 2) which accords best with the Sanskrit. The other readings are easily explained as attempts to turn the text into something which makes better sense in Chinese.


atha tau bhik~u etad avocatarp.1 prajfiadharo vinayadharo" 'yam upasakalf Ina tv ayarp. bhadantopalir yo bhagavata vinayadharal).am agro nirdi~talf I tav aham evarp. vadami I rna bhik!?u atra grhapatisarp.jfiam utpadayatam Itat kasmad dhetoJ;tb Itathi'igatarp. sthapayitva nasti kascic chravako va bodhisattvo va ya etasya pratibhanam acchindyat taqrsa etasya prajfialokalf Ie

There is no equivalent for this word in any Chinese translation or in the Tibetan. b Neither Q nor K has anything corresponding to the wording in bold, although it is represented in X and Tib. e Cf. Tib.: de nas dge slong de gnyis 'di skad ces mchi'o I khyim bdag 'di ni shin tu shes rab dang ldan te I bcom ldan 'das kyis 'dul ba 'dzin pa rnams kyi mchog tu gsungs pa btsun pa nye bar 'khor 'di ni de tsam ma yin no II de gnyis la bdag gis 'di skad ces bgyis so I dge slong khyed kyis 'di la khyim bdag snyam pa'i 'du shes ma skyed cig I de ci'i phyir zhe na I de bzhin gshegs pa ma gtogs par gang dag 'di'i spobs pa'i rgyun gcod nus pa'i nyan thos sam I byang chub sems dpa' de Igang yang med de I 'di'i shes rab kyi snang ba ni de dang 'dra'o II.

Experimental core samples


1. Zhi Qian (T 474, 14:523a26- 29)

~~]A!Ajt:tli : 'L~~! ~fl~~1JfHThree Editions insert: rfTFF&

ill f~f~L1fjffi::ff[Eft

fY;~: 'EH~~D*' *~~T&:g~f5$~1fT~' ~Di/:t[Three


~]&J~~:t3ill 0 J

2. KumarajIva (T 475, 14:541b29- c3)

~~= t:tli: 'L~~! ~fl~~~rfT::ff[E& f~1fZ Ljffi::ff[Eft
0 0

fY; ~D [tm omitted in Three Editions]~: 'El~~D*' *~~M&

:ft~- "'l::t/:.rl:tti:l3;Y'--'7 iz"iz

~Ji I1t:.rp~/,*iiJL.<:....tat / , E3;C; :Y:l .13':!,;;J::b.Ll-Ld..'.-'-' J


:tt~ptl 8R ;ti

"'" -++- M-J..!-r

3. Xuanzang (T 476, 14:563c14- 18) (Underlining: amplification beyond known Sanskrit text.)
8~= ~~Mft~8f4}*~~ , J9X:1'F~: I~~! ~JJ~~D~7-*rm
gli"iz s f~~i7kth~-r"'t: A ,,"",tat ' 7E !S't1b<.liJttnn' I1t:.Jx{;jp Y' -f:::f;/.~ EI

""':tt r.... r;;:;-r"'l::Y' a 7PiiJL1'Yl'i"-1&!,;;J/,-.LIIIJ/I'I1t:.iiJL oJ


fY; ~D"*: '~:ft~~1~iEe~;fJ!

a " ...

pfT tJ:t31DJ ? atp*~D* ' *~~M&


t~:g~ jffif[E~rJi/:t*~~~ , ~~~$ 8)j 7-*rm~D ~

and yet he is unable to speak!" Here X, like K, picks up Q's way of unpacking the implication of the text with words for which there are no direct equivalents in the Sanskrit (or the Tibetan).

VKN VI.6 a

aha babhUtaparikalpasya kirp mUlarp I aha abhUtaparikalpasyaC viparyasta sarpjfia mUlarp I aha viparyastaya1:l sarpjfiaya1:l kirp mUlarp I aha viparyastaya1:l sarpjfiayad aprati~~hae mUlarp I aha aprati~~haya1:l kirp mUlarp I aha yan mafijusrI1:lf aprati~~hanarp tasya g kirph mUlarp bhavi~yatii iti hy aprati~~hanamUlaprati~~hita1:l sarvadharma1:ll i


Sanskrit is that of the Ed., with variants in the Siks citation according to Bendall's edition (Bendall 1897- 1902), pp. 264.6- 9. Sik~ citation begins here. abhutaparikalpasya Ed.: Sik~ omits [not noted in Ed.]. viparyastayah samjiiaya Ed.: Sik~ omits [not noted in Ed.].


Paul Harrison

h i

aprati~thii Ed.: aprati~thiinaf(l Sik~ (in error?) [not noted in Ed.] mafijusrfh Ed.: mafijusrfr Sik~ [not noted in Ed.]. aprati~?hiinaf(l tasya Ed. (noting that MS reads aprati~thiinat (tasya)): aprati~?hiinaf(l na tasya Sik~ [not noted in Ed.]. Read MS as apratis?hii na {t}tasya and correct edition accordingly? But cf. Tib.: 'jam dpal gang rten med pa de'i rtsa bar 'gyur ba ci zhig yod de I de ltar chos thams cad ni rten med pa'i rtsa ba la gnas pa'o II. Although the switch between feminine and neuter forms (aprati~?hii, aprati~?hiinam) is a little awkward, the Sanskrit text as given in the Sik~ seems more in line with the Chinese versions, while Tib. is much closer to the MS. kim Ed.: kificin Siks [not noted in Ed.]. bh'avi~yati Ed.; Sik; omits [not noted in Ed .]. Cf. Tib.: smras pa I yang dag pa ma yin pa kun rtog pa'i rtsa ba gang I smras pa I yang dag pa ma yin pa kun rtog pa'i rtsa ba ni phyin ci log gi 'du shes so II smras pa I phyin ci log gi 'du shes kyi rtsa ba gang I smras pa I phyin ci log gi 'du shes kyi rtsa ba ni rten med pa'o II smras pa I rten med pa'i rtsa ba gang I smras pa I 'jam dpal gang rten med pa'i rtsa bar 'gyur ba ci zhig yod de I de ltar chos thams cad ni rten med pa'i rtsa ba la gnas pa'o II.

1. Zhi Qian (T 474, 14:528b20- 22)

jZJ:J~: 1::f~Z*IEA~*? J

E!: l::ff.t~*o 3<D~' e:t3! ::ff.tZ*~pfT~*' iJ::ff.t*.lz:-tJ] )t;: J


2. KumarajIva (T 475, 14:547c19- 22)


~~513UA~*? J

~E! : l aw1cr;t~~*

x.ro~ : I M1cr;f~A~* ? J
~E! :
I ~f.t~* 0


x.ro~ : 1~{.tA~* ?J ~E! : l~f.t ~U~*o ::xJ*gffJfU ! iJ~f.t*.lz:-tJ])t;:

3. Xuanzang (T 476, 14:573b17- 22) (Underlining: amplification beyond known Sanskrit text.)
x.ro~: I ~~5t3UA~* ?J

E! : I 1U;f~ [variant: ffi ] ~*

Experimental core samples


x.Fo, : I fiU;tJ(~.A~* ?J
El : I~{~*o J yrY-af-=F : I ~D~~{~J\~~* ?J
~Jf~ : 'WTFo'~FfI pfTtJ-:f3foJ ? X ~{:f3b!P~~*' ?fJ\~pfT{ EI:#~ ~* ' ~pfT{i!& ' b!Pg5~Jt-t)]Mf)~ J

VKN V1.16 a

aha aha aha aha

itas tvarp devate cyuta kutropapatsaye I yatraiva tathagatanirmita upapatsyate tatraivaham upapatsye I tathagatanirmitasya na cyutir nopapattil:l l evam eva sarvadharma1).arp na cyutir nopapattil:l l"

" Cf. Tib: smras pa Ilha mo khyod 'di nas shi 'phos nas gang du skye smras pa I de bzhin gshegs pas sprul pa de gar skye bar bdag kyang der skye'o II smras pa I de bzhin gshegs pas sprul pa la ni 'chi 'pho ba yang med I skye ba yang med do II smras pa I chos thams cad kyang de bzhin te I 'chi 'pho med cing skye ba yang meddo II.

1. Zhi Qian (T 474, 14:529a29- b2)

,@?fU9tFo'x: 1):ft)~lf:t~D~{DJ~? J xEl: I {~!HcPfT~:g~D1&:~ J El: I ~D{~{C~~F)~~lli J xEl: 'AR~ ".j@f~?fJ\/f"~~)~~:f3llio J


2. Kumarajlva (T 475, 14:548c9- 12)

,@?fU9tFo'x: 1):ft~tlt)~~~{DJPfT ? J xEl: I {~{cpfT~:g~D1&:~ J El: 1{~{CpfT~~F)~~lliO J ~ El 1"iP If-- 'f" 9ft WE ~ I:T If-- J..H ./\... AIA. :=:t:.:J@I \ \\ /\\\ /5Z~~ J


Paul Harrison

3. Xuanzang (T 476, 14:574c9- 13) (Underlining: amplification beyond known Sanskrit text.)
8~~fU-T ro5x:Y: : ,;ft:~\lf;b~[variant )1~]1ltfDJpJT ? J

x :Y:~ :
~fU-T :

I frD*pJT1 c1ltpJT~ft1lt1El 0 J I

:f(D*pJT1 c~;~[variant )1~ ] ~ ~1OJfffi1ltpJT~ ? J


xG: I ~~! t;;t;;1'~mHD3)\m ~;~ ~ ~1OJro5ft1lt1OJpJT ? J


Here Q implies the reading evam eva sarvasatviilJiiY[!, etc., as opposed to the evam eva sarvadharmiilJiiY[! of the Sanskrit text (and Tib.). This is followed by K , but oddly enough, X reflects both readings, suggesting that Xuanzang was consulting one or both of the earlier translations.

VKN IX.13'

atha tato bhojanat sarvab sa par~at trpta krtaC Ina ca tavad d bhojanarp. k~jyate I yais ca bodhisatvai}:l sravakai}:le sakrabrahmalokapalais tadanyais ca satvais tad bhojanarp. bhuktarp. tqarp. tadrsarp. sukharp. kaye 'vakrantarp. yiidrsarp. sarvasukhapratimal)~itef lokadhatau bodhisatvanarp. sukharp. I sarvaromakupebhyas ca te~arp. tiidrso gandha}:l pravati I tadyathapi nama tasminng eva sarvagandhasugandheh lokadhatau vrk~al)arp. gandha}:l Iii
Sanskrit is that of the Ed., with variants in the Sik~ citation according to Bendall's edition, pp. 269.l3- 270.3. b sarvii Ed.: sarviivat f Sik~. c krtii Ed.: bhiitii Siks. d tavad Ed.: tat Sik~ [sic Bendall's edition; Cambridge MS actually reads taJ . e sriivakaih Ed.: sriivakais ca Siks [not noted in Ed.J. f sarvasukhapratimanr/-ite Ed.: s'arvasukhamalJditiiyiim Siks [not noted in Ed.]. g tasminn Ed.: tasyiim Sik~ [not noted in Ed.J. h sarvagandhasugandhe Ed.: sarvagandhasugandhiiyiiY[! Sik~ [not noted in Ed.J. i Cf. Tib.: de nas zhal zas des 'khor thams cad ia tshim par byas kyang zhal zas de zad par ma gyur to II byang chub sems dpa' dang I nyan thos dang I brgya byin dang I tshangs pa dang I 'jig rten skyong ba rnams dang I sems can gzhan dag gis zhal zas zos pa pa de dag kyang ji ltar 'jig rten gyi khams bde ba thams cad kyis rab tu brgyan pa byang chub sems dpa' rnams kyi bde ba ci 'dra ba de Ita bu'i bde ba las ius las skyes so II de dag gi spu'i khung bu nas kyang 'di Ita bu'i dri 'byung ba ni 'di Ita ste dper na 'jig rten gyi khams spas

Experimental core samples

thams cad kyi dri mchog de na shing rnams las dri 'byung ba bzhin no II .



Zhi Qian (1474, 14:532c4- 8)

0 0

D~~ilit&&~HPJ~~' &&i!&/f'm t:f:giiJi:k%T7CWA-ttIlt&&e, ~ 5E$:;} , ~3<D-t)]$:.~cpt:f:giiJifu ~~pJTjlf:Jl ~$: 'W\3<D~


2. KumarajIva (T 475, 14:552c17- 20) D~~ffif&&~HPJ~~' 1@i!&/f'JWi[variant ~, read ;wr?]


~t:f:giiJiwM 7CA-ttIlt&&~ , ;}$: t3R~ ' ~3<D-t)] ~Mf~ ~t:f:giiJifu x.t:f~:rL ~ tfj~y~ ' W\3<D~~~ fH!fL~
0 0

3. Xuanzang


476, 14:580a29- b6) (Underlining: amplification


beyond known Sanskrit text.)


~~:k~ ~-ttll:t-tt~1~ft5Wi ' ffiJ fEj~~ B~t:fwM2SUt:f:giiJi #A7C t-Tl~nffi- ~r~t~<=I i t El,-'-' ;8i: JS$./t t-Tl -'-' ;8i:;t+- o -J-I-I..EE ::B::~+~- ~jAIA E3 ~.LJ- ~ LJ' /, ~3t * ' ' )(D- ~j 3t *il1lf>[ ~:n 0 ~Ji' 0

t)]$:~LPJT{ [variant {:IH~ 0 ;}t:f~:rL~ tfj~y~ , ~3<D-t)]~Y~tit

W~fiY~fM '*'tfj~;i:flfl~y~

Q here suggests a confusion of vrk~iiniilrl and some form of ak~af}a, possibly

as a result of using a text in Gandhari written in Kharo~!hi.


Paul Harrison

Analysis of the results The samples taken from our two Mahayana siitras are small, but sufficient to demonstrate a number of important points. In the Vaj, for which we have abundant manuscript evidence, we see a 'continuing development of the Sanskrit text. The trend is generally in the direction of enlargement and addition. Some aspects of this are documented in Harrison & Watanabe 2006, although the inventory given there (pp. 99-103) is far from exhaustive. 31 Two comments need to be made about this kind of textual development. First, we should never assume simple linear progression, as if all available witnesses can be placed on a single line, stretching from shortest (and oldest) to longest (and latest). Rather, we ought to expect multiple branching of the manuscript tradition, with enlargement and other textual changes not fully present in some of the branches, despite the late date of their witnesses. This presents the editor of texts like this with considerable problems which cannot be gone into here, but to put it in a nutshell, the idea that the wording of any Mahayana s(itra can be restored to some original and perfect state by text-critical processes must be abandoned: all lines do not converge back on a single point. Second, it is useful to think of s~ttra texts not as fixed quantities, but as prompt books or scores, which could be performed vistare7Ja or sa1'!1k~iptena (i.e. in amplified or condensed form), and therefore we might also expect this aspect of their character to be reflected in the manuscript tradition. 32 A

31 In broad terms we are dealing with the amplification of stock formulas, the insertion of the names of speakers and persons addressed, a much more liberal use of vocatives, and so on, all of which tend to increase the volume of the text without significantly altering its message. Mixed in with these changes, of course, are others which do make a substantive difference to the meaning. 32 A good example is the sequence of actions to be performed with a sCUra (learning, retaining in memory, reciting, mastering, etc.). While the Shorter Versions typically have only one or two verbs, are they to be taken as a genuinely shorter text or as cue words intended to evoke or trigger the longer sequence that we often find given in full in the Longer Versions?

Experimental core samples


further consideration relates to the distinction between what we might call "hard" and "soft" parts of the text, i.e. those portions (the "hard" ot "firm" parts) whose memorisation is not difficult, or which are so distinctive that little or no change can be expected, and those which are "soft" insofar as they can easily have other, equally plausible elements substituted, without any loss of overall coherence. All that said, the general trend is toward amplification of the text over time, or towards more extended performances, and we see this reflected in the later Sanskrit manuscript tradition of the Vaj, while the older mss, by contrast, normally carry a shorter, more compressed form of the text. By the later tradition we mean that reflected in Mi.iller's edition, which seems almost always to follow his Ch and T (whose readings are confirmed by our examination of T2), and to be generally, but not always, consistent with the Nepalese manuscripts (as, e.g., our Nel). Mi.iller, as we have seen, tended to set little store by the readings of his Japanese copies, not knowing that they would turn out to be surprisingly congruent with the older manuscript witnesses which were in his day still undiscovered (P, S, G and the Central Asian fragments).33 He outlines his approach in his introduction (1881: 17):
The text of the Vagrakkhedika, as handed down to us in China and Japan, is on the whole the same. Even what seem to be mere useless repetitions occur in all. When there is a difference, the Japanese text generally gives an independent and shorter form, as compared with the text of the Chinese and Tibetan books. But we must not ascribe too much importance to this, for it is known that some of the Chinese translators, Kumaragiva, for instance, shortened the Sanskrit texts of the Buddhist Sfitras in their translations, and this may have reacted on the originals. I have restored the text as well as it could be done, following chiefly the Chinese and Tibetan authorities, though occasionally giving preference to the Japanese text. I have not attempted to give all the various readings, many of which are misprints only, easily corrected by any one who is accustomed to the style of the Mahiiyana-sfitras.

33 This is abundantly evident even in the few short passages dealt with in this paper, as can be seen by the number of footnotes to the Longer Versions followed by asterisks.


Paul Harrison

So it was that, in virtually the same breath, Muller excl,lsed his edition from the need to meet the standard required for truly scientific work in this area and accused KumarajIva of lack of fidelity to his Sanskrit text, while in effect suppressing the evidence that could have been used to exonerate the great Kuchean translator of the charge. And that evidence, the testimony of the K6kiji copies, since Muller's day backed up by many other manuscript finds, confirms the existence of the Shorter Versions of the Vaj, copies of which had clearly reached China by the beginning of the 5th century, and continued to circulate there, at least until the Tang period, since it is then that the ancestor of the K6kiji copies was sent to Japan by Bnnin. 34 However, there are also several Chinese translations which reflect the Longer Versions, and are thus more consistent with the later Sanskrit copies of the Vaj, but this is not uniformly so. In some cases the Chinese translations contain material which we may assume was present in lndic versions still inaccessible to us, which may remain so indefinitely. This is especially true of X. However, there is another possibility, which is that Xuanzang in particular amplified the texts himself, i.e. "performed" them vistare1}a as he translated them. There need not be anything inauthentic about the versions of the text so produced, especially if he did this in Sanskrit first (or even perhaps if he did it in Chinese). He would thus have been part of a long tradition of lndic text recitation, according to which it was regarded as appropriate and meritorious to give the sutra one was reciting its most elaborate possible form, the "full monty." So much for the lndic text, an ever-flowing stream of variations which are never fully regular or predictable. In the Chinese translations, we see this variability reflected, but we also see a demonstrable tendency for some translators to go about their work with more than a backward glance over their shoulders at the work of their
34 Muller's editorial policy, which viewed in the light of our current knowledge seems astonishingly cavalier, condemned the K5kiji manuscripts and their valuable testimony to over a hundred years of oblivion. Had people known what was sitting on the shelves of the Bodleian, to say nothing of the holdings of K5kiji itself, the work on all the Vaj manuscripts discovered since 1881 would have been greatly facilitated.

Experimental core samples


predecessors, and we have given some particularly clear examples of this, where translators have borrowed their predecessor's wording wholesale, or modified it only slightly, to produce their own version of the text. We observe, for example, that Bodhi:r:uci was heavily indebted to KumarajIva, and that Paramartha also recycled much of his wording. It is clear at the same time that they both had access to copies of the Sanskrit text which were not quite the same as KumarajIva's exemplar, so that they sometimes modified his wording in the light of that text, or their different understanding of it. Dharmagupta's version, falling in the middle of the sequence of Chinese translations, is entirely different and cannot easily be compared with the others, although its word-for-word adherence to the Sanskrit text allows us to arrive at a reasonable approximation of what that may have been. That Sanskrit text cannot, however, be reconstructed on this basis, at least not with certainty, for various reasons. 35 Xuanzang's version is for the most part a genuine new translation of the Indic text. Although some of KumarajIva's wording survives in it (and thus in our samples quite a lot of red appears among the blue), this is almost always because his terminology had become the standard coinage by Xuanzang's day, and not because Xuanzang's text is derivative. Yijing's "translation," on the other hand, turns out to be the most unusual and derivative of the lot, and seems to have been put together with material taken from K and X in particular, often in an abbreviated or reworded fashion which we can assume has little to do with any Sanskrit sources, and much to do with Chinese notions of style and elegance. Yijing also seems to be ready to go to any length to maintain a four-character prosodic pattern. All in all, his translation of the Vaj is little more than a pastiche of previous versions, heavily reworked; its value for textcritical purposes is practically nil. 36
35 The two most important considerations in this regard are our inability to determine the degree to which the Sanskrit of Dharmagupta's copy of the Vaj had been regularized from the earlier Prakritic forms of the type we see in P and the Central Asian fragments, and the fact clear enough in our sample passages - that Dharmagupta did not supply a Chinese equivalent for every single word or inflection in his Sanskrit text. 36 This suggests that other translations by Yijing should be approached


Paul Harrison

Out of the six translations, then, we have to admit that only three can be trusted to any significant degree, K, Dh and X. As for the others, we simply cannot be sure of the extent to which the translators were paying attention to the lndic text in front of them in manuscript form (or being recited for them from memory), and the value of their testimony is therefore compromised. Now it would be rash to conclude that this is always the case. Each set of translations has to be assessed on its merits. Ho~ever, we cannot simply assume that a given series of Chinese translations reflects a corresponding series of lndic exemplars. This means that, unless proven otherwise, the evidential value of later Chinese translations is potentially undermined, so that, paradoxically, the most reliable translation, i.e. the one most likely to reflect its Sanskrit exemplar with minimal interference from other sources, is likely to be the first and the oldest. Even then there may be other kinds of interference. We can see in Kumarajiva's case how he was prone to inserting commentarial glosses into his translations, much as we. might nowadays (but he could not call on parentheses), to clarify the meaning of the text or make it read more smoothly. Thus his insertion of words like "definite," "fixed" (ding 5E) in 7 or "really" (shi . ) in other sections of the text can trap the unwary reader, who might take them as reflecting the wording of the lndic originalY Another example is his addition of the words dang zhi ~~O, "one should know that ... " in 12 (twice!). What this means is that the work of individual translators needs to be made the object
with caution, especially when earlier Chinese versions of the same texts are known to have existed. 37 So, for example, Alan Cole, in his Text as Father, makes a number of claims about the intentions of the Indian author of the Vaj on the basis of these interpolations by Kumarajlva (see, e.g., Cole 2005: 167-168, 183184, 186). More egregious still is his misconstrual of the last sentence of Vaj 7 in Kumarajlva's version (Cole translates: "All worthy sages are distinguished by taking lack (wu) as their teaching (dharma)."), which could only be excused if one were entirely unaware of the existence of the Sanskrit text. Unfortunately, this blunder is then put to work carrying a heavy analytical load which it has no hope of supporting (see pp. 183-185).

Experimental core samples


of systematic study, so that their particular modus operandi can be clarified. In this regard research like that of Jan Nattier on Zhi Qian or Daniel Boucher on Dharmaralqa is welcome. With all that, the attempt to reconstitute - or at least recognise the basic shape of - the Sanskrit exemplars lying behind the Chinese versions will never be an exact science, but I think we can gradually improve matters somewhat, even if this only means making the guesswork less wild. Taming our guesses may well be the most we can expect. Looking at the problem from the Sanskrit side, we see that the availability even of many manuscript witnesses of the Sanskrit text (as in the case of the Vaj) does not eliminate the usefulness of the Chinese. For example, we have seen some wording that must have been attested in some recensions, for which no Indian testimony survives. 38 Also in matters of interpretation, Chinese versions are extremely useful, since they indicate how a Buddhist reader of the 3rd or 4th or 5th century construed the text. The case of the VKN is quite different, in that, as far as the Sanskrit text is concerned, we have a codex unicus. It is, to be sure, a very exciting and important "find," but we cannot take its appearance as a reason to throw the Chinese and the Tibetan versions away. Indeed, we can see that far from reducing their usefulness, it increases it, since they become indispensable for the editing of the Sanskrit text and for working out what interpolations and scribal glosses have crept into it. It would in fact be most unwise to base all future discussion on this Sanskrit text, and to claim that this represents the VKN as an early Mahayana siitra, when it is quite clear from Zhi Qian's translation that the text has grown considerably over the centuries. However, once the edition we have now is translated into English, the danger is that this will then be taken as the VKN, and used as a basis for all sorts of claims about early Mahayana, the VKN as Nagarjuna read it, and so on. Again, Alan Cole's Text as Father shows that this is not a hypothetical situation, since he bases his discussion of what the Indian author of the VKN was about entirely on Kumarajlva's Chinese translation of the siUra.

38 A good example is the appearance of equivalents for Sanskrit sabrahmaciirin in Dharmagupta's and Xuanzang's versions of Vaj 12.


Paul Harrison

However, in this case too we see that the first translation of the text by Zhi Qian has thrown a long shadow, insofar as Kumarajlva often picks up its wording, and this should make us circumspect about relying too heavily on his version. Once again, Xuanzang's VKN seems to be a genuine retranslation, although even he appears t6 be repeating some of the wording of his two predecessors, over and above his use of established translation terminology, like shengwen M for sravaka and pusa :g:jll for bodhisattva, which is insignificant for our present purposes. As far as the establishment of a truly critical edition of the Sanskrit text of the VKN is concerned, the Chinese and Tibetan versions remain absolutely essential. The existing preliminary edition leaves something to be desired in terms of its editorial choices and in the application of its editorial conventions, although the manuscript's actual readings are - as far as I have been able to determine - recorded in the diplomatic edition with exemplary care. My impression is that this single late manuscript is a rather unreliable witness, with a considerable number of scribal errors, and that it also incorporates a fair amount of extra material, chiefly what appear to be marginal glosses which have crept into the body of the work. 39 At the same time it also omits portions of the text,40 and it would therefore be quite a challenge to edit it properly, with some serious methodological problems to sort out on the way. The conclusion to this paper will therefore come as no surprise, and is hardly likely to provoke disagreement. It is to reaffirm the utility of the Chinese translations as sources in their own right, but at the same time to emphasize the care required in their use,

39 Possible cases can be found at III.33 (Text reads tat saharthayu~mann, Ed. emends to utsahaya ayu~mann; construe as gloss and read tatsahartha ayu$mann > tatsaharthayu$mann?), VI.l5 (gatam a gloss on krta"fTl?), etc. One example of an enlargement unattested in any other version can be seen at III.36 (vinayadharo). 40 Clear cases of lacunae in the Sanskrit ms where the Chinese and/or Tibetan versions attest the missing text can be seen at, e.g., III.21 (emend to mahlivanasyanyatasmin), III.24 (possibly one folio line dropped out?), and III.45 (read sarvasa"fTlkhyavigataJ:t I fdrsasya kliyasya?).

Experimental core samples


and to draw attention to the importance of the earlier versions over the later, along with the need for their systematic study. It is also to remind Sanskritists that it would be a mistake to ignore these sources, no matter how many more so-called "Sanskrit originals" come into our hands in the next decades.

B Dh

Ch Cz Fragd

Frag e

G K Kl K2 MM Nel




Bodhiruci's translation of the Vaj. Dharmagupta's translation ofthe Vaj. Sino-Tibetan blockprint used by F. Max Milller (Sanskrit text only in Lafitsha script). Sanskrit text of the Vaj after Conze 1957 [1974]. Cat. No. 1910 in Heinz Bechert, ed., Sanskrithandschriften aus den Turfanfunden, Teil VIII (Wiesbaden: Franz Steiner Verlag, 2000), pp. 93-94. Cat. No. 1939+4194a in Heinz Bechert, ed., Sanskrithandschriften aus den Turfanfunden, Teil VIII (Wiesbaden: Franz Steiner Verlag, 2000), pp. 117-118. Gilgit manuscript of the Vaj, as edited in Schopen 1989. Kumarajlva's translation of the Vaj or the VKN. K6kiji Text 0162. K6kiji Text 0165-0167. Sanskrit text of the Vaj after Milller 1881. Nepalese manuscript of the Vaj, NAK (National Archive, Kathmandu) Acc. No. 5/186 (NGMPP B 90/16). Stein Ms of the Vaj, as edited in Pargiter 1916. Zhi Qian's translation of the VKN. Sch0yen manuscript of the Vaj, as edited in Harrison & Watanabe 2006. Sino-Tibetan blockprint used by F. Max Milller (Sanskrit in Lafitsha script & Tibetan) Bilingual woodblock edition of the Vaj kept in the library of the School of Oriental & African Studies, London. See Conze 1957 [1974]: 1, 17. Vajracchedikii Prajfiiipiiramitii. Vimalakfrtinirdda.

248 Intro.

Paul Harrison Study Group on Buddhist Sanskrit Literature (ed.), Introduction to Vimalakfrtinirdea and lfitintiloktila1Jlktira (Tokyo: Taisho University Press, 2004). Study Group on Buddhist Sanskrit Literature (ed.), Vimalakfrtinirdea and liitintiloktila1Jlkara: Transliterated Sanskrit .Text Collated with Tibetan and Chinese Translations, Part II: Vimalakfrtinirdesa: Transliterated Sanskrit Text Collated with Tibetan and Chinese Translations (Tokyo: Taisho University Press, 2004). Study Group on Buddhist Sanskrit Literature (ed.), Vimalakfrtinirdea: Transliterated Sanskrit Text Collated with Tibetan and Chinese Translations (Tokyo: Taisho University Press, 2006). Xuanzang's translation of the Vaj or the VKN. Yijing's translation of the Vaj. Paramartha's translation of the Vaj.




Bendall, Cecil (ed.). 1897-1902. (:ikshiisamuccaya: A Compendium of Buddhistic Teaching (Bibliotheca Buddhica I). St. Petersburg: Imperial Academy [reprinted Delhi: Motilal Banarsidass, 1971]. Chakravarti, N. P. (ed.). 1956. "The Gilgit Text of the Vajracchedika." In: G. Tucci (ed.), Minor Buddhist Texts (Serie Orientale Roma IX. I). Rome: IsMEO: 173-192. Cole, Alan. 2005. Text as Father: paternal seductions in early Mahtiytina Buddhist literature. Berkeley, Los Angeles: University of California Press. Conze, E. 1957. Vajracchedikti Prajiitiptiramitti, Edited and Translated with Introduction and Glossary (Serie Orientale Roma XIII). Rome: IsMEO. 2nd edition, with Corrections and Additions (Rome: IsMEO, 1974). Harrison, Paul. 2006. "Vajracchedika Prajfiaparamita: A New English Translation of the Sanskrit Text Based on Two Manuscripts from Greater Gandhara." In: Jens Braarvig (gen. ed.), Manuscripts in the Schyen Collection: Buddhist Manuscripts, Volume III. Oslo: Hermes Publishing: 133-159. Harrison, Paul. 2009. "Sanskrit Fragments of the Vajracchedika Prajfiaparamita in the British Library." In: Seishi Karashima & Klaus Wille (eds.), Buddhist Manuscripts from Central Asia: The British Library Sanskrit Fragments, Vol. Il.1. Tokyo: The International Research Institute for Advanced Buddhology, Soka University: 637-657 (+ Plates in Vol. Il.2).

Experimental core samples


Harrison, Paul & Shogo Watanabe. 2006. "Vajracchedika Prajiiaparamita." In: Jens Braarvig (gen. ed.), Manuscripts in the Schyen Collection: Buddhist Mq,nuscripts, Volume III. Oslo: Hermes Publishing: 89-132. Lamotte, Etienne. 1962. L'Enseignement de Vimalakfrti (Vimalakfrtinirdeia). Louvain: Institut orientaliste. Muller, F. Max (ed.). 1881. "Vagrakkhedikd [= Vajracchedika]." In: Buddhist Texts From Japan (Anecdota Oxoniensia, Aryan Series VoU, Part 1). Oxford: Clarendon Press [reprint Amsterdam: Oriental Press, 1972]: 15-46. Nattier, Jan. 2008. A Guide to the Earliest Chinese Translations: Texts from the Eastern Han *~ and Three Kingdoms =~ Periods. Tokyo: The International Research Institute for Advanced Buddhology, Soka University. Okukaze Eiko. 2008. "Kokiji-zo shinshutsu no bonbun Kongohannyagyo shahon ni tsuite." Koyasan daigaku daigakuin kiyo, Vol. 10 (2008): 33-42. Pargiter, E. F. (ed.). 1916. "Vajracchedika in the Original Sanskrit, Stein MS., No. D.III.13b." In: A. F. Rudolf Hoernle (ed.), Manuscript Remains of Buddhist Literature Found in Eastern Turkestan. Oxford: Clarendon Press: 176-195. Schopen, Gregory. 1989. "The Manuscript of the Vajracchedika Found at Gilgit." In: L. O. Gomez and 1. Silk (eds.), Studies in the Literature of the Great Vehicle: Three Mahiiyana Buddhist Texts. Ann Arbor: Center for South and Southeast Asian Studies, The University of Michigan: 89-139. - 2005. "The Mahayana and the Middle Period in Indian Buddhism: Through a Chinese Looking-Glass," Figments and Fragments of Mahayana Buddhism in India: More Collected Papers. Honolulu: University of Hawai'i Press: 3-24. [Reprinted from The Eastern Buddhist, New Series, Vol. 32, No.2: 1-25]. Yuyama Akira. 1967. Indic Manuscripts and Chinese Blockprints (NonChinese texts) of the Oriental Collection of the Australian National University Library, Canberra (Occasional Paper, 6). Canberra: Centre of Oriental Studies, Australian National University. ,Zacchetti, Stefano. 1996. "Dharmagupta's Unfinished Translation of the Diamond-Cleaver (Vajracchedika-Prajniiparamitii-satra)." T'oung Pao, LXXXII: 137-152.

Reopening the Maitreya-files

Two almost identical early Maitreya satra translations in the Chinese Canon: Wrong attributions and text-historical entanglements' Elsa I. Legittimo


The future Buddha Maitreya has long exerted an intense fascination and attraction to ancient and modern Buddhist civilizations. His popularity is attested in Buddhist art, literature, faiths and practices. As a matter of fact the great number of ancient sources and translations dealing with Maitreya's future buddhahood are complemented by just as many modern publications written on the Maitreya myth, its versions, its possible origin, the extant Maitreya texts and their affiliations.l However, only little attention was so
My heartfelt thanks go to Max Deeg for having organized the Symposium on Early Chinese Buddhist Translations in Vienna, in April 2007, and for his unfailing support in proofreading this paper. I also express my gratitude to the Institute for the Cultural and Intellectual History of Asia of the Austrian Academy of Sciences and to its director Helmut Krasser for having hosted and managed this event.
1 A selection of publications dealing with Maitreya's literature is given in the bibliography of the present paper. Regarding primary Maitreya text versions cf. the bibliography indicated for example by Baruch 1946, Deeg 1999, Demieville 1920, and Levi 1932. Moreover, Jan Nattier wrote a valuable appendix entitled "Major Canonical Texts Concerning Maitreya" that was unfortunately omitted from the publication (Nattier 1988). I am very grateful to her for giving me a copy of this unpublished appendix.

Journal of the International Association of Buddhist Studies

Volume 31 Number 1-2 2008 (2010) pp. 251-293


Elsa I. Legittimo

far paid to the linguistic features of one of the presumably oldest . Maitreya sutra translations extant in Chinese. We have in the present Taisho canon several S~ttras with titles, as well as content, referring to Maitreya. 2 Within the vast "Maitreya genre" five sutras focus on Maitreya's future buddhahood, its prerequisites and setting. The editors of the Taisho edition arranged these texts one after the other (no. 453-457) in volume fourteen. The first text is called the "Slitra on Maitreya's coming down to birth," Mile xiasheng jing 5i./itjff1:.*~. The following two scriptures are both called: the "Siltra on Maitreya's coming down to birth and buddhahood," Mile xiasheng chengfo jing 5i1f.Jff1:.pj(; 1!1M~. These are followed by the "Siltra on Maitreya's great buddhahood," Mile dachengfo jing 5i~*pj(;1!1M~, and the "Siltra on the time of Maitreya's arrival," Mile laishi jing 5i~*a~*~. Further Maitreya texts that mention Maitreya's future buddhahood focus on other issues. This is the case with the chapter dedicated to Maitreya in the "Siltra of the Wise and the Fool," Xianyu jing '~':J~',*~,3 the Maitreyaparip[cchii, Mile pusa suo wen benyuan jing 5i~i?f]i}im Fr:l~*m*~,4 and the "Siltra on Maitreya's Birth in the Tu~ita heaven," Guan Mile pusa shangsheng Doushuai-tian jing ilI.5i./itjJi?f]i}il::.
1:.g'la$3'::*~). 5

Of these eight texts no. 453 and no. 349 are attributed to Dharone of the best known early translators of Buddhist texts into Chinese. He worked in China between 265 and 313 AD. Thus these two sutras represent the supposedly oldest extant Maitreya sutra translations. However, as indicated in the title of the present paper, we have in the Chinese Buddhist Canon two virtually identical texts on Maitreya's future buddhahood. One of these is the "Siltra on Maitreya's coming down to birth" (no. 453) attributed to Dharmarak~a. The other one is included without a specific name in scroll forty-four of the Chinese Ekottarika-iigama
marak~a / Zhu Fahu ~:t~,

2 3

The full details are given in tabular form in Appendix 1. T4, no. 202. TI2, no. 349. TI4, no. 452.

Reopening the Maitreya-fiks


translation, the Zengyi ahan jing (t~~~PJ13*). 6 This text without any specific title is the third section of chapter forty-eight entitled the "Ten unwholesome [paths of the acts]" (dasa-akusala[-karmapatha] / shi bushan pin +/F~J'j:j). The text has no Pali equivalent in the Nikayas. The extant Ekottarika-agama translation is traditionally attributed to Gautama Sarp.ghadeva / Qutan Sengqietipo . .{~{iJD~~ and is said to have been produced in the year 397 AD. These twin scriptures were first noticed a century ago by Matsumoto Bunzaburo t~*Jt = Jj!~.7 The two texts are similar to such an extent that the possibility of two different translations can be ruled out. The setting and the content of the two slUras are the same. Whereas most of the Maitreya siitras begin with an account on Sariputra / Shelifu *fiJ5t, or at least have him as the Buddha's interlocutor, the twin texts start like an Agama siitra with the famous formula: "Thus have I heard ... " The narration is located at SravastI in the Jetavana Anathapil).<;iikarama / Shewei guo Qishu Jigudu yuan *1$:J~jftB;m*i5J1ll3~~ and Ananda / A'nan ~PJ~ft the Buddha's main interlocutor, inquires about the future Buddha Maitreya. The vocabulary of the two texts shows only minor variations and these are such that can be accounted for by copyist errors.

6 T2, no. 125, 787c2-789c28. The Ekottarika-iigama is a collection that has 51 scrolls and contains 476 siitras arranged in numerical order according to the sets of concepts or persons appearing in their subject matter. It is nominally equivalent to the Pali Aliguttara-Nikiiya but partly differs in terms of substance and content. In 1984 Thich Huyen-Vi started a serial translation of the Chinese Ekottarika-iigama. The translated sutras were published in 34 parts in the Buddhist Studies Review, from its very first issue onwards till vol. XXI, part 2. The translations appeared in sequence. The first six and half scrolls were published in French between 1984 and 1993 and the later five and half scrolls in English between 1993 and 2004. Huyen-Vi passed away in 2005. By then out of the total 51 scrolls of the Ekottarika-iigama the translation of the first twelve had been published. The English translations were made by Bhikkhu Pasadika in collaboration with Sara Boin-Webb. 7 Cf. Matsumoto's monograph on Maitreya's pure land: Miroku-ji5doron (fi1jl;!;:iiiijilll): Matsumoto 1911. His work was revised by N. Peri in the Bulletin de l'Ecole Jrant;aise d'Extreme-Orient, cf. Peri 1911.


Elsa 1. Legittimo

If the attribution to Dharmarak~a was correct, we would easily have concluded that the sutra in question was included into the Chinese Ekottarika-iigama version either by its translator or by a Chinese compiler shortly after the translation. This would explain wby a sutra that is generally not considered to belong to the oldest strata of Buddhist literature is nevertheless found in an Agama collection. But the attribution to Dharmarak~a is not beyond doubt. On the contrary, all evidence points to someone else being the tran~lator: the vocabulary, as we will see, is certainly the most convincing factor. Furthermore the old catalogues support hereto-related findings: the Kaiyuan shijiao lu (1*J:7C*'~~),8 a catalogue from the Tang period, states that no. 453 was extracted from the Ekottarikaiigama, a fact that was already noticed by Matsumoto. In Appendix IV of the present paper, I will provide a brief summary of further relevant findings from the catalogues. Moreover, already in the thirteenth century, Sugi ~~, the Korean editor-in-chief of the Koryo II canon noted that the attribution to Dharmarak~a was dubious. He added a postscript to no. 453 expressing his reservations regarding the attribution, arguing that the language is not characteristic of Dharmarak~a's time. He also compared the different canons available to him. Since the present Taisho edition is based on the canon edited under Sugi's supervision, its Maitreya sutra no. 453 (still) contains this note. Unfortunately Sugi's observations are not conclusive which might possibly be due to the fact that he had overlooked the Maitreya text in the Ekottarika-iigama. 9 'The translation of the twin Maitreya texts, in sum, appears to have been produced as part of the Ekottarika-iigama's translation, and I will thus try to tackle this problem from different perspectives, starting with the Ekottarika-iigama's translation issue.

8 Cf. T55, no. 2154, 656a9-17. This catalogue was composed by Zhisheng W~ in the year 730 AD. 9 Appendix III contains a translation of the postscript as well as an abstract of Robert E. Buswell's findings on Sugi's activities.

ReopenIng the Maitreya-fi1cs The Chinese Ekottarika-iigama translation


The Chinese Bkottarika-agama is preceded by an introduction written by DaoanJ!!3(,lO when the translation was completed in Chang'an :f:3(.11 A discrepancy exists however between this introduction that states that the Ekottarika-agama was expounded by Dharmanandin / Tanmonanti ~ and translated into Chinese by Buddhasmr:i / Zhu Fonian -tt1~~ in 384 AD, and the note within the text itself, which appears directly beneath the title and states that it was translated by Gautama Saqlghadeva in 397 AD. On this issue the ancient Chinese catalogues contain contradictory information. And the thorough investigations undertaken so far in modern times, mostly by Japanese scholars, have not yet settled the question. 12 A few years ago, however, a new era of terminological search options was inaugurated owning to the creation of the electronic database of the Chinese Canon (CBETA). This tool allows us to support linguistic and terminological observances with scientific data, and it can help to solve the translation problems related to the Ekottarika-agama. In fact, while searching hundreds of terms which Buddhasmrti had employed in his translation of the Womb siltrap I repeatedly encountered these terms in his other translations, as well as in the Ekottarika-agama. Certain wordings are only or nearly exclusively used by Buddhasmrti, no matter whether these constitute technical terms or "normal vocabulary." We even find in the Ekottarikaagama expressions that were unmistakably created by him and not taken over by later translators. In the case of vocabulary forged by Dharmarak~a, he is sometimes the last translator to have used certain terms, and such items can also be seen in the EkottarikaDaoan lived between 312 and 385 AD. 11 Chang'an is present day Xi'an W'ti: in the Shanxi ~W province. 12 Cf. Mizuno 1956 and 1989, Warita 1973, and Enomoto 1984 and 1986. 13 The Pusa chutai jing tfroi~n~*~ (T12, no. 384, 1015a-1058b), cf. Legittimo 2006b.


. Elsa I. Legittimo

iigama. In previous works I put forth the hypothesis that either the extant Chinese Ekottarika-iigama is still the first translation by Buddhasmrti (in this case the second one by Sarp.ghadeva was lost), or that the greatest part of it is Buddhasmrti's translation, and that a veritable second translation by Sarp.ghadeva never took place. The collection instead might have simply been amended or enlarged by Sarp.ghadeva.14 The question whether the greatest part or even all of the Chinese Ekottarika-iigama could still be Buddhasmrti's first translation from Dharmanandin's (most probably) oral exposition is of crucial significance for our understanding of this important but yet "unaffiliated" Agama collection and for all subsequent research related to it. Regarding the affiliation of the Chinese Ekottarika-iigama scholars mostly agree that it was not translated from an Indian version belonging to the Sarvastivadin or MUlasarvastivadin schools. Modern secondary literature generally proposes an affiliation to the Mahasamghika school or in rare cases to the Dharmagupta schooP5 Ho~ever, already in 1967 Etienne Lamotte rightly noted that there is no consistent proof for any of these assumptions,16 and his assertion is still valid today, since no significant data has been generated in the last few decades. Be that as it may, to ascertain the actual translator of the aforementioned extant Maitreya text is certainly an important step in the right direction and might help to clarify the origin of the Chinese Ekottarika-iigama.

14 Cf. Legittimo 2005: Synopsis Part I, 3, and Legittimo 2006b: 80-81. Independently of my findings, Jan Nattier also noticed that the Chinese Ekottarika-agama contains terminology typically found in Buddhasmrti's translations (personal communication). Her inferences are based on investigations she carried out on the terminology found in the "Sutra of the ten stages" (the Shizhu duanjie jing +11T~*~, T 10, no. 309, 966a41047b13), also a translation by Buddhasmrti. 15 Cf. for example the overview published by Mayeda Egaku about Japanese research on the Ekottarika-agama's school affiliation: Mayeda 1985: 102-103. 16 Cf. Lamotte 1967: 106.

Reopening the Maitreya-files


The Chinese Ekottarika-iigama collection remained unchanged since its tran~lation at the end of the fourth century. The collection preserves a lot of material, about one third of its corpus, that could not yet be put in relation to other Agama, Nikiiya or para-canonical sources, and deserves investigation to enable conclusions about the early and middle phases of Indian and Central Asian BuddhismY The vocabulary of the twin Maitreya sutras By means of careful investigation of the results obtained through electronic searches, a Chinese Buddhist translation might be successfully attributed to a particular translator on the basis of its vocabulary. I have extensively investigated all the vocabulary of the first section of the twin Maitreya texts18 within the other translations of the Chinese Canon. 19 Basically every term or formulation appearing in the twin Maitreya texts as well as at least in one other translation was taken into account. Those formulations that appear . over sixty times were left aside, as their connection with a particular translator Or group of translators cannot be established. I have then categorized the texts in which the terms are found according to their translator, if known, or to the epoch of their translation. In a further step the texts are arranged in their chronological order, and set in relation to the Maitreya text. These terminological investigations reveal that a great number of the linguistic features of the Ekottarika-iigama Maitreya text and the Maitreya sutra no. 453 - both specific Buddhist termini as well as common language expressions - are not typical for
17 The final results of my ongoing research project called "Comparative Studies on the Buddhist Canon: Analysis of the Chinese Translation of the Ekottarika-agama, the Zengyi ahan jing :tti'])iiJ13.*~" will be published in 2010. I thank the Swiss National Science Foundation for the financial support granted to this project. 18 In the Taish6 edition this corresponds for both versions to the first twenty-two lines of text, in the case of the Chinese Ekottarika-agama: T2, no. 125, 787c2-22, and for the Maitreya sutra: T14, no. 453, 421a627. 19 The 1690 texts included in Taish6 vols. 1-32.


Elsa!. Legittimo

Dharmarak~a, or even for Sarp.ghadeva, but reflect a Chinese translation idiom mainly found in Buddhasmrti's translations. 41.7% of all the occurrences in other texts of the searched terms can in fact be found in translations by Buddhasmrti. While Dharmarak~a's translations account for 0.9% of the occurrences, none of the terms are found in a translation attributed to Sarp.ghadeva. The details of this investigation are given in Appendix II. This result constitutes a small piece of evidence and supports my longstanding hypothesis that the extant Chinese Ekottarikaiigama was translated by Buddhasmrti.

Buddhasmrti's connection to Dharmarak~a and Kumarajlva

It is certainly reasonable to say that Dharmarak~a cannot have trans-

lated a scripture that contains as much vocabulary, which is incongruent to his own linguistic preferences and which overlaps with Buddhasmrti's terminological habits. The few occurrences of the searched items in Dharmarak~a's translations should be considered as terminological borrowings by Buddhasmrti from Dharmarak~a. Other translations by Buddhasmrti actually show a higher number of borrowings from Dharmarak~a than the investigated opening section of the Maitreya text. The low percentage is probably due to the fact that the beginning section has relatively few doctrinal terms that usually constitute the core of the borrowed vocabulary. It is no exaggeration to say that borrowings represent a significant aspect of the translation process and the translation history of the Chinese Canon. Dharmarak~a who translated during the second half of the 3rd and the beginning of the 4th century AD has created many important terms and formulations and it was common practice for later translators to make use of his terminology until certain wordings were reformulated by KumarajIva in the first decade of the 5th century. Even the famous scholar-monk Daoan had a particular interest in Dharmarak~a's translation corpus. And since manuscripts of Chinese Buddhist translations were hard to obtain in the middle of the fourth century, Daoan is credited with having learned numerous scriptures by heart. At a rather early stage of his career he started to

Reopening the Maitreya-fiks


collect the extant Chinese translations of Buddhist texts and related data. In 374 Daoan composed the first catalogue of the Chinese Canon, the Zongli zhongjing mulu ~;@~*~1~.20 In 379 he arrived in Chang'an where he set up a translation team. Buddhasmrti arrived at Chang'an at about the same time and started to work as a translator under Daoan's supervision. Buddhasmrti seems to have had a predilection for terminology created by Dharmarak~a. Many expressions in Buddhasmrti's translations are discernable as vocabulary created by Dharmarak~a. This tendency might be due to Daoan's educational influence. In any case Buddhasmrti's literary style and vocabulary presuppose that he was well-acquainted with Dharmarak~a's translations. It is safe to assume that Buddhasmrti knew all the Dharmarak~a translations available in Chang'an during the last quarter of the fourth century, and that he either had direct access to the texts or had learned them by heart. The use of Dharmarak~a's terminology is therefore one of the characteristic features of Buddhasmrti's translation corpus. When Buddhasmrti's vocabulary appears in Kumiirajlva's translations, however, the circumstances are different and more difficult to comprehend. The fact that certain translations by Kumiirajlva contain a great number of formulations and vocabulary that is mostly, but not exclusively, used by Buddhasmrti deserves our fullest attention. Buddhasmrti actually lived and worked in Chang'an at least between 378 and 413. This means that during Kumiirajlva's whole stay in Chang'an, roughly the first decade of the fifth century,21 Buddhasmrti was also living there. It is significant that we have - with one exception - no record of Buddhasmrti's translation activity during these ten years. The catalogues only men20 It is generally assumed that he made additions to the catalogue until he passed away in 385. Although his catalogue was lost, most of its data is included in the still extant Chu sanzang jiji Il =jijG~[l~, T55, no. 2145. composed in 515. The data borrowed from Daoan's catalogue are specified as coming from the Anlu ~~, the "catalogue of (Dao)an." 21 No matter which tradition regarding Kumarajlva's stay in Chang'an we assume as correct (arrival in 401 or 402, and death between 409 and 413), during the whole time Kumarajlva lived and worked Chang'an, Buddhasmrti was also residing there.


Elsa I. Legittimo

tion him once as KumarajIva's collaborator.22 Yet de'spite this fact a certain number of other translations attributed to KumarajIva display Buddhasmrti's linguistic influences. On several occasions I have detected, for example, that the Chinese version. of the Mahiiprajfiiipiiramitopadea, the Da zhidu fun *~BUj'ij,23 that has always and exclusively been attributed to KumarajIva, bears in fact considerable traces of Buddhasmrti's terminology.24 It cannot be excluded that in the first decade of the 5th century, Buddhasmrti worked "backstage" as one of KumarajIva's translating assistants together with other scholar monks, translators and scribes. The fact that scriptures translated by others or with the help of others carry solely the master's name - in this case KumarajIva's - is a common feature of Chinese Buddhist translation data.

22 The only explicit mentioning of a collaboration between Kumarajiva and Buddhasmrti concerns the PaiicaviJ?1satisahasrika-prajfiaparamita (Mohe banruoboluomi jing .~*~ti'rBt~IlH~:*~, T8, no. 223). The Lidai sanbao ji JB!H'(;=Jlffo2 (T49, no. 2034, 77b-79a) says that this text was expounded by Kumarajiva (Shi zhi fanwen 1t}1.~Jt), translated by the Indian Buddhasmrti (Zhu Fonian chuan-yu ~{iJIj~1t~) and written down by Ruizhao (Ruizhao bi-shou ~~$t). Cf. Hureau 2006: 98, note

23 T25, no. 1509, 57c6-756c19. The full title is Mohe banruo boluomiduo jing shilun .~~ti'rBtMBl$~~~iiil. 24 There is no reason why Kumarajiva should not have adopted vocabulary and termini used by Buddhasmrti or by other preceding translators, but ordinary usage of preexisting vocabulary cannot explain why certain translations attributed to Kumarajiva contain a relatively significant proportion, i.e. a higher number of expressions characteristic of Buddhasmrti. As the Mahaprajfiaparamitopadesa is a large scripture consisting of hundred scrolls that cover seven-hundred pages in the Taish6 edition, it is no easy task to determine whether Buddhasmrti influenced the translation on particular occasions, i.e. whether his traces are to be found only in certain chapters, whether he has "collaborated" throughout, or whether he even translated the whole text on behalf of Kumarajiva.

Reopening the Maitreya-files


The relatively early insertion of a Maitreya sutra in an Ekottarika-iigama

The fact that the Chinese Ekottarika-agama contains a Maitreya sutra might come as a surprise. The same can be said about the discovery that the Maitreya sutra in question was extracted from this Agama to be circulated as an individual translation. It would be interesting to trace the history of this extracted s utra throughout the various canons. In this respect the postscript by Sugi represents the beginning of the scholarly investigation. 25 Also the Japanese manuscript transmissions still need to be investigated. So far I could only access the data on the extant twin Maitreya texts of the Amano Kongo-ji .:R!I!Yii'i:IlUJIj~ and the Nanatsu-dera -f:; ~. 26 The canon of the Amano Kongo-ji has the Maitreya sutra corresponding to no. 453 as an independent sutra, but the canon of the Nanatsu-dera lacks this sutra. It is further noteworthy that the version of the Kongo-ji does not mention any translator's name. In both canons the Ekottarika-agama contains the Maitreya text. The principal question remains why a text of Maitreya's future buddhahood was incorporated into an Ekottarika-agama before or during the fourth century AD. Texts related to this Maitreya sutra exist in various Indian languages as well as in translation,27 thus we have several Sanskrit manuscripts of a scripture called Maitreya-

Cf. Appendix III. 26 The photocopies of this and the other Maitreya sutras included in the Kongo-ji Canon Were kindly made available to me by Prof. Ochiai Toshinori, Tokyo, of the Japanese manuscript project FRONTIER, which aims to catalogue and photograph the ancient manuscript of the Nara and Heian periods kept in manuscript collections of Japanese temples (Gakujutsu Furontia "Nara-Heian koshakyo kenkyil shoten no keisei" Purojekuto ,:pr-q:7 P ~T-17 r~R3:jZ*ll1J~1iJf5:e#lJ,,~O)%pj(;J 7PY::c. ~ t-), Tokyo .. 27 Maitreya-texts are extant in a great variety of ancient languages: PaIi, ("Hybrid") Sanskrit, Tibetan, Iranian, Tocharian, Uyghur, Turkish, and of courSe Chinese. Cf. footnote no. L


.Elsa 1. Legittimo

vyiikaralJa. 28 In Hili, the Aniigatava1!1sa is a similar text. In the Theravada tradition it is classified as para-canonical and exists in various versions. 29 The Maitreya siitra recited by Dharmanandin in Chang'an in 385 as part of his Ekottarika-iigama transmission predates the known Sanskrit and Pali versions. The same can, of course, be said of the Maitreya siitra no. 349, the Maitreyapariprcchii / Mile pusa suo wen benyuan jing 5i~lfliimFp~*M~*, translated by Dharmarak~a between 265 and 313 AD. Whereas in China the several Maitreya siitras are included in the Chinese Canon and are therefore considered canonical, and the one under discussion is even included in an Agama, the Sanskrit and Pali Maitreya siitra versions might never have been part of a canon. 30 In the absence of other complete Indian canons besides the Pali Canon, the canonical or para-canonical status of the Indian versions might just as well be undeterminable. A large number of sections of the Chinese Ekottarika-agama still remain unattested in other traditions. This is not only the case with the Maitreya text. The fact that the Indian or Central Asian Ekottarika-iigama known to Dharmanandin includes material excluded from the Pali Nikayas, should not belie the archaism of the collection. Before 385 this Agama was most probably transmitted orally, which allowed for the inclusion of material pertinent to its holders. Although more recent entries could easily be set
2~ The latest discovery regarding the MaitreyavyiikaralJa is a manuscript fragment contained in the Schoyen Collection. Its transliteration is given together with a concise overview of all the previous critically edited manuscript versions by Jens-Uwe Hartmann 2006: 7-9. 29 For a brief note on the Aniigatava1!lsa cf. von Hiniiber 2000: 98, 200. The author notes that texts concerning Metteyya/Maitreya seem to have been more popular in Buddhist schools other than Theravada. This is certainly true, but the Aniigatava1!lsa and its commentaries have nonetheless been handed down in the Buddhist traditions of Sri Lanka and South East Asia up to the present day in various versions. On the basis of Jacqueline Filliozat's cataloguing efforts of South-east Asian manuscript collections it appears that a considerable number of Aniigatava1!lsa commentaries are still extant: cf. Filliozat 1993. 30 Cf. Hartmann 2006: 7.

Reopening the Maitreya-files


within the numerical context of an oral Ekottarika-iiga111:a, neither the Maitreya sutra, nor any of the other sutras of the Ekottarikaiigama that also mention Maitreya, can demonstrate that the collection is a younger compilation. Rather, these sutras signify that Maitreya gained popularity at an early stage of Buddhist development, a proposition further attested by Maitreya's portrayal at the very early stages of Buddhist art from Mathura and' Gandhara. 31
A note on Maitreya in the Chinese Agamas

Except for the Maitreya text under discussion, Maitreya appears another thirty-four times in the remaining Chinese Ekottarikaiigama. These other occurrences are found in eleven of the total fifty-one scrolls of the collection. What we see in the Ekottarikaiigama is that Maitreya is mentioned in the introduction as well as in twelve different sutras. With this much data at our disposal an investigation on Maitreya's role in the Ekottarika-iigama is without doubt a meaningful undertaking. 32 When searching for Maitreya in the other Chinese Agamas it is easy to detect the far greater number of occurrences within the Ekottarika-iigama. In the Dfrgha-iigama 33 Maitreya is only mentioned once and this instance corresponds to Maitreya's occurrence in the PaIi Cakkavatti-sfhaniida-sutta. 34 In the Madhyama-iigama he appears throughout the later part of a section called the "Sfitra expounding the origin," Shuo ben jing IDt**~, that could so far

31 Cf. the chapter "Sieben Buddhas und Maitreya" in Zin 2003, p. 457470, in particular footnote 62 and 63, p. 464. The oldest Maitreya portrayals date from the first half of the 2nd century AD. 32 An evaluation of each of these sutras can yield results that may help to understand the overall circumstances of how "Maitreya found his way" into the Ekottarika-iigama. Such an investigation is planned as part of my presently ongoing Ekottarika-iigama project, for which see footnote no. 17. 33 Cf. the Chang ahanjing *1lJiJ~*~ (T!, no. 1, expounded by Buddhayasas and translated by Buddhasmrti between 412 and 413). 34 Dfganikiiya XXVI, PTS edition, vol. 3, section 25-26, p. 76.


. Elsa I. Legittimo

not be traced in the Pali Canon. 35 Further, the two extant translations of the Sa1!lyukta-agama 36 do not mention Maitreya. Among the extant Chinese Agamas the Ekottarika-agama thus stands out as the one which is most fond of Maitreya. It goes without saying, however, that the Agamas now extant in Chinese are of different school affiliations, and that "the responsibility" for any particularities found in the Chinese Ekottarika-agama could lie with the denominational transmission of this particular collection prior to its translation.

35 Cf. the Zhong a hanjing 'P~j;l2;t (Tl, no. 26, scroll 13, SlOb-SUc). The extant version is said to be a translation by Sa:rp.ghadeva from the very end of the fourth century on the basis of a manuscript belonging to the Sarvastivada school. The collection, however, had been translated by Buddhasmrti thirteen years earlier on the basis of Dharmanandin's (apparently oral) exposition. Sa:rp.ghadeva arrived in Chang'an immediately before this first translation was undertaken. Due to political troubles, the first translation had to be finished in a hurry and under difficult conditions. In later years, when Sarp.ghadeva was able to read Chinese, he is said to have realized how bad the translation was and that it contained many inaccuracies. He was then able to retranslate it on the basis of the aforementioned manuscript. It would have been an extreme coincidence had both sources, the (probably) oral transmission line on which the first translation was based and the Sarvastivada manuscript that generated the second translation, been identical. Sarp.ghadeva thus might have reused those parts of the older translation that were not found in the newly obtained manuscript. Only a thorough investigation of the vocabulary and the linguistic features can reveal whether the satra in question, the Shuoben jing iJ1.**, might still be part of the older translation or whether it was indeed translated by Sarp.ghadeva. Analayo (forthcoming) has investigated the extant Chinese translation of the Madhyama-agama and discovered certain irregularities (personal communication) that might support this hypothesis. 36 The Za ahan jing *I~PJ-2;"* (T2, no. 99, translated by GUl].abhadra / Qiunabatuoluo 3:j(13~jijtWtiJi in the middle of the fifth century), and the Bieyi za ahan jing}jIj~~*I~PJ-2;"* (T2, no. 100, an anonymous translation from the end of the fourth or the beginning of the fifth century).

Reopening the Maitreya-files


A Maitreya passage in a Chinese Dharmapada The opening 'section of the Chinese Dharmapada (or Udiinavarga) called Chuyao jing tBlliff37 strikes as an extraordinary way to start such a verse compendium. The collection, in fact, begins with a condensed Maitreya stltra said to have been expounded by the Buddha while he was staying in VaraIJasI.38 Surprisingly (or not) this translation from the end of the fourth century was also made by Buddhasmrti. It provides a further trace of a Maitreya siitra, or rather an extract of it, contained in an older collection. Whereas the rest of Buddhasmrti's Dharmapada version is composed in a style that alternates between long explanatory prose sections and one or two giithiis, its beginning section takes the literary form of a "normal" short Nikaya or Agama siitra. After the passage on the future Buddha Maitreya, the text continues with a short passage on a siitra called "the six indriyas of the kiimadhiitu" (you jing ming yue: Liugeng ledao 1fff1b EI:t;"~~J!l). 39 In the next

T4, no. 212, 609c-776a.

38 I am referring here to the very first part of chapter one called

"Impermanence" (Wuchangpin diyi zhiyi ~~6'b~-Z-) (609c). 39 T4, no. 212, 61Oal1. This might be a reference to a satra that exists in two versions in the TaishO Canon: the Modeng nit jing .R~:td~ (T14, no. 551, tr. attributed to An Shigao *i:!t~) and the Modeng niijiexing zhong liushijing .~:fr"~~%q:rt\**~ (T14, no. 552, an anonymous translation from around 317 - 420). The attribution to An Shigao is doubtful, and the text is not mentioned in Stefano Zacchetti's recently published list of ascertained translations of An Shigao (cf. Zacchetti 2007). We do not know from when this translation is and whether it precedes or postdates the other. No. 551 has 1273 characters (incl. dots), and no. 552 has 1235. The divergences between the two translations are so minimal, that the two texts cannot be said to represent two different translations of the same original. It remains uncertain which text was written first and taken as a basis for the other. The texts contain a short exposition on the six indriyas ("sense-organs") and their defiled nature as part of the Buddha's attempt to turn away the attention of a young woman from Ananda's beautiful . appearance. The girl had fallen in love with Ananda, but was strongly "discouraged" from focusing on those sense-organs for which she loved Ananda (his eyes, his mouth, etc.).


Elsa I. Legittimo

passage, probably the proper beginning of an earlier version of this Dharmapada the Buddha is said to dwell in VaisaIi. The Maitreya passage and the section on the "six indriyas" seem to be interpolations. It is yet unclear when these were added to the collection, and whether there might be further unusual coincidences between Buddhasmrti's Dharmapada and Ekottarika-iigama translations.

Final remarks
When considering content-related resemblances among scriptures that were translated by the same translator, there are a few final points that I wish to bring up. A natural (and frequent) conclusion is to claim that such similarities are due to their being products of the same translator. Notwithstanding this, we should not exclude the possibility that similarities can also be due to - yet unknown common school affiliation. The Chinese canon contains hundreds of texts of which we do not know the lndic school affiliation. In many of these cases we have no corresponding Indic source text. A certain number of these Chinese translations could be interrelated in regards to their place of origin, despite their generic difference. Intertextuality in Chinese texts translated by the same person or the same group of persons in one particular Chinese locality might indicate that the texts were brought from one particular place through the same route to the same destination, i.e. in this case to Chang'an. Texts of different genres translated by the same persog that share certain contextual motives and/or doctrinal views might have belonged to the same - yet unidentified - school or the same Buddhist community and might have even been transmitted from the same canon. Thus, it is not excluded that the community which possessed the Ekottarika-iigama now extant in Chinese also had the Dharmapada, now no. 212 in the Chinese canon, and that there could be further scriptures in the Chinese canon that once belonged to the same Buddhist school. Although at the moment it is too early to propose or try to prove precise intertextual relations, the question may be raised with reason. The cultural, religious and social Chinese environment is often assumed to have had a strong influence on the translation process of Buddhist texts and on their contents as well. It is well known

Reopening the Maitreya-files


that Daoan, who acted as the key figure in the dissemination of Buddhism in China in the second half of the fourth century, was a fervent believer and worshipper of Maitreya. 4o Buddhasmrti might have shared the same predilection. On the other hand, Daoan is the uncontested pioneer of the establishment of Buddhist canonical orthodoxy in China and the first who fought against apocrypha in a systematic way.41 Towards the end of the fourth century Maitreya worship had got a foothold in China since not long ago, and it can be presumed that at that time belief in Maitreya was more widespread in Central Asia, than in China. 42 Also the numerous scriptures treating Maitreya's future buddhahood that were brought to China from abroad before (or shortly after) the year 400 were probably more popular in their place of origin than in China. Thus, even though a canonical Indian or Central Asian Maitreya siitra was never found, the canonical, or "agamic," status of the Chinese Maitreya siUra discussed in this paper does not necessarily reflect a Chinese peculiarity.

40 The Gao seng zhuan iI1ii{~~ (T50, no. 2059, 353b26-28, composed by Huijiao, 497-554) tells us that Daoan held special repentance sessions during po~adha days and that this practice was later carried out in every temple in China. On this occasion he and his disciples would gather in front of their Maitreya statue and express their wish to be reborn in the Tu~ita heaven near Maitreya. 41 Cf. note no. 20. 42 Colossal Maitreya statues, for example the one seen by Faxian and Baoyun around the year 400 in modern Dardistan further support this hypothesis. Cf. Deeg: 2005, 111-115.


Elsa I. Legittimo

Appendix I The Chinese siitra translations on Maitreya's future buddhahood

In the table the translators' attributions are given as indicated in the Taish6 edition.


Attributed translator and date Dharmarak~a


i Beginning narration of the scripture

The scripture starts Agama sutra with the famous formula: "Thus have I heard ... " The narration is located at Sravastl in the Jetavana AnathapiJ;u;likarama. Ananda inquires about Maitreya, the future Buddha. The text starts with a praise directed towards Sariputra: He is great and wise, is able to follow the Buddha, and turn the wheel of the dharma. He is a great leader in matters of the Buddhist teachings, and it is due to his kindness towards all the living beings and for their sake that he addresses and questions the Buddha. His questions immediately refer to the future arrival of the Buddha Maitreya. The satra begins like this: "Thus have I heard . i is set on the Grdhr~ kiit~

453 (T14)

i "Sutra on Maitreya's coming down to birth" Mile xia sheng jing 5Mf~T1=o~ (also called Mile chengfo jing 5ffl



end of the 3'd, beginning of the 4th century Kumarajlva (Jiumoluoshi !\~~ ~1t) - first decade of the 5th century

454 (T14)

"Sutra on Maitreya's coming down to birth and buddhahood" Mile xia sheng cheng fo jing 5ffl


(also called Guan Mi shoujue jing ilt5ffl5t ~~ and Mile danglai chengfo jing 5Mf4\JJ-&*iJX;

455 (T14)

"Sutra on Maitreya's coming down to birth

YiJing (~


i'ii-) - first
decade of


Number of characters including dots.

Reopening the Maitreya-files and buddhahood" Mile 'xia sheng cheng fa jing fi the 8th century



near Rajagrha. After a brief introduction to the setting, Sariputra is introduced as the foremost among the leaders of the dharma. He then speaks to the Buddha in verses and immediately addresses the subject of the future Buddha. 44

456 (T14)

"Siitra on Maitreya's great buddhahood" Mile da cheng fa jing fi:tb:k


Kumarajlva - first decade of the 5th century

(also called Mile chengfa jing fi:tbP;\(;{~


At the beginning we find the formula "Thus have I heard ..." The Buddha is staying in Magadha on a mountain, the place where all the Buddhas of the past have subdued Mara. 45 Sariputra is among those spending the summer retreat together with the Buddha on top of the mountain. 46 In this version the prolegomenon is longer than in the others. Sariputra requests the Buddha to talk about the future Buddha Maitreya. Without any indication on the location the text starts by saying that Sariputra is the Buddha's foremost


"Siitra on the time of Maitreya's arrival"

Anonymous (317-420)


44 Although this sCUra has the transcription for Maitreya (Mile fi:tb) in its title, the scripture itself does not contain this transcription for Maitreya but only the translation Cishi ~.E.I;;. 45 Mara is the lord of the world of desire kamadhatu, the highest of the six heavens. To conquer Mara (xiang Ma ~1l'~) means to subdue passions and desires. 46 The mountain name is spelled Rasha-shan YEay J1J. The text of the Taish6 edition adds within brackets the translation of the Indian name and says: "This is the 'Mountain of the solitary end'" (Gujue-shan ye tJ.MfJ1Jili). I thank Max Deeg for the suggestion that the Chinese transcriptions Rasha YBllYand Rashana BllYll~ might refer to the Indrasaila mountain.


Elsa 1. Legittimo disciple and that due to his mercy he reflects on those who dwell under heaven. He approaches, pays respect and questions the Buddha on Maitreya. Huijue (7!;t ft), middle 5th century 7146 The chapter has a "classical" satra beginning: "Thus have I heard. Once the Buddha was staying on the Grdhrakilta at Rajagrha. He was accompanied by 1250 disciples." At that time the king of BaraI,lasI was Brahmadatta. To him a goldencoloured son is born endowed with the thirty-two major and all the minor signs of a great being. The newborn clearly bears the insignia of Maitreya and thus his parents are extremely happy. The story then develops in an unusual way. Towards the end of the chapter the Buddha intervenes and states that this is a past story and then tells the future story about Maitreya. This last part resembles the other Maitreya-satras, but seems to be a shorter version.

Mile lai shi jing


202 (T4)

The chapter called Pllvari (?) (Bopoli yEt ~Iill) of the "Siltra of the Wise and the Fool" (Xianyu jing 'if,~*)Y

47 The satra has thirteen scrolls. The chapter concerning Maitreya's future buddhahood is contained in scroll number twelve. According to Demieville 1920: 163, the title BopoH Hl~Ji\jE is a transcription of PravarI. It might also transcribe the name Pavllri. Elsewhere in no. 202, Maitreya only appears once in each of the following scrolls: one, four, and thirteen.

Reopening the Maitreya-files 349 (T12) "Sutra on Maitreya's inquiry," Maitreyapariprccha, Mile pu sa suo wen ben yuan jing 5Jl!iji\\JJlfiiJim




-end yd, beginning 4th century

(also called Mile pusa benyuanjing 5Jl!iji\\JJ lfiiJi*~*, Mile wen benyuanjing 5Jl!iji\\JJ F,,~*~* and Mile benyuan


The text begins with "Thus have I heard. Once the Buddha travelled to the country of the Vrjis.'''19 Although the "Sutra discusses Maitreya's future buddhahood, this theme is by no means the only main topic of the scripture. The sutra acknowledges both: the bodhisattva Maitreya, the disciple of the Buddha, as well as the future Buddha Maitreya. The sutra is mostly devoted to highlighting the (present) Buddha's qualities and merits. The first interlocutor to the Buddha is the bodhisattva Maitreya himself, the second but not less important one is A.nanda. The sutra begins like no. 453 with the standard formula on the Buddha being in the Jetavana park in Sravastf, but then the text continues with the Buddha's miraculous golden light emISSIOns. This is followed by the enumeration of the Buddha's worthy disciples, such as Sariputra, and the number oftheirretinues. In this text also the Buddha speaks

452 (T14)

"Sutra on Maitreya's Birth in the Tu~ita heaven," Guan milepusa shangsheng doushuai-tian jing n5Jl!iji\\JJlfiiJi

Juqu Jingshengal3.~ JR~),mid-


dIe of the 5th century

(also called Mile shangsheng jing 5Jl!iji\\JJ


48 The Chu sanzang jiji supports the attribution to Dharmarak~a, cf. T55, no. 2145,8bl0. 49 Pi tit in Piqi guo t1tjji.a;~ is either an alternative or a mistaken character for ba:f:ft as found in Baqi guo tftffUJ;~, the country of the Vrjis.


Elsa I. Legittimo to Ananda. The scripture discusses Maitreya's future buddhahood, but differs from the other sUtras as it rather focuses on the marvellous performances and qualities of a Buddha. Moreover this sutra also belongs to the "genre" of scriptures that know "both Maitreyas": the bodhisattva, disciple of the Buddha, listening to the present discourse and the future Buddha of the same name.

Appendix II The vocabulary of the twin Maitreya texts The aim of the second appendix is to show the details of the investigation on the vocabulary of the first section of the text. A total of nineteen terms or formulations comply with the abovementioned prerequisites, i.e. they appear at least once in another translation and not more than sixty times in total. The items appear:
11 times in translations preceding Buddhasmrti and out of these, two are found in translations by Dharmarak~a, 9 times in texts with yet unclear chronological order in relation to Buddhasmrti's translatisms, i.e. in texts that were translated either before, during or shortly after Buddhasmrti's working period (378-413), 94 times in Buddhasmrti's translation corpus, 14 times in Kumarajlva's translation corpus, and 74 times in translations postdating Buddhasmrti.

None of the terms are found in a translation attributed to Satp.ghadeva. Excluding the Maitreya text the terms are contained another 23 times in various passages of the Chinese Ekottarika-iigama.

Reopening the Maitreya-files


Since the purpose of the investigatiQn is to ascertain the translator .of the twin Maitreya sutra, and as there are good chances that he is the translator of the other parts of the Ekottarika-iigama, these 23 occurrences in the Ekottarika-iigama have not been counted for anyone particular translator. In total the nineteen terms occur 263 times within the translations of the Chinese Canon. Considering that the Maitreya sutra was duplicated we can count 244 distinct textual occurrences, and 225 occurrences without the Maitreya sutra. As noted at the beginning of the paper, this means that Buddhasmrti's translations account for 41.7% and Dharmaralqa's translations for only 0.9% of the 225 occurrences. The searched terms are underlined and the results explained in the footnotes. The characters differing in the two Maitreya texts are underlined with dots, and the readings of the Japanese Kongo-ji manuscript are given in bold characters when coinciding with one or the other version. In case the manuscript contains a third variant reading this is put into squared brackets within the text version of the Ekottarika-iigama. The characters in brackets refer to the foregoing characters. A question mark indicates that the manuscript has an unreadable character, and a minus sign that the foregoing character is missing in the manuscript.


Elsa 1. Legittimo The "Slitra on Maitr)!ya's Descend and Birth," Mile xiasheng jing 5ffliflJ'J

Chinese Ekottarika-iigama, Zengyi ahanjing \Y~fliiJBr5o (T2, no. 125, 787c2 - 789c28, EA no. 414) section 3 (~)

(TI4, no. 453, 421a5 - 423b13)

flfltzot&, -~if, 1~1:E'i%i'w~jjiJJ;;j:[.J"*;5"Rli 1f1ll~, ~*Jt.frtf/t3isAi!l,o


ootmt&, -~if1~, 1:E'i%i'~J~fftB;iM*;5"tJJl 1f1ll~, ~*Jt;frtf/t3isAf~o

*~iffliiJ~ft~~::t:lm::t:l~t*:l:1J1. Sf~.
":tzo*~~, j!!PJ~~,

53": tzO*~~5\


~*i@\:'l.1:E-=-ilt"&~1lFl To ~i@\~1~fr~*~~56o *-T~ji)i JtfJt:?tp57~~~Z, -t;iJst;iJ,

~1!!Ii~t;iJ, ~U~~o

~~*~A~~i*: Wrll~7i%iJo tzo4-:'l.1:E~ffT.~To ~~IlFlT,


~*i@\:'l.1:E-=-ilt"&~IlFlTo i@\~1~fr~*~~o *-T~ji)i KiJt:?tp "&~IkDZ, -t;iJst;iJ, ~1!!Ii~t;iJ, ~~~~o ?iJ'1;j[p'o.~ :

tzo4-:'l.1:E~ffT.~To ~1;j[IlFlT,

~::E*~AJ3.';fr~*, ~IJll~)j3IJo

ff~*Rii5fljljwtl:\:'l., .~lEJl;
~flflA~o *~'<lIiN.58,1~:%'I:~,

ff~*Rii5ffliflJ'JtI:\:'l., .~lEJl;o
~flflA~o tE.~'<lIil:t 1~:%!I':~

50 Scroll 44, chapter 48. Translation attributed to Gautama Sarp.ghadeva (Dongjin Jibin sanzang Qutansengqietipo yi *~lJjl~~a:{~ilJot:~~). 51 Translation attributed to Dharmarak~a (Xijin Yuezhi sanzang Zhu Fahu yi

j~ fj~~~~$~~).

52 T2, no. 140, 862c17 (the Anabindihua qizi jing jlDJjj~5:J~~~1t.t:T~~, tr. attributed to An Shigao '*'i:!t~, second half of the 2nd century); TI3, no. 397, 212b22 (the Saf[1.nipiitasiitra, the Da fangdeng daji jing *:1J~*~~&, section 12 - the Ak~ayamatinirdeSa - tr. by Zhiyan 9&~ and Baoyun .~ in the early middle 5th century); 44 times in T22, no. 1428, (the Sifen lii [gJj-ff, tr. by Buddhasmrti), <total of 48 appearances>. 53 It is a little but remarkable difference between the two versions that Ananda addresses the Buddha with bhagavat (shizun ilt.) in the Ekottarika-iigama, and Buddha (fo 1~) in no. 453. 54 T4, no. 212, 650b17 (the Dharmapada, tbe Chuyao jing tl:\1l1U, tr. by Buddhasmrti), <total of 3 appearances>. 55 Chin. Ekottarika-iigama (EA), T2, no. 125, 661cl4; T2, no. 128a, 839b25 (Xumoti nU jing ~~~:tcr, tr. by Zhi Qian jz~ and others in the first half of the 3rd century); T4, no. 212, 717c3, <total of 5 appearances>. 56 EA, 790a22; TIO, no. 309, 100lb17 (the "Siitra on the Ten Stages," the Shizhu duanjie jing +1:tlWTiffir, tr. by Buddhasmrti); TI6, no. 656, 34b2 (the Pusa yingluo jing ~ji)i~l1.~, tr. by Buddhasmrti), <total of 5 appearances>. 57 EA, 708b28 and 708b 29; T4, no. 212, 683c23, 684a18 and 746c16; TIO, no. 309, 1027cl4, <total of 8 appearances>. 58 EA, 758b12 ; T4, no. 212, 677b13 and 689b27; TIO, no. 309, 1012b17; TI6, no. 656, 77bI6, <total of 7 appearances>.

Reopening the Maitreya-files


59 EA, 710b4, 786c4, 787b20 and 787c10; two different translations of the KarU1;tiipuTp!arfkasutra - T3, no. 157, 169a22, 185blO, 215a8 and 216b7 (the Beihua jing ?~i!iM~, tr. by Dhannak~ema / Tanwuchen B;~~ between 414 and 421, and T3, no. 158, 235b26 (the Dacheng bei fentuoli jing **~0-~tflj~, anonymous translation, 350-431); T4, no. 200, 255c4 (the Zhuanji bai yuanjing, juan disi ~#'B~~~mlm, tr. by Zhi Qian ~~ in the first half of the 3rd century); T6, no, 220, 803a7 and 851c21 (the Da banruo boluomiduo jing *~ffJii ~~~~, tr. by Xuanzang j;;:!!il in the middle of the 7 th century); Tl2, no. 360, 270a6 (the Wuliang shou jing ~ji~~H~, tr. by SaIp.ghavarman I Kang Sengkai ~{!!:l~, in the middle of the 3rd century); Tl3, no. 397, 111a9 (the SaY(lnipiitasutra, the Da fangdeng daji jing *:/J~:;k#,*~, section 8, tr. by Dharmak~ema, 414-421); T27, no. 1545, 360c23 (the Abhidharmamahiivibhii~ii / Apidamo da piposha lun JliiJmJl~*m~~~, tr. by Xuanzang); T28, no. 1546, 267a16 and 336b16 (the Apitan piposha lun JliiJm.m~79~, tr. by Buddhavarman / Futuobamo n~tJ/;b(:. and Daotai ~* between 427 and 439); T29, no. 1558, 74a3 (the Abhidharmakosasiistra / Apidamo jushe lun JliiJmJl~{!'!,~~, tr. by Xuanzang); T29, no. 1562, 550bll (the Apidamo shunzhenli lun JliiJmJl~JllfilEl'!~, tr. by Xuanzang); T29, no. 1563, 867a21 (the Apidamo zang xianzong lun JliiJmJl~' .mHj~~, tr. by Xuanzang); T31, no. 1598,411b3 (the She dasheng lunshi ~~, tr. by Xuanzang), <total of 23 appearances>. 60 Tl, no. 1, 101c2 (the Chinese Dfrghiigama, Chang ahan jing iii:JliiJ-@I*~ tr. by Buddhasmrti); Tl, no. 7, 220c28 (the Miihaparinirvii1)a sutra, the Da banniepanjing *~!E~t&, tr. by Faxian rt.D between 414 and 420); EA, 787c12; T3, no. 156, 128b20 (the Dafangbianfo baoenjing *:/J~{~~,~*~, anonymous translation of the 1st or 2nd century); T3, no. 159, 292c14, 294a12 and 330c15 (the Dasheng bensheng xindi guan jing ***1:A:,,!fufit&, tr. by Prajiia / Banruo ~ff, end of the 8th or beginning of the 9 th century); T3, no. 163, 391alO (the Miaose wang yinyuan jing f!'P-@.:X:j2;J~~, tr. by Yijing ~f-- between 700 and 712); T4, no. 200, 217b17; T7, no. 220, 995c1; T9, no. 278, 500a2 and 512c20 (the (Buddha]avataY(lsakasutra, the Da fangguang fo huayan jing *:/Jll~i!j!i ,IJ~, tr. by Buddhabhadra / Fotuobatuoluo {~~tJ/;b(:~t~ between 408 and 429); Tl2, no. 339, 100a3 (the Dewugou nil jing q~~W*~, tr. by Gautama Prajiiiiruci / Qutan Banruoliuzhi .~~ffmt~ in the first half of the 6th century); Tl4, no. 482, 663c21 (the Chi shi jing ~ilt*~, tr. by Kumiirajlva in the first decade of the 5th century); Tl5, no. 614, 274b3 (the Zuochan sanmeijing ~jj/!Ii-=-~~, tr. by Kumiirajlva); Tl6, no. 665, 432b5 (the Jinguangming zuisheng wang jing 4i:7tfj}j ~M13:-*~, tr. by Yijing); T17, no. 721, 160blO (the Zhengfa nianchu jing lErt.~ ~t&, tr. by tr. by Gautama Prajiiiiruci); T22, no. 1425,386c17 (the Mohesengqi li.i ..g.r{~frlf;1$, tr. by Buddhabhadra); T23, no. 1442, 869b3 and 873c5 (the Genben shuo yiqieyoubu pinaiye ;fH:*~--IWff:g:~m~Jf~, tr. by Yijing); T24, no. 1448, 68b19 (the Genben shuo yiqieyoubu pinaiye yao shi ;fH:*~--IWff:g:~m~~~*,



. Elsa I. Legittimo

i ~}l!!,~~, ~:(EIL"tg61o ~~, i ~iiJn111~~Wc62, E!P~tJmo

Ii~~~ iiii~~~s~-~~*~ii------! -Jia~: ~.*~~-s-:-ff~*~3ii-------'

I Jf[i+- il1O, ii~t-l:;il1O, i lliJl~~, ABC~~, WltfiX;1To

! ~n111~~Wc, E!P~t.*.o

. ~,Il!!,~~, }\:(E'L,'tgo ~~,

i Jl:t~J'f-o ~1f~*G;t,S~[~]lm630 I Jf[i+- il10 6 ii~t-tiE1065, \ I lliJl~~66,AJ3!Jf&~, WltfiX;1To

I Jl:t~J'f-o ~1f~n1SS~.llJio

~_~~!~_~~_~~~~!~7~7p~7~___________L~~_!~!!jf~~~. S **'0


tr. by Yiing); T25, no. 1509, 536c7, 540c14 and 736a1 (the MahCiprajfia.pa.ramitopadesa, the Da zhidu lun tr. by Kumarajlva I Jiumoluoshi ~"ff-l1t) ; T32, no. 1646, 352c2 (the Chengshi lun fiX;1rllffil, tr. by Kumarajlva), <total of 27 appearances>. 61 EA,708c11 ; T3, no. 155, 115a15 (the Pusa benxing jing ~iil*1T~, anonymous translation 317-420); T9, no. 263, 128c25 (the Lotus, the Zhengfa huajing lEl't~~, tT. by Dharmarak~a in 286); Tl4, no. 432, 77c10 (the Shijixiangjing tftH~, anonym. translation 350-431), <total of 6 appearances>. 62 EA, 767b7 and 769a3 ; T2, no. 149, 874b15 (A'nan tongxuejing ~iiJnlRJ~~~, tT. by An Shigao *ift~ in the second half of the 2nd century); T4, no. 194, 143b1O (the Sengqieluocha suo jijing {~{jjD~~ljm~~ tr. by Buddhasmrti); Tl7, no. 814, 782b29 (the Xiangyi jing ~mH~, tr. by Dharmamitra I Tanmomiduo tt"~$ between 424 and 442), <total of 7 appearances>. 63 The town name found in the EA (Jitou ~lm) appears also in other translations by Buddhasmrti as well as in a few texts by other translators, but none by Dharmarak~a. The town name found in no. 453 (Chitou lf3]lm) is elsewhere only found in the two Maitreya sutra translations by Kumarajlva: no. 454 (the Mile xiasheng chengfo jing 5mtJJT1:.fiX;{~~~) and no. 456 (the Mile da chengfo jing 5m :tJJ*fiX;{~~). These town names have not been taken into account for the final count. 64 Tl, no. 1, 120a1; EA, 609b27; T3, no. 190, 659c27 and 664a16 (the Fo benxing ji jing {~*1T~~, tr. by Jfianagupta I Shenajueduo f1E~~JUl$ in the second half of the 6th century); TIl, no. 310, 430a9 and 465a14 (the Ratnakuta, Da baoji jing *Jf~~, section 16, tT. by Narendrayasas I Naliantiyeshe JJGig!J~lf~* in the second half of the 6th century); Tll, no. 320, 974a17 (the Pita.putrasama.gama, Fuzi hejijing :X:-=fil-~~~, tr. by Richeng f3~ in the 11th century); T22, no. 1428, 782b1 and 91Oc27, <total of 11 appearances>. 65 The three occurrences coincide with three appearances of the preceding expression: Tll, no. 310, 465a14, T22, no. 1428, 782b1 and 91Oc27, <total of 5 appearances>. 66 EA, 609b28 and 731b29; T4, no. 194, 121b27 and 135c21; TlO, no. 309, 1030b6; T22, no. 1428, 782b2, <total of 8 appearances>. 67 Tl3, no. 402, 553a26 (the Baoxing tuoluoni jing Jlwtff-lJj~,*, tT. by PTabhamitra I Boluopomiduoluo 7$:ff-l~~$ff-l in the first half of the 7th century), <total of 3 appearances>.


Reopening the Maitreya-files


An inversion of two characters has occurred in one of the texts. EA, 688b25, 688b28, 827c6 and 827c8 ; T4, no. 194, 116b21; TlO, no. 309, 970a4 and 970b4; T20, no. 1134A, 576clO (the lin 'gang shouming tuoluonijingfa ~jiQJIjS~~ta:JBff'~, tr. by Amoghavajra / Bukong /F~ in the 8th century); T27, no. 1545, 29c29 (the Abhidharmamahlivibhli~ii, the Apidamo da piposha lun Ili Jm ~m*m~i'j>~, tr. by Xuanz~g); T28, no. 1546, 21a17 (the Apitanpiposha lun IliiJm~m~i'j>~, tr. by Buddhavarman / Futuobamo l$~tJlJt*, 424-453), <total of 12 appearances>. 70 EA, 623b12 ; TI5, no. 627, 416b14 (the Ajiitasatrukaukrtyavinodana / Wenshushili puyue sanmei jing x~ajjlfIJ.~-=-~ff', tr. by Dharmarak~a); TI8, no. 898, 776b23 (the Pinaiye jing m~lfIlff', an anonymous translation 618-805); TI8, no. 901, 839a2 (the Tuoluoni jijing ItEa:JB~*,'!, tr. by Adiquduo 1liiJ!I!!~ ~ in 653-654); T20, no. 1180, 779b23 (the Liuzi shenzhou jing 1\"f:;pjlnff,, tr. by Bodhiruci / Putiliuzhi tfm:~$ in the first half of the 6th century); T20, no. 1184, 785 (the Bazi wenshu gui )\."f:X~tJL, tr. by Bodhir~i / P}ltixian tfm: 1W in 847); T21, no. 1331, 497b17 (the Guanding jing iinU,, tr. by Srlmitra / Bo Shilimiduoluo 1.,F'~.~a:, in the first half of the 4th century); T22, no. 1428, 783c20, <total of 10 appearances>. 71 T8, no. 228, 675b23 and 675b24 (the Fomu chusheng san fazang banruo boluomiduo jing -Ml-J:I\1:-=-~i(U~l1frBta:.~~, tr. by Shi Hu hfti:~, at the end of the 10th or the beginning of the 11 th century), <total of 4 appearances>. 72 TI, no. 1, 101c11; EA, 616a14 and 799a28; T4, no. 212, 386a18, 635a13, and 734c3; TIO, no. 300, 907a5 (the Dafangguang fa huayan jing busiyi fa jingjie fen *J:fJl'1?Il.~~/F,ID~1?Il~~:S, tr. by Devaprajfia / Tiyunbanruo m:~~11f at the end of the 7th century); TIO, no. 301, 91Ob19 (the Dafangguang rulai busiyi jingjie jing *J:f.frO*/F,ID~~~ff', tr. by Sik~ananda / Shichanantuo .X~ ~t between the end of the 7th and the beginning of the 8th century); TIO, no. 309, 1015c12, 1015c27, 1016alO, 1016a14, 1016a25, 1016b9, 1023b17 and 1040c27; Tll, no. 310, 449c7, (the Ratnakiita, the Da baojijing *ftftff,, section 17, tr. by Kumarajlva in the first decade of the 5th century, section 23, 503a20, 505b4, and 508a15, tr. by Urdhvasiinya (?) / Yueposhouna ~~1H!B in the 6th century); TI2, no. 374, 377b11 (the Mahliparinirviir;.asiitra / Da banniepan jin,g *~?~~~, tr. by Dharmak~ema, 414-421); T12, no. 384, 1024bl, 1036c27, 1040c29, 1041al, 1050b25, 1053c14 and 1054a3 (the "Womb satra," the Pusa chutai jing tfNlJJljt
68 69


Elsa I. Legittimo

~io{f,I, tI. by Buddhasmrti); T13, no. 397, 112c8 (section 8, tr. by Dharmak~ema); T16, no. 613, 245a4, 253bll, 253c25, 267a24 and 267a27 (the Chan miyaofa jing Jii!Il.iii~~It;;~, tr. by Kumarajlva); T16, no. 617, 299c4 (the Siwei liieyaofa }l!'Htllli3-~ It;;, tI. by Kumaraj.); T16, no. 619, 327c3 (the Wumen chanjingyao yongfa 1LF~iii' ~~Jtllt;;, tI. by Dharmamitra); T16, no. 643, 692b14 (the Guanfo sanmeihaijing &1~.=J*$~, tI. by Buddhabhadra) ; T16, no. 656, 19b29, 93c25 and 120a16 ; T16, no. 671, 533c26 (the Lmikiivatiirasutra, the Rulengqie jing Ai7HfJo~, tT. by Bodhiruci); T16, no. 672, 599c25 (the Lmikiivatiirasutra, the Dasheng rulengqie jing **Am{!m~, tr. by Sik~ananda) ; T17, no. 823, 853 (the Yiqiefa gaowang jing --IWIt;;ii11i.:E~, tr. by Gautama Prajfiaruci) ; T19, no. 997, 526c25 (the Shouhu guojie zhu tuoluoni jing "f~@IDJ1!.j:Jt~JE.~, tI. by Prajiia and MunisrI / Mounishili .$JE.*flJ between the end of the 8th and the beginning of the 9 th century) ; T20, no. 1060, 109a14 (the Qianshou qianyan ai dabeixinjing T-'FTJli:li;E *;lJ;,[.'~, tT. by Qiefandamo {fJo1i/';it.,); T20, no. lO72A, 169b18 and 169c29 (the Matou niansong yigui fa pin }iljlJ:$~fii1f.Ui}L)t;;6b, tI.. by Amoghavajra) ; T21, no. 1230, 162a3 (the Dalun jin 'gang zongchi tuoluoni jing *~jfr:IillJIj*J!(:fif~t~JE.~, anonymous translation 618-847) ; T21, no. 1332, 536c14, 537c18 and 539a23 (the Qifo suo shuo shenzhoujing -t{~mIDH$'j!\'J, anonymous translation 317-420) ; T25, no. 1509, 105b ; and T28, no. 1547, 519c21 (the Bingposha lun ~~lj>fAil, tr. bySarp.ghabhadra / Sengqiebadeng {~{fJoJl&l'ft in 383), <total of 55 appearances>. 73 T7, no. 220, 780a27 and 876b28); two translations of the [Adhy]ardhasatikii: T8, no. 240, 776b7 and 778a22 (the Shixiang banruoboluomi jing ":t~~tlUIH~: ~, tf. by Bodhiruci), and T8, no. 241, 779a6 (the lin'gangding yujia liqu banruo jing jfr:1ffiJ1jJ}1j)m({fJo~51j!U5t~~, tr. by Vajrabodhi / Jin'gangzhi jfr:IffiJIJ~ in the twenties or thirties of the 8th century); T12, no. 385, 1064blO (the Antariibhavasutra, the Zhongyin jing r:p~,;~, tI. by Buddhasmrti); T19, no. 946, 179a2 and 179b22 (the Da foding guangju tuoluoni jing *{~TJ:'H~~t~JE.~, anonymous tr. 618805); T19, no. 1007, 664al (the Mmtli mantuoluo zhou jing if-~~~t~'j!\,,*, anonymous tI. from the first half of the 6th century); T20, no. 1080, 189c5 (the Ruyilun tuoluoni jing tw~~~t~JE.~, tI. by Bodhiruci); T20, no. 1083, 202a2 and 202a6 (the Guanshiyin pusa ruyi moni tuoluoni jing &tftl'~ji)ifro~"JE.~t ~JE.~, tr. by Ratnacinta / Baosiwei .!~,'It, between the end of the 7th and the beginning of the 8th century); T21, no. 1335, 575c17 (the Dajiyi shenzhoujing S~t$'j!Et, tr. between 453 and 462 by Tanyao .art) ; T32, no. 1670 (B), 709bl (the Naxian biqiu jing 13B%J:tliJ, anonymous tr. 317-420), <total of 16 appearances>.

Reopening the Maitreya-files


Appendix III Annotated translation of the postscript74 Thanks to Robert E. Buswell's thorough study on the editorial process of the Koryo II Tripi!aka, the seemingly anonymous postscript that follows the Maitreya sutra no. 453 can clearly be attributed to Sugi ~~, the Korean editor-in-chief of the Koryo II canon.75 "Completed in 1251 after sixteen years of labour by thousands of scholars and craftsmen, the entire set (of the Koryo II canon) consisted of some 1,514 texts in 6,815 fascicles, carved on 81,258 individual blocks. All texts appearing in previous editions of the canon were included, making it the most comprehensive collection of East Asian Buddhist literature assembled up to that time."76 Sugi strongly suspected that the Maitreya sutra in question was wrongly attributed to Dharmarak~a, but since he did not have enough evidence for a different attribution, he left the attribution intact, but added a scholarly postscript to the text.

74 Cf. T14, no. 453, 423b14-423cl. I thank Jan Nattier, International Research Institute for Advanced Buddhology (IRIAB), Tokyo, for going through a preliminary version of this translation with me, for her valuable suggestions, and for introducing Robert E. Buswell's research to me. I am also very grateful to Christoph Anderl, Institutt for kulturstudier og orientalske sprak (IKOS), (the Department of Culture Studies and Oriental Languages), Oslo University, for revising one of my last versions of the postscript's translation, and I wish to thank those who commented on the translation when I presented it in Vienna in April 2007 at the Symposium on Early Chinese Buddhist Translations. 75 Cf. Robert E. Buswell 2004: introduction and 156-157, for the postscript in question. I quote from the introduction: "Sugi's work appears in cases chun,i, and mil of the xylographs of the Koryo canon. It has been reprinted in Koryo taejanggyong, vol. 38 (1976), p. 512-725; it appears as K. 1402 in the reprint. The text was also included in the Pinqie edition of the canon (Shanghai: Pinqie Qingshe, 1909-14), case jieh, nos. 9-10; vols. 397-98." 76 Cf. Buswell (2004), introduction. Parenthesis added by the author of the present paper.


Elsa 1. Legittimo

WOOjGWlff~1!\li J:j:lffi't:~~5fijj JtlJpMjll*~-:t 5fijj1j\;l]



I checked [the section on items] for which there is a [known] translator but no [extant] work in the Kaiyuan catalogue. 77 In [this section] there is the "Sutra on Maitreya's buddhahood" translated by Dharmarak~a, also called the "Sutra on Maitreya's future coming down to birth." At a first glance this sutra is just that lost work and we [may think we] have regained it. But in reality tpis is not so (i.e. this assumption is incorrect). Why is that so? [Because in the Kaiyuan catalogue] the note after the title of the "Sutra on Maitreya's buddhahood" translated by KumarajIva says: "[This text] is a different text from the "Sutra on the coming down to birth" but it is the same text as the "Siltra on Maitreya's buddhahood" translated by Dharmarak~a. There are two translations and one is missing." So that lost text is not this (i.e. the present one) "Siltra on the coming down to birth." It is evident that [this text] is one of the three lost translations among the six translations. Moreover, according to Gushan Zhiyuan's78 reedition (i.e. collation) of the Vajrapraj/iii (:&/ffiIIH9:~r9 the following is stated in the postscript: "When the ancient virtuous ones were distinguishing Slltras they used [the method of] counting the pages. 80 One page consists of twentyfive lines containing seventeen Chinese characters each."

'FW'lJl:t*~o ElPflt:'ft;: *lliJ~1~Zo~.


1~U~ ~ rl';Z3;o ~T

1:.*~~*~jt;;~~ 5fijjJtlJpx1~*~[PJ*0


:A~-=-:'ft;:z-*13Fl ~o

:&/ffiIiJ~9:~~IT;Z3; 0 ;:j1f1!.\)j*~.I'Jfl *S:l!<

*0 -iTS:ff=+3i1T -1T+-l::;+0

The Kaiyuan shijiao iu OOj[;~~~, T55, no. 2154. Gushan Zhiyuan 1J.llrl!~1tiI is a Chinese master (967-1022). I could not find that special notice which Sugi attributes to him, but the information he is said to have reported regarding the ancient manuscripts seems correct and is evidenced by existing manuscripts. The number of lines might vary slightly and the last page might just contain a few lines. Besides other methods, the ancient Chinese way to check the size of the extant texts for verifying whether the title and lengths of a text fit its description found in the catalogues is indeed an appropriate approach. 79 He must be referring to one of the versions ofthe Vajracchedikii, the fin 'gang jing :&/ffiIIJ*~, in its shortened Chinese title as found in T7, no. 220(9) and T8, no. 235-239.

Reopening the Maitreya-files


4'~*=*5m1f!J* Tff:i3:o-+-t*Jl; fllJnfj-tT=S=+ JJU.R1f':::' T-S-t+:f:;*o fllJf<Ij*A"o :'l!t~



If we critically compare [this information with] the note [found] under the title of the lost Maitreya satra that says: "seventeen pages," then [we can] estimate [that this text] consisted of seven thousand two hundred and twenty two characters. This (i.e. the present) satra only has three thousand one hundred and seventy six characters. Since this is less than half that size, how could it possibly be that satra?

1~0 ~JlU;tMH~ ~lf*ffo JZ1f~ :i3:::Z~o~~Jlt~.:::.

*=*q:;m-*o~ :i3:~mW1f*II=o

Although the Khitan Canon (Danzang ftit) does not have this satra, the style of this scripture rather resembles that of the satras and commentaries of the Han ~ and Jin 1f dynasties. In addition, it contains words spoken (i.e. used) during the Han dynasty.81 I also suspect that this is the first (anonymous) translation (lit. text) among the three lost translations (lit. texts) [of the Maitreya satra that was translated six times]. [In support of my assumption] the [Kaiyuan] catalogue states [about this translation]: "Now it is attributed to the Western Jin period."82

80 I.e. they counted the paper sheets, which when glued together constituted a scroll (or several scrolls). Particularly in the case of scriptures that have variant versions, various translations, and that are known under different titles - especially when it comes to shorter texts - only the number of pages might help to differentiate similar texts. Cf. note 78. 81 As noted by Ch. Anderl the formulation han yun zhi yan ~:i3:::Z i3 is indeed particular. Within the Chinese Canon yun zhi yan :i3:::z i3 is not found a second time. And han yun r~:i3: is mostly used in the catalogues for indicating the Chinese meaning of a foreign name or term that has been transcribed into Chinesephonetically. It is a typical feature of Buddhasmrti's translations that they contain older vocabulary. Sugi has correctly noticed this feature regarding this text although he did not link the translation to Buddhasmrti. 82 The Kaiyuan catalogue has the following entry; "The Mile danglai sheng jing 5m1f!J'&3f<:1::*, one scroll: the Chu sanzangjiji (i.e. the catalogue by Sengyou) says that [this text] was [already] in Daoan's catalogue, included in [the section containing] the texts for which the translators were unknown (lit. lost). Now it is recorded as the first translation of the Western Jin period (265-317)," cf. T55, no. 2154, 629c28.

*~~1~rm~~}\.:z ~1~:Z~o rm=~ .illz~T1:.*~;lHt:~ ~*o 4-:Z3:i't:~~ *{iiJJf~o

Elsa I. Legittimo To the [editors of the] Song canon it was also available and having obtained it they included it. But both catalogues (the Chu sanzang jiji and the Kaiyuan iu) did not list the "Sutra on the coming down to birth" as being translated by Dharmarak~a. How can it be then that it is nowadays regarded as a translation by Dharmarak~a? I will submit this to the wise ones. 83

Appendix IV An overview of further relevant findings from the catalogues The overview is focused on our main Maitreya satra no. 453 called Mile xiasheng jing sm1~lff:~!:J&. In regard to this title, however, it is important to keep in mind that it might also refer to the satras no. 454 and 455. For the sake of correctly presenting the data of the catalogues, I will primarily use the Chinese names. The data presented in this appendix is drawn from the Chu sanzang jiji Il '='~~[i,:* (T55, no. 2145) composed in 515 by Sengyou ({~tt), the oldest extant critical catalogue, and the Kaiyuan shijiao lu OO]1;*~ ~~~ (T55, no. 2154) composed in 730 by Zhisheng ~*. Before making this choice, I have also searched other Chinese secondary sources of the 6th and 7th century, and verified whether they contain important supplementary data. An investigation on Maitreya sfitra related information, for example in the Lidai sanbao ji HiHi: - . ff-c (T49, no. 2034) composed by Fei Changfang Jtjjtw in 597, did not yield relevant findings: This and later sources have hence been excluded. The author of the Chu sanzang jiji mentions the Mile xiasheng jing sm1fiJfF1J& twice. First, he includes it together with a further Maitreya s~ttra, the Mile chengfo jing sm1fiJJpjG1*~&, in a list of thirtyfive translations attributed to Kumarajiva. 84 Note that the title of
83 84

Which means: This matter awaits the judgement of future scholars. For the mention of the text cf. T55, no. 2145, lla5-6, for the expla-

Reopening the Maitreya-fiJes


the other Maitreya text probably refers to no. 456. Later the Chu sanzang jiji mentions the Mile xiasheng jing fi~iJT1J~ as a different (anonymous) translation in a section on newly gathered extant scriptures of which the translators' names have been lost. 85 Among the scriptures that were extant and of which the translators' names were lost it also lists a sutra called Mile shoujue jing fi1WJ~tR:*~. 86 The author of the Kaiyuan shijiao lu at first mentions the sutra in question while referring to another one. The listed text is actually the Mile laishi jing fi1WJ3fJ~*~ and it is said to be the third translation of the scripture called Mile xiasheng jing fi~jff~~EJ~, translated by KumarajIva as well as by others.8? The Mile laishi jing fi1WJ*S~*~ is mentioned in the list of the extant translations of which the translators are unknown. 88 Further on in the Kaiyuan shijiao lu the Mile xiasheng jing fi1WJT:'*~ also appears in the list of translations attributed to KumarajIva. 89 The catalogue says:
The Mile chengfo jing ~.Ij\]]px;f~*~ in one scroll is the second translation of the same Mile chengfo jing ~.Ij\]]px;f~*~ text that had been translated by Dharmarak~a. It was translated [by KumarajIva] in the fourth year of (the era) Hongshi 5.L#1 (402). The Mile xiasheng jing ~.Ij\]]T::1J~ in one scroll is also called the satra on "Maitreya receiving the vyiikaralJa" (Mile shoujue jing ~.Ij\]]$t:(;R; *~). The text starts with the great Sariputra's inquiry. It is the same text as the Mile laishi jing ~.Ij\]]*aif*~. It is the fourth translation. It is further called Mile chengfo ~.Ij\]]px;f~ as well as "(Maitreya's) Future Descent and Buddhahood" (Dangxia chengfo '&Tpx;1~).90

nations on the text cf. lla26-29, and for the introductory explanations to this section of the catalogue cf. 21b17-21c9. 85 Cf. T55, no. 2145, 22b29 for the text's listing and 21b17-21c9 for the introduction to the list. 86 Cf. T55, no. 2145, 32c8 for the text, 37b13-16 for explanations on. the listed texts, and (like above) 21b17-21c9 for the introduction to the list. 8? Cf. T55, no. 2154, 509c24.
88 Cf. T55, no. 2154, 503al for the title of the list and 51Obll-16 for the final comments on the listed scriptures. 89 Cf. T55, no. 2154, 513c-515.3 for the comments on KumarajIva. 90 Cf. T55, no. 2154, 512cll-12.


. Elsa I. Legittimo

Further, the Mile xiasheng jing 5m1(WTj::*~ is also said,to have been translated by Yueposhouna Jl ~Jj~ in 554. His translation is said to be the fifth translation of the sutra, and is identified with the text translated by KumarajIva. Yueposhouna is said to have translated eleven scriptures in the middle of the sixth century, all listed with their titles. Out of these eleven texts six were extant when the Kaiyuan shijiao lu was composed and five were lost. Yueposhouna's Maitreya sutra is listed among those works that were already lost. This is what we learn from the catalogue. 91 This translation might have disappeared as an individual text, but it is in my opinion not lost, as in the Ratnakuta we have a sutra translated by Yueposhouna that is probably the Maitreya version Zhisheng, the author of the Kaiyuan shijiao lu, believed to be lost. 92 By no means is it the same sutra that KumarajIva translated. Although it is also a Maitreya sutra, it is, compared to the other extant texts, a generously extended Mahayana version. The sutra is known as the Maitreyamahiisirrrhaniida. With over 20000 characters in total, it fills two scrolls of the Ratnakuta and is probably the longest available Maitreya sutra. In scroll one it contains, among other narrations, long discussions between the Buddha and Maitreya, and in scroll two it includes some of the core events also narrated in the canonical Maitreya versions: the great cakravartin king, .the increased lifespan, the arrival of the future Buddha and so forth.

91 Cf. T55, no. 2154, 538b1O for the mentioning of the text and 538a2227 and 538b15-16 for the comments. 92 Cf. the Da baojijing *.~*~ (TIl, no. 310, scrolls 88 and 89, section 23 (Mohejiaye hui ~~iiJ.lti!!~it), 501b-514b). The Ratnakuta was compiled by Bodhiruci on the basis of already extant as well as new translations at the end of the 7th or the beginning of the 8th century, and at least two decades before the composition of the Kaiyuan shijiao lu. For some reason Bodhiruci did not keep or use a Maitreya-title when he inserted the sutra into the great Ratnakuta collection. Zhisheng's failure to trace this version in the Kaiyuan shijiao lu is understandable. As the sutra had been known as a Maitreya sutra in China, how could he have identified it without its name?

Reopening the Maitreya-files


As the Kaiyuan shijiao lu is an extremely well structured and meticulous catalogue, the same text appears under several headings. We find 'the Mile xiasheng jing ~:lf9ff1:.f~ also in the section that lists the satra translations of which the translators are known, the texts extant, that belong to the 'bodhisattva tripitaka catalogue' and that are Mahayana scriptures translated more than once. 93 In the list below, the texts II), III), and IV) are identified as different translations of the same text. Moreover, the catalogue explains that the text was translated six times in total, that these three are the only existing versions, and that all the others are lost. 94
I) The Mile chengfo jing B:i\'lbJJX;#b*~ translated by Kumiirajlva is said to be a different sutra from the three following ones. Further, it is also said to be the second of two translations, the first translation being lost. II) The Mile laishi jing B:i\'lb*~*~ is said to be a translation by an anonymous translator, and the third translation of the sutra in question. III) The Mile xiasheng jing B:i\'lbT1:.ff~ is said to have also been called the sutra on "Maitreya receiving the prophecy (vyiikaralJa)" (Mile shoujue jing B:i\'lb~lS\!:*~). Its exposition starts with Siiriputra. The translation is attributed to Kumiirajlva, and it is said to be the fourth translation of the Mile xiasheng jingo IV) The Mile xiasheng chengfo jing B:i\'lbT~JJX;1?!l~ translated by Yijing ~7 is said to be the sixth translation.

The Kaiyuan shijiao lu further lists those Mahayana sUtras that have been translated mOJ;e than once of which the translators are known but the texts lost. 95 Under this section we find the seventeen pages long Mile chengJo jing ~:lf9JPX:{~~, also called Mile danglai xiasheng jing JlJ:If9J'&*r1:.f~, which was translated by Dharmarak~a. This text is mentioned as the first (lost) translation of a Maitreya-text that was translated a total of two times. The catalogue further lists the lost translations of another Maitreya-text: (I) the first anonymous translation called Mile danglai
93 94 95

Cf. T55, no. 2154, 595a8. Cf. T55, no. 2154, 595b15-29. Cf. T55, no. 2154, 626a5.


Elsa I. Legittimo

sheng jing 5jl~]J'&*::1:*~, (II) the second anonymol,lS translation called Mile zuofo shishi jing 5j1fiM'F1ifljaif**~, and (III) the fifth translation by Paramartha / Zhendi ~~ called Mile xiasheng jing 5j1fijJT::1:*~. These three translations are said to be the three lost translations of the siitra that was translated a total of six times. 96 Of great interest is the section of the Kaiyuan shijiao lu that lists those texts that were extracted from larger siitras or collections. 97 The section has been divided into several subgroups. Under the subcategory of texts belonging to the small vehicle, i.e. Agama and Agama-related texts,98 we find the following information:
The siltra called Mile xiasheng jing 5JlIij1 19fF1:.* was extracted from scroll number forty-four. In this siltra the Buddha is said to dwell in SravastI (Shewei guo ~lttr~) and is asked by A.nanda to hold an exposition. The text contains seven pages. It was extracted from the Chinese Ekottarika-agama (Zengyi ahan ~-/lPJi3).99

Finally, one of the Kaiyuan shijiao lu's most valuable sections is the 'Catalogue of what is contained in the Canon,' the Ruzang lu (A ~~) as it represents the detailed table of contents of the Buddhist Canon of its time. Also in the Ruzang lu, data has been repeated as certain sections were re-entered in full length from elsewhere. The opening section of scroll nineteen deals with the siitras, vinaya texts, and commentaries attributed to the great vehicle (dasheng jingliilun ***~1f,:~Hll).100 In scroll twenty, its second part, the scriptures belonging to the small vehicle (xiaosheng ruzang lu xia /J\ ~A~~T) are listed. 101 However within scroll twenty after the small vehicle section has ended we find once more the information given in scroll nineteen. lo2 As shown below in the translation of the
Cf. T55, no. 2154, 629c25-630a4. Cf. T55, no. 2154, 651a16 for the heading of the section. Cf. T55, no. 2154, 655a8 for the heading of the subsection. Variant writing for the Zengyi ahan jing :f:~:;:~iifi3*, cf. T55, no. 2154, 656a9-17. 100 Cf. T55, no. 2154, 680a27-b11. 101 Cf. T55, no. 2154, 691all-22. 102 Cf. T55, no. 2154, 700c24-701a8.

96 97 98 99

Reopening the Maitreya-files


relevant passages of the two scrolls, scroll twenty contains supplementary details on the translators.
Scroll nineteen 103 The Mile chengfo jing 5i1flJJM{~*&1 in one scroll has seventeen pages. The Mile laishi jing 5i1flJJ:w:a~*&1 in one scroll has three pages. The Mile xiasheng jing 5ithT1: *&1 in one scroll is also called Mile shoujue jing 5P\lith&lOS~J&1, as well as Mile chengfo jing 5ithM{~ *&1 and Dangxia chengfo 'litTM{~ and Xiasheng chengfo T1=.g)(;{)f#. It begins with Sariputra. It has seven pages. The Mile xiasheng chengjo jing 5i1flJJ T1:M{~*&1 in one scroll is a translation by Yijing ~:lii-. It has five pages. Scroll twenty104 The Mile chengfo jing 5P\li1flJJM{~*&1 in one scroll has seventeen pages and was translated by Kumarajlva. The Mile laishi jing 5P\lith:w:aif*&1in one scroll has three pages and is an anonymous translation. The Mile xiasheng jing 5ithT1: *&1in one scroll is also called Mile shoujue jing 5P\li1flJJst~*&1, as well as Mile chengfo jing 5P\lithM{~*&1, and Dangxia chengfo 'litTM{~, and Sheng chengfo 1:M{~. It begins with Sariputra. It has seven pages. It is a translation by Kumarajlva. The Mile xiasheng chengfo jing 5P\li 1flJJT1:M{~*&1 in one scroll has five pages. It is atranslation by Yijing ~


As in the case of most Buddhist scriptures that existed or exist in several translations, although the catalogues say that various translations were made of the same Maitreya siitra (Mile xiasheng jing BifYJT1:.ff), in reality all "true" translations were made on the basis of different manuscripts, and thus represent at least slightly differing versions. Siitras with similar names often, but not necessarily, belong to the same genre and expound similar content matter. Nevertheless, when comparing such texts, there are two main distinctions to be taken into account: translations of essentially different siitras, and translations of various versions of a certain siitra. It goes without saying that this distinction is not in

Cf. T55, no. 2154, 684a5-9. Cf. T55, no. 2154, 705c12-16.


105 The character shou ~ might be a mistake for shou ~ The Taish6 edition has no footnote regarding this anomaly.


Elsa I. Legittimo

every case an easy one. Even in the case of several translations that are said to have been made of the same sutra, the variations found between these translations are not simply due to different translation methods, but are mostly based on divergences found in the source texts. Thus, even if this is written in the Kaiyuan shijtao lu, we should not take for granted that six translations of exactly the same Maitreya sutra were ever made, but rather consider that the author of the catalogue, in this case Zhisheng, considered certain translations to represent the same text. This kind of classification needs to be understood in relation to other, more divergent translations that clearly do not derive from the same text. Contents and length, i.e. the number of pages, of Chinese translations are formal indicators for grouping translations as having been made from the same sutra. If we return to the still extant five Maitreya sutras, the Mile chengfo jing :I(g]::kiJX;1"'M~, now no. 456 attributed to Kumarajlva is by far the longest (8383 characters) of the five Maitreya sutras in question. Since it is not mentioned in the Kaiyuan shijiao lu under this name, the present title should be regarded as a later invention. The text corresponds to the sutra that has seventeen pages. In ancient times it was most frequently called Mile chengfo jing 5I:I(g]iJX; 1~*~, and sometimes it was also called Mile danglai xiasheng jing 5I:I(g]~31Cf1j. Kumarajlva's translation is said to have replaced Dharmarak~a's first translation. Whether both translations ever existed at the same time, or whether Dharmarak~a's translation was lost before the new translation was produced, is not clear. It cannot, of course, be verified how similar these two versions actually were. The catalogues agree however in saying that this sutra was "only" translated twice, that Dharmarak~a's translation did not survive and that this sutra is different from the other Maitreya sutras. Moreover, although the title Mile xiasheng chengfo jing 5I:I(g]T 1:.iJX;1~* is used for two Maitreya sutras in the present TaishO edition (no. 454 and 455), the Chu sanzang jiji never refers to this title. In fact, when a sutra was translated several times, its title could change considerably. The Kaiyuan shijiao lu uses this title only to refer to the sutra translated by Yijing. Besides the abovementioned entries, the Kaiyuan shijiao lu lists it under Yijing's translations

Reopening the Maitreya-fi1es


and says that it has one scroll, that it is the sixth translation of the
t~xt called Mile xia jing 5i1fg]--n~, which was also translated by

Kumarajlva, and that Yijing's translation was completed in the year 701 (Dqzu yuannian *..@:5i;1f).106 The extant TaishO no. 455 has 2258 characters and coincides well with the five pages it is said to have filled. Next, the Mile laishijing5i1fg]*a~*~ does not appear in the Chu sanzang jiji under its present title either. But the catalogue lists a sutra called Mile shoujue jing 5i1fg]~#!:~ among those scriptures that were extant but their translators' names 10st. 107 The Kaiyuan shijiao lu notes that the Mile laishi jing 5i1fg]*~*~ is an extant anonymous translation, the third translation of a group of similar sutras, and that it is the smallest one among those that survived. This text can be identified as no. 457 in the TaishO edition. We further have in the TaishO edition the Mile xiasheng chengjo jing 5i1fg]r1:.pjG~*~ (no. 454), a title which, as stated above, is not mentioned in the Chu sanzang jiji. This sutra is thought to belong to the same group as Yijing's translation (no. 455). The author of the Chu sanzang jiji however mentions two Maitreya sutras attributed to Kumarajlva: the Mile chengfo jing 5i1fg]pjG~*~ and the Mile xiasheng jing 5i1fg]r1:.*~. The first one could already be identified as no. 456, and the second one is no. 454. The sutra no. 454 is a translation of the Maitreya sutra which the Kaiyuan shijiao lu tells us was translated a total of six times. Three translations were and are still extant and can be identified: the third is the anonymous translation no. 457, the fourth is Kumarajlva's no. 454, and the sixth is Yijing's no. 455. The three lost translations (the first, second and fifth among the total of six) might have been: an anonymous translation of a sutra called Mile danglai sheng jing 5i1fg]'&*1:.*~, an anonymous translation called Mile zuofo shishi jing 5i1fg]fF{~IH!~$*~, and Paramartha's translation of a sutra entitled Mile xiasheng jing 5i1fg]T1:.*~. To this, we should add that Yueposhouna is also said to have produced the fifth trans106 Cf. T55, no. 2154, 567bl-2 for the text and 568b4-569b4 for the passage on Yijing. 107 Cf. T55, no. 2145, 32c8 and 21b17-21c9.


Elsa I. Legittimo

lation, and that this was believed to have been lost, although it is still extant within the Ratnakuta. In sum, the Mile xiasheng jing 5plllfyff::1;*~ (no. 453), the sutra wrongly attributed to Dharmarak~a, is in fact only mentioned in the catalogues as a text extracted from the Ekottarika-iigama, and the Maitreya text translated by Dharmarak~a was lost already at the time of the earliest catalogues, and belonged to the other Maitreya sutra that was translated six times.

Abegg, Emil. 1928. Der Messiasglaube in Indien und Iran. Berlin / Leipzig: De Gruyter. Akanuma, Chizen $rt?9&*,. 1929. The Comparative Catalogue of Chinese Agamas & Pilli Nikilyas. Nagoya: Hajinkaku-ShobO 1ilt~OO

Analayo (T.P. Steffens). forthcoming. A Comparative Study oftheMaJJhimakilya. Habilitationsschrift des Fachbereiches Fremdsprachliche Philologien der Philipps-Universitiit Marburg 2006. Baruch, Willy. 1946. "Maitreya d'apres les sources de Serinde." Revue de l'Histoire des Religions 1946: 67-92. Buswell, Robert E. 2004. "Sugi's Collation Notes to the Koryo Buddhist Canon and Their Significance for Buddhist Textual Criticism." The Journal of Korean Studies 9/4: 129-184. Deeg, Max. 1999. "Das Ende des Dharma und die Ankunft des Maitreya. Endzeit- und Neue-Zeit-Vorstellungen im Buddhismus mit einem Exkurs zur Kasyapa-Legende." Ze itsch rift fur Religionswissenschaft 7: 145-169. --- 2005. Das Gaoseng-Faxian-Zhuan als religionsgeschichtliche Quelle. Wiesbaden: Harrassowitz. Demieville, Paul. 1920. "Compte-rendu de Leumann 1919." Bulletin de I'Ecole franfaise d'Extreme-Orient XX: 158-170. --- 1954. "Maitreya l'inspirateur." Bulletin de ['Ecole franfaise d'Extreme-Orient XLIV: 376-387. --- 1954. "Le paradis de Maitreya." Bulletin de ['Ecole franfaise d'Extreme-Orient XLIV: 387-395. Emmerick, Ronald E. 1968. The Book ofZambasta, London: Oxford University Press.

Reopening the Maitreya-files


Enomoto, Fumio ;tj)js:Jt~. 1984. "Setsuissaiubu-kei Agama tenkai _ Chiiagon to Z6agon wo megutte" m-tw:ff:a:~*7~jJ-y~~ W9='~iiJ~J1.!:: W*~~J1 :a::1iJ(,',,-::>'""C [The development of the Agamas in the Sarvastivada lineage - An approach to the Madhyama-ilgama and the SaYf'lyukta-ilgama]. Journal of Indian and Buddhist Studies XXXII/2: 51-54. --. 1984. '~gonky6ten no Seiritsu" ~~*-IAO)px:lz: [The formation of the Agama scriptures]. Toyogakujutsukenkyu .n~*1l3fJi:; [Journal of Oriental Studies] 23/1: 93-108. --. 1986. "On the Formation of the Original Texts of the Chinese Agamas." Buddhist Studies Review 3/1: 19-30. Filliozat, Jacqueline. 1993. "The commentaries to the AnagatavaIp.sa in the Pali manuscripts of the Paris collections." Journal of the Pali Text Society XIX: 43-63. Filliozat, Jean. 1950. "Maitreya, l'invaincu." Journal Asiatique 238: 145149. Hardacre Helen; Sponberg, Alan (eds.). 1988. Maitreya, the Future Buddha (Proceedings of the Princeton Conference on Maitreya Studies, May 1983). Cambridge / New York: Cambridge University Press. Jaini, Padmanabh S. "Stages irrthe Boddhisattva Career of the Tathiigata Maitreya." Hardacre, Sponberg 1988: 54-90. Hartmann, Jens-Uwe. 2006. "Maitreyavyiikara:q.a". Jens Braarvig, Paul Harrison, Jens-Uwe Hartmann, Kazunobu Matsuda, Lore Sandere (eds.). Buddhist Manuscripts, vol. III. Manuscripts in the Sch0yen Collection. Oslo: Hermes Publishing: 7-9. Hiniiber, Oskar von. 2000. A Handbook of PilIi Literature. Berlin / New York: De Gruyter. Hureau, Sylvie. 2006. "Preaching and translating on po~adha days: KumarajIva's role in adapting an Indian ceremony to China." Journal of the International College for Postgraduate Buddhist Studies X: 87-118. Lamotte, Etienne. 1967. "Un Sutra Composite de l'Ekottarika-ilgama." Bulletin of the School of Oriental and African Studies 30: 105-116. Legittimo, Elsa I. 2005. "Synoptic Presentation of the Pusa chu taijingPart I (chapters 1-14)." Sengokuyama Journal of Buddhist Studies 1llJ ;;PrlIffijQ~ II: 1-11l. --- 2006a. "Synoptic Presentation of the Pusa chu tai jing - Part II (chapters 15-38)," Sengokuyama Journal of Buddhist Studies 1llJ;;PrlI ffijQ~m: 1-175. - - 2006b. Analysis of the Pu sa chu tai jing fi\l~Bil'~ (T12, no. 384). Tokyo 2006, [unpublished doctoral thesis, International College for


Elsa I. Legittimo

Postgraduate Studies (ICABS); a publication including a full translation is planned in the series Melanges Chinois et Bouddhiques of the Institut BeIge des Hautes Etudes Chinoises, Bruxelles 2010]. Leumann, Ernst. 1919. Maitreya-samiti, das Zukunftsideal der Buddhisten - Die nordarische Schilderung in Text und Ubersetzung. Strassburg. Levi, Sylvain. 1925. "Le sfitra du sage et du fou dans la litterature de l'Asie Centrale." Journal Asiatique207: 305-332. --- 1932. "Maitreya Ie consolateur." Leroux, Ernest (ed.). Etudes d'Orientalisme publiees par le musee Guimet it la memoire de Raymonde Linossier, vol. II. Paris: 355-402. Majumdar, Ramesh Chandra. 1959. "Arya-Maitreya-vyakaraJ;la." Dutt, N. (ed.). Gilgit Manuscripts, vol. IV. Calcutta: 185-214. Matsumoto, Bunzaburo t~*Jt=Jj!~. 1911. Miroku-jodo-ron fi:lffl1$~ilii [A Monograph on Maitreya's Pure Land]. Tokyo: Heibonsha .IJLfLti. Mayeda, Egaku. 1985. "Japanese Studies on the Schools of the Chinese Agamas." Heinz Bechert (ed.). Zur Schulzugehorigkeit von Werken der Hfnayana-Literatur, Erster Teil. Gottingen: 94-103. Meddagama, Udaya; Holt, John Clifford. 1993. AnagatavaT[lsa Desanti The Sermon of the Chronicle-To-Be. Delhi: Motilal Banarsidass. Minayeff, Ivan P. 1886. "Anagata-varpsa." Journal of the Pali Text Society II: 33-53. Mizuno, Kogen 7.klllf515I';. 1956. "Kanyaku Chuagon to Z6ichiagon to no yakushutsu ni tsuite" 1~~r:pIlPJi3U~-~iiJi3~0)~IH~./)1t\'-C [On the Chinese Translations of the Madhyama-agama and the Ekottarikaagama]. Proceedings of The Okurayama Oriental Research Institute, vol. II. Yokohama: 41-90. --- 1989. "Kanyaku no ChUagongyo to Zoichiagongyo" 1~~0)r:p1lPJi3*I U~-~iiJi3*') [The Chinese Versions of the Madhyama-agama and the Ekottarika-tigama]. Bukkyo Kenkyu / Kokusai Bukky6to Kyokai 18 {b~:jjJfJ'E / OO~{b~1tm~: 1-42. Nattier, Jan, 1988. "The Meanings of the Maitreya Myth: A Typological Analysis." Hardacre, Sponberg 1988:23-47. Peri, Noel. 1911. [Untitled book review of Matsumoto's Miroku-jodo-ron" (under heading: Bibliographie - Japon)]. Bulletin de l'Ecole franfaise d'Extreme-Orient XI: 439-458. Warita, Takeo i!fIjIIlIlUlIJ~. 1973. "Dommanandai-yaku Zoichiagongyo no genzaikyo ni tsuite" fi*~fHJl!:~ r~-IlPJi3*IJ O)m1:E*If~./) 1t \'"( [On the Extant Sutra of the Ekottarika-agama Translated by Dharmanandin]. Journal of Indian and Buddhist Studies XXI, 2: 148-149.

Reopening the Maitreya-fil~s


Zacchetti, Stefano. 2007. "The Earliest Chinese Buddhist Translations: A Preliminary Attempt to Define the Corpus of An Shigao's Translations and Review the State of Art in the Relevant Research." Language Identity and Language Change in Collision and Dialog Between Civilizations (II). Forum: The Harmony of Civilizations and Prosperity for All - Diversity in the Development of Human Civilization. Beijing: 1-17 [unpublished conference paper} Zin, Monika. 2003. Ajanta - Handbuch der Malereien / Handbook of the Paintings 2: Devotionale und ornamentale Malereien, Vols.l-2, Wiesbaden: Harrassowitz.

Who produced the Da mingdu jing *~~~ (T225)?

A reassessment of the evidence*

Jan Nattier

The Da mingdu jing has long been considered one of the most solidly attributed texts in the. corpus of Zhi Qian :~ (fl. 222-252 CE). Credited to Zhi Qian already in the earliest extant catalogue of Buddhist scriptures, the Chu sanzang ji ji I\.:::.it[.* compiled by the eminent scholar-monk Sengyou {~:ffi (completed c. 518 CE),t the Da mingdu jing has accordingly appeared on virtually every roster of Zhi Qian's works published in scholarly studies. Based on external evidence alone - that is, on the title of the text and its treatment in early scriptural catalogues - there would seem to be no reason to doubt its authenticity. Several decades ago, however, an American scholar of Buddhism, Lewis R. Lancaster, published an article in which he argued that the Da mingdu jing is not actually Zhi Qian's work, but

* This paper was originally presented at a conference on "Early Chinese Buddhist Translations" held in Vienna on April 18-21, 2007. I would like to thank the conference organizer, Max Deeg, for his kind invitation to participate and our host, Helmut Krasser of the Austrian Academy of Sciences, for the sponsorship of this event. This paper has benefited from comments and suggestions made by a number of the conference participants, in particular Christoph Harbsmeier, Paul Harrison, and Stefano Zacchetti. Any errors that remain, of course, are my own.
See T2145, 55.7a8. In the main entry the title is given as Mingdujing Sengyou also gives the alternate title Da ming duwuji jing *~)jt Sengyou describes the text as consisting of four fascicles (IITI:'), though later catalogues give the alternatives of four or six.

~)jt~; ~fEE~J[.

Journal of the International Association of Buddhist Studies

Volume 31 Number 1-2 2008 (2010) pp. 295-337


Jan Nattier

was instead produced by the Han-period translator An Xuan '1l:1i: (Lancaster 1969). So far as I have been able to determine, this was the first article published in English devoted to any of the works of Zhi Qian (in Wade-Giles transcription, Chih Ch'ien), and this . pioneering study has since been widely quoted. 2 According to Lancaster, the "avoidance of transliteration" (that is, the preference for translation) that characterizes the Da mingdu jing is not at all typical of Zhi Qian's other work (p. 248), but on the contrary, serves as evidence that this text is the work of someone else (p. 249). More specifically, Lancaster proposed that chapters 2-27 of the Da mingdu jing should be considered the work of An Xuan, while chapter 1 (which contains an interlinear commentary) is a revision of that text by another hand. 3 Writing some fifteen years later, a Japanese scholar, KATSUZAKI Yugen ElfjlfIjJ?<frg, took precisely the opposite position. 4 In an article published in 1985, Katsuzaki described the Da mingdu jing as "the most Zhi Qian-ish" (t-:J c!::: t3Z:~~"J) of Zhi Qian's translations (pp. 68, 91). For Katsuzaki - and indeed, for Japanese scholars in general - the preference for translation (Katsuzaki uses the term giyaku ~~) rather than transcription (onsha 1if~) is taken as one of the very hallmarks of Zhi Qian's translation style.

Methodological issues (1): Lancaster's approach

The fact that these two scholars could produce quite opposite arguments is a product, to a large extent, of their very different methodologies. To begin with the earlier of the two, in Lancaster's pa-

2 Lancaster's argument is apparently accepted by ZUrcher in his "A New Look at the Earliest Chinese Buddhist Texts" (Zurcher 1991, p. 294, notes 4 and 9). The paper is also cited (though not necessarily always with agreement) in numerous other studies. 3 For reasons that are not stated directly, Lancaster does not include chapters 28-30 (containing the story of Sadaprarudita and Dharmodgata) in his discussion of the vocabulary of the Da mingdu jingo 4 See Katsuzaki 1985. Katsuzaki does not refer to Lancaster's article, and presumably it was not available to him.

Who produced the Da mingdu jing *~~}jt~ (T225)?


per all fifty-three texts credited to Zhi Qian in the TaishO canon are accepted as genuine, and Lancaster refers to this entire body of material as a valid source of evidence concerning Zhi Qian's translation style (p. 247). A careful evaluation of the testimony provided by scriptural catalogues, however, shows that over half of these attributions were unknown to Sengyou and to his illustrious predecessor, Dao'an m~, whose earlier catalogue (no longer extant as a separate work) was incorporated into the Chu sanzang ji ji. In many cases these "newly discovered" Zhi Qian translations were introduced for the first time in the Lidai sanbao ji JfIHi; - .~c compiled by Fei Changfang .~E (T2034, completed in 597 CE). As Fei is known to have introduced many false attributions into his catalogue, the degree of confidence that can be placed in such items is extremely low. s Thus Lancaster's working list ofZhi Qian's translations was corrupted by the inclusion of dozens of texts that are certainly not his.6 Conversely, though Lancaster refers to having consulted all fifty-three of Zhi Qian's supposed works, virtually all of the terms which he says appear in the Da mingdu jing but not in any of Zhi Qian translations (p. 249) can in fact be found here and there in other texts, even when the list is narrowed to the two dozen or so that can be considered to be his genuine works? In the following chart the column on the left contains those terms which, according to Lancaster, appear in chapters 2 through 27 of the Da mingdu jing but not elsewhere in Zhi Qian's work. The column on the right
S For an extensive critical discussion of Fei Changfang's catalogue see Tan 1991; pp. 3-246; a pioneering discussion in Japanese can be found in Hayashiya 1941, pp. 82-84 and 300-302. For brief overviews in English see Tokuno 1990, pp. 43-47 and Nattier 2008, pp. 14-15 and nn. 25 and . 26.

6 It should be pointed out, however, that Lancaster has recently been an active participant in a project to update existing scriptural catalogues to reflect more reliable attributions, and that largely as a result of his efforts far better resources should be available online in the near future. 7 For a discussion of the translations that can reliably be attributed to Zhi Qian at the present state of our knowledge see Nattier 2008, pp. 121-148.


Jan Nattier

gives the TaishO text numbers of translations reliablY,attributed to Zhi Qian in which these terms also occur. Where the Chinese term in question is being used to translate a different Indian word than in the latter part of the Da mingdu jing (i.e., chapters 2-30 of the text; henceforth T225B), the text number is given in brackets together with the alternate lndic-language equivalent. Terms from ch. 2-27 of T225B
for bhik~u JE (for samadhi) 51H~ (for upasaka) ~ (var. Iffl) (for bodhisattva)

Other Zhi Qian texts where attested

T210 B T6, 76, 87, etc. 9 T6, 76, 198, 361, 533, 581, 790 T76, 493, 533

B See T21O, 4.567a24. The fact that this term is indeed being used as a translation of bhik~u can be confirmed by consulting the parallels to this verse in the Udanavarga (XVII.8) and the Maitreyavadana (Vaidya 1959, p. 34, lines 14-15). Previous publications have treated this verse as being without any close parallel; see Mizuno (1981, p. 284), who suggests Udv XI.l5 as a partial parallel, and Dhammajoti (1995, p. 184, n. 14) who follows Mizuno in offering the same suggestion. 9 The character JE appears in virtually every text translated by Zhi Qian; in the texts listed above, it is certain that it is being used to represent samadhi. There may be many other cases as well, but a thorough examination of every occurrence of the character JE lies beyond the range of this paper. For a representative passage in which the equivalent is certainly samtidhi see T6, 1.181c27. 10 Lancaster treats this as part of a single long expression, viz., ~P!T IEJIm::I&IEt';, which he considers to be a translation of anuttarasamyaksarrtbodhi (p. 249). In fact, however, the context - where the word 3z0 31<: "tathagata" appears just before (8.482b21) - makes it clear that this group of characters consists of two distinct epithets of the Buddha: ~ P!T, used as a translation of arhat, and lEJIm:m'tlEt'; corresponding to samyaksarrtbuddha (consisting of lEJIm:, used by several early translators for samyaksarrtbodhi, and :l&lEt';, presumably originally derived from *abhisarrtbuddha). On these epithets see Nattier 2003, pp. 217-219 and 222-223.

Who produced the Da mingdu jing *~J:j1l~~ (T225)?


* (for mahiisattva)
~PJT~ (fo~ arhat)10

IE1{mmIEI: (for

(for srotaiipanna)

Tl69;474, 532 [T76, for mahiipuru~a] Tl69,532 [T361 and 474, for samyaksaT{lbodhi; cf. also T281 and TlOll]l1 T6,87,198,474,790

Lancaster's paper was written, of course, before the advent of the digital versions of the Chinese Buddhist canon (of which he himself was an early supporter), which now make it possible to search through large quantities of material at lightning speed. Moreover, the growing scholarly consensus on Zhi Qian's authorship of two of the texts given above - a non-Mahayana MahiiparinirviiIJa-sutra (T6) and a version of the Larger Sukhiivatzvyuha (T361) - was not yet in effect in the late 1960s. 12 Even if we take these factors into consideration, however, it is clear that Lancaster's statements concerning which terms are, or are not, found in Zhi Qian's work are contradicted by what we actually find in a number of genuine Zhi Qian texts. In evaluating the claim that the Da mingdu jing is the work of An Xuan (or more properly, the work of An Xuan and Yan Fotiao, since the two men produced the Fa jing jing 5ti:~~ together), it is important to note that no early catalogue attributes a prajfiiipiiramitii translation of any kind to An Xuan. Of course the silence of the catalogues does not, in itself, prove that such a translation did not exist. But in the absence of such bibliographic support it is all the more important to compare the vocabulary of the Fa jing jing (T322) thoroughly and systematically with that of the Da mingdu jingo Lancaster's article, however, does not include such a systemIn all of these cases this expression is prefaced by the characters and this combination is clearly intended as a translation of the term anuttarasamyaksaT{lbodhi. For slightly different phrasing cf. T281, 1O.450c22-23 (tJtli!:1~~?1{~~..tIEJrZm~mIEI:) and TlOll,. 19.680b22-23 (~~~..tIEJrZm~mH). 12 For the attribution ofT6 to Zhi Qian see Nattier 2008, pp. 126-128 (with reference to an earlier analysis of the text in Ui 1971, pp. 517-523). On the authorship of T361 see Nattier 2008, p. 139 and the further references given there.



Jan Nattier

atic comparison, but focuses only on terms that are shared by the Da mingdu jing and the Fa jing jingo When we consider the vocabulary of the Fa jing jing as a whole, however, we find many terms and expressions used by An Xuan and Yan Fotiao that do not match those used in chapters 2-30 of the Da mingdu jing (T225B). (In cases where an expression appears only rarely in T225B, and where there is reason to think that its appearance is the result of editorial emendation rather than part of the original translation, the item is given in brackets with an explanatory note.) A few representative examples are the following:
Sanskrit tenn
pratyekabuddha sarvajiia bhagavat kulaputra

Equivalent in T322

Equivalent(s) in T225B




3(CP3(, [3(#, tit


;g;, ~

punar apararrz . tat kasya hetoJ:t




{OJ tzii)(, [pff tz~{OJ]15

13 The use of the transcriptionfa {~ as a translation of bhagavat is discussed in Nattier 2006. ,,' 14 The expression "Heaven-honored One" (tianzun 3(#) occurs nine times in T225B, but these occurrences are clustered together on pp. 490b-492a (with all but two occurring on 490b-491b), a distribution that almost certainly points to the editorial emendation of this section (on this epithet see Nattier 2003, pp. 232-234). When the word bhagavat is not simply translated as fa {~"Buddha," the translator ofT225B overwhelmingly prefers the expression "god of gods" (tian zhang tian 3(CP3(), which occurs ninety-three times. On the latter expression, which is especially frequent in the vocative use, see Iwamatsu 1985, Boucher 1996, pp. 210214, and Nattier 2003, p. 234. The term "World-honored One" (shizun "tlt#) occurs only once in T225B (at 488b7), and is surely the result of a copyist's alteration. 15 There are only five occurrences of the expression suayizhe he f5jfL-)

Who produced the Da mingdu jing *~Jjt~~ (T225)?


Even this brief selection makes it quickly evident that several key technical terms (as well as certain ordinary expressions) used in the Fa jing jing do not match those found in the Da mingdu jingo Even more important than the differences in these individual lexical items, however, are fundamental differences in what we might call "translation policy." First of all, one of the noticeable features of the Fa jing jing - which is, unfortunately, the sole work known to have been produced by the team of An Xuan and Yan Fotiao - is the relative consistency of its terminology; for example, the text contains only one term for bhik~u (chujin ~i), one term for nirviiIJa (miedu ~~), and one name for Maitreya (Cishi ~~). The Da mingdu jing, by contrast, exhibits considerable variety in its terminology, using both chujin ~i and biqiu ~tli for bhik~u, miedu )~~ and niehuan YJt;lli for nirviiIJa, and Cishirm:~ and Mile 5j1(ij] for Maitreya. Even in the case of the word bhagavat, which is rendered into Chinese in a variety of ways (including simply fo {~ "Buddha") and which has multiple equivalents in the latter part of the Da mingdu jing as shown in the chart above, the translators of the Fa jing jing appear to have made an effort to be consistent, using the term zhongyou ~:ft "Mass of Blessings" wherever the corresponding Tibetan text indicates that the underlying term was a form of bhagavat "Blessed One." Most significant is the fact that throughout the Fa jing jing the translators consistently attempted to translate the meaning (rather than transcribing the sound) of all names and Buddhist technical terms, with the exception only of a few ancient words that had already come into widespread use before their time, i.e., the wordfo {~ (ONWC/EMC but) "Buddha," the personal name Anan IftiJB '~nanda," and the deva-names Shi ~ "Sakra" and Fan ~

fffO] in T225B, and these are clustered together in just one section of the
text (8.482c-483b), thus again presumably revealing an editorial emendation of that passage. The expression he yi gu fOJ-tU&, by contrast, appears no fewer than ninety times, and is the standard formula in this text. Conversely, in T322 the usual form is J5JTt.l.fffOJ (thirteen occurrences). fOJ t.l.i&: appears once in T322, but only as part of a more complex question (12.l9c13, fOJj))i&:[ltz1.l,~J1:I!? "Why are they called sacred texts?").


,Jan Nattier

"Brahma."16 Indeed, the Fajing jing is such an extreIl!e example of a "translation-only" policy that it is fair to say that its translators made every effort to avoid the use of transcriptions. The translator of the Da mingdu jing, however, does not follow suit. In T225B (as is also the case in T225A) we find a significant number of transcriptions, coexisting quite comfortably with a wide range of translated terms. 17 Thus the Da mingdu jing does not conform (pace Lancaster) to the single most distinctive feature of the Fajing jing: an almost total absence of transcription terms. In sum, there is no solid evidence that the Da mingdu jing is the work of An Xuan and Yan Fotiao, while on the contrary several factors, both external and internal, militate against it: (1) The absence of any reference in early scriptural catalogues to the production of a prajfiaparamita text of any kind (let alone of the Da mingdu jing in particular) by An Xuan (and/or Yan Fotiao); (2) Thefactthat chapters 2-30 of the Da mingdu jing (i.e., T225B) exhibit considerable variety in translation terminology,18 while the Fa jing jing generally employs a single Chinese equivalent for Buddhist names and terms; and (3) The fact that the Da mingdujing (again, referring specifically to T225B) contains a significant number of transcriptions, while such terms are strenuously avoided (with the exception of a small handful oflong-accepted names) in the Fajing jingo

16 The word seng fIt a transcription of salflgha, occurs only once in the text (12.l6b6), and it seems likely to be the result of a scribal emendation. "assembly." Elsewhere the word is consistently translated as zhOng 17 There are far too many transcriptions in T225B to list them all here; for convenient access to this data see the extensive table compiled by Katsuzaki (1985,pp. 69-89). For his discussion of the issue of transcription itself see pp. 89-90. 18 In documenting this variety it will be important to exclude those variations that seem to have resulted from subsequent scribal emendation (cf. above, notes 14-16).


Who produced the Da mingdu jing *E)jll~ (T225)?


Even without considering the question of what - if anything - might be described as Zhi Qian's "usual translation style" (a thorny issue to which we will return below), it is clear that there is no reason to assign the Da mingdujing to An Xuan and Yan Fotiao. Methodological issues (2): Katsuzaki's approach KATSUZAKI Yugen employs a quite different method in analyzing the Da mingdu jing, taking as his point of departure descriptions of the work of Zhi Qian by scholars going back to the time of Sengyou. Drawing on the work of KAGAWA Takao II:tt (1984) and others, Katsuzaki proceeds on the basis of the assumption that the use of translations rather than transcriptions is a distinguishing feature of Zhi Qian's work. Accordingly, he finds the Da mingdu jing - which does indeed abound in translated terms - to be the most typical of Zhi Qian's style. On this basis, he recommends that the Da mingdu jing be used as a fundamental point of reference in studies of Zhi Qian's work. Katsuzaki's description of Zhi Qian's "typical style" is well grounded in traditional bibliographic sources, and as such it offers an accurate reflection of the views of Chinese Buddhist scholars from at least the sixth century CE onward. Conversely, however, the fact that his discussion reflects these traditional views means that it also shares in their shortcomings. In particular, neither the sources he quotes nor Katzusaki himself take into account the troublesome fact that Zhi Qian's corpus exhibits a high degree of inconsistency. Thus while some translations solidly attributed to him - e.g., the Fanmoyu jing ~=5WJ~ (T76) and the Weimojie suoshuo jing ~fE=apfT~~ (T474) - do abound in translated terms and are composed in an elegant literary style, others - e.g., the Yueming pusa jing A ~ifMi~ (T169) and the Huiyin sanmei jing ~fP ~J[ (T632) - are less polished in style and contain a large number of transcribed terms. Thus if there is anything that could be described as a "characteristic feature" of Zhi Qian's corpus, it is the very fact that no set of characteristics appears consistently throughout his work. On the contrary, texts that are solidly attributed to Zhi Qian from the time of Dao'an onwards exhibit a wide range of variation in both terminology and style.




Jan Nattier

To be fair, however, it should be pointed out that it was not Katsuzaki's objective in this article to establish the attribution of the Da mingdu jing to Zhi Qian; instead, his purpose was to provide a comparison of the terminology used in the Da mingdu jing with the corresponding terms found in Lokak~ema's JZ:~~~ Daoxing banruo jing mf'J~~;fi~ (T224). And in so doing he has made an important contribution to the study of early Chinese translations, for his article contains a valuable table of Buddhist terms and proper names found in each chapter of the Da mingdu jing, with the corresponding terms in Lokak~ema's Daoxing banruo jing given for comparison (1985, 69-89). A close look at this list, however, reveals another issue that requires our attention. Katsuzaki (again following traditional East Asian scholarly practice) treats the whole of the Da mingdu jing as a single text, drawing his examples from all of its chapters, from 1 through 30. But the sutra as we have it consists of two quite dissimilar parts. Chapter 1 (henceforth T225A) differs in numerous respects from chapters 2-30 (T225B), including, but not limited to, vocabulary and style. Thus, as Lancaster rightly pointed out (p. 247), the first chapter is not of the same vintage as the others, and it should properly be dealt with separately.

A hybrid creation: Components of the Da mingdu jing

Even a cursory glance at the text of the Da mingdu jing contained in the present TaishO edition of the canon quickly reveals that the first part of this work, the "Practice" chapter (f'J), is very different from the rest. Not only does it contain a translation of the first chapter of the sutra itself, but it also includes an extensive interlinear commentary which was clearly produced not in India but in China. The commentary explains various words and concepts found in the sutra translation, quoting from a number of scriptures previously translated into Chinese. 19 In some cases the explana19 The texts cited in the commentary to T225A are referred to there by the titles Anban ~Bt (cited at 8.478c7; cf. T602, the Da anban shouyi jing *~Bt~:~Jil., but also the newly discovered manuscript discussed by Stefano Zacchetti in this issue), Liaoben y;zr;: (480a26; cf. T708, the

Who produced the Da mingdu jing *~~~ (T225)?


tions are prefaced by the words "The master says ..." (gi!i7i), but neither the name of the master nor that of the author of the commentary is provided. Many of the texts cited here, however, are also cited in an early commentary on An Shigao's Yin chi ru jing ~~:tA~ (T1694), and the latter uses much of the same language, including the expression gf!i7i. In a recent study Stefano Zacchetti has presented compelling evidence that the Yin chi ru jing commentary was the product of a community of Buddhists in the Wu ~ kingdom that included the Chinese layman Chen Hui [~JI as well as the Sino-Sogdian monk Kang Senghui gjt{~ff (Zacchetti forthcoming). Given the striking similarities between this text and the interlinear commentary to T225A, I believe it is highly likely that the latter was composed in the same milieu. It is not only the presence of the commentary that distinguishes chapter 1 from the remainder of the text, however, for its vocabulary is different as well. There are a number of cases where the latter part of the text (T225B) uses vocabulary that does not occur in chapter 1 (T225A). The following are some representative examples: 20#
bodhisattva kulaputra

Chapter 1 (T225A)

Chapters 2-30 (T225B)


Liaoben shengsi jing 7*~t1E~~, but the material cited here does not have a parallel in that text), Faju $;1:i] (480b2; see T21O, the Fajujing 5! 1:i]m.), Dun zhenjing ~J1r~ (480b3; see T624, the Dun zhendouluo suowen rulai sanmeijing f~Jt~B~p.ITrJ:I53m*.=.lI?[c~), andHuiyinjing ~t:P~ (480b4; see T632, the Huiyin sanmei jing ~t:P.=.lI?[c~!). It is noteworthy that all of the above texts - though not always the same passages - are also cited in T1694, a commentary on An Shigao's Yinchiru jing ~mA ~~ (on which see Zacchetti forthcoming). 20 Though the names and technical terms discussed here are given in Sanskrit for ease of recognition and convenience of reference, it is assumed throughout this discussion that Zhi Qian's translations (and in fact most if not all of the Chinese translations produced from the second to fourth centuries CE) were based not on Sanskrit but on Prakrit originals.


Jan Nattier


#, t!t#F1




As is immediately evident from this chart, T22SA uses terms for bodhisattva (pusa ~~), kulaputra (zuxingzi ~~r), and 'bhik:ju (biqiu tcli) that are widely attested in other early translations, while T22SB uses a number of highly idiosyncratic translation terms, including "opener, revealer" (kaishi 1Wl) for bodhisattva, "good sir" (shanshi ~) or "exalted sir" (gaoshi !%) for kulaputra, and "famine discarder" (chujin ~~~i) for bhiko?u, some of which had already appeared in the work of An Xuan and Yan Fotiao. In certain cases a word used in T22SA as the sole equivalent for a given Indic term is also found in T22SB, but there it alternates there with other forms (for instance, in T22SB both ~~~i and tcli are used for bhiko?u, while T22SA uses tcli alone). But it is not only in Buddhist terms and proper names that we can find differences between the two parts of the Da mingdu jing. If we examine the pronouns used in each part of the text, for example (following the lead of Matsue 200S), we find that T22SA freely uses the word wu -R "I, me, my" for the first person pronoun (19 times in the non-commentarial part of chapter 1), while the word wo f1<; is somewhat less common, occuring only 14 times in the same chapter. In the much-longer T22SB, by contrast, the proportions are reversed, with -R occuring only 20 times versus 219 occurrences of f1<;. The second-person pronoun qing g~p "you" occurs five times in T22SB, but never in T22SA, while conversely the pronoun ru 5fr. occurs four times in T22SA, but only three times in the whole of T22SB.22 The demonstrative pronoun ci rtt "this" is used 44 times
21 The expressions 5'C# and t!t# rarely appear in T225B and are probably the result of scribal emendation; see above, note 14. 22 T225B generally uses the word ruo E to express the second-person pronoun. In T225A, by contrast, all 38 occurrences of E appear to be in

Who produced the Da mingdu jing *l)IjJjt~ (T225)?


in the non-commentarial portions of T225A (and another 41 times in the commentary), but only 28 times in the whole of T225B. In other words, the two parts of the text differ not only in their treatment of Buddhist names and terms, but in their choice of pronouns as well. Other differences in ordinary (i.e., non-Buddhist) terminology can also be found. The oft-used question "Why is that?" (in Sanskrit, tat kasya hetoJ:t) generally appears in the form suoyizhe he pJTtJ:t:ff6j in T225A, while in T225B he yi gu 1iiJ~t)(: is almost always used. In introducing ordinary quoted speech - for example, statements made by Subhuti to Sariputra or vice versa - T225A routinely uses the verb yue B, while in T225B the verb yan is overwhelmingly preferred. To introduce a reply, T225A usually uses the verb da ~, while in T225B the standard form (with only a few exceptions)23 is dui ft. When it is the Buddha who is speaking T225B often uses yu R, but this verb never appears in T225A at alJ.24 These pervasive differences make it quite clear that T225A and T225B were not produced by the same person, but were originally separate texts that were "pasted together" at some point, with the first chapter of the original T225B presumably being lost in the process. If additional chapters of T225A were ever completed - and this is not at all certain - they were presumably lost at the same time. When did this amalgamation take place? That is, when did the Da mingdu jing as we have it - consisting of chapter 1 of T225A together with chapters 2-30 of T225B - come into being? It is not possible to answer this question with precision, but it is clear that it had already occurred by the time a version of the text came into the hands of the great lexicographer Xuanying 1r~ (fl. 645-656). If we examine his discussion of the vocabulary of a text called (Da)
the sense of "if," a usage which of course also occurs in T225B. 23 More than half of the occurrences of da ~ in T225B are clustered together in one brief section of the text (8.482b-483a), suggesting that the word has been introduced in the course of revision of this passage. 24 In T225A yu ~ occurs only as a noun.


Jan Nattier

ming duwuji jing (*)~Ii~f~~J[ in the Yiqie jing yi'[lyi -W~1if ~, we find that it begins with material now found in T225A, then proceeds to discuss material from T225B, in the same sequence in which these terms are found in the TaishO edition of T225 today.25 Thus by the middle of the seventh century CE at the latest - and in all probability, long before that time - the hybrid Da mingdu jing was circulating as an integral text in China. Based on the data presented above, we can already see that the way in which the question of the attribution of T225 .is usually phrased - "Does the Da mingdu jing conform to the usual translation style of Zhi Qian?" - is flawed in at least two respects. First, there is no such thing as "the" Da mingdu jing; instead, the text consists of two parts (T225A and 225B) which were clearly produced at different times and under different circumstances. And second, as we shall see, there is tremendous variety in both vocabulary and style within Zhi Qian's corpus, and thus there is no such thing as "the" usual translation style of Zhi Qian. Like a mathematical equation containing too many variables, the problem of the authorship of T225 cannot be solved when it is stated in the above terms. As an alternative, therefore, I would like to propose that we refine our approach by dealing with three aspects of the problem separately: first, an evaluation of the relationship of T225A and T225B (considered individually) to Lokak~ema's T224; second, a comparison of the vocabulary and style of T225A and B to one another; and third, a comparison of the vocabulary and style ofT225A and B (again considered separately) to other texts in the corpus of translations reliably attributed to Zhi Qian.

25 See T2128, 54.364a24-c13. The text is called Da ming duwuji jing

in the table of contents (T2128, 54.362c10) but only Ming duwuji jing Ej.EjJJt~t~~-![ in the headings to the individual entries themselves (e.g., Ej.EjJJt~t~~%-offi at 364a24). Note that the version of the text used by Xuanying was in four juan offi, a figure which corresponds to that given by Sengyou for Zhi Qian's translation of the Da mingdu jingo The text also circulated in a version divided into six juan (Ej.EjJJt~7\offi[ :g*Ej.EjJJt~t~~W:l21loffi]); see Fajing's Zhongjing mulu, T2146, 55.l19b6), as is the case with the current TaishO edition of the text.

.Who produced the Va mingdujing *~J3Jjt~ (T225)?


The Da mingdu jing *1JI31l~ (T225) and the Daoxing banruo jing iif.TM~ (T224) Thus far we have considered only the two distinct segments of what today is known as the Da mingdu jing, i.e., chapter 1 (T225A) and chapters 2-30 (T225B), respectively. Our understanding of the history of both parts of the text can be greatly enhanced, however, by comparing them with the corresponding sections of the earliest extant Chinese translation of the text, Lokak~ema's Daoxing banruo jing (T224). It is commonly said that the Da mingdu jing is a revised version of Lokak~ema's translation, but as we shall see, this statement is true only if it is worded with great precision. A close comparison of these texts yields the following results. (1) T225B is noticeably abbreviated with respect to T224. Even a cursory glance at these two texts shows that there is a noteworthy difference in their length. T224 - or rather, chapters 2-30 of that text, the portion which parallels T225B - occupies just over fortynine pages in the TaishO edition (8.429alO-478bI4), while T225B is only about half as long, totalling just twenty-six pages (482b6508bI3). While a certain amount of this difference in length can be explained by the replacement of long transcriptions of Indian terms with much shorter translation terms, it is due above all to the compression of Lokak~ema's often verbose and repetitive mode of expression. Compare, for example, the following renditions of a passage from Chapter 16 (corresponding to Chapter 18 of the Sanskrit text):


~J[:g:m: "{~)t '::fPJ~t' 0 @. , ~jf , ,~~ , ~1E ' ~1J\::fPJ {~~~J[:g:m: ")3t:pJTrp~~ , 1ffcJJZSI15e:@. , ~jf , ,~~ , ~1E' ~::f

PJ~t '::for.,;. 0


T225B: ~~: "{~)t '::fPJ~t' liJ~1J\~?" {~: ";fipJTrp~~' 1f fflJZSl15e:liJ~orH";' 0" (8.496a23-2sr-

It is easy to see that in T224 the standard list of "form, feeling, conceptualizing' conditioning forces, and consciousness" is given in full as 15' ~$ , ,~~ , ~% ,~, while in T225B it is abbreviated to simply "the five skandhas" (E~~).


Jan Nattier

But it is not only lists that are abbreviated, for T22~B compresses other rhetorical elements as well. See for example the following passage from Chapter 2: .
T224: ~JHffJH1H~UytfB[ZSJ1'f: "foJjf! ;l'&;fi.:rTf-it~-T~~~~&,

)t 0" ~J[::g:mft{~~t$t~{~j]fJt$s~~T)tmt:;fi)~~~: "1OJpff~-T *1=r::g:~m' ~*1=r~~'&'1=r tzf~~J[~B)gm ::fOJ11~::g:~m

0 0 0"

(8.429a18-23) T225B: ~*E!: "~~-T~M~~~)t [ZSJt~{~j]fJt)t~N {OJ ~-T*>J(:Mm~~,&,>J(: j;M~)#~m~::fOJ1f~Mm {} :-;; .. (8.48213-15)26
0 0 0

Here the name of the person to whom SubhUti is speaking, the god Sakra (~~m:tglZSl in Lokak~ema's rendition, including the epithet deviiniim indra "lord of the gods"), has been eliminated; likewise his epithet Kausika, presumably used in the vocative in the underlying Indian text (t1i]~ in T224), is also absent from T225B. A long reference in T224 to those gods (devaputra) who do not yet practice the bodhisattva path ({P]?fiji~.:r-*1'T:g:Willi' ;tt:*ij-:ff) has been shortened in T225B as well ({P]5'i::.:r-**Mlli:ff). Sometimes several types of abbreviation are employed in a single passage, and in such cases the resulting difference in length can be even more extreme. A good example can be found in Chapter 3:
T224: [g~xB{~1'f: "~Jli3~i;l'&~~-T~:ftA~mt:;fi)~~~~
ft~m~ 0" JtP1J=m*~&Jt~~*-A1~B{~1'f: "~Jli3~i;l'&~ ~T~:ftA~mt:;fi)~~~~' t~~, m~ 0" !f~mfB[ZSJB{~1'f: "~i3i
;l'&~~T' ~:ftA~mt:;fi)~~~~' t~~, m~o"


T225B: [g~x !f~ Jt&~~T~'&B{~1'f: "~&'i;l'&~ftm~ 0" (8.483clO-11) -

Here the statements made separately in T224 by the four lokapiilas (lm5'i::.=:E in Lokak~ema's translation), Brahma Saharp.pati and other gods of the Brahmaloka (~. -flf4s:5'i::.&~5'i::~5'i::A), and Sakra, Lord of the Gods (~~m:if'''IZSI) are compressed in T225B into a single statement made by this entire group (lm5'i::.=:E~~~.&~5'i::.:r :W) The statement itself - in T224, "We will protect those good
26 The character enclosed in braces is presumably a scribal emendation, and should be removed.

Who produced the Da mingdujing *~EjI~J (T225)?


men and good women who study the prajiiiipiiramitii, uphold it,27 and recite it" (IZE!~}H~:~:13-=r' ~frA~~bfi5Elm~lf' t~lf' ~f If) - is also starkly abbreviated in T225B, which reads only "We will protect those who study, uphold, and recite [it]" (IZ~~~~ Nflf). As a result, a passage that occupies eight lines in T224 takes up less than two full lines in T225B. This pattern obtains throughout T225B, and there is no need to belabor the point by citing additional examples here. 28 T225A, however - as we will see in the following sections - has a quite different relationship to T224. (2) T225A is not abbreviated with respect to T224. It is not as straightforward to compare the length of T225A with that of T224, for as noted above; T225A includes an interlinear commentary. Thus it is only by subtracting the space occupied by this commentary that a genuine comparison with T224 can be obtained. If we do so, however, the result differs dramatically from what we saw in the case of T225B. A typical example, drawn from a passage at the beginning of the sutra, is the following:
T224: {~1mJ;tJt~' ~I*JIf!IlliLlJr:p ~~iiJttJi{IlFFPJH' ~~r ' ~fU 9t, ~JfB'm~ , ~~iiJ~i~~~I5i: ' 5jj]i~ 'y:J*gfflfUi~~ J=j +

1iB~:tlX;a; 0

0 0
0 0

T225A: M:ftD~ -aM~}Qt:li1'3::~~ '~.LlJ ,~*ttJi~::fPJH ~

r~*~- &*Xf;.i~~~lo ~11.wJ:1o ~a;+1i]\1fBJ=j~


Subsequent passages display a similar ratio, for example the following:

T224: 1PJP)ttl: ?{~PfT~)t:, 5t:r:ppfT~\'~~
*~ 5t:r:p~*::f;tt*'

\,1li5t: E"*fx: E" f61tl.ttl:? a;ffiJ~5t:, ~::f~~ffE!;E\ ~:!5I5r'

0 0 0 0

~:tt::AffiJ~ (8.425c14-17) T225A: pfTP)ff1PJ? ~{~~)t:, iiJ:~5t:~~ffr' ~ff:tt::1~)t:~tl.1.w

~ ;t::!;1.w~ffpfT~' PfT~' pH"' -t)]:ftD)t:~*'


pfTP)ff1PJ? :ftD*~)t:

1.wwr~ff ~~r1'f*~:ftD~~~pfT*,

(8.478c1-8, with commen-

tarial material removed)

27 I.e., who bear it in mind. 28 For one other comparison, drawn from the story of the bodhisattva
Sadaprarudita toward the end of the satra, see Zurcher 1991, p. 281.


. Jan Nattier

In a few cases, T225A is even slightly longer than th~ corrresponding passage in T224. For example:
{'F~~~jis)g:i~ , :g:ilJIlifJ~)G\T'liig 'T!; , T'f~ , TifE ' T ~ , ~w.f~~{'F~~ ~f'F~f ~~{'F~~ ArpJ~\/f~~~~
0 0 0

T225A: s:fto~5t, :g:ilJI~~T~' T~' T. 'Ttg 'T.~J,!;~ 'T

jf' T}~, 'T~ifElltf~~*,Y~Ii' :WLf~jffitJ~1=r ' ~U~PJ~~f5z*m 5Z.:g:ilJI*1=r~Nm~~~~llt :ftO~llt*T~~~: ":tltmill:~
0 0 0"



In sum, the translator of T225A does not show any signs of having attempted to condense the wording of his Indian source-text. (3) T225B follows the non-technical wording of T224 very closely. While the difference in overall length between T225B and the parallel portion of T224 is quite striking, in other respects what is noteworthy is the degree of similarity between the two texts. If we set aside for the moment the question of Buddhist names and terms, restricting our inquiry to the ordinary (non-technical) vocabulary used in the two texts, we find that the author of T225B drew heavily from the wording of T224 even as he abbreviated its prose style and replaced many of Lokak~ema's transcriptions of Buddhist names and terms with Chinese translations. We have already seen several examples of this phenomenon in the passages cited above. A particularly vivid instance, however, can be found in Chapter 16 (corresponding to the beginning of Chapter 19 in the Sanskrit):
T224: ~~:g:~B{~: ":g:ilJIt~1)]~~31I[liiJwi7i':::'f~.:::.:g:, st~ik
~~31IL ?" {~: "1)]~~1k*~ ~m~m~JJ
0 0"

*~W~~m~JJ {P}~:9J1}~i:Jj-&* ?" {~: '1i:ftOf~f;JH:t m1)] i:JjfYf~tt? mik*fYf~tt?" ~~:g:~a: "~F1)]~fY~tt 'T~ ~fYf~tt ' 2Jj\~Fik8Af~tt~1f\T~fHkfYf~tt 0" {~Fo5~~:g:~: "-z:;:{P}? :fto~ 'T?" ~~~~: ":fto~' 5\:cp5\: !" {~: ":g:ilJITm1)]~1~[liiJ
0 0



~NJi7i':::'f~':::'~' T~fH)]~1~' Tm1k~1~' T~lik~1~m 0"

(8.457a15-25) T225B: :g*Fo5: "Ml.l;J,W~3ITmLJE~ill:f~? tJ1k*~31I:s:F?WT m~mJt {PJ~:9J1}i?:(i:Jj-&*L*?" {~: "V:ftO;l;[1if~--:jff1)]i:JjfY


j~tt:s:F? ik*fYl1~?"


"~F1)]fYTitl11)]fY 0 ~FikfY:1]\TitllHk

Who produced the Da mingdujing *~J5rm: (T225)?


BJ'jo" {Jt": "3<O~/f"mM~1~~l:iE~m3J\/f"~lH)]x~> ~F1Jt~ 3J\/f" il1Jt~1~ (8.496b16-22) 0

In terms of Buddhist technical terminology and general rhetorical style, these two passages are as different as they could possibly be. The word anuttarasamyaksa1J1bodhi is transcribed as anouduoluosanyesanpu llaJ*fiJ~H - Jf~ -i' in T224, while it is translated as wushang zheng zhen dao ~J:lE1J;m in T225B. SubhUti appears as Xuputi ~~i'm: in T224, but as Shanye ~* in T225B, and the key term bodhisattva is given as pusa i'i\i in T224 but as kaishi M in T225B. In addition to these differences in Buddhist terms, it is also clear that T225B is substantially shorter, eliminating repetitions and condensing the rhetorical style of T224, as we have already seen in the other examples given above. Yet in other respects the two texts exhibit striking parallels. In the above passages I have underlined some of the instances in which the two contain identical terminology (though T224 sometimes supplements them with additional words). Such a thoroughgoing resemblance in non-technical or "ordinary" vocabulary - especially when viewed in light of the dramatic differences in rhetorical style and in Buddhist names and terms - cannot be accidental. On the contrary, given that this phenomenon is attested throughout chapters 2-30 of the Da mingdu jing, it seems clear that T225B is not an independent translation, but rather a revision of T224. The author of T225B thus maintained a great deal of the non-technical . terminology found in Lokak~ema's translation even as he radically altered its proclivity for the transcription of Buddhist names and technical terms and its repetitive and verbose style. The same cannot be said, however, about T225A. Here we have a far smaller body of material to deal with, since T225A consists of only a single chapter of the text. Based on a comparison of this material with the corresponding portion of T224, however, we can immediately see that the relationship between these two texts is very different than that between T224 and T225B. (4) T225A does not adopt the terminology of T224. In sharp contrast to the case of T225B, the wording used in T225A does not show signs of reliance on Lokak~ema's text. The opening nidana


.Jan Nattier

found in these two translations has already been intrqduced above; here, instead of noting what the two texts have in common, I have underlined instances where - if T225A were dependent upon T224 - the two texts might be expected to agree, but they do not:
{1In~~ttt ' ~IJUflliLlJcp *~OJttli{tFFPJt' ~5I'S-T ' '@ffU ~ , ~JHfm~ , *~OJ~:g:~m:Sf;(-1ft ' 5ij]:g:~ ')(J*gffifU:g:ilJi~ fj 1iBJGftl(;~--:- (8.425c6-9) 0 0 0


T225A: I4l"ftD~ -B~{~5Qi~.:E'@f~ ,3-tnLlJ ' ~*ttli4i)FPJgt o 51'S -T~*~- R*~:g:~m:Sf;(-1ft :fJt~J:i ~B~+1iJ,WB' fj)M'ij (8.478b23-25) 0 0 0

First of all, in T224, as is typical of Lokak~ema's genuine translations, there is no equivalent of the famous formula beginning with "Thus have I heard" (evaY[l maya srutaY[l), T225A, however, does not follow suit, but uses the standard pre-Kumarajlva form "Thus it was heard [by me]," wen rushi M:5<O~, followed by yishi -8~ "at one time,"29 The terminology used to indicate the Buddha's location likewise does not match, being expressed with "was staying at ..." (zai 1E ... zhong cj:I) in T224 but with "was traveling about in ..." (you yu ijfE}j~) in T225A. Another noteworthy difference is in the treatment of what was surely the word pramukha, a term which means "at the head" in the sense of either "foremost" or simply "and so on." While T224 takes it in the latter sense, using deng (here meaning "et cetera") in both cases, T225A has "number one" (di yi m-) in the first instance and "foremost" (shang shou...t 1) in the second. The terms used to refer to the time at which the d.iscourse took place are also different: according to T224, the sutra was preached on the fifteenth day of the month, "at the time when the precepts were pronounced" (shuo jie shi g)tffZ8~). T225A, by contrast, uses the term "abstinence day" (zhai ri ~B), adding also that "the moon was full" (yue man A)iiili). Such cases can be found throughout T225A, and as we will see in section (6) below, many of the differences involve not only the wording but the content of the text as well.


29 For a discussion of this and other treatments of the opening formula in early Chinese Buddhist translations see Nattier forthcoming.

Who produced the Da mingdujing *~~~~ (T225)?


There are, to be sure, some vocabulary items that do match: for example, the expressions "incalculable" (bu keji ::fCJH) and "innumerable" (wuyangshu f*~IQ:) are the same in both texts. This is, however, only to be expected if the two translators were working from similar originals and there was no obvious alternative available in Chinese. Compared with the close resemblance we have seen abovebetween T224 and T225B, the degree of difference between T224 and T225A is striking. (5) T225B follows T224 very closely in content. In the examples given in sections (1) and (3) above, we have seen that, even as T225B condenses the text of T224 by eliminating repetitions and summarizing well-known lists (e.g., by referring to "the five skandhas" rather than naming the five items individually), it does so without altering its overall content. The pattern illustrated by the examples already given above can be seen by comparing virtually any section of T225B with its parallel in T224, and it is not necessary to adduce additional illustrations here. There are, to be sure, occasional discrepancies between the two; these can probably best be explained by postulating that the author of T225B also made use of an lndic-Ianguage manuscript that differed slightly from Lokak~ema's source-text. 30 The overall pattern, however, is one of extremely close correspondence to T224. (6) T225A often diverges from T224 in content. We have already seen that, unlike T225B, T225A differs from T224 not only in its rendition of many Buddhist names and terms, but also in its ordinary (i.e., non-technical) vocabulary. But it diverges from T224 in more substantive ways as well. In the opening lines of the siitra, cited in section (4) above, T224 singles out two disciples (Subhiiti and Sariputra) and two bodhisattvas (Maitreya and MafijusrI) for special mention. T225A, by contrast, mentions only one character in each category (Subhiiti and MafijusrI, respectively). Several other differences in the content of this passage have also been not-

30 For examples of such slight divergences see the lists of non-human beings on pp. 329ff. below.


. Jan Nattier

ed above. These are matters of substance, not simply variations in mode of expression. Thus it seems reasonable to assume that these discrepancies reflect differences in the source-texts that served as the basis for T224 and T225A. Though T225A consists only of a single chapter, numerous other divergences between its content and that of T224 can also be found. Indeed, some passages are so different that - were it not for their location in the sequence of the narrative - it would be difficult to determine that they are parallel to one another.31 In sum; despite the fact that it is comparable to T224 in length, T225A does not exhibit a close relationship to T224 in any other respect.

The examples given above demonstrate clearly that T225B is not an independent translation, but instead is a revision of Lokak~ema's Daoxing jing (T224). Though the author of T225B shortened the text dramatically by eliminating much of the repetitive prose of the Indian original, and though he replaced most of Lokak~ema's cumbersome transcriptions with Chinese translations of Buddhist names and terms, he also carried over a great deal of the non-technical vocabulary found in Lokak~ema's text while, in most cases, reproducing the overall content of T224 (albeit in condensed style). Thus, despite the radical differences between T224 and T225B in Buddhist terminology and in literary style, the direct dependence of T225B on T224 can clearly be discerned. The fact that the author of T225B made deliberate changes in the terminology and style of T224, however, means that these two texts can be used together to highlight the distinctive features of each. In particular, a systematic study of which elements the author of T225B did and did not change in the process of revising T224 may cast additional light on which portions of Lokak~ema's text sounded too colloquial to be acceptable to an author seeking to recast the text in more elegant and more literary Chinese. 32 31 See for example T224, 8.426b24-26 vs. T225A, 8.479b26-29, and T224, 8.426c21-25 vs. T225A, 8.480a7-1O. 32 On vernacular elements in Lokak~ema's translations see Zurcher

Who produced the Da mingdujing *~JJt~ (T225)?


T225A, on the other hand, manifests no direct connection to T224, and as such it is valuable in a different way. As an independent translation (of which unfortunately only a single chapter survives) it can serve as a witness to an Indic-Ianguage text clqsely related to, but slightly different from,. that used by Lokak~ema. Here a worthy project would be a systematic comparison of T225A not only with the first chapter of Lokak~ema's T224, but with the corresponding section of the second translation of the text produced by Xuanzang 1r~, preserved in his Da banruo boluomituo jing ~~;fi)Elm~37m: (T220[5]). The latter, which to my knowledge has received almost no scholarly attention to date, is the only one. of the post-Zhi Qian Chinese translations of the A~tasiihasrikii that may turn out to belong to the same recensional family, broadly conceived, as T225A and T224.33


Intertextuality: The relationship between T225A and T225B This is not the end of the story, however, for before we can determine which part - if either - of T225 might be the work of Zhi Qian, we must first deal with the relationship between the two parts of the Da mingdu jing itself. Above we have seen that the terms used in T225A and T225B - not only specifically Buddhist expressions, but also ordinary non-technical words - exhibit differences that are too great to allow the hypothesis that the same person could have been responsible for both parts. In certain cases, it is true, the two parts use identical Buddhist terms:

1977 and 1996. 33 Xuanzang knew of two differentIndian versions oftheA~tasiihasrikii prajfiiipiiramitii, which he evidently considered different enough to warrant separate translations. Of these T220(4) is the more developed version, while T220(5) appears to be based on an older recension. In particular, the similarity of its opening nidiina, which mentions the presence of Subhiitiand Sariputra as well as Maitreya and MafijusrI in the audience (7.865c7-11) is similar enough to that found in T224 and (to a lesser extent) T225A to suggest that these three texts may belong to the same branch of the textual family tree.


. Jan Nattier Sanskrit sriivaka pratyekabuddha mahiisattva tathiigata prajiiiipiiramita avaivartika / avinivartanfya T225A andB

* :f{O*



Such matches as these are hardly decisive, however, for these same terms also occur in many other Buddhist texts, including (but not limited to) works translated by Zhi Qian. Thus the presence of such widely used Buddhist terminology in both T225A and T225B cannot tell us very much. There are, however, a few shared terms which are extremely rare, and whose occurrence in both T225A and B is of great significance. Especially noteworthy are the following proper names:
Sariputra Subhilti Grdhrakilta

Of these Qiuluzi fXlmT (var. ~JtT) for Sariputra appears only in five other texts,35 while Shanye ~* as a translation of "Subhuti" does not appear - to the best of my knowledge - anywhere else in the Chinese canon. Occurrences of lishan (or its variant Yaoshan ~~W) are likewise extremely rare. 36 The fact that both


34 T225B also uses the term ~N(~t]E) in a few cases, but these are clustered in just two places in the text (four occurrences on 482b, two on 485b), and thus it seems virtually certain that they are the result oftextual emendation. 35 For a discussion of these occurrences (which are found, aside from their appearance in T225A and B, only in Tl45, 152,500,507, and 769, in addition to texts citing or commenting upon these works) see Karashima and Nattier 2005, pp. 362-365. 36 The term appears in what was probably its original form, Iishan Llr "Chicken Mountain," in Tl6 (p}Mm.~7\jffl~, 1.250c14), Tl01 (*l 1lnJ13-~J[, 2.496b14), Tl50A (t)}l&=W~J[, 2.880bll), and T507 (*:5:~~, 14.774b26), in addition to the occurrences in T225A and B. (There are also a few occurrences in commentarial sources quoting from the texts

Who produced the Da mingdujing *IJJ3N~ (T225)?


T225A and T225B contain these highly unusual names thus seems to require one of two scenarios: either the translator of T225B had access to a copy of T225A (whether a complete or an incomplete version of the text we do not know), or the translator of T225A had access to a copy of T225B. The only other alternative would seem to be that both translators had access to an unknown third source a "catalyst text," as it were - which contained this vocabulary and from which both of them could have borrowed. No such text is extant, however, and surviving lexicographic sources do not provide any evidence for the existence of such a work. The most prudent course, therefore, would be to assume that there was direct contact between T225A and T225B. If this was in fact the case, it would imply that one of these two translators invented the name Shanye - and perhaps the name Qiuluzi as well - while the other, in the course of producing his own version of the text, simply adopted these terms. All three of these unusual names are used exclusively - that is, without any alternative translation or transcription - in both parts of T225, and it would seem at first glance that there is no viable way to determine the direction of their transmisssion. Yet a few clues may be available nonetheless. First, they are translations of terms that already had well-established transcriptions in Chinese by the time of Lokak~ema (late second c. CE), viz., Shelifu ~ fU5lt for Sariputra (already attested in the works of An Shigao),

just listed.) In the more elegant-sounding form Yaoshan ',~LlJ "Hawk Mountain" it occurs in T5 ({JI;f,Jt5Jb5.:*,Ilf, 1.160b8, c26 and 28), T6 (f,Jt5Jb5.: ~!, 1.176a5 and 13, b16), T152 (7\)j{~~, 3.1a7), T511 (14.779a9 and c8, both with the variant reading Iishan $iLlJ), T528 CifiliWi~!, 14.803all), and T536 ($ Bx'*~!, 14.819b29). The latter form is discussed in one treatise (T1766, 53'!.~~~~5~t~~, 38.l8b28) in connection with its use in T5, and it is regisetered in the Yiqie jing yin yi (T2128, 54.672cll) in the section dealing with vocabulary used in T511. With the exception of T528 and 536 (both of which are listed as anonymous in Sengyou's catalogue and, pending a detailed study, must be regarded as of uncertain date) and the treatises which quote from the above works, all of the texts in which these two renditions of Grdhrakuta occur appear to have been produced in the second or third century CEo


. Jan Nattier

Xuputi ~&~m for Subhiiti, and Qishejue ~~Ilfal for Grdhrakiita. Thus they represent a deliberate attempt to find an alternative for these transcribed forms. Another feature that these terms have in common is that, etymologically speaking, they are clearly erroneous. Qiuluzi appears to be based on an interpretation of the alternative name Siiradvatzputra as consisting of sarada "autumn" plus a Prakrit form of either pathi "road, path" or dadhi "yoghurt, fermented milk."37 As to Shanye for SubMiti, the first character (shan :g "good," for su- "good, well") is unproblematic, but the use of ye "work, business" to translate -bhuti "existence, wellbeing, prosperity" is not at all expected. Likewise the translation of Grdhrakiita as lishan "Chicken Mountain" is puzzling, reflecting perhaps a confusion between Prakrit forms of Grdhrakuta and kukkuta "cock." We may begin, therefore, by formulating the question in this way: In which part of the Da mingdu jing as we have it - T225A or T225B - would such creative but mistaken translations appear to be more at home? That is, can we find, in either part of this hybrid text, other translations that appear to be of a similar type? It is immediately evident that both parts of T225 contains a rich array of translation terms that may be relevant to our inquiry. But an essential principle in understanding the'modus operandi of any Chinese Buddhist translator is that we must first distinguish between those terms that appear to have been newly introduced in the text in question and those that were already present in earlier translated texts. That is, to understand how a particular translator worked, we must distinguish between terms that he himself seems to have invented and those that were already in circulation and which he could have simply borrowed from another source. In the case of the Da mingdu jing this means that, before saying anything at all about the terminology used in either section, we must first compile a list of those terms found in T225A and T225B that had previously appeared in other Chinese texts. When we do so, it becomes immediately apparent that both parts of T225 are drawing on a substantial reservoir of pre-existing


37 See Karashima and Nattier 2005.

- Who produced the Da mingdu jing *fJJ3Jjt~ (T225)?


Buddhist vocabulary_ The terminology used in T225A has anteced':' eI1ts in a variety of sources, among them the works of An Shigao (who uses Wangshe 3:.~ for Rajagrha, xianzhe Jf~ for ayu~mat, and biqiu bcli for bhik~u), Lokak~ema (whose translations are the earliest extant works to use pusa ifWi for bodhisattva;though the term was surely in circulation orally well before), and Kang Mengxiang (shizun t!:!:# for bhagavat, zuxingzi ~M:.y for kulaputra). It also employs numerous terms introduced by An Xuan and Yan Fotiao (e.g., Jingshou iff)(:t for MaiijusrI, duwuji &f*~ for piiramitii, shanze LWi for ara1}ya, and wushang zheng zhen [zhi] dao f*J:.lEJHZ]m for anuttarasamyaksaf!lbodhi). T225A gives the impression, in sum, of having been produced by a translator familiar with a wide range of translations produced through the early third century CE, and of having drawn his terminology from a number of these sources without discrimination. T225B, by contrast, shows a particular preference for the terminology of An Xuan and Yan Fotiao, adopting even two of their most idiosyncratic renderings (kaishi M1 for bodhisattva and chujin ~ I for bhik~u) as well as a host of more ordinary terms (including yingyi JJ!{~ for arhat, Cishi ~~ for Maitreya, and so on). In this regard Lancaster was correct in calling attention to the large number of terms used in T225B that have counterparts in T322. But while it is not possible (as we have seen) to say that T225B was produced by An Xuan and Yan Fotiao themselves, the fact that the creator of T225B drew heavily from their work is undeniable. Indeed, it is no exaggeration to say that T225B was produced by an author who was under the spell of the terminology and style of the Fajing jingo What this implies for the topic of discussion here is that it is the author of T225B, and not that of T225A, who would have been most likely to follow An Xuan and Yan Fotiao in their practice of translating (rather than transcribing) proper names, thus jettisoning the well-established forms Shelifu ~fU9t "Sariputra," Xuputi ~i:fm "SubhUti," and Qishejue ~Mil!Iffi "Grdhrakiita" and replacing them with the imaginative neologisms Qiuluzi fJca.y. Shanye ~~, and Jishan ~LlJ.


Jan Nattier

If this is the case, it would imply that the author of T225A had access to a copy of T225B, and that he adopted the renditions of the names used in the latter in some cases, but not all. A good illustration of this scenario is the treatment of the name of a figure who plays only a minor role in this sutra, but who appears (fortunately for our purposes) in both T225A and B. His name, which appears in Sanskrit texts as Purl.lamaitrilyal.llputra, is translated in T225A as Maneizi ~iiili~T, which can easily be identified with the Sanskrit, with man .~ "full" serving as a translation of piinJ-a (id.), while maitraya1Jz has been interpreted as a derivative of maitrz "lovingkindness" (hence its translation as ci ~ "kind, loving"). The version of the name found in T225B, by contrast - where it is given as Manzhuzi ~f5LT "son of fulfilled wishes" - is unexpected. It is surely the case that, as KARASHIMA Seishi has suggested, the translator of T225B did not see a form of the name resembling the Sanskrit Piin;amaitraya1Jzputra (which would have allowed its association with maitrf) , but rather something closer to the Pilli PU1Jftamantaniputta.38 It is less certain, however, that his contention that the translator interpreted the middle segment of this name as if it were the a form of the word mantra (cf. Pilli manta) is correct. If this were the case, the translator of T225B would surely have used the easily available equivalent zhou JE "spell," a term that actually appears several times in the same text as a translation both of vidya (in the sense of "magical formula") and of dhara1Jf. 39

. 38

See Karashima 1992, p. 277

For zhou in T225B see 8.484a2-b15, where it occurs five times as a translation of vidya, and 8.506b1 and 507b25, where the corresponding Sanskrit text has dharalJ-f. One additional occurrence of zhou is at 495b24, where it occurs in a list of practices to be avoided by the bodhisattva (;fi {:fn~::fiTZ). The corresponding Sanskrit text contains a longer list of items including mantras, recitation Uapa), herbs (au~adhi), spells (vidya), and medicine (bhai~ajya), making it difficult to align precisely with the text of T225B. It seems likely, however, that the character zhou is being used here, as before, to translate a form of vidya, while fu {:f corresponds to mantra, and yao to au~adhi or bhai~ajya. The corresponding passage in Lokak~ema's text (8.455c2) also has three items, viz., f,f, tJG (previously used several times as the equivalent of vidya) and iT~.

Who produced the Da mingdu jing *8JjI!r~ (T225)?


It is true that both dhiira1'}1 and vidyii are translated by Lokain T224 as zhu :tJ? (sometimes with the addition of another character),40 but it is noteworthy that the author of T225B never follows suit, but consistently changes this reading tozhou n. Thus it seems virtually certain that, if he had perceived a form of the word "mantra" in Pfir:r;ta's name, he would have used the character zhou here as well. As to zhu :tJ? itself, its most basic meaning is '(good) wishes," especially wishes directed toward another, hence the additional connotations of "blessing" or "benediction." If, as is often the case in Indian manuscripts, the letter -n- was represented simply by a dot (i.e., an anusviira), it would be a simple matter for the translator to overlook it and to interpret a Prakrit form *manta as if it were mata, meaning "thought, intention, wish.''41 Be that as it may, what we have here is a case in which the forms of the name found in T225A and B are parallel in their overall structure, but differ in their renditions of the middle character. If the translator of T225A had access to a copy of T225B, but was also working from an Indian manuscript in a different dialect - one closer to Sanskrit, in which an echo of the word maitrf could be discerned in this name - it would have been easy for him to adopt the name found in T225B while "correcting" :tJ? to ~ in light of his own Indic-Ianguage source.

40 The character zhu tJL occurs in T224 at 8.431c18(2x), 19, and 21, and at 433b20(2x), 21, 22(2x), 23, 24, 25, 26, 27 and 28, where the corresponding Sanskrit text indicates an underlying form of vidyli; it also occurs at 474c26, 475b29, 475cl, and 477a7, where the Sanskrit text points to a form of dhlira1}l. The sole occurrence of zhu in a negative sense that is, as part of a list of practices to be avoided by the bodhisattva) - is found at 455c2, where it seems likely to be another example of the use of the character as a translation of vidyli (see the previous note). Two other occurrences of zhu (at 471a9 and 12) are in the transcription of a proper name, and are not relevant to the discussion here. 41 Here we may compare the rendition of Piin~a's name as Manyuanzi )jIjlj~-T "Son of Wishes-Fulfilled" which appears in the works of a number of other translators, including Dharmarak~a and Zhu Fonian. (Note that the term yuan, which eventually came to be used as a technical term for "vow," generally means simply "wish" or "desire" in early translations.)


Jan Nattier

The picture that emerges from the above considerations, in sum, is that the author of T225A produced a retranslation of the text, relying heavily on a (different) Indic-Ianguage original but also consulting the lost first chapter of T225B. In the process he took over a few translation terms found in T225B, including the renditions found there of the names of Sariputra, Subhuti, and Grdhrakuta. In many other respects, however, the translator of T225A showed his independence, jettisoning some of the peculiar vocabulary borrowed by the author of T225B from An Xuan and who wrotYan Fotiao and reverting to the use of alternatives already well established before his time. Zhi Qian and the Da mingdu jing: Re-framing the question At this point we must finally return to the question posed in the title of this paper: Who produced the Da mingdu jing? As we have seen, the text consists of two quite distinct parts, and so we must ask the question separately concerning each one. Since this hybrid text has long been attributed to Zhi Qian - not just in modern editions of the canon, such as the Taisho Shinsha Daizokyo, but at least since Xuanying encountered it in the seventh century CE - it seems reasonable to begin our inquiry by asking which of the two parts of the Da mingdu jing, if either, can legitimately be counted as his work. A major challenge, however, immediately confronts us in attempting to answer this question, for Zhi Qian's translations are .. extremely varied in both vocabulary and style. Some manifest a four-character prosodic pattern, while others are composed in nonmetric form. A few (including T169 and 632 and, to a lesser extent, T361) abound in the long transcriptions introduced by Lokak~ema, while another group (T76, 474, and 532) features the distinctive translation terminology introduced by An Xuan and Yan Fotiao. Most contain a mixture of vocabulary of both kinds, but even here we find distinctive subgroups of other kinds. If we examine the patterns in Zhi Qian's treatment of the word arhat, for example, we find a mutually exclusive distribution of the terms aluohan IfriJm5l and luohan m5l in his works; we also find that some texts belonging to the luohan m~ group also use vocabulary belonging to the

Who produced the Da mingdu jing *I!f3Jjt~ (T225)?


zhenren Ja:A group, while those belonging to the aluohan ~iiJ~5J: group do not 42 Some of this variety is surely due to the evolution of Zhi Qian's own stylistic preferences during his thirty-year translat.ion career. Recalling that as a young man he was a student of one of Lokak~ema's disciples in Luoyang, but that most (possibly all) of his translations were produced after his move to the Wu kingdom in the South, it is possible that the texts in the "Lokak~ema-like" group are his earliest productions, and that in subsequent years he abandoned that mode of translation in favor of a more literary and elegant style. But another factor was surely at work as well. It is well known that Zhi Qian revised the work of many other translators, and in so doing, he carried over various elements of their termi:p.ology and style. The fact that the texts he revised were themselves composed in a wide range of styles could thus have contributed to the diversity that we see in his work. In addition, however, we must take into consideration his own evident preference for variety, for Zhi Qian's terminology does not only vary from one text to another, but even within individual translated texts. 43 Given both the substantial size of Zhi Qian's translation corpus and the wide-ranging variety in his terminology, it is perhaps not surprising that - with the exception of proper names of people and places who happen not to appear elsewhere in Zhi Qian's corpus, virtually all of the Buddhist names and terms used in T225A and B can be found somewhere in the corpus of Zhi Qian. 44 They are

Nattier 2003, p. 235, and cf. the following note. See for example the translation of briihmalJ-a both as shixin Wiil,' and as Janzhi ~;t;; in T76, T198, and T21O, or the rendition of arhat both as yingyi H!{~ and as yingzhen H!Jr, in T76, and in no fewer than three different forms - as yingzhen J!!JJ;, zhenren JJ;A, and zhizhen ~Jr, - in T6, T87, and T474. 44 The one item that I have been unable to locate elsewhere is shanshi '& (usedin T225B as a translation of kulaputra). It seems likely, however, that this term was coined specifically as a modification of Lokak~ema's shan nanzi '&~-T (replacing ~-T with ), and thus was tied to this specific context alone.


. Jan Nattier

also used, however, in texts produced by many ot1!er translators, above all those of Dharmarak:~a (Zhu Fahu ~)*~), who borrowed extensively from Zhi Qian's vocabulary and style. If we begin with Buddhist names and terins, in other words - as has been standard in virtually all studies of translator attributions produced to date - we will find that either T225A or T225B could be described as "looking like" the work of Zhi Qian. What I would propose to do at this point, therefore, is to approach the problem from a different angle, examining a broad sample of ordinary expressions as well as a few unusual Buddhistic terms, to see whether they do, or do not, appear in Zhi Qian's translation corpus. T225A. We may begin with T225A, listing a sampling of items that appear here but never in any text that is solidly attributed to Zhi Qian. In cases where it is only a part of the usage that is unusual, the part which does not occur in Zhi Qian's work is underlined. Where there is an obvious variant in punctuation or wording that could be expected, I have searched for these alternatives as well (they are given in brackets). Where such variations actually occur in the text of T225A, they are given without being enclosed in brackets. (I have excluded material from the interlinear commentary, which is clearly of a different vintage and will be discussed briefly below.) Expressions found in T225A, but not in Zhi Qian's corpus
~:E~i'm~~LlJ L.'~m ~~LlJ]45 foJ~~... -0

foJPJT~ ... f~{~~ [f~{~gg, f~~L~, f~{~Lgg]

~ff [1tff]

fL~ m~ ~~

45 What is unusual here is the use of the character qi ~ to introduce the second piece of geographical information.

Who produced the Da mingdu jing *fJF.!Jjt~ (T225)?


The fact that so many ordinary expressions found in T225A - as well as a few specifically Buddhist formulations, such as "ascend the Great Vehicle" (sheng yu dasheng ~m-**), "receive a prophecy (vyiikaraIJa)" (shou bai tff), "six roots" (liu gen 7\*~, for the six indriyas or sense organs) - do not appear anywhere in Zhi Qian's quite extensive corpus makes it quite certain that T225A cannot be the work of Zhi Qian. The fact that T225A also contains some Buddhist terms that do occur in Zhi Qian's translations need not deter us from drawing this conclusion, since these are terms that can also be found in a wide range of texts by other translators. In sum, at this point in our inquiry we can confidently remove T225A from the list of possible translations by Zhi Qian. T225B. What, then, of the remaining part of the Da mingdu jing, i.e., T225B? Here we have quite the opposite situation from what we saw in T225A, for it is difficult to find any terminology in T225B that is alien to Zhi Qian's work. There are, of course, a number of cases where such expressions can be found, but virtually all of them fall into one or the other of the following categories: (1) terms that were borrowed directly from T224, or (2) terms whose absence from other translations by Zhi Qian can be explained by their content. The second of these categories includes the names of people and places who happen not to occur in other texts translated by Zhi Qian (e.g., SreI;lika, purI).amaitrayaI).Iputra, Dharmodgata, and
46 The character sheng ~ does appear in Zhi Qian's corpus, but never in conjunction with dasheng **or any other translation of mahayana.


. Jan Nattier

Sadaprarudita, as well as the city of GandhavatI); it also includes terms related to topics which are not treated elsewhere, e.g., f~WX ~t~ti (for pU1Jya-anumodana "rejoicing in [another's] merit"), M riSr (a certain type of box in which a jewel is placed), or f;Jm (for "magician," a term taken over from T224). But it is not only general conformity with Zhi Qian's work that we find here; in fact, T225B contains several extremely unusual terms that are virtually unique to the translations of Zhi Qian. In the following sections, therefore, I will focus on some extremely rare Buddhist terms which occur in T225B, but not in T225A, and which can help to clarify the identity of the translator of this portion of the text.

Ca) The best of bipeds: Siikyamuni in T225B

Above we have encountered several epithets of the Buddha that are used in T225B, among them tianzhongtian 7:CP7: (used as a translation of bhagavat), wusuozhuo ~FJT~ (for arhat), zhengzhen dao zui zheng jue lE~m:l&lEt: (for samyaksa/!lbuddha), and of course fo {5f# (for buddha, as well as for bhagavat). Of these, only the ubiquitous fo also occurs in T225A as an epithet of the Buddha. 47 As we have seen, all of these expressions - though most of them are quite unusual - can be encountered in the translation corpus of Zhi Qian. The most peculiar term used in T225B, however, is not an epithet but a proper name: the term Nengru fgg1[ffi "Capable of Being Scholarly" as a translation of Sakyamuni. The name occurs in T225B in a passage in which the past Buddha Dipamkara confers a prediction on the future Buddha Siikyamuni. Th~ passage in question reads as follows: tE:W:{5f#B't[~tlfZ;)~~: ";fiirt'i'

EJAcP :t-if~}l{5f#~ , fF{5f#ilfgg1[ffi , ftO* ' ~FJT~ , lE~m:l&lEt: ' Jf.:I&#" (8.483b29-c2). The term Nengru is quite unexpected as
a translation of "Sakyamuni,"48 though the similar term Nengren


47 There is one occurrence of wusuozhuo ~PfT~ in T225A, but there it means simply "unattached," and is not being used as a translation of arhat as an epithet of the Buddha (see 8.480c15-17: f-*KTBFp~: "Mit!:
iHI*?J}\1Et~9;Ojffi~/F~?" ~*B: "j;)~~it!:n~1EZ~J;Djffi~pfT~").

48 The use of the term ru 11m is especially peculiar in light of the fact

Who produced the Da mingdu jing *~,Ej1l~ (T225)?



("Capable Df Humaneness") o.ccurs in a number Df texts. 49 The name Nengru, by cDntrast, is much rarer, Dccurring in Dnly two. Dther places in the Chinese Buddhist canDn (aside frDm discussiDns Df the term in lexicDgraphic wDrks and qUDtatiDns Df it in Chinese treatises): in Zhi Qian's Pusa benye jing lfiii*~U~ (T281), in a list Df highly sinified epithets Df the Buddha,50 and in Zhi Qian's biDgraphy Df the Buddha (Taizi ruiying benqi jing Jlfflll*JEQ~, T185), in a glDSS explaining that ~ means "Sakya" and f~ means "muni."51 The Dnly Dther Dccurrence Df this name in a suppDsedly translated text is in BaDyun's .~ Fo benxing jing {~ ::zjs:1T~ (T193), which draws Dn a wide range Df early translations and here appears to. be retelling Zhi Qian's version Df the StDry Df the future Sakyamuni's predictiDn (4.93bl). In sum, the name Nengru is attested Dnly in texts translated by Zhi Qian Dr in Dther wDrks qUDting directly from them. 52


(b) NDn-human beings: gandharvas, mahoragasand their ilk GDds (devas), nagas, and asuras appear with great frequency in early Chinese translatiDns, but Dther members Df the list Df the sD-called "eight kinds Df beings" (J\fi5) appear Dnly rarely. 53
that it means not only "scholarly" in a general sense, but also (in certain contexts) "Confucian." 49 The earliest occurrences of the term Nengren are in Kang Mengxiang's (revised) biography of the Buddha (T184), Zhi Qian's Vimalakfrtinirdesa (T474), and Kang Senghui's collection of jataka tales (T152). There are also well over a hundred occurrences in the translations of

~1f~1115~*~A.0 ~1f~1115~*5j;>F~ a ~5JJtmtb a ~5JJt1$A. ~f*~

(10.447a14-17). 51 )ft:~{'F1115 a ~,*~:x:(*~~,*~~ij5:x:~f;~~g5f~) (3.473a23). 52 In addition to Baoyun's apparent citation, the name appears only in treatises composed in China in which earlier translations are quoted or the term itself is discussed (e.g., T1763, a commentary on the MahiiparinirvliIJ-a-satra [T374], which quotes the gloss in Zhi Qian's T185). 53 The expression}\$ itself does not occur in the work of any translator of Zhi Qian's time or before

~ a ~f*t!t~ a ~~ij51~ a ~~~{llJ a ~~*m a


Jan Nattier

Fortunately, however, extended lists of various categories of living beings occur in two different places in T225B, which allows us to examine their terminology in detail. The first one is the following:
+1J;fAAtfz{~WJ~72:A' ~'$J~ '~**1$' ttI~1$' fih<jJij~1'T1$' {J;lA~1$

(485a7-8) "All the gods (72:A), yakijas (~), nagas ($J~), asuras (~gg1$), gandharvas (ttlWH$), mahoragas (ijh<jJij~1'T1$), and kirrmaras ({J;lA~1$) of the innumerable buddha-lands of the ten directions ...."

Characteristically, the parallel passage in Lokak~ema's T224 is almost exactly twice as long:
+1J;fAA~tfz{~WJ~7::A' ~$J~' poJ~JH1ffi' ~~JZJ1M$' ~}Q[f!Ui~ ~,~gffia~~'~~ffia~~'~.~~~~~A~~~A

(8.434c28-435a2) "All the gods (7::A), nagas ($J~), asuras (poJ~~{1ffi), yakija-spirits (~ JZ.~1$), garua-spirits (~;fJa)~J$), ki'?1nara-spirits (gW8a~1$), gandharva-spirits (~W8a~1$), mahoraga-spirits (.!!f~1j]~1$), and human and non-human beings of the innumerable buddha-lands of the ten directionss ...."

A glance at the terminology used in these two versions also shows clearly how the author of T225B worked: in every case where Lokak~ema's version used a transcribed term - even such a familiar one as axulun /5iIJ~JJ{fffi for asura 54 - he has opted for a translated term instead. Another such list appears in a later section of T225B:


A~B. J=j. 1Ef. ~g:f$. t~*. A~:f$.1b1-A*:f$. ~Jit1T:f$. !l!t~&; vJ'A~Wt. ~*. J-IMM. ~IHI~~1ffl~. ~1TPfM}~" (506b25-27)

"He enters into [the realm of] the sun, moon, stars, asuras ~**1$), naga kings ($J~~.:f.) [or naga- and yakija-kings?], gandharvas (ttI~ 1$), kif!maras ({J;lA~1$), mahoragas (jfij~1'T1$), and snake-bodied spirits (!I!t~1$); he also enters into [the realm] of the birds and beasts

54 The transcription axuluo poJ~~a does not appear in translations produced during Zhi Qian's time or before; instead, the final syllable is regularly written -lun {1l!i (var. lJfff). I suspect that the latter forms reflect a spoken Iranian plural form that ended in -t, but this remains to be documented.

Who produced theDa mingdujing *~13Jjt~ (T225)?



for the animal realm), hungry ghosts (ftftJll), hell-[beings] (:tlli - [in short, into the realms of all] things that flit and wriggle, [all] things that-crawl and breathe (!DMffM!f~ItJ, ~~1Tl1ffif}~D." 55

Cf. T224, which again is twice as long and uses a substantial number of transcribed terms:
11'A~B 'F1 '11'A~1iiW' 11'A~JliiJ~~{~' 11'A~Mg' 11'A~JIlt$' 11' A~MWem '11'A~J!l!!Wilf)] '11'A~IL~Bm '11'A~)@UI~f}] '11'A~m ~U '11'A~ft~:tl!' 11'A~i$~ '11'A~~!t '11'A~)Jb~ '11'A~!DM~ ,
11'A~~1tJ '11'A~~~1T '9JV\~l1ffif,~, (8.475b16-21)

text generally offers a close match to the passage just cited, .though it contains two items - rak~asas (mif.:U) and kumbha1Jqas (N~fs) - that have no equivalent in T225B. Once again the author of T225B has both abbreviated the text and replaced these foreign-sounding transcriptions with translations. Though the two lists found in T225B are quite different from one another, they share four extremely rare terms: zhiliang shen jf ~t$ "spirits whose nature is indulgent" (?) for asura, zhiyue shen Wt~t$ "spirits who carry musical instruments" for gandharva,56 si renxing shen {.r;CAJ~t$ "spirits that take human form" for ki1'(lnara, and xiongyixing shen H~!JI1'Tt$ "breast-going spirit" for mahoraga (written xieyixing shen !JIfft$ "flank-going spirit" in the second passage). Making use of the wonderful search capabilities made possible using the CBETA edition of the canon, we can eas-


55 Here Zhi Qian uses a pre-existing expression for "all living beings" (sarvasattva), viz., !DM~~ItJ, which appears widely in Lokak:~ema's corpus as well as in the specific parallel passage here (I have borrowed the felicitous English translation "all things that flit and wriggle" from Paul Harrison). The expression ~1Tl1ffif,~ "things that crawl and breathe," by contrast, is not used as a compound by Lokak~ema; in his corpus these elements appear only here (i.e., in T224), where they appear as two separate components (~HT and l1ffif)~J 56 I am grateful to Paul Harrison for his assistance in interpreting this expression, in particular for for pointing out that the character ~ should be read as yue "music" or "musical instrument" here rather than Ie "pleasure, delight," reflecting the identity of the gandharvas as celestial musicians.


Jan Nattier

ily determine that these four translation equivalents o~cur together only in one other text in the entire Chinese Buddhist canon: the Si1'f'Lhamati-sutra (T532, Sihemo jing fbJrnJ**~)57 translated by Zhi Qian. 58 . This text, like the Da mingdu jing, contains a long list of human and non-human beings, in which all of the members of the list are rendered in translation rather than transcription:
-:fi:!([D~~z:1fft@ffU1~{;!!;. 0 ~)1( , ~ , ~.t$ , jf~t$ , tfi~t$

, 3i:


fttAJ~t$' g~HjfJt$ 'A' ~FA' ..

~*{;!!;*~fU' 1.&{'FtI~1'I

(T532, 14. 812a25-28) . First, after the Tathagata's extinction, his relics will be worshipped. The various gods ()1(), nligas (~), yak~as (~.t$), asuras (ff~t$), gandharvas (tfi~t$), garu4as (3i:,~t$), ki1'[Lnaras ({J;I,A~t$), mahoragas (g~HjfJt$), human and non-human beings (A~FA) will all come to make offerings to the relics; they will pay homage to them without limit.

Once again there are notable differences in content and sequence between this list and those found in T225B, which we can assume reflect differences in their underlying Indic-Ianguage originals. Yet the similarity in terminology is striking. There can surely be no doubt that these two texts are the product of the same milieu. The attribution of the Sihemo jing to Zhi Qian, as it happens, is one of the most solid in his entire repertoire: not only is the text credited to him by Sengyou (drawing, in turn, on the catalogue of Dao'an), but it contains a substantial number of lines in six-charatter verse, a style that does not appear in the work of any other

57 Var. fl)IilJ,*~ Sihemei jing; also called ~i\imW~ according to Sengyou, 55.6c23. 58 Some of the individual terms can be found in other places; of these zhiliang jf~ (not always with the added character shen t$) is probably the most common, appearing in a number of translations by Dharmarak~a as well is in other later texts. Interestingly, in the Fahua yishu $~iffrf (TI721), the term jf~ is explained as corresponding not to asura, but to sura, a back-formation of the word asura formed by taking the initial a- as a negative prefix; see 37.465b29ff. I would like to thank FUNAYAMA Toru for bringing this passage to my attention.

Who produced the Da mingdu jing *~l!t~ (T225)?


translator of Zhi Qian's time or before. The sutra is also mentioned in Sengyou's biography of Kang Senghui as one of the three texts on which Kahg Senghui is known to have written a commentary, which establishes the date of the text as prior to Kang Senghui's time. 59 According to the Zhongjing mulu ~~tT< compiled by Fajing )*~~ et al. (T2146), Kang Senghui also composed a preface to the text, which unfortunately has not been preserved. 60 An examination of the vocabulary of the Sihemo jing shows that, among the texts reliably attributed to Zhi Qian, it is one of those that most closely resemble the style of An Xuan and Yan Fotiao. With only a few exceptions, all Buddhist terms are translated nlther than transcribed, and a number of terms found in the Fajing jing are also attested here. If we assume (as I believe we should, in the absence of any countervailing evidence) that this is indeed the work of Zhi Qian, it might be classified as one of the translations he produced during the period in which he was entranced by the Fajing jingo Just as an art historian might have difficulty in ascertaining that the paintings produced during Picasso's "Blue Period" were the work of the same artist who produced his "Rose Period" or "Cubist Period" works, just so we may now consider T532 - and, I would suggest, T76, T474, and T225B as well - as belonging to Zhi Qian's ''An Xuan Period." Even though they differ in many respects from the texts produced at other times during his career, they also have many elements in common, and may be considered to belong to a single sub-category of Zhi Qian's texts.
The remaining piece: T225A

What, then, can we say of the remaining piece of the current Da mingdujing, i.e., T225A? As we have seen, the disjunction between parts A and B of the received text of T225 is too great to conceive of them having been produced by the same individual. If this is the case, though, who might be the author of T225A?



See T2145, 55.97a15: jZ5.1:~~~t5Tg$~mW=~. See T2146, 55.l47a9.


. Jan Nattier

One strong candidate for the authorship of T22,5A would be Kang Senghui, for according to Sengyou's catalogue he was the translator of atext called the "Wu pin" (~).61 This text is considered by ZUrcher and others to be a version of the smaller Prajiiiipiiramitii; could it be, then, that T225A is a remnant of Kang Senghui's translation of this text? Once again, because T225A is so short, it is possible to conduct . a fairly thorough comparison of its vocabulary with that of Kang Senghui's work. The following is a representative list of' expressions found in T225A that are unknown in the sole authentic translation by Kang Senghui (T152):



{iiJPJT~ 1~{~~

fL~ m~~

61 See T2145, 7a26. Sengyou describes the text as consisting of five juan and says that, in his time, it was lost (JmI). 62 As noted above, what is unusual here is the use of the character qi ;It to introduce the second.piece of geographical.information.

Who produced theDa mingdu jing




Indeed, a search not just of early Buddhist translators, but of the TaishO Daizokyo as a whole, shows that T225A is unique in many respects. The unusual terms it employs, which in many cases are not found in any other Chinese Buddhist translation, point to the likelihood that it was the work of a translator who produced little or nothing else. At this stage, therefore, it appears to be most reasonable to state simply that T225A is the work of an anonymous translator, and that it cannot be associated with any other work still extant today. Matters are quite different, however, with the interlinear commentary to T225A, which exhibits numerous similarities to the Yin chi rujing zhu (TI694). The two commentaries quote from many of the same texts - that is, they share a common "canon" (see above, n. 19) - and they use similar language, including the repeated statement "the teacher says" (gffl~). According to a recent study by Stefano Zacchetti, the "Master Chen" ~*B::; who is said to have annotated T1694 (presumably to be identified with the Chinese layman Chen Hui) is probably its primary author, with the comments attributed to the "Teacher" belonging to Kang Senghui (Zacchetti forthcoming). Given the very strong resemblance between T1694 and the interlinear commentary to T225A, I would suggest that the latter is a product of this same Wu-kingdom community. Conclusions In light of the evidence presented above, it is clear that T225A cannot be the work either of Zhi Qian or of Kang Senghui. Conversely, it seems quite certain that T225B (which shares a number of very rare terms with other works in Zhi Qian's corpus) should be considered one of Zhi Qian's genuine translations. As to the interlinear commentary to T225A, it seems very likely that it is the product of the community headed by Kang Senghui. We have also seen that T225B is a revision of Lokak~ema's T224, produced with at least a cursory reference to a different Indian manuscript. The authorship at T225A, however, remains unclear. Because of the large number of terms that are unattested in any other early Chinese translation, it seems quite possible that


Jan Nattier

this is the only work ever produced by the translator in question. For the time being, therefore, it seems most prudent to regard it simply as an "anonymous',' text.
Cowell, E.B., and R.A. Neill, eds., 1886. The Divyilvadilna, a collection of early Buddhist legends, now first edited from the Nepalese Sanskrit mss. in Cambridge and Paris. Cambridge: Cambridge University Pre~s. Dhammajoti, Bhikkhu Kuala Lumpur, 1995. The Chinese Version of Dharmapada, Translated with introduction and annotations. Kelaniya, Sri Lanka: Postgraduate Institute of Pali and Buddhist Studies, University of Ke1aniya. . HAYASHIYA Tomojir5 **~bz:~~~, 1941. Kyoroku kenkyu ~~:o.tf~. Tokyo: Iwanami shoten. IWAMATSU Asao it~~7;:, 1985.. '''TenchUten' k5" r i:f:1XJ::;: [An Examination of the Epithet 'Devatideva'). ToM J10'j", 1, pp. 201-219. KAGAWA Takao il?J[[;f:$IE, 1984. Muryojukyo no shohon taisM kenkyu 1Iiii:i:: %~0mf*M~:o.tf~ [A Comparative Study of the Texts of the Larger Sukhavatfvyuha-sutra). Kyoto: Nagata Bunshodo. KARAsHIMA Seishi ~~~M>~, 1992. The Textual Study of the Chinese Versions of the Saddharma-pw:u!arfka-sutra in the Light of the Sanskrit and Tibetan Versions. Tokyo: Sankibo. ___ and Jan Nattier, 2005. "Qiuluzi fkSr: An Early Chinese Name for Sariputra." Annual Report of The International Research Institute for Advanced Buddhology at Soka University for the Academic Year 2004 [ARIRIAB), vol. 8, pp. 361-376. KATSUZAKI Yfigen M1~m~, 1985. "Shiken no yakugo no kenkyfi / Daimyodokyo no yakugo bunseki" :$l~0m.~0:o.tf~ (-) - i"*f!I3J}t ~.lI 0m.~:7ttlT - . Taisho daigaku sogo bukkyo kenkyujo nenpo *IE* ~~*~{b~~:o.tf~~~i,no. 7,pp.67-93. Lancaster, Lewis R., 1969. "The Chinese Translation of the A~tasahasrika -prajiia-paramita-sutra attributed to Chih Ch'ien :$l~." Monumenta Serica, Journal of Oriental Studies, vol. 28, pp. 246-257. MATsuE Takashi t~)I*, 2005. "J5ko kango ni okeru ninsh5 daishi no 'kakkutsusetsu' 0 megutte .l.~5l~'':'i3~tQA:.f${-t~i!J0'mJffitff'~Ob<"? -C." Totetsu i"V~.lI13. MIZUNO K5gen 7.K!f5.j[;, 1981. Hokkukyo no kenkyu )*{g~0:o.tf~ [A Study of the Dharmapadas). Tokyo: Shunjusha. Nattier, Jan, 2003. "The Ten Epithets of the Buddha in the Translations of Zhi Qian :$l~." Annual Report of The International Research Institute

Who produced the Da mingdu jing *I.IJjt~ (T225)?


for Advanced Buddhology at Soka University for the Academic Year 2002 [ARIRIAB], vol. 6, pp. 207-250. _ _, 2006. "Masquerading as Transcription: Substitution Terms in Chinese Buddhist Translations." [Unpublished; read at the] International Symposium on Aspects of the Language of Chinese Buddhist Translations, International Research Center for Advanced Buddhology, Soka University, Tokyo, Japan, November 2006. _ _ , 200S .. A Guide to the Earliest Chinese Buddhist Translations: Texts from the Eastern Han *~ and Three Kingdoms -=~ Periods. Bibliotheca Philologica et Philosophica Buddhica, X. Tokyo: The International Research Institute for Advanced Buddhology, Soka University. ___ , forthcoming. "Now Y--ou Hear It, Now You Don't: The Phrase 'Thus Have I Heard' in Early Chinese Buddhist Translations." In Tansen Sen, ed., Buddhism across Asia: Networks of Material, Intellectual and Cultural Exchange. Singapore: Institute of Southeast Asian Studies. TAN Shibao ~t!tf*, 1991. Han Tang Foshi tanzhen ~1l'!f{~J!:~l{. Guangdong: Zhongshan daxue chubanshe r:pw*~t/jJjBH. TOKUNO, Kyoko, 1990. "The Evaluation of Indigenous Scriptures in Chinese Buddhist Bibliographical Catalogues." In Robert E. Buswell, Jr., ed., Chinese Buddhist Apocrypha (Honolulu: University of Hawaii Press), pp. 31-74. Ur Hakuju ~#{B., 1971. Yakukyoshi kenkyii ~~J!:1iJf~. Tokyo: Iwanami shoten. Vaidya, P. L., ed., 1959. Divylivadiina. Darbhanga: Mithila Institute of Postgraduate Studies and Research in Sanskrit Learning. Zacchetti, Stefano, forthcoming. "Some Remarks on the Authorship and Chronology of the Yin chi ru jing zhu ~:f,fA~>.1 T1694: The Second Phase in the Development of Chinese Buddhist Exegetical Literature." In Mauro Maggi and Silvio Vita (eds.), Buddhist Asia 2. Papers from the Second Conference of Buddhist Studies Held in Naples in June 2004 (Universita di Napoli "L'Orientale." Centro di Studi suI Buddhismo), Italian School of East Asian Studies, Kyoto. Zurcher, Erik, 1977. "Late Han Vernacular Elements in the Earliest Buddhist Translations." Journal of the Chinese Languages Teachers Association, vol. 12, no. 3, pp. 177-203. ___ , 1991. "A New Look at the Earliest Chinese Buddhist Texts." In Koichi Shinohara and Gregory Schopen, eds., From Benares to Beijing: Essays on Buddhism and Chinese Religion in Honour of Prof. Jan Yilnhua (Oakville, Ontario: Mosaic Press), pp. 277-304. _ _, 1996. "Vernacular Elements in Early Buddhist Texts: An Attempt to Define the Optimal Source Materials." Sino-Platonic Papers, no. 71, pp. 1-31.

A new attribution of the authorship of T5 and T6 MahllparinirvlllJa-siltra

Jungnok Park (t)*,l

In my thesis submitted to the University of Oxford, I examine how

Buddhist translators in China interpolated certain Chinese terms into the canonical body of Buddhist translations. There, my arguments mainly rest on textual attestation; for this attestation, the first step to be taken is to verify the traditional attributions of translatorship. In a chapter of this thesis, I illustrate how such a verification is performed, presenting a case study of T20 Fo kaijie Janzhi aba jing {5tIffl~,jt~IloJJff~~ / Amba:jtha-sL7tra, which has traditionally been attributed to Zhi Qian 3Z~ (fl. 222-253).2 In this paper, I adopt the same method for the attribution of the authorship of two ParinirvQf}a-sL7tra translations, T5 Fo bannihuanjing {5t~~5Jb5g~! and T6 Bannihuanjing ~~)Jb)g~!, to Zhi Qian. 3 Jungnok Park's tragic passing away prevented the author's own final revision and proof reading of the paper; he could, for example, not add translations to some of the sources quoted as I asked him to do. In respect for Jungnok Park's authorship I only undertook some minor changes in style and added some Chinese characters [Max Deeg].
1 I sincerely thank Prof. Paul Harrison and Prof. Jan Nattier for their kind comments on and corrections to my preliminary paper on this topic. I also thank Prof. Richard Gombrich and Mr. Lance Cousins, who supervised my research and drew my attention to the importance of T5 and T6. 2 In this paper, all dates are c.E., unless designated as B.c.E. 3 Since the thesis is planned to be published in book form, where the chapter in which I attest the translatorship of T20 will be omitted, and since, on the other hand, my unpublished thesis may be inaccessible to

Journal of the International Association of Buddhist Studies Volume 31 Number 1-2 2008 (2010) pp. 339-367


Jungnok Park

Following the traditional attributions, the editors 9f the Taisho shinshu daizokyo (hereafter T) attribute T5 to Bo Fazuo ffi5iHli (ft. 290-306) and T6 to an anonym during the Eastern Jin C*S': 317419) dynasty. Some time ago Ui (1971) and recently Nattier (2003) argued that T6 is probably a translation by Zhi Qian. However, in the course of attesting the traditional attribution of T20 to Zhi Qian, I found that, while both T5 and T6 appear to be translations by Zhi Qian or a successor in his circle,4 the translation of T5 can hardly be later than T6. Then, I undertook a critical reading of T5 and T6 in comparison with the Sanskrit and Pali recensions of the MahiiparinirviilJa-sutra (hereafter MPS), as well as the corresponding Chinese translations, and found that T5 and T6 are rare materials from which we can extract information about how archaic Buddhist texts developed into the later, standardised sectarian texts. These two texts are valuable for the following reasons: firstly, they are complete translations of the MPS, which is one of the most voluminous texts of early Buddhist literature, containing much that is of doctrinal and literary interest; secondly, given the popularity of the sutra, there are plenty of corresponding texts that demonstrate its chronological development; thirdly, and most importantly, these two texts clearly demonstrate the process of how the archaic Buddhism of T5 developed into T6, and then into later standardised Buddhism with a sectarian affiliation, because the original text of T6 is a revised version of T5, and T5 and T6 are probably affili-

the public, a certain degree of overlap with my previous work is unavoidable in this paper, in order to present my basic method in as complete a manner as possible. 4 I use the expression "Zhi Qian's translation circle" etc. in order to designate those whose translation works exhibit virtually identical lexical and stylistic features to those of Zhi Qian. For example, judging from the lexical and stylistic features, it is difficult to distinguish Kang Senghui's J~IH!!!'~ translation work from that of Zhi Qian. Considering that there might have been other translators like Kang Senghui who were close in time and space to Zhi Qian, we should not attribute a translation to Zhi Qian merely because its composition style is virtually identical to his own.

The authorship ofT5 and T6 Mahiiparinirviir;.asatra


ated to Sarvastivada in close connection to MPS(S) and T1451, the Vinaya of the Mi1lasarvastivada. As the designated length of this article does not allow me to demonstrate all of the information that I found regarding. T5 and T6, I restrict myself in this paper to arguing that T5 is probably one of the earliest translations by Zhi Qian, and that T6 is a retranslation based on a revised original, maybe by Zhi Qian himself at a later date, but, more probably, produced by one of his successors during the Wu (~: 222-280) dynasty. The textual and doctrinal development of the MPS as reflected in T5 and T6 will be conveyed in a separate paper.

1. Previous research on Zhi Qian's work

Among modern scholars, Uis was the first to argue that T6 is probably a translation by Zhi Qian. Nattier summarises his argument as follows: "Ui's argument takes as its point of departure the testimony of the Chu sanzang jiji tB - ~g2.~ (see T2145, 55.6c15; a text by this title is also credited to Zhi Qian in his biography, 97clO-ll). Ui then adduces a number of citations from a two-fascicle Mahiiparinirvii1Ja Satra !!;ffi*~t:>Jb>g~i found in Sengyou's {~fti (445-518) Shijia pu ~}t[[~ (T2040), showing that - of the seven extant and non-extant texts entitled 'Mahiiparinirvii1Ja SiUra' registered in Sengyou's catalogue - this two-fascicle text can only correspond to the scripture attributed there to Zhi Qian. Finally, Ui demonstrates that the terminology used in these citations corresponds closely to what is found in the extant 'anonymous' text (i.e. T6) and not to the language of any other known version. On this basis, Ui concludes that T6 is in fact Zhi Qian's translation.''6 UF presents a list of 22 extant translations by Zhi Qian among his 36 translations listed by Sengyou ({~fti: 445-518) in the Chu sanzangjiji (T2145.55.6clO-7a24, hereafter CSZJJ): T6, T54, T68, T76, T87, T169, T185, T198, T225, T281, T362, T474, T493, T532,

Ui 1971: 519-523. Nattier 2004: 176, n. 34. Ui 1971 :530-532.


Jungnok Park

T556, T557, T559, T581, T632, T708, T735 and T790. In general, Ui follows the attributions by the cataloguers of T. Apart from T6, the other 21 texts among the 22 on Ui's list are identified as Zhi Qian's work by the T cataloguers. It remains unclear why he excludes from his list TlOll, which is attributed to Zhi Qian both by Sengyou and the T cataloguers. In the case of T533, which is also attributed to Zhi Qian by the T cataloguers, this may have been excluded from his list because the title Chamojie jing ~~ tr~~~ in CSZJJ is different from the title Pusa shengdi jing :g:ffii1:. :ttt~ in T. However, major Buddhist catalogues have traditionally identified them as the same text with different titles. It also appears that Ui failed to include T210 in his list, since it is attributed by the T cataloguers not to Zhi Qian, but to Wei Qinan ~fEfE\;B and others. However, as Nattier pointed out,S CSZJJ lists T210 as one of Zhi Qian's translations, and adds the information that Wei Qinan merely brought the text to China, while the actual translators were Zhu Jiangyan ~M~3R. and Zhi Qian (T2145.55.6c, 50a, 96a). Discussing the renderings that are peculiar to Zhi Qian, Nattier includes 26 extant texts in the "provisional list of the authentic works of Zhi Qian": T6, T54, T68, T76, T87, Tl69, Tl85, Tl98, T21O, T225, T281, T328, T361, T474 , T493, T532, T533, T556, T557, T559, T581, T632, T708, T735, T790 and Tl011.9 To the 22 texts on Ui's list, she adds T21O, T533 and Tl011 for the reason mentioned above. She also adds T328, which is attributed to Bai Yan (8g: fl. 254-259) by the T cataloguers. Pending further study of this text, she points out that, while two translations with the same title, Xulaijing ~&*j~, are mentioned in CSZJJ, Bai Yan's translation was not extant at the time of CSZJI Following the attribution by the T cataloguers, Ui attributes T362 Amituo sanye sanJo salou Jotan guodu rendao jing [5i1J~W8 - J1~ -{JI5WiifJ{JI5fl~JjtA.m:~gf to Zhi Qian; however, agreeing with Harrison (1998), Nattier points out that the attribution of T361 Wuliang qingjing pingdengjue jing ~~5~5:w~~ to Lokak~ema (3(~}lli[~: fl. 178-189) and that


Nattier 2003: 241, ll. 119. Nattier 2003: 208-209, 241-242.

The authorship of T5 and T6 Mahaparinirviir.zasiitra


of T362 to Zhi Qian have been mistakenly transposed. 10 She adds .four other texts (T20, T27, T507 and T511) to the list of "additional candidate texts." In order to verify the authorship of T5 and T6, I will temporarily make use of Nattier's list, including the "Gandidate texts," except for T6, which is in question, and T328, whose new attribution by Nattier requires further research. 2. Verification of the authorship of T5 and T6
In order to pinpoint the translation date and place of T5 and T6, adopting the method of Zurcher (1991), I will proceed in four steps. First I internally compare the writing style, renderings and transliterations of T5 and T6 and argue that the former precedes the latter. Secondly, comparing these features of T5 and T6 with those of Zhi Qian, I argue that T5 and T6 are compatible with other works of Zhi Qian or of a successor in his circle. Thirdly, comparing those features of T5 and T6 with other possible translators' work, I argue that it is unreasonable to attribute these texts to a translator whose translation work does not belong to Zhi Qian's translation circle. Finally, examining more minute details of the writing style of T5 and T6, I conclude that T5 is probably one of the earliest translations of Zhi Qian, and that T6 is a later retranslation, possibly by Zhi Qian himself, but more probably by a successor in his circle.

2.1 The precedence ofT5 to T6

As Nattier points out,l1 there is no doubt that either T5 or T6 is dependent on the other, or that one of these two is a retranslation of the other. As for "retranslation," we may categorise this into three types. The first would be a retranslation of the same original in order to correct mistranslations or revise awkward expressions; the second would be a retranslation based on a revised original within the same textual tradition; and me third would be a retranslation

10 For this attribution, refer to Harrison 1998: 556-557; Nattier 2003: 242, n. 121); Ono 1936: spec. vol., 33-34; and Hirakawa 1968: 76, 89. 11 Nattier 2003: 241, n. 118.


Jungnok Park

based on a new original which belongs to a different ~extual tradition. According to my examination of the textual development of T5 and T6 in comparison with the extent Pali and Sanskrit recensions as well as other corresponding Chinese translations of the MPS, the relationship between T5 and T6 seems to belong to the second type of retranslation. A comparison between the writing style of T5 and T6 quickly reveals that T5 cannot be the result of a retranslation of T6. For example, T5.l.l65b has: 12
{~E~ttli: "7("T~m~1!!Ll A~~3:.YE ' T>J<:JJrii!:ill:tf ' ~~. o5Z &1'~JlUlilt, 1f~gz~' A"i'El}~,~~M1~~' 7("T~3:.TYE:tf fZ
0 0

*~!Eg)t: '3:.:tfI'~YE' YE:tf~3:. "i'El~g ,~{7fc,~J~ '~~BmLlrf6jM :t~ , 7(J:~7(ij]\YE ' 1'F.:tfij]\YE ' ~~jU~T~lf3:. ' ~3:.TYE:tf ~
0 0 0

tI{~tPfi=R~~t5Jb5Q! "

Whereas T6.l.181a has:

{~Effittli: "ii!:Fs'~m, ~~2f1!!Ll ' l'~lilt!N ' ~m1:E:tf )G\~PJT1'T ' @~Elft,X m,~-EI~' AillE1~:7,. 7(Jtt~~5fi' f6j~M:t~' 55FfAt?B']ffiWZ -ftff: ?3:.YE~=a=, PJ~E~ {~fi=R ' ~~~5Jb5Q ~m~~! "

In the above, the classical Chinese composition of T6 is in an elegant style, strictly following the rule of four syllables per phrase. In comparison, the composition of T5 appears to be relatively coarse and archaic. It appears absurd that anyone would have revised the composition of T6 into that of T5. An overall examination of the writing styles of T5 and T6 reveals that, while T6 generally follows the elegant style of classical wenyan x:~, i.e. a regular prosodic pattern of four or six syllables, that of T5 is comparatively archaic and coarse, so we should regard T6 as the result of the retranslation ofT5. The style of the verses (giithti) confirms that T5 predates T6. When there is more than one translation of the same scripture, the later ones tend to contain more verses. T5 contains only 14 verses in a single location, whereas T6 has 48 verses in 17 locations. Furthermore, the form of the verses in T5 shows that it could not
12 I leave these passages untranslated; what is in question is not their meaning but their prosodic style.

The authorship of T5 and T6 Mahiiparinirvii1Jasutra


be a result of a retranslation of T6. The number of syllables in each line of the quatrains in T5.1.174a9-b7 show the following irregu1arities: 5/4/5/5, 5/5/5/5, 5/5/5/5, 5/5/5/5, 5/5/4/6, 4/6/5/5, 5/5/5/5, 5/5/5/5, 4/5/4/5, 4/5/5/5, 4/6/5/5, 5/4/4/6, 5/5/5/5, 4/6/5/6 .. On the contrary, all of the lines of the corresponding verses in T6 (1.189c190a) regularly consist of five syllables. T5 corresponds to K653 (19.182a-204c) of the Tripitaka Koreana ~..m*~~! (hereafter K) and Q670 (18.544a-58b) of the Qisha dazangjing {I1iy*~~! (hereafter Q). T5 preserves the irregu1ar verses, following K653 (19.202a21-b16); whereas, Q670 (18.556c13-57a1) presents regular verses of five syllables per line, which is passed on to the Yuan and Ming Tripitakas (see T5.1.174 notes 4-20). In order to understand this discrepancy in the transmission of the verses, we have to understand the history of woodblock Tripitakas in East Asia and the Buddhist method of presenting verses in Chinese prints. The first woodblock Tripitaka, widely known as the Kaibao dazangjing I*J.*~~, was engraved during 973-983 by order and with the support of Emperor Taizong the founder of the Song dynasty.13 Based on the Kaibao dazangjing and collecting more texts, the Khitans of the Liao dynasty (~: 916-1125) pro-: duced their own blockprint-Tripitaka, probably before 1063, and the Koreans of the Goryeo (Koryo) dynasty produced their first one during 1011-1087. The Koreans edited their second edition during 1237-1251, which is the present K.14 After the first engraving of the Kaibao dazangjing, which was burnt around 1120 during the invasion of the Tungut Jurchens, there were a series of private (sometimes produced with partial governmental support) blockprint-editions during the Northern (960-1126) and Southern Song dynasties (1127-1279); the private engraving of the present Q began during the reign of Emperor Lizong (1225-1264) of the Southern Song and was completed in 1349 under the Yuan dynasty. While the present T is based on K, the Tripitakas of China,


13 For details of the Kaibao dazangjing, refer to Tong, 1991: 1-16.

14 For details of the two Korean Tripitakas, refer to K: vol. 48: 1-17, and Buswe112004: 129-138.


Jungnok Park

including the governmental Tripitakas of the Yuan, Ming and Qing dynasties, are based on the private engravings of the Song dynasty. As for Indian verses, metres are measured by the number of syllables or by morae; in any case, the cadence decisive to the metre is determined by a particular pattern of heavy and light syllables. On the contrary, having no distinction between heavy and light syllables in classical Chinese, the Chinese had fixed forms of verse that were mainly determined by the number of syllables and the combination of rhyme, tone and antithesis. However, since it was impractical to follow all of these rules while translating the tens, hundreds or thousands of verses in a Buddhist scripture, Buddhist translators in China ignored all the other factors except the fixed number of syllables per line. From the "embryonic" translation period of An Shigao ~t!ti% (fl. 148-171) and Lokak~ema, they translated Indian verses into quatrains of lines of four, five, six or seven syllables. Besides, the Buddhist translators "visualised" the verses, as it were, by writing them in separate lines from the surrounding prose, forming rectangular horizontally running columns or blocks. Hence, while the Indians perceived verses by metrical reading, Chinese Buddhists could perceive them visually in the written text. In archaic translations, however, we find that some verses are not translated into quatrains of the fixed number of syllables, although such cases are very rare. For example, both K745 and Q763, corresponding to T101 (the one-fascicle version of the Saf!lyukttigama) preserve the irregular translations of the verses. There, while K745 indicates the existence of verses, merely locating them on separate lines, Q763 visualises them by arranging them in the shape of two rectangular columns, using varying space widths. Visualisation is very important in identifying verses: for example, if K793 (19.907c18-21) and Q812 (19.l92b29-c3), which correspond to T785 1~:ill:1~mj~H.tm., did not visualise the seven-line verses, the first line of which has one more syllable, there would be no way to detect that they were translations of verses: they do not comprise of eight lines, the number of syllables per line varies, and there is no combination of rhyme, tone or antithesis.

The authorship of T5 and T6 MahliparinirvlilJ.asutra


Considering the tradition of visualising verses in Chinese Buddhist prints, it appears virtually impossible for scribes or engravers to' ruin the regular verses as beautifully presented at Q670.l8.556c13-:57a1, by adding or removing a syllable here and there: the varying space widths for the purpose of maintaining the rectangular columns must have been too irritating to their sight to be ignored. Moreover, the proofreaders must have caught such manifest misprints, unless the original text preserves suchirregular verses. Therefore, we should conclude that, while K653 preserves the verses as archaic, as they were in its first translation, Q670 reflects a later, revised version. Therefore, I conclude that it is absurd to retranslate the verses of T6 into those of T5. As for renderings, we also find T5 containing more archaic renderings than T6. For example, the archaic rendering of pilJtjapiitika asJenwei 5t~at T5.l.l63a23 ff. is replaced by the later standard rendering of qishi.z~ in T6.1.178b14 ff.; the~rendering of briihmalJa as shixin WiJl,' at T5.1.169b27 by Janzhi ~;E; at T6.1.185alO; and the rendering of bhik~u as chujin Il*Hi at T5.l.l69c4 by biqiu t:tli at T6.1.185a3. In particular, the renderings jingxin ~*Jl,' ("mind of purification"), sixin ,'G',Jl,' ("the mind of thought") and zhixin ~Jl,' ("the mind of wisdom") at T5.l.l66all f. are rendered as jiexin f.!l(; Jl,': ("the mind of precepts"), dingxin IDl,' ("the mind of concentration") and zhixin ~Jl,' ("the mind of wisdom") at T6.l.l82a4 f. The rendering sixin ,'G',Jl,' in T5 is especially odd and archaic, compared to the dingxin )EJl,' of T6. In sum, a comparison between the writing style and renderings of T5 and those of T6 shows that the translation of T5 obviously precedes T6, considering that these two texts are in the relationship of retranslation in the second sense. Ui's (1971) arguments for attributing T6 to Zhi Qian may well be re-considered. I agree with his identification that T6 is the two-fascicle MPS that is quoted in Sengyou's T2040 ~~~. However, his identification of T6 with Zhi Qian's MPS in two fascicles is implausible. In his arguments, Ui makes use of the seven translations of Mahayana and "HInayana" MPS listed in CSZJJ


Jungnok Park

(T2145.55.14a5-1O); however, in this specific case, a~ he himself points out,15 we cannot trust Sengyou's testimony with reference to Zhi Qian's MPS. In CSZJJ, Sengyou designates four of the seven translations as extant in his time: Zhi Qian's Da bannihuan ling *m[5Jb5~! in 2 fascicles, Zhu Fahu's Fangdeng nihuan ling 15%fi. 5Jb5~! in 2 fascicles,16 and the two translations of the Mahayana Mahaparinirva;lJa by Dharmak~ema and Faxian. 17 The MPS quoted in T2040 cannot correspond to the translations of Zhu Fahu, Dharmak~ema or Faxian, which are Mahayana recensions. Furthermore, since Sengyou states that Zhi Qian's MPS is similar to that of Zhu Fahu 18 and that Zhi Qian's MPS is not from the Dfrghagama, as Daoan claims,19 the MPS quoted in T2040 cannot be Zhi Qian's, either. However, given the fact that the writing of T2040 began before the completion of T2l45,z we reach the contradictory conclusion that the two-fascicle MPS that he documents in T2040 was not extant at the time when he wrote the respective passage in T2145. As mentioned above, T6 corresponds to K654 and Q671. Following K654, T6 is entitled Bannihuan ling m[5Jb5~, whereas Q671 is entitled Fangdeng nihuan jing 15%fi.)Jb5~!, which is passed on the Yuan and Ming Tripi!akas (see T6.1.176, n. 1). If what Sengyou designated as Zhu Fahu's Fangdeng nihuan jing 15 %fi.5Jb5~ is in fact Q671 (=T6/K654) rather than T378 Fangdeng bannihuan ling 15%fi.m[5Jb)~, the above problems can be solved entirely: the MPS quoted by Sengyou in T2040 is what he called 15 Ui (1971): 519-520. 16 Both Kawano 1986, and Suzuki 1995 identify T378 jJ~B~5Jb5g~ as
Zhu Fahu's Fangdeng nihuan jing jJ~5Jb5g~g;. 17 The three translations that were not extant at his time are Faxian's Fangdeng nihuanjing jJ~5Jb5g~~, Zhimeng's '&1lh (fl. 424) Nihuanjing 5Jb5g~g; and GUI:tabhadra's Nihuan jing 5Jb5g~. Faxian's translation is identified as T7 by the T cataloguers, and the latter two translations are no longer extant. 18 A~~ {*B~5Jb5g} ,~t1J~5Jb5g} *fEJ (T2145.55.l4a8-9). 19 *B~5Jb5g~=;ff :!i:0~: "tl:Ji't1fiJ1f'"f:(i~4i't1fiJ~!MJlt~" (T2145.55.6.c15). 20 The contents of T2040 are quoted in CSZJJ (T2145.55.87b-88a)

The authorship of T5 and T6 MahiiparinirviiT}asiitra


Zhu Fahu's Fangdeng nihuan jing h~)Jb)g~, which is the present T6 Bannihuan jing ~)Jb)g~ (=K654/Q671); whereas, what he called Zhi Qlan's Da bannihuan jing *~~5Jb5g~ is T5 Fo bannihuan jing {1M~5Jb$g~ (=K653/Q670). In this case, we should interpret Sengyou's statement that Zhi Qian's MPS is not from the Dfrghagama, contrary to Daoan's testimony, not in the sense that Zhi Qian's MPS is a Mahayana recension; it would rather mean that his translation is quite different in content from the text in Tl, the Dfrghagama, as I will illustrate in a separate paper. However, the crucial fact remains uncertain: to decide if what Sengyou called Zhu Fahu's Fangdeng nihuan jing h~5Jb5g~ in CSZJJ is indeed T6/K654/Q671 will require further. philological research. 2.2 The compatibility ofT51T6 with the authorship ofZhi Qian's translation circle As the first generation of Chinese Buddhist translators, An Shigao ~ilt~ (fl. 148-171) and Lokak~ema laid the foundation for future Buddhist translation, such as the basic renderings, the system of transliteration and the structure of the translation teams. However, being full of unknown technical terms and exotic transliterations, written in a clumsy style, their translations are difficult to read, even when one is equipped with specific knowledge about Buddhism. Hence, in that embryonic period, when a basic knowledge of . Buddhism was hardly to be expected of the readers, the pressing mission for the next generation was to produce translations that enabled the readers to understand the texts by themselves, even without any specific knowledge. In this respect, Zhi Qian's work is conspicuous: his translations were probably far more readable by Chinese intellectuals than the works of his predecessors. Furthermore, his classical Chinese composition reflects a sense of literary style; although to be located in the middle of the archaic translation period, many of his works are


. Jungnok Park

as polished as the works of the following old tran~lation period. 21 The composition of T5 and T6 demonstrates this characteristic of Zhi Qian's work: the relatively polished classical Chinese in T5 and T6 does not hamper readers from understanding its meaning because of awkward sentence structures, exotic writing style, etc., as much as in the works of his predecesHowever, such improved linguistic standard as refined composition, appropriate renderings and an adequate proper choice between Chinese names and transliterations were still insufficient to produce a readable translation per se, since the very content to be conveyed by the linguistic rendering was extremely exotic to the Chinese at that time. To solve this problem, Zhi Qian utilised or interpolated Chinese concepts in his translations, sometimes to the extent that it distorted the original content of the Indian texts. For example, in T76 *BrahmiiyuJ:tsutra Y!tJ*5tu~!lf, when Brahmayus seeks refuge in the Buddha, Zhi Qian introduces the five precepts in the following Chinese style: I want to be a Buddhist follower. Maintaining benevolence, I will not kill living beings; being content, I will not commit theft; being chaste, I will not have inappropriate sexual relationships; being faithful, I will not lie; being pious to parents, I will not drink intoxicants. 22 Here, above all, the justification of sobriety in terms of filial piety must be a Chinese interpolation; such a justification is not found within the context of Indian Buddhism. Zhi Qian's interpolations in his translation work, on the one hand, helped the Chinese readers to understand exotic Buddhist ideas in the familiar terms of ,Chinese thought, and, on the other hand, transformed bizarre Indian values, so that Buddhism became compatible with the pre-

21 I follow Ono's distinction of three periods in the history of Chinese Buddhist translation: that of archaic translation (before 375), that of old translation (376-617) and that of new translation (after 618): Ono, 1936: spec. vol., 7-9. 22 Jjt~~m' ~e1'iil~ , ~)E1'~ , ~~1'~ , ~{1'W\' ~;f:1'M (T76.1.886a.) In the Pali Brahmliyu-sutta (M.II.145) and the Chinese Fanmajing of the Madhyamligama (T26.1.689b), corresponding to T76, only the threefold taking of refuge is mentioned, without reference to the five precepts.

The authorship of T5 and T6 MahiiparinirviilJasutra


established Chinese moral principles, particularly filial piety and loyalty. 23 T5 and T6 show a similar tendency towards interpolations reflecting Chinese thought and values. For example, while the Sanskrit Mahiiparinirvii~1a-satra (hereafter MPS(S and the Pali Mahiiparinibbiina-sutta (hereafter MPS(P24 mention the Vrji people's refusal to have sexual relationships by force as one of the seven reasons why the republic of Vrji is invulnerable to the attack of other countries,25 T5 and T6 introduce this virtue as follows:
Have you heard that the Vrji people are cultivated and polite, that there is distinction between the sexes, and that the elder and the younger people [look after and] serve one another?26

Indeed, the "distinction between the sexes" ~:R~JjU, a famous Confucian slogan, implies far more than merely not having sexual relationships by force. Among educated people, i.e. the expected readers of T5 and T6, men and women (if we accept the tradition literally, when over seven years old) were not supposed to be in the same space, much less mingling together, unless they were parents

23 Not to mention leaving one's own family by entering monkhood, for which the Confucians have blamed Buddhism throughout Chinese history, even the Indian Buddhist idea of donation was accused of ruining the value of the Chinese family system. The Mouzi lihuo lun $T:El!!~~Rij (T2102.52.la-7a) discusses this topic. 24 In this paper, references to MPS(S) include MPS_ST.I and MPS_ ST.H., and references to MPS(P) include the Mahasudassana-sutta (D.II.169-99) 25 MPS(P) (D.II.74): kin ti te Ananda sutarrz: Vajjf ya ta kulitthiyo kulakumiiriyo ta na okkassa pasayha vilsentf? ti. Mostly reconstructed from the Tibetan *Malasarvilstivilda-vinaya-k~udraka-vastu, MPS(S) ( 1.26) has: (kirrz nu tvayananda srutarrz yils til vrjfnilJ!l vrjiprajapatyo vrjikumilrikas ca pitrrak~itil miltrrak~itil bhratrrak~itil bhaginfrak~itil/:l svasurarak~ita svasrurak~itil jfiiltirak~itil gotrarak~itil/:l saparidaJ:uj.il/:l sasvilmikil/:l kan)ya/:l paraparigr(hfta antaso millilgUlJaparik~iptil api tadrupasu) na sa(hasil cilritram ilpadyante).
25 ~ftM!: ~t~, tI1c~i~~>~fr:1f53U' -tH;iJif~/f"? (T5.1.160c); ~t~, tI1c~i:(;)9:' ~fr:1f53U '-&~t:Ff~? (T6.1.176b)



Jungnok Park

and children or husband and wife. The translation in question is an interpolation, in that it conveys extra meanings that are absent from the original context. In this way, with regard to the tendency towards Chinese interpolations, T5 and T6 are in accordance with the translation work of Zhi QianY A comparison between the renderings in TS/T6 and those of Zhi Qian quickly reveals that T5 and T6 are indeed compatible with Zhi Qian's authorship. For example, the renderings of the four achievements of Buddhist practitioners, i.e. srota-iipanna, sakrdiigiimin, aniigamin and arhat, are respectively translated as gougang 5~;:g ("stream"), pinlai ~* ("visiting in repetition"), buhuan ::f~ ("not returning") and yingzhen H!Jt ("truly worthy"), found both at T5.1.164a and T6.1.179c. For four out of the five aggregates reconstructed from the list of twelvefold dependent origination, both T5.1.163b and T6.1.178c use se 5, tong ;Iffi, xing f=r and shi ~ respectively for riipa, vedanii, sarrzskiira and vijiiiina. For the seven treasures of the wheel-turning king, i.e. cakraratna, hastiratna, asvaratna, malJiratna, strfratna, grhapatiratna and parilJiiyakaratna, TS.1.170a has huangjinfeilun ;;::ili::m;,~, shen/ibaixiang t$JJS*, ganseshenma ~f:f5t$)~, mingyuezhu fYjJjf3K, tianyunuqi X33.:tz::W, zhubaoshengchen ._~~ and dianbingshengchen :!I=!i:~~~, and T6.1.185c has jinlunbao :ili:~_, baixiangbao S*_, ganmabao ~f:f,~_, yunubao 33.:tz::W, shenzhubao t$~_, lijiabao and xianjiangbao Ji#~_. These peculiar renderings of T5 and T6 are in accordance with the renderings in Z11i Qian's other works. 28


27 For a relevant example unique to T5, we find an interpretation of equanimity (upek:;ii), one of the four immeasurable states of mind, saying " ... [The king] contemplated the great practice of filial piety in order to liberate [those who were his] parents during [past] incalculable eons. [Thus, he] observed his five body organs and the nine body opening [filled] with discharges": ... ,'is\*~fj, tlXJJrmIJtWZ~~ , 11~Hl1ijji1L fL~Jit (T5.1.171a) 28 Between T5 and T6, only the latter contains the renderings of the ten epithets of the Buddha (1.187b) and the twelve divisions of scriptures (188a). For the peculiarity of these two sets of renderings by Zhi Qian, see Nattier 2003 and 2004.

The authorship of T5 and T6 MahliparinirvG]Jasiitra


In sum, based on the writing style, the interpolations reflecting Chinese thought and the peculiar renderings of T5 and T6, we hardly find evidence to reject the attribution of either text to Zhi Qian's translation circle.

2.3 The incompatibility ofT51T6 with the work of other translators

As mentioned above, T5 and T6 are complete translations of the longest s~ltra from the early period of Buddhist literature. A comparison of several of the peculiar renderings in T5 and T6 reveals that T5 and T6 may be attributed to Zhi Qian, but at least were probably not translated outside Zhi Qian's translation circle. For example, as Nattier pointed out,29 we find in T5.U67a and T6.1.182b the list of 28 Buddhist heavens that are unique to Zhi Qian's translation circle; among Zhi Qian's other works, we often find the same lists, with trivial differences, in T198.4.185b, T225.8.485a, 487a and T28U0.447ab. As another example of such parallels, the collocation of the renderings yintai fr~f'iJ ("lustful state (of mind)"), nutai r.9f'i",~ ("angry state (of mind)"), and chitai .f'i",~ ("ignorant state (of mind)") is only found in T5 (U63b ff.), T6 (U77a) and furthermore in Zhi Qian's T54 (1.848b ff.) in the huge amount of translations of canonical texts in the Chinese Tripi!aka. These few renderings that are particular to Zhi Qian's translationwork strongly support the view that T5 and T6 are probably works by Zhi Qian's translation circle. However, this is not yet sufficient to attribute these two texts to Zhi Qian's translation circle. To make this more certain, we may as well verify that the lexical and stylistic features of T5 and T6 are incompatible with the authorship of any other translator outside of Zhi Qian's translation circle. In order to achieve this certainty, I adopt the following approach: 1) I restrict the examination of the compatibility of T5 and T6 with other translators' works to the period of archaic translations (-375). The renderings used in T5 and T6 imply that

Nattier 2003: 241,




Jungnok Park

these texts were hardly translated during the period of old or ' new translations. 2) I do not consider the first two translators in China of whom we know, i.e. An Shigao and Lokalqema, as possible authors of T5 and T6. The writing style and renderings of T5 and T6 are distinct from these translators' work. However, T5 contains some archaic renderings that are common to Later Han translations (unti1219); therefore, I also consider other Later Han translators than the mentioned first two as candidate authors for T5. 3) I consider only those texts that are attributed to specific translators by the CSZJJ. This restriction is imposed in order to prevent any incorrect conclusions being drawn from renderings or composition styles that later cataloguers may have arbitrarily added. 3D I therefore am only concerned with the extant texts cited in the CSZJJ. Other candidate translators who may fulfill the above conditions are: An Xuan tJ;i:1:; (c. 181) and Yan Fodiao ~{iJ#~; Zhi Yao 5tHI (c. 185); Kang Mengxiang ~jfu$ (fl. 194-199); Kang Senghui ~{~ 1t (?-280); Bai Yan a1Jf (fl. 254-259); Zhu Fahu ~5*~i (fl. 266308); Nie Chengyuan l.\jfc}j; Wuchaluo ~JZ.H (fl. 291) and Zhu Shulan ~ zt-,'iZiVl'; Faju 5*)tE (fl. 308) and Fali 5*fL. From An Xuan to Kang Mengxiang, I will examine whether these are compatible with an authorship of T5 alone, and, for the remaining translators, if they are possible candidates for both T5 and T6. An Xuan t:Jl:-p: and Yan Fodiao Ii~~ CSZJJ (T2145.55.6c3-4) attributes the Fajing jing 5*~~J[ and the Shihui +~ to An Xuan and Yan Fodiao, among which the former is identified with T322 by the T cataloguers and the latter is not ex-

30 For the arbitrary attributions of translatorship according to Fei Changfang's JIt~m catalogue T2034 Lidai sanbao ji mHt=W~c (hereafter LDSBJ) from 597, see Tokuno, 1992 [1990]: 33-35.

The authorship of 1'5 and T6 Mahaparinirva1J.asiltra


tanto As Ztircher31 pointed out, the most conspicuous characteristic of their translation is that "virtually all proper names and technical terms have been translated." Due to this characteristic, they are excluded from the list of possible authors ofT5. For exampl~, while T322 has only the semantic rendering of chujin ~i (11 times) for bhik~u, T5 has both the semantic rendering chujin (3 times) and the transliteration biqiu teE:. (239 times), (while T6 has only the transliteration biqiu (91 times)). The frequent occurrence of trans- . literations in T5 leaves little possibility that An Xuan or others are the translators of T5. Zhi Yao :stili CSZJJ (T2145.55.6c1-2) attributes the Chengju guangming jing RlG ffil:.*EJ..E3~ to Zhi Yao, and this text is identified with T630 by the T cataloguers. The classical Chinese composition style of T630 is as refined as that of T6, and appears superior to T5, transliterations are rare. For example, while T5 uses the transliteration pusa :gH for bodhisattva, T630 uses mingshi EJ..E3, which never occurs in T5 or in Zhi Qian's other works. This clearly weakens the possibility that Zhi Yao is the author ofT5. We hardly find a reason to attribute T5 to Zhi Yao, given the close similarity between T5 and other works attributed to Zhi Qian's translation circle. Kang Mengxiang mii~ CSZJJ (T2145.55.6c7-9) attributes the Zhong benqi jing tp*m~ to Kang Mengxiang, a text which is identified with T196. Again, the classical Chinese composition style of Kang Mengxiang in T196 is compatible with the authorship of T5; However, one conspicuous characteristic of T196 is that we seldom find semantic renderings of proper names. This, again, weakens the possibility of his being the author of T5, where we often find semantic renderings of proper names. Furthermore, some transliterations of T196 fail to coincide with thOse of T5: T5 transliterates Magadha and Vrji respectively as mojie )!J.ftflJ and yuezhi ~t~, whereas T196 uses mojieti )!J.ftflJ:f'!

ZUrcher 1991: 283.


Jungnok Park

and baqi W~. In addition, we find a discrepancy between the renderings in T5 and T196. For example, while T5 usually translates briihmalJa as shixin W[1l,\, T196 always translates it as Janzhi ~~. To sum up, it also seems to be unrealistic to attribute T5 to Kang Mengxiang, while at the same time ignoring its greater similarity with the works of Zhi Qian and Kang Senghui. Kang Senghui (.ffh~: 1-280) CSZJJ attributes the *$atpiiramitiisamiisa-sutra, Liudu ji jing 7\& ~~J. and the Wupin :!:R:~i::l to Kang Senghui; the former is identified with T152 by the T cataloguers and the latter is not extant. As for the style of composition, the prose of T152 is as elegant as T6, and appears superior to T5; both T152 and T6 follow the regular style of classical wenyan )2::3 more than T5. The forms of the verses in T152 are also similar to those found in T6, being quatrains of four, five, six or seven syllables; T152 has no quatrain and instead containing a mixed number of syllables as found in T5. T152 also shows the same tendency to use terms which reflect Chinese values as found in T5 and T6: all three texts are full of basic Confucian terms, such as ren f= ("benevolence"), yi ~ ("righteousness"), xiao 2f: ("filial piety"), etc. In addition, the three texts are interpolated with the Chinese term hunshen 01;:):$ ("spirit") which denotes a permanent agent that goes through sa"f!lsiira (T5.1.162a15, T6.1.177a26 ff.; T152.3.35c6, 48c26). Besides, in T152, we also find the Chinese style justification of the five precepts, which is very similar to the aforementioned statement in T76 *BrahmiiyuJ:t-sutra (1.886a): Firstly, maintaining benevolence, I will not kill living beings but do them a favour. Secondly, being humble, I will not commit theft but abandon my things to aid people. Thirdly, being chaste, I will not have inappropriate sexual relationships but keep to celibacy. Fourthly, being truthful, I will not tell a lie but speak sincerely. Fifthly, being filially pious, I will not get drunk but behave faultlesslyY
32 -~, ~C~*& ' }~t&M~; =~ )~~~~ mal%fJR ; =~ , J;l[~~ fr~ , :::f)8~~ ; 1Z1l:tf ' ~~{~~!tj: , 1~~fp ; E:tf ' *#:::f/j~ , 1T~)~)'


The authorship of T5 and T6 Mahiiparinirviif/asiitra


We also find that the renderings of T152 are compatible with both T5 and T6.The writer of T152 renders the correspondent Pali cakkavatti dhammiko dhammaraja as feixinghuangdi ~f'J~w (3.1c ff.) as T5(1.169b3 ff.), instead ofT6's zhuanlun (sheng)wang *Rll(~).x. (1.185a7 ff.); sotapanna and sakadagamin as gougang )~ ~ and pinlai (3.2b ff.) as in T5 (1.163b4 ff.) and T6 (1.178b25 ff.). The five aggregates are given as se -ES, tong 1i, xiang f~, xing f'J and shi ~ (3.43b), as we can reconstruct four of them from the list of twelvefold dependent origination (T5.1.163b, T6.1.178c). For the renderings of the seven treasures, T152 presents three slightly differing lists (3.21c, 48c, 52a); they are, however, similar to those of T5.1.170a and T6.1.185c. In sum, an examination of the writing style and the use of renderings in T152 reveals that the authorship of T5 and T6 is compatible with that of T152. Furthermore, we hardly find any substantial difference in the writing style and the use of semantic renderings or transliterations between T152 and Zhi Qian's translations; only historical records, such as CSJZZ, inform us that this is Kang's translation rather than Zhi Qian's.



Bai Yan

SJJ (fl. 254-259)

The T cataloguers attribute T328 Xulai jing ~~*Jl~ to Bai Yan, which is also one of the three translations attributed to him by CSZJJ (T4215.55.7b). His brief biography in CSZJJ indicates that he was not a creative but a revising translator (T2145.55.96a) [specify?]; his Xulai jing may also be a retranslation of Zhi Qian's previous work, probably in the first category of the three kinds of retranslation. However, as mentioned above, Nattier rejects the traditional attribution of T328 to Bai Yan and attributes it to Zhi Qian instead. Indeed, the renderings of T328 are in accordance with Zhi Qian's other works. However, in terms of writing style, we find a few factors that weaken Nattier's attribution. Most of all, in T328, we cannot detect interpolations reflecting Chinese values, which are so conspicuous in T5 / T6 and Zhi Qian's other works. We cannot find a single occurrence of yi ~ ("righteousness") or xiao ~ ("filial piety") in T328. In the case of ren 1=


Jungnok Park

("benevolence"), it is used only as a rendering of karuTlii in place of bei rJ, or as an honorific, meaning "gentleman." On the contrary, advocating the concept of impermanence in accord with the Indian Buddhist context, the translator directly attacks the Daoist value of fostering life, deriding the Daoist express zibaobusi f* /f":9E ("[aiming at] the art of self-protection and immortality"), and also directly attacks the Confucian concept of filial piety, using the expression ~=;~L~~ ("I do not hold hair and bodily hair dear").33 This attitude towards Confucianism and Daoism is odd in a translation work of Zhi Qian's. It appears more reasonable to regard T328 as Bai Yan's retranslation of the original by Zhi Qian. Judging from its renderings, the writer of T328 is compatible with the authorship of T5 and T6. Many renderings peculiar to Zhi Qian are found in T5/T6 and T328. However, the reflected attitude towards Chinese values which is indirectly expressed in T5 and T6 seems to rather question his translatorship of T328. The probability that the translator of T328 also wrote T5 and T6 is far lower than the probability of Zhi Qian and Kang Senghui having translated it. Hence I omit his name from the list of possible candidates for the authorship of T5 and T6.

Zhu Fahu ~5*~ (239?-316?)34

CSZJJ attributes 154 texts to Zhu Fahu ~5t:~. Among the extant texts in T, 95 translations are attributed to him. Having researched 33 To understand what the quoted passage means to the Chinese, consider the following attack by a hypothetical Confucian in the Mouzi lihuo lun: "The Xiao jing (~!: Scripture on Filial Piety) announces, 'Any part of the body, even the hair and the skin, should not be damaged, since it is given by the parents.' And, at his final moments, Zengzi ~r [summoned his sons and] said, 'Uncover my hands, uncover my feet. [Is there any part damaged?]' Now, Buddhist monks shave their heads. How could they violate the saints' words and not commit themselves to the duty of filial piety?" ~!~: "~~~Ji" ~LXm, ::fruzfE9:f~ "~r~5~: "~T~' ~TJE "4-5j>F~wUlm' f6J;\'t:3t~.AL~' ::f~rLmfu? (T2102.52.2c) 34 Zhu Fahu's biography in CSZJJ (T2145.55.97c-98b) records that he died on the road while trying to escape a revolt, because of which
0 0

The authorship of T5 and T6 Mahtiparinirvii1J.asiitra


the prologues, colophons and Buddhist catalogues, Suzuki (1995) declared 40 texts to be definite translations by Zhu Fahu, and classified them into five types according to their similarities in terms of renderings and.transliterations. In order to inspect whether Zhu Fahu could be a possible candidate for the translatorship of T5 or T6, I restrict my examination to the 40 texts researched by Suzuki. Her list and classification of the definite translations of Zhu Fahu are as follows: Type A (T222, T588, T636), Type p; (Tl86, T263, T266, T285, T291, T292, T310 (fascicles 8-14), T310 (fascicles 117-118), T345, T398, T403, T460, T461, T565, T606, T627, T817), Type B (T585), Type B' (T338) and Type C (Tl03, T170, Tl82ab, Tl99, T283, T315ab, T317, T342, T349, T378, T399, T425, T435, T459, T481, T589, T598, T737). Among the five types, we find that Zhu Fahu copies many renderings of Zhi Qian in works of Types A', B, B' and C. However, we find that some renderings in T5/T6 fail to accord with any of the forty texts by Zhu Fahu given above. As for the four achievements of the Buddhist practitioners, neither of the renderings gougang ).!If ~ for Pali sotapanna, or pinlai ~* for sakadagamin of T5 and T6 appear in Zhu Fahu's works.35 There are several occurrences of yingzhen ~J!f: (T263, T266, T398, T403 and T481) for arhat; however these occur independently, never in the sequence of the four achievements. For these four achievements, Zhu Fahu uses the transliterations xutuohuan ~&Wfm[, situohan wrW8"'6-, anahan IniJ}j~~ and aluohan roIm5l (T222.8.150b, T460.14.449c, T481.14629b and
Emperor Hui}l* moved his residence to Chang'an -&$.:-. Since this happened in 304 and since he is said to have been 78 when he died (according to the Chinese calculation), he may have been born in 227. However, listing Zhu Fahu's translations, Sengyou himself annotates that Zhu Fahu had been engaged in translation from the era Taishi *fr.1 (265-274) to the second year of the era Yongjia 7kB (308) (T2145.55.9bc). The colophon to T588 Xuzhen tianzi jing ~1t*T~~ (T2145.55.48b) records that the text was translated in 266; and the colophon to T186 Puyao jing ff8l~ (ibid.) records that it was translated by Zhu Fa:hu in 308. Therefore, the record of his death year in his biography appears to be incorrect. 35 Pinlai appears once, at T425.l4.43c. However, there it is not a technical rendering for sakada.ga.min.


Jungnok Park

T588.15.107a), or the renderings daoji mfi}Jji (or wanglai 11 bU/1Uan ::f'~ and wuzhu ~~ (T263.9.118a, T266.9.206a-21Oc, T342.12.146b ff., T345.12.164c, T398.13.419c ff., T403.l3.592c ff., T585.15.9b and 598.15.143c). Besides, while both T5 and T6 use lijia fJ.* to render grhapati, none of the forty texts of Zhu Fahu uses lijia; Zhu Fahu normally uses changzhe -6t1l for this. This apparent inconsistency between the renderings of T5 and T6 and those of Zhu Fahu reveals that Zhu Fahu could not be the writer of T5 or T6. Nie Chengyuan U:m Nie Chengyuan is a Buddhist layman who participated as a scribe (bishou ~)1t) in Zhu Fahu's translation teams for T222, Guangzan jing :Jf[;~~, T263, Zhengfahua jing IE5t~~, T285, lianbei yi. h'd .. ;';!;r;J4!i_+llj!;O/.ii\;/.( T398 D qle Z I e Jlng 1'1'1111'11 ~J s j,c;;h, ,am... Jlng --l--~b( T585 , /,-~h, Chixin fantian suo wen jing t:~H/'l'!t:7CpJTFp~~, T588, Xuzhen tianzi jing ~&Jt:7CT~! and the lost translation of the Suraf!lgama-sutra, Shoulengyan jing -gr~~~!. CSZJJ attributes the Chaoriming jing ~B~~! to him and notes that it is a revised translation, probably of the first category of the three types, of Zhu Fahu's original version. T638 is identified with Nie's Chaoriming jing by the standard catalogues of the T. T638 copies the renderings of Zhu Fahu: for example, for the renderings of the four achievements of the practitioners, it follows Zhu Fahu's renderings: daoji mfi}Jji, wanglai 11*, buhuan ::f'~ and wuzhu ~~ (15.535a, 536b). As for the rendering of grhapati, he also uses changzhe -6t1l instead of lijia f!.*. Therefore, given that he is the writer of T638, Nie Chengyuan should be excluded from the list of possible writers of T5 and T6. Wuchaluo ~x.~ and Zhu Shulan ~lIi The Chinese monk and first documented traveller to the Western Regions Zhu Shixing left China around 260, received the Fangguang jing 1J9::Jf[;~i. (Paiicavif!lsati-siihasrikii-prajfiiipiiramitiisutra) in Khotan and sent it back to China. It arrived in Louyang 5~~ in 282 and was translated in the Suinansi 7.Km~in 291. The Khotanese monk, Wuchaluo ~x.~, recited the original text




The authorship of T5 and T6 MahiiparinirviilJasutra


(zhi huben J\t!if1:zt;::), and Zhu Shulan~*jZlfl!, a Chinese layman of Indian origin, translated / interpreted it (kouchuan pi'). The text was revised again by Zhu Shulan and Zhu Faji ~;t;W in 303-304. Considering the typical procedure of Chinese Buddhist translation, we may regard Zhu Shulan as the substantial translator of the text. T221 Fangguang banruo jing ]])l:71:mt:;:fi~~ is identified by the T cataloguers as the Fangguang jing attributed to Zhu Shixing and others in CSZJJ. An examination of the renderings in T221 quickly reveals that Zhu Shulan and others from his circle are no candidates for the translatorship of T5 and T6. Instead of the gougang ;.~, pinlai g;~*, buhuan :f"~ and yingzhen ff!Jj;. of T5 and T6, T221 uses the transliterations xutuohuan ~J'lWB;, situohan wr~t-a-, anahan ~"iiJ tI~-a- and aluohan ~~;~ for the names of the four achievements of practitioners. Furthermore, T221 is full of transliterations, which do not appear in T5 or T6, as for example, aweisanfo ~"iiJ'rt -i~: abhisaf!lbuddha, anouduoluosanyesanpu ~*m7~-=If~ -:g: anuttarasamyaksambodhi, boluomi ;~~~: piiramitii, and ouhejusheluo 5~~J~~~: upiiyakausalya are not in accordance with the respective vocabulary of T5 or T6. Due to the inconsistency of renderings between T5 / T6 and T221, I exclude Zhu Shulan and hIs circle from the list of possible candidates for the translatorship ofT5 or T6.

Faju (5!j:g: fl. 308) and Fali 5!iL The colophon to T186, Puyao jing tf81~~, states that Faju was one of the scribes (bishou ~'t) in Zhu Fahu's translation team in 308 (T2145.55.48bc). CSZJJ attributes four texts, i.e. the Loutan jing ;fI~~, the Faju benmo jing ;t1:D:zt;::*~, the Futian jing 1~B3~I[, and the Dafangdeng rulaizang jing *:1J:WfrD*~~I[, to Faju,36 and notes that Fali ;tfL is the co-translator of the second and the third of these. The first three are identified as T23 *;fI~, T211 5t1:D ":'otu~ and T683 ~1}i;H~B3~~ by the T cataloguers.

36 It is striking that the number of Faju's translations increases to 132 in LDSBJ.


Jungnok Park

As for the writing style, I could find no decisive difference between T5 / T6 and Faju's three extant translatibns. The prose and verse styles of Faju's translations are similar to those of T6 and superior to those of T5. As for the terminology reflecting Chinese values, we find the Chinese-style emphasis on filial piety intermittently throughout T23 and T211. However, we find some conspicuous differences between the renderings of T5/T6 and those in Faju's works. Instead of the Chinese renderings of gougang )~~, pinlai ~*, buhuan ::f"~ and yingzhen B!E for the four achievements of practitioners in T5 and T6, Faju's works have the transliterations xutuohuan ~J[~t)g (T211.4.575c ff.), situohan WT~t-2i (T211.4.581a) and aluohan 1lc1~~ (T23.1.290b ff., T211.4 ff.). Also, for the rendering of grhapati, all three texts of Faju use changzhe -&1 instead of lijia. The writing style of Faju's works is in accordance with T5 and T6, but there are several differences between the renderings of Faju's works and those of T5 and T6. As a result, compared to Zhi Qian and Kang Senghui, it is not very likely that Faju is the writer ofT5 and T6. Thus, a comparison of the writing style and renderings reveals that it is unreasonable to attribute T5 or T6 to anyone else except Zhi Qian or one of his successors in his translation circle. 3. The translation date and place of T5 and T6 About the end of the Han emperor Xian's reign (190-219), there was nationwide upheaval, and Zhi Qian moved (probably from Henan )aJli province)37 to the territory of Wu ~ (222-280) in southern China. Sun Quan m~m: 182-252), the lord of Wu, invited him to the capital, Jianye 9l~, and supported his translation work. Early records (T2145.55.49a28, T2059.50.325bl) reveal that Zhi Qian was engaged in Buddhist translation from 222 during


37 Daoan's prologue to the Liaobenshengsi jing 7*1:~~'![ designates Zhi Qian as HenanZhi Gongming foJMJZ:~~ (T2145.5545b21). Henan is the province in which Luoyang, the capital of the Han dynasty, was located.

The authorship of T5 and T6 MahiiparinirviilJasatra


the Jianxing ltJI! era (252-53). Kang Senghui was born in Jiaozhi x.d: (now in northern Vietnam), Wu :SR:. He became a monk at the age of ten, after losing his parents, and moved to Jianye in 247. With the support of the Wu dynasty, he produced several translations and commentaries, dying in 280 (T2145.55.96b-97a). The Wu dynasty was overthrown in 280 by the Jin if dynasty (265-419), which resulted in the reunification of China. After the collapse of the Wu dynasty, the centre for Buddhist translation activities moved to Chang'an -&~ and Louyang ~~~, the two major cities in northern China. Considering these historical factors and the above examined philological information, we should locate the translation of T5 and T6 in southern China around Jianye It*, during the Wu dynasty. From the internal evidence of T5 and T6, we can trace more detailed translation dates: while T5 is probably one of the earliest translations by Zhi Qian, T6 is a work by a successor in his circle rather than by Zhi Qian himself, and I will bring further evidence for this conclusion. First of all, the archaic form of the verses in T5 indicates that we should locate the translation of T5 prior to that of the Dharmapada translation T21O, Faju jing $1o]~Ilf, which is composed entirely in verse. T210 was translated by Zhi Qian and Zhu Jiangyan ~X~3t around 224, having been revised and enlarged shortly thereafter. 38 As mentioned above, the number of syllables in each line of the verses varies in T5.1.174ab. On the contrary, T210 has only regular quatrains of four, five or six syllables; a quatrain never occurs with a mixed number of syllables. It is improbable that Zhi Qian produced such a clumsy composition, after having composed a large number of quatrains in regular form. Therefore, the composition of T5 should be located prior to the translation of T21O, i.e. around 224. Considering that Zhi Qian started his translation in c. 222, this explains why we find more archaic renderings in T5 than in T6. .


38 For the translation date ofT21O, see T2145.55.50a2-28, in particular lines 9-10. It tells that the original text arrived at Wuchang :lEt~ in 224.


Jungnok Park

As for T6, an internal comparison of the renderings in Zhi Qian's work reveals the probability that T6 is not a retranslation by Zhi Qian himself, since some of the renderings in T6 are not in accordance with his other work. For example, instead of buwangyan ("not uttering lies"), qiyu ~~~ ("frivolous words"), liangshe C" ("double tongue"), and ekou ~D ("harsh words") in T5 (1.162b), T6 (1.177b) renders respectively as buqichan =1":wt~ ("flattering"), ningshi mfff ("false embellishment"), and ema ~,~ ("reviling"). While the four renderings of T5 are found intermittently in Zhi Qian's other works, the three renderings of T6 fail to feature at all. In addition, a comparison of the attitudes towards Chinese interpolations between T5 and T6 suggests that the writer of T6, i.e. the retranslator of T5, is not Zhi Qian himself. As mentioned above, the frequent use of terms reflecting Chinese values is one of the important characteristics of Zhi Qian's work. Such interpolations, particularly the emphasis on filial piety, are conspicuous in the translation of T5: as the Chinese expanded the concept of filial piety to the concept of loyalty, so that it came to function as a political ideology, the author of T5 utilises filial piety as a religious ideology. With reference to these expressions of filial piety, the writer of T6leaves them in a few cases but removes them in the majority of cases: Table 1: Interpolations of expressions I passages related to filial piety in T5 and T6

=1"* m

Only in T5 1.164a19-20, 164b7, 169b19, 169b25, 169c1, 169c20, 170a9, 170c24, 171a5, 171a16, 172a6, 172a09, 172a14, 172b19, 172b24, 172b29, 173all, 173b23.

Both in T5 and T6 TS.1.160c1S-16 :=; T6.l.l76bS-6,17Sb8 :=; 191a22.

Only in T6


Given that we can find expressions of filial piety throughout Zhi Qian's other works, the tendency of T6 to remove such expressions from T5 reveals that the author of T6 is probably not Zhi Qian. Surveying the use of the relatively older and later renderings in T5, T152 (by Kang Senghui), and T6, we can obtain meaningful

The authorship of T5 and T6 Mahaparinirviil}-asiitra


information that narrows down the possible period during which T6 was translated. Comparing the peculiar renderings of the older and later couplets, we find the following examples: Table 2: Use of archaic renderings in T5, T152 and T6
pil}-Ijapiitika briihmal}-a Samiidhi

riijan cakravartin

7tf&lT z1t Will,\

T5 T152 T6
7 0

;t~ ,~A~J\ 5EJ~\


239 112 91


13 0


1 2

2 125 18

0 0

1 2

3 14

13 13 0


In the above table, the terms in the left-hand column for each couplet pre-date the ones in the right column. As expected, T5 represents the oldest stratum using archaic renderings most frequently. Then, although being a work of a generation after Zhi Qian, T152 shows that it still preserves a considerable number of archaic renderings. On the contrary, T6 replaces all of the archaic renderings with later "standardized" ones. This implies that T6 is possibly more recent than Kang Senghui's T152. Hence, we should regard T6 as one of the latest works in Zhi Qian's translation circle, i.e. possibly a work from about 280. Conclusion From the above examination, I draw the following conclusions: I) An examination of the renderings in T5 and T6 reveals that both texts are translations by Zhi Qian or by one of his successors from his translation circle. 2) The archaic writing style, particularly the style of the verses, and the renderings that are unique to Zhi Qian indicate that T5 is probably the work of Zhi Qian and that it may be one of his earliest translations. T5 appears to have been composed earlier than T21O, which was translated around 224. 3) The tendency of T6 to remove the interpolated expressions


Jungnok Park

of filial piety in T5 and replace the archaic rend~rings of T5 with later standard renderings indicate that the writer of T6 is probably not Zhi Qian himself, but a successor who produced the work possibly around 280.

1. Primary sources
[D] [K]
Dfgha-nikiiya: Rhys Davids, T. w.; Carpenter; J. Estlin (eds.). 1890. Dfgha Nikiiya, 3 vols. London: Pali Text Society. Tripitaka Koreana: Dongguk University (ed.). 1957. Goryeo Daejanggyeong raJ~*~~J[, 48 vols., Seoul: Dongguk University.

[MPS] Mahaparinirviir;a-satra: [MPS(P)] Mahaparinibbana-sutta (D.II.72-168) and Mahasudassana-sutta (D.II.l69-99) [MPS(S)): Waldschmidt, Ernst (ed.). 1950-51. Das Mahaparinirviir;asatra, Berlin: Akademie-Verlag. [MPS(ST.I)]: Waldschmidt, Ernst (ed.). 1961. "Der Buddha preist die Verehrungswtirdigkeit seiner Reliquien: Sondertext I des MahaparinirvaJ;lasutra." Nachrichten von der Gesellschaft der Wissenschaften zu Gottingen: Philologisch-Historische Klasse. Gi:ittingen: Van den Hoeck & Rupprecht: 375-385. [MPS_ST.II]: Waldschmidt, Ernst (ed.). 1948. "Wunderkrafte des Buddha: eine Episode im Sanskrittext des MahaparinirvaJ;lasutra." Nachrichten von der Gesellschaft der Wissenschaften zu Gottingen: Philologisch-Historische Klasse. Gi:ittingen: Van den Hoeck & Rupprecht: 48-91. [T]
TaishO shinsha daizokyo: Takakusu, Junjiro; Watanabe, Kaikyoku (eds.). 1924: TaishO shinsha daizokyo *lEJfJT{~*~~I, 85 vols. Tokyo: Taish5 shinshu daiz5ky5 kank5kai *lEJfJT{~*~~fU1'J~. Qisha dazangjing: Yanshengyuan dazangjing ju M~~J'G*~~J[fi'V (ed.). 1987. Songban qisha dazangjing *M&~i:fiij>*~~![. 40 vols. Taibei: Xinwenfeng chuban gongsi JflTX:fli\M&0BJ.


The authorship of T5 and T6 Mahliparinirvar;asutra


2. Secondary sources:
Buswell, Robert E. 2004. "Sugi's Collation Notes to the Koryn Buddhist Canon and Their Significance for Buddhist Textual Criticism." Journal of Korean Studies, 9,1: 129-184. Harrison, Paul. 1998. "Women in the Pure Land: Some Reflections on the Textual Sources." Journal of Indian Philosophy, 26, 6: 553-572. Hirakawa, Akira .ljZ} r [~. 1968. Shoki daijo bukkyo no kenkyu WM*~IH~~ O)liJf~. Tokyo: Shunjusha lff..kt. Kawano, Satoshi )i:if!f~.III. 1986. "Jiku Hogo no yakukyo ni tsuite" ~5*~Q) ~~CJ~ ~"'(. Journal of Indian and Buddhist Studies fpJjt${~~Jt~~Jf~ 35, 1: 72-74. Nattier, Jan. 2003. "The Ten Epithets of the Buddha in the Translations of Zhi Qian 3Z:~." Annual Report of the International Research Institute for Advanced Buddhology at Soka University EU1illi*$~~~{~~$fc\l~liJf~ pJTiptf 6: 207-250. _ _ _ . 2004: "The 1Welve Divisions of Scriptures (+- .g~~) in the Earliest Chinese Buddhist Translations." Annual Report of the International Research Institute for Advanced Buddhology at Soka University EU1illi* $~~~{~~$fc\l~liff~pJTiptf 7: 167-196. Ono, Genmyo IJ\!f~fr); Maruyama, Takao julJ~~. 1936-1978: Bussho kaisetsu daijiten ~flM~*iWf!ll!. Tokyo: DaitO shuppansha **:tf:\fl&t. Suzuki, Hiromi n**~. 1995: "Koyaku kyoten ni okeru yakugo ni tsuite: , Jiku Hogo yakushutsu kyoten wo chUshin toshite" il~~!11!'.::})~t~~~ '':-:J~ ~"'(: ~$~~t:f:l~!II!~.pI~\t l",(, Journal of Indian and Buddhist Studies fpJjt~{~~~liJf~, 43, 2: 198-200. Tokuno, Kyoko. 1992. "The Evaluation of Indigenous Scriptures in Chinese Buddhist Bibliographical Catalogues." Buswell, Robert E. (ed.). Chinese Buddhist Apocrypha, ed. Delhi: Sri Satguru Publications (Original publication: University of Hawai'i Press. 1990): 31-74. Tong, Wei ]t:Et, 1991. Beisong "Kaibao dazangjing" diaoyin kaoshi ji mulu huanyuan :It* Gf'!li*~~> mfPsf:f~.&13~:lIJ]t Beijing: Shumu wenxian chubanshe -=f=;13:XilIkt:f:lfl&t. Ui, Hakujyu ~#{B . 1971. Yakukyoshi kenkyu ~~~liJf~. Tokyo: Iwanami shoten :E5~fl~. ZUrcher, E. 1991. "A New Look at the Earliest Chinese Buddhist Texts." Shinohara, Koichi; Schopen, Gregory (eds.). From Benares to Beijing: Essays on Buddhism and Chinese Religion in Honour of Prof Jan YiinHua. Oakville: Mosaic Press: 277-324.

The Jifayue sheku tuoluoni jing Translation, non-translation, both or neither?

Jonathan A. Silk

In respectful memory of Antonino Forte, scholar and friend

I. Theoretical considerations
It is probable that there have been questions about the authenticity of scriptures from the very earliest days of Chinese Buddhism, although our available evidence does not stretch back quite that far.1 Modern scholars have also been intrigued by similar questions of origins, although sometimes for dierent, even perhaps quite opposite, motives. For the arbiters of orthodoxy in Buddhist China, one of the principal criteria for the authenticity of a scripture was its legitimate Indian (or Western) origin; a text was valid or genuine if it had been translated, rather than written or composed in China. What was crucial was that the text be authentic, and authenticity rested with the Buddha, in India.2 For many modern scholars, in
A summary of this paper was presented at the conference Early Chinese Buddhist Translations sponsored by the Austrian Academy of Sciences and held in Vienna April 1821, 2007. I thank the organizer, Max Deeg, and the participants for their helpful comments and advice; in particular I acknowledge with appreciation my debt to Stefano Zacchetti. I am grateful to Nobumi Iyanaga, Ksei Ishii and Paul Harrison for valuable comments. 2 Of course, the case is much more complicated than this simple characterization suggests. For instance, even orthodox cataloguers were
Journal of the International Association of Buddhist Studies Volume 31 Number 12 2008 (2010) pp. 369420


Jonathan A. Silk

particular those who focus on East Asian Buddhism, on the other hand, it has been the scriptures composed in China which are the true treasures, since these are felt to reveal a genuine Chinese religiosity, absent from, or at the very least less evident in, translations.3 My concerns come from another direction. As a student of Indian Buddhism, I am interested in Buddhist scriptures in Chinese primarily from the perspective of the use to which they may be put in elucidating the Buddhism of India. The questions of greatest interest to me in this context revolve around how I may most legitimately and authentically make use of works in Chinese. To address such questions, we have to think about just what such works represent and reect. Can we, in fact, use them to shed light on Indian concerns at all and if so, how should we do this? Or do they reect Chinese problematics to such an extent that their applicability to Indian questions is either eaced or so far hidden as to be beyond recall? Is it possible to balance these two extremes? On the other side is a concern for the Sinologist: how Chinese can a text be which, in part or as a whole, comes from, or is motivated or inspired by, a foreign creation? What might such an import or transplant have to say about domestic Chinese concerns? One thing is sure: whether establishing a viable standpoint either of the student of Indian Buddhism who would attempt to make use of Chinese evidence, or of the Sinologist who would refer to foreign-inspired works, considerable care and nuance is required. The broad central question here, then, is: to what sorts of uses may we legitimately put Chinese versions of scriptures? Setting aside the Sinologists concerns as best dealt with by genuine Sinologists, from the point of view of Indology, part of this question is easy or at least, easier to answer. If they are translations of Indian works, as may be veried through comparison with extant Indic texts, or through coordinated examination of independently produced Tibetan translations, for example, Chinese translations

able to accept that genuine revelation which is to say, transmission of Indian Buddhist scripture could be possible through dreams, visions and the like. See the very interesting discussions in Campany 1991, 1993. 3 A good survey is Buswell 1990; see too Kuo 2000.

The Jifayue sheku tuoluoni jing


may oer us an interpretation of, or viewpoint on, an Indic text although, to be sure, the diverse problems that accompany the eort to make use of such translations are only now beginning to be explored seriously.4 But how are we to proceed when we are unsure of the origins of a text? This question becomes especially acute when we take note of the recent, important ideas of Funayama Tru, who has introduced in a particularly clear way the idea of a type (or types) of scriptural production which is (are) neither pure translation nor pure native creation. Funayama has eloquently brought to the fore the following observation: some works which claim for themselves, or have claimed on their behalf, Indian origins can be demonstrated to have been composed elsewhere.5 But the use here of the world composed conceals a multitude of possible variations. The key (moving) point along the arc of possibility hangs on the extent to which the content of the text might have originated in an Indian, or perhaps better Indic, environment, reecting Indic concerns, and the extent to which Chinese agendas, expectations and assumptions penetrate the work (simplifying, for the moment, the complication that the binary opposition of Indic and Chinese is also more than a little problematic). Setting aside works composed or compiled by Indians in China (or in the Sinitic sphere),6 and concentrating on scriptures more narrowly understood,7 the types of works which result from what might, in some circumstances, be
Some of the papers presented at the conference referred to in note 1 are good examples of recent work moving, in my opinion, in the right direction. 5 See Funayama 2002, 2006. 6 I am reminded in this context of the situation attendant on the later Indian Buddhist transmission to Tibet, in which we know that texts composed more or less to order by Indian paits were accepted by Tibetans as genuine. A comparison of the two cases, removed as they are by centuries, should prove very interesting, the more so since the creation of a number of works within the Chinese cultural sphere also had the active cooperation or supervision of foreign authorities (on which see recently Funayama 2006). 7 I leave out of consideration here stric compositions, usually classied in Chinese as ln , which should perhaps be dealt with separately.


Jonathan A. Silk

called trans-creation range from abbreviated summaries or bestof collections, as it were, to works inspired by or based on Indic sources, whether those sources themselves had attained some tangible form or not. The former type of works, the best-of collections, might be considered those the literal content of which can be traced to works having an Indic origin, although the arrangement of that content has been altered to some extent, usually by excision. Somewhere else along this continuum would lie a work like the Stra of the Wise and the Fool (Xianyu jing ), clearly Indian to some extent, but not Indian as such. As for the class of works inspired by, I have used the language of Hollywood here intentionally. For most of us are familiar with lms which claim themselves to be inspired by real events, or based on a true story. I have come to think of some Chinese scriptures in this way, as located along a continuum, or even better, as distributed in a multi-dimensional space, rather than as divisible into one of two categories.8 In this light, if we can no longer state the problem as one of deciding dichotomously between a work being either a translation or an apocryphon, what are we to do? In fact, I already approached this general question some ten years ago, although I failed to articulate it within the same framework at that time. In studying the origins of the Guan Wuliangshoufo jing , the conclusion I came to and it is not mine alone is that while the text was compiled or brought together in China or Chinese-speaking Central Asia, it nevertheless contains genuine Indic elements which must have been derived directly from Indian traditions.9 In particular, I argued that the frame story providing the setting for the meditative visions of the text was of thoroughly

It may even be that, just as some philosophers speak of all language as metaphor, we should speak of all Chinese scriptures, by their very nature as translations or even simply as inspired by Indic mentalities, as in some fundamental sense Chinese. I leave this discussion for another occasion. For my thoughts on the multi-dimensional space within which we might locate scriptures, see Silk 2002. 9 See Silk 1997.

The Jifayue sheku tuoluoni jing


Indian origin. Simultaneously, it is perfectly evident that the text is not Indian as a whole. I did not myself extend my researches beyond the frame story, but based on the work of others dealing with the text in extenso, I accept that many elements cannot but have been composed originally in a Sinitic environment. So what is the text? Is it Indic, or Chinese, or something else? Is it a translation, or a trans-creation, or an apocryphon, or something else? These are particular questions about a specic work. But they are at the same time part of a much larger issue. One way perhaps the only way to work toward a generalizable answer to such questions of identity or origins, one or indeed, the only way to develop a method for evaluating and considering such cases, is to see what other types of examples one can nd. One must try, that is, to plot the arc or distribution of such creations by careful examination of relevant works, one by one, leaving until later a more far-reaching evaluation of the range of evidence to be produced by such investigations. The present paper is intended, then, among other things, as a small contribution in this general direction.

II. Sources
The Jifayue sheku tuoluoni jing , The Dhra-stra on Collecting the Joy of the Teachings and Getting Rid of Suering, is extant only in a Chinese version, the date of which I will discuss in a moment.10 Although it now appears in Chinese canons embedded within other texts, it was evidently transmitted in China as an independent text at one time.11 Catalogues, beginFor helpful hints toward the resolution of some of the problems dealt with in this section, in particular regard to T.1332 and 1336, I am grateful to James Benn, James Robson, Robert Gimello, and especially Nobumi Iyanaga. I owe my initial acquaintance with the Jifayue sheku tuoluoni jing to Kat 1950: 39. 11 The Bussho Kaisetsu Daijiten (Ono 19321935: 5.223a) refers to a manuscript of the text now in the Kyoto University Library (Kyto daigaku fuzoku toshokan ), registered as 1//6 (z 1/shi/6). According to the kind information of Funayama Tru, z here points to the Zkyshoin , the collection of drafts and


Jonathan A. Silk

ning with the Zhongjing mulu (I) of 594, cite it as a one juan work, here classied as belonging to the category of separate compilations of Mahyna scriptures, dshng zhngjng bishng .12 For the Lidai sanbao ji of 597, it is a Mahyna stra of an unknown translator, dshng xidulu shy ,13 while the Zhongjing mulu (II) of 602 calls it a separately compiled abbreviated extract of a Mahyna scripture, dshng bishng cho .14 The Dazhou kanding zhongjing mulu , compiled in 695, again refers to it as a Mahyna scripture by an anonymous translator dshng shyjng .15 As is well known, while the attribution to an anonymous translator is not, in and of itself, necessarily a sign that a work is not a genuine translation, in the sense of being based upon some foreign original, whatever this might mean, it can suggest the possibility of some origin outside the domain of ocialdom. The classication of the text as a separate compilation or abbreviated abstract, bishng cho , however, is interesting. Almost all the works so classied are very short; the Zhongjing mulu (II) lists as dchng bishng cho fully 117 works in 137 juan,16 indicating that virtually all of them are no longer than one juan. A number of these works, but by no means all, are dhras. In fact, bishng cho, or simply chojng , is an interestingly contested emic category, closely related to that of apocryphal or doubtful scriptures (wijng or

manuscripts used for the compilation of the Zokuzky , including eventually unpublished texts. Hoping it might indeed be an independent manuscript of the stra, with the help of Funayama I obtained a copy. Unfortunately, I discovered that it is merely a transcript of the text as transmitted in the QDSJ (see below). For further information on the works transmission in Japan, see below. 12 T.2146 (LV) 125c10 (juan 2). 13 T.2034 (XLIX) 114a2 (juan 13). 14 T.2147 (LV) 164b22 (juan 3); identical in T.2148 (LV) 199a28 (juan 3). 15 T.2153 (LV) 437a24, b17 (juan 11). 16 T.2147 (LV) 163c15165a16 (juan 3).

The Jifayue sheku tuoluoni jing


yjng ).17 An extensive study of the works classed as dchng bishng cho should help us move toward an understanding of the signicance of this category, and in turn illuminate one traditional location of the Jifayue sheku tuoluoni jing. While the Jifayue sheku tuoluoni jing (hereafter JSTJ) has not been, then, independently transmitted within the known Chinese canons, and appears to be unknown at Dunhuang , at least as an independent work, it does exist embedded in other works, or in some versions of other works. In the Jin and Second Kory canons, upon the latter of which the Taish edition is based, the JSTJ appears within the Guan xukongzang pusa jing (Stra on the Contemplation18 of the Bodhisattva kagarbha hereafter, GXPJ), the Qifo bapusa suoshuo datuoluoni shenzhou jing (Spirit Spells Spoken by the Seven Buddhas and Eight Bodhisattvas hereafter, QDSJ), and the Tuoluoni zaji (Dhra Miscellany hereafter TZ).19 In the Song period Qisha canon, on the other hand, it is found only in the third of these, the TZ. This is interesting because, as Stefano Zacchetti has clearly pointed out, the Jin and Kory editions belong to a lineage separate from the Qisha, that of the Kaibao canon.20 Moreover, according to a note appended to the GXPJ in the Kory edition, the dhra was also not found therein in the Qidan or Liao edition, which is not now known to be extant.21 The Qidan / Liao edition is closely

See Tokuno 1990, and earlier Okabe 1971. As far as I know, the most thorough account of chojng so far is that in Ono 19321935, bessatsu , 300369. The materials collected there should serve as the basis for any future study of this interesting category. 18 Or perhaps visualization; I do not wish to enter here into the debates over the signication of the term guan in such titles. 19 The translations of the last two titles I borrow from Strickmann 2002: 312, n. 39. See for QDSJ Strickmann 1996: 7376, for the GXPJ Tsukinowa 1971: 112119. On the TZ see Strickmann 1996: 7678, and now Ochiai 2003, who notices our dhra on p. 73. 20 Zacchetti 2005: 74140, esp. the stemma on 133. 21 See the note to the text edition, below. Although small portions of


Jonathan A. Silk

related to the Fangshan stone canons , whose copies of both the GXPJ and the QDSJ likewise do not contain the JSTJ (and TZ itself is apparently not represented in the Fangshan collection). Here we have, therefore, a rather clear set of lineages of printings of the Chinese canon, within some of which the JSTJ is attached to the GXPJ and QDSJ, while in others it is not; it seems to appear everywhere within the TZ. This distribution may have implications for the history of the JSTJ. At the same time, we also need to deal with the issue of the relations between the three homes of the JSTJ themselves. Catalogues attribute the GXPJ to the translator *Dharmamitra / Tanmomiduo ,22 a foreign monk who came to China and died there in 442. Whether or not this attribution is to be accepted,23 the dating of the text is nevertheless probably generally correct. But what is the extent of this work? As noted above, some versions of the GXPJ do not contain the JSTJ, which is, moreover, being rather awkwardly appended near the end of the text, from a structural point of view obviously an intrusion. The dating of the GXPJ to the mid-fth century, then, even if solid, does not necessarily help us securely determine the date of the JSTJ itself, or help us trace its origins. The QDSJ is cited by reference works as an anonymous work belonging to the period 317420 (Eastern Jin ). However, here again there are complications. Although there are a number of references to a one juan text with a similar title, what appears
the Qidan edition have been found, I do not know that our text is among them at this time. 22 T.2151 (LV) 361b1314 (juan 3); T.2153 (LV) 384b1921: ; T.2145 (LV) 12bc (juan 1) 12b28, c34 . 23 Tsukinowa 1971: 123 opines: There is not one true example of something which could be termed a translation of Dharmamitra, going on (pp. 123124) to argue this on the basis of Dharmamitras biography. See the biography translated in Shih 1968: 140143, and the consideration of Dharmamitras translations in Hayashiya 1945: 444453. I have not investigated the matter, and therefore do not necessarily accept Tsukinowas opinion as fact.

The Jifayue sheku tuoluoni jing


to be this text more or less as we now know it is mentioned rst several centuries later, in the Datang neidian lu of 664, where it appears as an anonymous translation in four juan, 70 folios.24 The Kaiyuan shijiao lu of 730 lists a Qifo suoshuo shenzhou jing in 4 juan,25 of which it says The rst juan calls it Qifo shiyi pusashuo datuoluoni shenzhou jing26 The same catalogue subsequently refers to a Qifo suoshuo shenzhou jing , likewise in four juan, going on as follows: An anonymous translation of a Trepiaka of the Jin period,27 specifying in a note: At present this is catalogued among works of the Eastern Jin (317420).28 Moreover, it then oers a lengthy comment as follows:29
The preceding Qifo suoshuo shenzhou jing is registered in the Great Zhou Catalogue [= T.2153] as a retranslation. [This] states that [this Qifo suoshuo shenzhou jing] has the same original as the Qifo shenzhou jing in one juan translated during the Wu period [222280] by the foreign upsaka Zhi Qian.30 [However,] at present, since this single juan version has been lost for quite some time, and the number of juan is also dierent [four as opposed to one], there are insucient grounds for identifying the two (?). At present, relying on (the) old catalogue(s), I register it among unique texts [not as a retranslation].

T.2149 (LV) 314c4 (juan 9): ; also 287b18 (juan 6), 303c7 (juan 8). The Dazhou kanding zhongjing mulu of 695 says the same, but listing 71 folios, T.2153 (LV) 465a3 (juan 13): . Tokiwa 1938: 793 refers to Sengyous Chu sanzang ji ji of 515 as recording this text (T.2145 [LV] 31b14 (juan 4): , with the note jilzh ybn . See below. 25 . 26 T.2154 (LV) 510a8 (juan 3): . 27 . 28 . 29 T.2154 (LV) 510a8 (juan 3): . 30 Compare the following: T.2153 (LV) 400b2124 (juan 5):



Jonathan A. Silk



Indeed, as do other catalogues, the Kaiyuan shijiao lu also lists a Qifo shenzhou jing in a single juan (with explicit appeal to Sengyous Chu sanzang ji ji of 515), attributing it in a note to the translator Zhi Qian of the Wu Yuezhi .31 On the other hand, the catalogue also reports the existence of a Qifo ba pusa suo shuo shenzhou jing in a single volume, of which it says that according to the Chen catalogue this is an abbreviation of the Qifo jing.32 It is hard to know what to conclude from this information. As Zhisheng himself, the author of the Kaiyuan shijiao lu, concluded, there is virtually no chance that the QDSJ as we know it in four juan has anything to do with Zhi Qian or with such an early period. On the other hand, given its structure as a collection of dhras, an expansion of a single volume text into a compendium of four volumes would have been easy to accomplish, at least from a mechanical point of view. Probably Zhisheng is right that the earlier text, while coincidentally sharing a name similar to that of our QDSJ, is otherwise unrelated. While we can probably, though not certainly, accept an attribution of the QDSJ in four juan to the fourth or fth century, we have, once again, no assurance that it contained the JSTJ in that period. The TZ is attributed to a slightly later period, 502557. The basis for this appears to be once again the Kaiyuan shijiao lu, which says that the compiler is unknown, and that it is catalogued among works of the Liang (502557).33
31 T.2154 (LV) 633a2021 (juan 14): . 32 T.2154 (LV) 654b10 (juan 16): . I have so far not been able to identify this Chen catalogue . 33 . This dating is stated as fact in the Hbgirin catalogue, for instance (Demiville 1978), almost certainly on the authority of the Kaiyuan shijiao lu. Recently, however, some reasoning was

The Jifayue sheku tuoluoni jing


It then continues:34
This spell collection is listed in the Great Zhou catalogue [T. 2153] as an independent Mahyna stra,35 and it also states the name of the translator to be lost, but this is not accurate. Examining the style of writing, [we learn that] this is a locally produced abbreviated compilation, and not an alternate translation of a Sanskrit original. We know this because works such as the Qifo shenzhou jing and the Tuolinnibo jing were translated having been brought from abroad.36 The Hu zhutongzi tuoluoni jing was originally translated by Bodhiruci of the Wei. Moreover, the Tuolinnibo jing is the same work as the Zuishengdeng wang jing. Because such scriptures as these were all gathered together in this work, transmitted without a Sanskrit original, the work must be a local compilation. Because it is not yet known who compiled it, I mention it here (?).

Zhishengs argument is that since TZ incorporates works known to have been otherwise translated, the compilation must be secondoered by Ochiai 2003:13, who wrote as follows: Since the very similar Tuoluoniji jing was translated by Atikta between the fourth and fth years of the Yongzheng period (653654), we can probably place TZ, which appears in the Liang catalogue, about a century earlier. (Atikta) ~(~) . This reasoning, however, is far from rm, and the dating must remain unsure. 34 T.2154 (LV) 624b47 (juan 13). 35 T.2153 LV) 380a20 (juan 1): ( 460b22 [ juan 13]). 36 The expression rcho su fn remains unclear to me. Chinese rcho appears to indicate, basically, the arrival from abroad of ambassadors to have an audience with the emperor. This is the only instance of this term in Zhishengs work.


Jonathan A. Silk

ary and domestic. While Zhishengs conclusion is certainly right, his reasoning is problematic, since one can well imagine Chinese translators doing what we know Tibetan translators did: in rendering compilations, these translators as a rule borrowed pre-existing translations, instead of translating the passages anew. If TZ were an Indian compilation, and its Chinese translators had access to and made use of earlier Chinese renderings of (some of) its contents, the result would be a Chinese rendering of a genuinely Indian text which, nevertheless, fullled Zhishengs conditions for a local composition. For reasons to be detailed below, I believe that while his reasons are wrong, Zhishengs conclusion is nevertheless correct, and TZ is a secondary and local Chinese compilation, some of the sources of which are very clear. However, even accepting the sixth century date for the collection, this does not help us with the date of JSTJ, since there is no assurance that the original TZ contained the JSTJ. Now, one reason to maintain the secondary status of TZ is its large scale citation or incorporation of the QDSJ, of which the quotation of the JSTJ appears to be a part. TZ therefore seems to be entirely dependent upon the QDSJ.37 But how could this be the case when the JSTJ does not appear within the QDSJ in the Qisha or Fangshan editions? Were the JSTJ in the TZ dependent on QDSJ, we would seem to be compelled to conclude that the JSTJ had once been part of QDSJ whence it was borrowed into the TZ but was subsequently removed from the QDSJ in the Qisha and Fangshan editions; the Kaibao tradition canons alone, in this scenario, preserved the original (or: a more original) format of the QDSJ containing JSTJ. TZ clearly post-dates the QDSJ, since it subsumes it; this would seem to rule out the TZ as the original canonical home of the JSTJ. Nevertheless, such collections clearly were able to grow over time, such that QDSJ could have borrowed back JSTJ from TZ although why this would have happened only in the Kaibao lineage of the canon, and not in the Qisha and Fangshan

TZ incorporates large portions of QDSJ, particularly, but not only, in the rst four juan; for details of the correspondences see Ochiai 2003 (already noted in Strickmann 1996: 76).


The Jifayue sheku tuoluoni jing


canons, remains unexplained. As unclear and, frankly, confusing as all this is, the individual histories of the QDSJ and the GXPJ are more problematic still. As extended units, both the QDSJ and the GXPJ were obviously heavily edited, if not outright compiled or composed, in China. Strickmann, for example, characterizes the QDSJ (as we have it) as an obvious melange of prototantric elements and Chinese practices. Already, he explains, in passing from the bodhisattvas to the planetary spirits, we begin to notice, despite the Indian trappings, that we are truly under the skies of China.38 He does observe, however, that it is in the fourth and nal chapter that Chinese elements begin to appear overtly; our dhra, on the other hand, occurs earlier, in the second chapter. Therefore, although the composite nature of the text is manifestly evident, it is at the same time certainly possible that the JSTJ is preserved in the QDSJ as a genuinely Indic element (I will explore below just what this expression might mean). The dhras status as somehow originally independent of the QDSJ a collection of diverse materials does seem evident. Therefore, identifying the composite nature of the text as a whole does not move us very far toward addressing the problem of the origin of the JSTJ itself. What, then, might we learn from the GXPJ?39

Strickmann 1996: 7374. See also Xiao 1994: 386390 (I owe the latter reference to James Robson). 39 Modern treatments of GXPJ generally do not take account of the JSTJ. It is, however, briey noted in De Visser 1931: 33, who says the following: Here begins a new part of the stra entitled Stra on the dhras [sic] for collecting the joy of the Law throwing away suerings (shhetsu shaku daraniky). He then gives, in a new paragraph and without explanation, the following: Nama Buddhya! Nama Dharmya! Nma Saghya! Nama Vivadhacya (?)! Nama gakhabucya (?)! Nama Mahsattva Bhagali (?)! It is curious that he then continues: The remaining text is evidently a repetition of the contents of the kagarbhastra given above. He appears here to ignore the narrative, which is clearly the most characteristic element of our short text. This narrative is summarized by Kamibayashi Ryj in Ono 19321935: 4.340d341a, in his discussion of the QDSJ. Likewise, Kuo Li-Ying 1994: 137138 summarizes the story. In saying that the GXPJ is


Jonathan A. Silk

There are some indications that, at least in the later tradition, the JSTJ was closely identied with the GXPJ. The JSTJ consists, as we will see, of its dhra and an explanatory story. The narrative is quoted, or perhaps better paraphrased or abbreviated, in several works of East Asian authors. In the Dasheng fayuan yilin zhang of Kuiji (632682),40 which appears to preserve the earliest such reference to the story, the passage is attributed to the Daji jing , that is to say, to the Mahsanipta collection of scriptures. Now, the GXPJ itself is not formally considered to belong to that collection.41 But, at least in the Dazhou kanding zhongjing mulu scripture catalogue compiled in 695,42 which is to say, only a few years after the death of Kuiji in 682, the translation is listed along with other texts which do belong to the Mahsanipta formally speaking. Evidently the same association reected in this catalogue lies behind Kuijis attribution of his citation to this collection, suggesting that for him the passage belongs to, or with, the GXPJ. On the other hand, very soon after Kuiji, the Korean scholar ijk cited the text in his Posal kyebon-so ,43 without reference to any collection and calling it by a shortened name, Jifayue jing , suggesting that the connection with the Mahsanipta collection was not necessarily always asserted. Later still, and perhaps in light of the same tradition as that followed by Kuiji, the Khitan-Liao monk Feizhuo (d. 1063) in his Sanbao ganying yaolelu explicitly sources his citation as from the Jifayue sheku tuoluoni jing, a separate transmission of a work found within the Mahsanipta.44 Further association comes from the context
the only place in which the JSTJ is found, however, she overlooks its two other sources. 40 T.1861 (XLV) 307a1522 (juan 3); see below. More correctly his name should be given as merely Ji. 41 See the brief discussion of the supplementary Mahsanipta texts in Braarvig 1993: xxxxxxi. 42 .2153 (LV) 384b19 (juan 2). 43 T.1814 (XL) 657a620 (juan 1). 44 T.2084 (LI) 839c6 (juan zhong):

The Jifayue sheku tuoluoni jing


in which the JSTJ is found in the QDSJ. In the QDSJ found in the Kory edition, the JSTJ is preceded by a small character note which reads:45 In the Song edition, there are found here twelve lines of the Xukongzang pusa dhra. We checked it, and it is what is given above in juan one, leaf twenty-two,46 so here we omit it. This again suggests some association between the JSTJ and the GXPJ in the minds of the compilers of this recension of the QDSJ, and / or the editors of the Kory canon. There is a nal possible reason for some association between the QDSJ and GXPJ. The Kaiyuan shijiao lu has an entry on the Xukongzang pusa wen fo jing in one juan, concerning which it then says:47 This is also called the Xukongzang pusa wen Qifo tuoluoni jing, and again the Qifo shenzhou jing . Here the names of the QDSJ and the GXPJ seems to be almost fused, and it is possible, then, that for a reader who had been unable to compare the actual texts, some confusion may have arisen due to the alternate names by which the two works, QDSJ and GXPJ, were known. Despite the preponderance of often contradictory or just unclear evidence, the simple conuence of certain, albeit not entirely independent, pieces of information, including the presence of the JSTJ in the GXPJ, QDSJ and TZ in one canonical lineage, seems to suggest that JSTJ was established already in the fth century in China. The only really rm date we have to work with, however, is signicantly later, the rst catalogue reference in the Zhongjing mulu (I) of 594, in which the text is recorded independently, and not as forming a part of any of the three texts within which it is now to be found. If this independent version were an extract from an earlier
. 45 K.433 (XIII) 1084a18: ; see the same at T.1332 (XXI) 541b17 (juan 1). 46 In the reprint edition this is found at 1081a12. 47 T.2154 (LV) 539a16 (juan 6): , See also 600c14 (juan 12), 708c22 (juan 19 alternate version).


Jonathan A. Silk

version transmitted only within another work, however, while 594 would necessarily remain our terminus ante quem, the true date of origin could be signicantly earlier. Even setting aside the problem of its date, there remain signicant unexplained questions. How could it be that the JSTJ is transmitted only in some recensions of the QDSJ if, as certainly seems to be the case, TZ borrowed the former from the latter? Or is TZ the original home of the JSTJ? There is some philological evidence for this in the readings of the text themselves; although such a judgement must be to some extent subjective, TZ appears to preserve a more readable text of the JSTJ than do either QDSJ or GXPJ. This might suggest that the JSTJ was borrowed, albeit imperfectly, from TZ by GXPJ and QDSJ but once again, the chronological problems this entails are not trivial. For the moment we must be content, it seems, to catalogue the substantial problems with the history of the text, and then move on.

III. The text: Edited and translated

Before we proceed further, let us see what the JSTJ itself looks like.48 I edit it here as it appears in the sources presently available to me:49
Guan xukongzang pusa jing, T.409 (XIII) 679c29680b23. Korean 64 (VII) 824c16825c6 Jin not reproduced Qisha not included in this version Fangshan not included in this version Qifo ba pusa suo shuo datuoluoni shenzhou jing, T.1332 (XXI) 544b5 c26 (juan 2). Korean 433 (XIII) 1084a181085a11 (juan 2) Jin 466 (XXIII) 893b7894a22 (juan 2, leaves 46) Qisha not included in this version My understanding of the text owes much to the kindness of Christoph Harbsmeier, although he is, needless to say, not responsible for my misunderstandings (or punctuation!); thanks also to Stefano Zacchetti for excellent advice. 49 I cite the Taish locations for reference; my edition relies directly on (reproductions of) the blockprint sources.

The Jifayue sheku tuoluoni jing


Fangshan not included in this version Tuoluoni zaji, T. 1336 (XXI) 631a4b27 (juan 9). Korean 1051 (XXX) 1282c51283b22 (juan 9) Jin 1142 (LIII) 462c4463b21 (juan 9, leaves 1113) Qisha 1072 (fasc. 445) 76a1977a11 (juan 9) Fangshan text not included in published collection

5051 52 5354 5556 57 58 596061 62

Some write not further noted. K 433 adds in small type , This dhara is not contained in the [Qi]dan edition, meaning it was not found in what is also known as the Liao canon . In the notes, I use the following: ZH = Zhonghua Dazangjing (Beijing 19841988); Sixi , Puning , Ming . 52 K 433, J 466 , J 1142, K 1051 , Q 1072 , and so below for all instances of . With the exception of the rst ve or six words, and the last one, only the rst three of which I understand, I have no condence in the correctness of the word divisions within the dhra. 53 J 466, K 433, Q 1072 54 J 1142, K 1051, Q 1072 omit . 55 J 466, K 433 small type for ; J 1142, K 1051, Q 1072 omit . 56 K 64 adds small type ; J 466, K 433 add small type 57 J 1142, K 1051, Q 1072 58 K 1051 , J 1142, Q 1072 59 K 433 small type for ; J 1142, K 1051, Q 1072 omit 60 K 64 adds ; J 466, K 433 for followed by small type 61 K 64 omits 62 J 1142, K 1051, Q 1072



Jonathan A. Silk

6364 65 66 67 6869 70 71 72 7374 75 76 77787980 8182


J 466, K 433 small type for ; J 1142, K 1051, Q 1072 for

K 64 ; J 466, K 433 for followed by small type ; J 1142, K 1051, Q 1072 for 65 J 1142, K 1051, Q 1072 for 66 J 1142, K 1051, Q 1072 for 67 J 1142, K 1051, Q 1072 for ; writing may have led to confusion 68 K 64 adds small type ; J 466, K 433 add small type ; J 1142, K 1051 , Q 1072 for ; Taishs note to T. 1336 indicates that Sixi, Puning and Ming read . 69 J 466, K 433 70 J 466, K 433 omits 71 J 466, K 433 for 72 J 1142, K 1051, Q 1072 73 K 1051 for ; 74 Following J 1142, K 1051, Q 1072; others omit 75 Following J 1142, K 1051, Q 1072; others add 76 Following J 1142, Q 1072 (K 1051 ); others 77 J 466, K 433 for 78 J 1142, K 1051, Q 1072 79 Taishs note to T. 409, ZH indicate is missing in Sixi. 80 Following J 1142, K 1051, Q 1072; others omit 81 J 1142, K 1051, Q 1072 add 82 J 1142, K 1051, Q 1072 for

The Jifayue sheku tuoluoni jing


83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98

J 466, K 433 for J 466, K 433 (and according to Taishs note and ZH Sixi also): others for ; therefore J 466, K 433 have the common word im . 85 ZH says Sixi omits 86 J 466 for 87 Following J 1142, K 1051, Q 1072; others omit 88 J 1142, K 1051, Q 1072 add 89 J 1142, K 1051 for , but Q 1072 90 J 1142, K 1051, Q 1072 for 91 J 1142, K 1051, Q 1072 add 92 Following Q 1072 (Taish T. 1336, ZH note that Sixi, Puning and Ming and several others share this reading); all other texts read for 93 Following J 1142, K 1051, Q 1072; others 94 Q 1072 adds ; Taish T. 1336, ZH note the same for Sixi, Puning and Ming 95 Following J 1142, K 1051, Q 1072; others 96 Taish note to T. 1336, ZH note that Sixi reads for 97 Taishs note to T. 409, ZH indicate missing in Sixi 98 J 1142, K 1051, Q 1072 add



Jonathan A. Silk

99 100 101102 103104105 106 107 [Dhra] At that time the Buddha spoke to the members of the great assembly, saying: When, during innite aeons, I was still at the stage of being an ordinary person (*pthagjana), my name was Zhetatuo.108 Living in the land of Jiatouluo I engaged in sales and peddling. I was dishonest, lied, and did all manner of evil deeds, which are impossible to recount. My sexual perversity and my immorality are impossible to fully detail. At that time, stupidly insensitive, I killed my father and made love to my mother.109 Over a number of years the people of the

J 1142, K 1051, Q 1072 add J 466 for 101 Taishs note to T. 409 here says that is missing in Sixi, which makes no sense to me. 102 J 466 for ; ZH says Sixi omits 103 Following J 1142, K 1051, Q 1072; others (J 466, K 433 ) 104 Following J 1142, K 1051, Q 1072; others omit . See the note to the translation here. 105 J 466 for 106 Following J 1142, K 1051, Q 1072; others 107 K 433, K 1051, Q 1072 108 Kamibayashi in Ono 19321935: 4.340d suggests Cathadha? which I cannot understand. 109 The rhetorical device here is interesting; with the phrase the text references the so-called three poisons, moha, dvea and rga, delusion, hatred and lust.


The Jifayue sheku tuoluoni jing


entire country all came to know of this, and loudly proclaimed: Now its been a number of years since this Zhetatuo killed his father and made love to his mother. At that time I pondered [the fact that I] was no dierent from the beasts; [what I did] wasnt the act of a human being. Then at night I jumped over the city wall at Jiatouluo, ed and hastened toward a deep marsh. At that time the king of that country was called *Vija. He issued a proclamation to the people in his state: This fellow, Zhetatuo, has committed acts of sexual perversity, and his immorality extends to committing this oense. Whoever can lay hands on this person will be handsomely rewarded. Then each and every person in this country responded to this appeal and was eager to get hold of me. Much alarmed, I left the state, and became a ramaa in another country. I cultivated the ten good [precepts], practiced seated meditation, and studied the Way. I wept day and night for thirty-seven years. Because of the obstacle of having committed the ve sins of immediate retribution,110 my mind was never at rest, and I could not nd peace. For thirty-seven years I lived in a cave in the mountains, always crying out Oh, how painful it is! Oh, how painful it is! With what mental [technique?] should I get rid of this pain? When, sobbing with grief, I went down from the cave to beg for alms, on the road I found a large bowl. Within it there was a stra box, but only one stra inside: the Dhra on Collecting the Joy of the Teachings and Getting Rid of Suering. It is said that in the past Buddhas as many as the sands of the [Ganges] river, at the time of their nirva, always lived in the land of Piyueluo, preaching this dhra, bestowing it upon the great bodhisattvas. Later there was someone who was able to hear this dhra. This person in a past age practiced upholding the ve restrictions and the ten good actions, and now he did hear it. If there is someone who, although he hears it, still does not take it seriously and does not practice, such a person is called one without a karmic link.111 This dhra can These ve murder of a father, mother, arhat, drawing the blood of a buddha, and causing a schism in the monastic community are the most serious crimes catalogued in Buddhist literature; see Silk 2007. 111 The sense of wyun seems to be that if, despite being presented with the opportunity to prot from the text, one still fails to do so, this is due to the burden of past karma and the absence of a necessary karmic conditioning from previous lives.


Jonathan A. Silk

remove the great transgression produced by the ve sins of immediate retribution committed through hundreds of thousands of aeons of rebirth. If there is a person who upholds, reads and recites it, he will never fall into the three unfavorable realms, hells, hungry ghosts or animals. Why? The Buddhas of the past, when they were about to enter nirva, devoted themselves to preaching it. They venerated and praised its merits as incalculable, and bestowed it upon bodhisattvas. Later, if there will be beings who have the opportunity to hear this dhra, practice [it] and take it seriously, their positive rewards will be dicult to calculate, like an ocean of gems [as great as Mount] Sumeru. Ordinary people will not be able to reckon it. If there is someone who performs all sorts of evil acts, [but] surreptitiously hears the name of this dhra, even without practicing [it], as soon as he holds it in mind he might fall into hell. [Then] all people in hell would benet from this persons benecial inuence, and their suerings would not be active (?).112 If there is someone who can practice this and in the present body energetically cultivating obtain [the dhra], he will see hundreds of billions of buddha elds, and the merit he acquires will be limitless and inexpressible. Only buddhas and bodhisattvas are able to fully exhaust it. Auditors and those belonging to the second vehicle (of the Lone Buddhas) cannot understand. Why? This dhra was not preached by [only] one buddha or two buddhas, but by the buddhas of the past as many as the sands of the Ganges. At that time, I picked up this scripture and without begging for food immediately returned to the cave joyously. Once inside the cave, I burned incense and oered worship.113 Piteously weeping and venerating, I practiced the recitation of the scripture inside the cave. After one year, I was rst able [to understand it], but because of the obstacle of my sinful actions, I was not able to get it to enter my mind. At that time, on an autumn moon-lit night I washed and practiced for a whole seven days. Like a beginning student worried quite a lot, I practiced again for seven days. I was still as unsettled as before. I was disturbed These two sentences are dicult, and I am not condent I have understood them well. 113 The expression shoxing lbi is attested in Buddhist works of the early fth century (T.397 [XIII] 136c4 [ juan 19]; T. 643 [XV] 696a1 [ juan 10]), and in secular works at least as early as the Weishu of 551554 (see Morohashi 19551960: 7.524b [19420.18]). Is this signicant for the dating of JSTJ?

The Jifayue sheku tuoluoni jing


in my mind, and I did not know what to think. Contemplating the written form of this dhra, after many perturbations,114 my mind was suddenly settled. Then I was delighted. Like a person who nds a hundred thousand gold jn on the ground, that others did not know were there, inside I was endlessly joyous. I then was also like that man. Practicing for many years, I became able to y without obstacle and see the buddhas of the three worlds in the ten directions. Later there will be practitioners who follow this practice.

Although some obscurities remain, the general sense of the text and the trajectory of its narrative are quite clear. What are we to make of this text? Let us begin with the dhra itself. This presents, for the time being at least, insuperable diculties. While the restitutions of the rst few words are obvious enough, the remainder is indecipherable in terms of its (putative) Indic original. All we can be certain about is the following: Nmo buddhya, nmo dharmya, nmo saghya, and then the nal svh! Even word boundaries are far from clear, and while the printed editions do separate the characters spatially into units, their separations are not consistent, suggesting, as we would expect, that the respective editors likewise had little idea what shape the dhra should take. Is it, in fact, a genuine dhra in that is, transcribed from some Indic language? There is simply no way to know with the information available at present. As with the dhra itself, the proper names in the story which follows in modern Chinese pronunciation Zhetatuo, Jiatouluo and Piyueluo in their turn also defy reconstruction. The name Pishe (if it is not to be read Pisheluo, with the variant recorded in TZ) appears to reect *Vija, which might suggest itself as a name for a king, although even this is far from sure. So the transcriptions if that is indeed what they are oer little help to us, since they could simply represent either badly transmitted forms or irregular tranOr after repeating it many times? But this does not necessarily t with the idea that it was the written form of the text which was contemplated. It is very dicult both to establish and understand the text here. Few clear parallels are to be found in other texts, and for the time being both the correct reading and interpretations must remain elusive. I thank Iyanaga Nobumi and Ishii Ksei for their advice.


Jonathan A. Silk

scriptions, or mere pseudo-Indic inventions. At present there is no way to further rene our appreciation of their original(s). To focus back on our central question, then, neither the dhra itself nor the transcriptions in the story nudge us one way or another toward any particular point on the arc spanning the gulf between translation and local composition. The narrative core of the work does, nevertheless, lead us in a particular direction, or rather, provides a solid point of reference. For, unlike the case with the dhra, in the story we are able to identify a clearly Indian precedent, and moreover one not known to have been otherwise transmitted to China before the time of Xuanzang in the mid-seventh century, long after catalogues assure us the JSTJ was already circulating in China. The basic story of our text, a Jtaka of the Buddha, has him as a dishonest peddler, who kills his father and has sex with his mother. He escapes, eeing the wrath of his fellows, ending up in another land, where he becomes a ramaa. He lives in a cave for thirtyseven years, in despair. During this time and this is crucial he habitually cries out Oh, how painful it is! Oh, how painful it is! The story goes on, but this is the portion of central relevance for us here. For it is to this that we can compare Indic versions of the story of the notorious Mahdeva, the putative instigator of the fundamental schism between the Sthaviras and Mahsghikas, the putative cause of the dissolution of the Buddhas previously unied monastic community.

IV. A parallel
The core version of the story of Mahdeva is that found in the *Abhidharma Mahvibh, which is now known only in Chinese translation (Apidamo Dapiposha lun ). This story is also relatively widely known in a variety of, generally quite abbreviated, forms in Indian Buddhist literature. The various ways in which the story is cast, however, do not elsewhere in known

The Jifayue sheku tuoluoni jing


Indian texts involve the integration of an exclamation of pain, as does the Vibh, whose account reads as follows:115
Long ago there was a merchant in the kingdom of Mathur. He married while still a youth and soon his wife gave birth to a baby boy. The child, who had a pleasing appearance, was given the name Mahdeva. Before long, the merchant went on a long journey to another country taking with him rich treasure. Engaging in commercial ventures as he wended his way, a long time passed without his return. The son, meanwhile, had grown up and deled his mother. Later on, he heard that his father was returning and he became fearful at heart. Together with his mother, he contrived a plan whereby he murdered his father. Thus did he commit his rst sin of immediate retribution. This deed of his gradually came to light, whereupon, taking his mother, he ed to the city of Paliputra, where they secluded themselves. Later, he encountered a monk-arhat from his native land who had received the support of his family. Again, fearing that his crime would be exposed, he devised a plan whereby he murdered the monk. Thus did he commit his second sin of immediate retribution. [Mahdeva] became despondent. Later when he saw that his mother was having sexual relations with another, he said to her in raging anger: Because of this aair, I have committed two serious crimes. Drifting about in an alien land, I am forlorn and ill-at-ease. Now you have abandoned me and fallen in love with another man. How could anyone endure such harlotry as this? With this excuse he also murdered his mother. He had committed his third sin of immediate retribution. Inasmuch as he had not entirely cut o the strength of his roots of goodness, [Mahdeva] grew deeply and morosely regretful. Whenever he tried to sleep, he became ill-at-ease. He considered by what means his serious crimes might be eradicated. Later, he heard that the kyaputra ramaas [Buddhist monks] were in possession of a method for eradicating crimes. So he went to the Kukkurma monastery. Outside its gate he saw a monk engaged in slow walking practice. The monk was reciting a hymn:

The basic translation is that of Mair 1986: 2025, which I have modied. The full account is in T.1545 (XXVII) 510c24512a19 (juan 99), with the portion quoted here found at 510c24511b28.



Jonathan A. Silk

If someone has committed a serious crime, He can eradicate it by cultivating goodness; He could then illuminate the world, Like the moon coming out from behind a screen of clouds. When [Mahdeva] heard this, he jumped for joy. He knew that, by taking refuge in the Buddhas teachings his crimes could certainly be eradicated. Therefore he went to visit the monk. Earnestly and persistently, [Mahdeva] entreated the monk to ordain him. When the monk saw how persistent [Mahdevas] entreaties were, he ordained him without making an investigation or asking any questions. He allowed him to retain the name Mahdeva and oered him admonitions and instructions. Now Mahdeva was quite brilliant and so, not long after he had renounced the world he was able to recite the text and grasp the signicance of the Tripiaka. His words were clear and precise and he was skillful at conversion. In the city of Paliputra, there were none who did not turn to Mahdeva in reverence. The king heard of this and repeatedly invited him into the inner precincts of the palace. There he would respectfully make oerings to Mahdeva and entreat him to lecture on the teachings. Later, [Mahdeva] left [the capital] and went to dwell in a monastery where, because of impure thoughts, he had wet dreams. Now, he had previously declared himself an arhat, but when he ordered a disciple to wash his soiled robes, the disciples spoke to him saying: An arhat is one in whom all the outows have been exhausted (*ksrava). How then, Master, is it possible that you still have such a thing? Mahdeva spoke to him, saying: I was aicted by Devaputramra. You should not think this strange. Now, the outows may broadly be classied into two categories: one due to delements (*klea) and the other due to impurities. The arhat has no outows due to delements, but he is yet unable to avoid those due to impurities. Why? Although the delements of arhats are extinguished, how can they be without urine, feces, tears, spittle, and the like? Now, the Devaputramras always hate the Buddhas teachings. Whenever they see someone who is cultivating goodness, they invariably attempt to ruin him. Even an arhat is aicted by them, and therefore I had an outow. They caused it. You should not be skeptical about this. This is termed the origin of the rst false view. Again, second false view. third false view. fourth false view.

The Jifayue sheku tuoluoni jing


Mahdeva had, indeed, committed a host of crimes. However, since he had not destroyed his roots of good, during the middle of the night he would reect upon the seriousness of his crimes and upon where he would eventually undergo bitter suerings. Beset by worry and fright, he would often cry out, Oh, how painful it is! His disciples who were dwelling nearby were startled when they heard this and, in the early morning, came to ask him whether he were out of sorts. Mahdeva replied, I am feeling very much at ease. But why, asked his disciples, did you cry out last night, Oh, how painful it is!? He proceeded to inform them: I was proclaiming the noble path (*ryamrga). You should not think this strange. In speaking of the noble path, if one is not utterly sincere in the anguish with which he heralds it, it will never become manifest at that moment when ones life reaches its end. Therefore, last night I cried out several times, Oh, how painful it is! This is termed the origins of the fth false view. .

In the translation above I have underlined the crucial phrases linking the story in the JSTJ with that in the Vibh. Despite this similarity, within a Chinese context we cannot see the source of the former in the latter. It is important to establish this, since otherwise one might well see the JSTJ as a purely native production. However, the Vibh which contains this story was not to be translated into Chinese by Xuanzang until 659, at the very least sixty-ve years after the JSTJ counting from the 594 terminus ante quem for the JSTJ in the Zhongjing mulu and the translation may postdate the JSTJ by as much as two centuries, if a mid-fth century date for the latter were to be accepted. It is also very dicult to imagine that the story could have been borrowed by the JSTJ from the earlier translation of the Vibh produced by Buddhavarman / Futuobamo in the rst half of the fth century, the last forty of whose one hundred juan were destroyed in a re, it is said. There are two reasons for this diculty. In the rst place, the remaining sixty juan of the earlier translation correspond to the rst 111 juan of Xuanzangs version, and the story of Mahdeva appears in Xuanzangs juan 99, well within the scope of the overlapping portion. Moreover, even in the unlikely event that the Buddhavarman translation originally did contain an account of


Jonathan A. Silk

Mahdeva as an Oedipal criminal, in order to accept that version as the basis for the story in the JSTJ one would have to assume not only that the name of the protagonist was changed, which is a trivial matter, but that a calumnious story associated with one of the villians of Indian Buddhist history was subsequently intentionally applied to the Buddha himself, albeit in a past life. This, I believe, is hardly credible. If, then, it is most unlikely, if not impossible, that the JSTJ found its inspiration in the Vibh in Chinese, because the story almost certainly did not appear in Buddhavarmans version and because Xuanzangs version clearly postdates the JSTJ, what of the Indian sources of the Vibh? I have argued, in a recent book, that the Vibhs narrative account of the schismatic Mahdeva is directly related to the story of Dharmaruci as preserved in the Divyvadna,116 itself unknown in China, and no other likely sources are now known to be extant. What is important, moreover, is that reference to the utterance of the phrase Oh, how painful it is! is absent from the story in the Divyvadna. Although in the Vibh this ejaculation plays a part in the recitation of the fth of the heretical Five Theses, and is related to the protagonists failure to overcome his delements, in the JSTJ the motivation is much more direct despair over his sinful state. The existence of the JSTJ appears to provide evidence either for a different Indian transmission of the basic story, probably unrelated to the episode of Mahdeva, or for positing an otherwise unknown Indic tradition upon which Xuanzangs Vibh translation was based, or to which it is related. However, there is some additional evidence that a structurally similar version of the story circulated, perhaps even in India, in which the story element of the verbal expression of despair played a part, although the evidence for this tradition comes very much later, and from far away.

See Silk 2008a. This also provides, inter alia, additional detail on other Indian versions of the same basic story, and some related, but brief, Chinese references. It is interesting, although almost certainly nothing other than dumb chance, that the Chinese fyu which forms part of the title of the Jifayue sheku tuoluoni jing, could represent Dharmaruci, as Paul Harrison pointed out to me.


The Jifayue sheku tuoluoni jing


Thirteenth century Tibetan presentations of the story of Mahdeva contain reference to this episode. A good example for our purposes is the rGya bod kyi chos byung rgyas pa (Extensive History of Buddhism in India and Tibet) of the rNying ma pa author mKhas pa ldeu:117
Then, 110 years after the passing of the Teacher, there was a Venerable Mahdeva who was born in a merchant family. While his father was gone on trade, he slept with his mother. When his father returned, having deliberated with his mother, he killed his father. Concerned about their bad reputation, they ed to another country. There was an arhat-monk whom they had earlier patronized. When they met him there, out of concern that he might have spread their bad reputation, through a stratagem they oered him an invitation and killed him by giving him poison. Then after the mother slept with another, [Mahdeva] became jealous, and killed his mother as well. Thus did he commit three of the sins of immediate retribution. Still, his outlook was not inverted. Having removed the impediments to his serious religious practice, going to another country he then requested initiation in the monastic communities, and this being given he was fully ordained [as a monk]. Since his intelligence and drive were great, he applied himself to religion, and thus he grew full of wisdom, such that the king of the land and all of the people honored him greatly. He then became lustful, and pridefully he lied, saying: I have obtained the fruit of arhatship. His merit increased, and the king offered him an invitation [to attend him]. There [at court] he became enamored of the kings consort. Since [she] saw him ejaculate, [she] asked: If one is a saint, one has cut o the delements, and thus does not produce semen, yet how is it that you produce semen? I am tormented by Mra. Even though I have become an arhat (*aaika), Devaputramra places obstacles in the way of my goodness. Because his disciples were given to idle chatter, he said to several of them: You have obtained the status of Stream Winner, or Arhat, Lone buddha or Renunciant.

Chab spel tshe brtan phun tshogs 1987: 98.20. See also, for example, Sa skya Paitas sDom pa gsum gyi rab tu dbye ba in Rhoton 2002: 325326, trans. 172174.



Jonathan A. Silk

Since he said that, his retinue asked: We dont know anything at all, so how are we able to obtain these great fruits? [He replied] Sure you have obtained them! and said many such things. On another occasion, having repented since he had lied in giving inverted teachings to his disciples, at night he was aicted, and called out Alas, alack, the great suering! The assembly heard this, and said What is the trouble? There is no trouble at all. Then why did you say alas, alack they asked. He said: I was thinking of the Noble Path. If one does not call out, it will not be clear to one. .

Now, it is possible that thirteenth century Tibetan authors may have based themselves on, or been inuenced by, traditions from the east, from China, as much as upon Indian legends.118 Therefore, even setting aside their late date, these stories do not necessarily constitute evidence for the Indian origins of this Oedipal tale in this conguration. In this light, the remarkable parallelism between the narrative in the JSTJ and that in the Vibh must be due either to the reliance of the former on some version of the Vibh (or, in turn, its source(s)) circulating in India or Central Asia, or to a parallel transmission of this story which, nevertheless, was less closely aligned with the Dharmaruci story traditions than with that associating the same basic story with Mahdevas fth thesis, that concerning the arhats nocturnal exclamation of pain.

V. Thinking about classication

What the presence in the JSTJ of the story of Mahdeva, under the guise of the mysterious Zhetatuo, tells us about the origins of the scripture appears to be the following: although it is possible that the story was known in China before the translation of the Vibh by Xuanzang, in a form that was either not written or subsequently became entirely lost, such that now we can nd traces of
See Silk 2008b for a consideration of the Tibetan materials relevant to the legend of Mahdeva.

The Jifayue sheku tuoluoni jing


it nowhere other than in the JSTJ, there can be little question but that the narrative core of the JSTJ more or less directly reects a genuine Indian tradition. And this is so no matter where the text as a whole was ultimately compiled or created. I will return to this point in a moment. We should consider here a number of related concerns. First is just how the JSTJ is to be classied, in terms of genre. It calls itself, or is called by the sources which now preserve it, and many but not all works which refer to it, a jing, generally indicating in Buddhist technical usage, a stra. Yet, it lacks the formal structural features which characterize a stra, the stock opening and closing, above all, something which is, probably not incidentally, characteristic of the class of separate compilations or abbreviated abstracts, bishng cho, mentioned above. It is possible that single juan independent versions do or did have these formal opening and closing formulas, which were removed when the content the dhra and narrative were embedded in other contexts, the works within which we now nd the JSTJ.119 At the same time, as we will see below, at least one later scholar, the thirteenth century Japanese monk Gynen, explicitly raises the point that this dhra is not a stra. The core of the work is, in fact, a Jtaka-style narrative, the central point of which is a basically typical self-promotion of the text itself. (It is worth pointing out that there is nothing tantric about the text at all, which is not surprising since there is nothing inherently tantric about dhras.) The point of the story comprising the text the story the text tells is the ecacy of the text itself or the dhra alone to extinguish sin. It argues for this ecacy, interestingly, in two ways. First is the conventional approach of simply promising to eradicate evil karma for one who upholds, reads and recites the text itself. In addition, however, it also argues in a more philosophical, or at least sophisticated, manner, oering the removal of obstacles to meditative development, leading to peace of mind and magical
In fact, such transformations were noticed traditionally. For instance, Huiyuan of the Jingyingsi noted how the opening of Kumrajvas translation of the Daabhmika was modied when it was placed in the larger Avatasaka collection; see Weimo yiji T.1776 (XXXVIII) 425b2426 (juan 1), cited by Funayama 2007: 3.


Jonathan A. Silk

attainments, the ability to y and to see all buddhas throughout the universe. It also, once again typically, grounds the authority of the text in the historical fact that it was preached by buddhas of the past; it is from these buddhas that our buddha, kyamuni, the narrator of the account, was able to receive the text during a distant previous life when he was an Oedipal criminal. The dhra beneted him, and thus he preaches it or we had better say, less technically, relates it to us, to anyone whom it might likewise benet. I do not think that the fact that the text as we have it is not structurally a stra had any implications for the authority with which it might have been vested, such that, for instance, its failure to contain a prefatory Thus I have heard gave license to Chinese compilers to treat it with less reverence that they might have treated scriptures that looked more canonical. Comparison with similar works allows us to think about questions like this, and we do have a number of such works, the presentation of which does not follow the canonically enshrined format of a stra. Perhaps the most famous of these is the so-called Platform Stra (short form, Tan jing ), but this is not the only example. Therefore, it does not seem likely that the mere structural form of the JSTJ necessarily had, in this respect at least, any direct impact on its reception. A more interesting question may be how Indian this work is, not from the perspective of traditional Chinese receptions of the text, but from a modern historical viewpoint. Now, the question of the status of works containing both Indic and local elements is not a new one, either to modern or traditional scholarship. Seventy years ago Alexander von Stal-Holstein discussed the status of the so-called Larger ragama-stra (Dafoding rulai miyin xiuzheng liaoyi zhupusa wanxing shoulengyan jing ). In his study he noticed the introduction provided this scripture by the Qianlong emperor in 1770 (or much more likely, in his name).120 The emperor argued (or assumed?) that since the *Sugatoadhra found within the stra entirely agrees with an Indian text,
Stal-Holstein 1936. I was reminded of this article by a remark in Kapstein 2007.

The Jifayue sheku tuoluoni jing


the stra is manifestly authoritative (dei nang gi bde gshegs gtsug tor gyi gzungs rgya gar gyi dpe dang shin tu grig pas || mdo di tshad ma yin par mngon). The question was cast as one of genuine, Indian scripture or Chinese forgery. In considering the arguments of a Chinese scholar published in 1934,121 however, Stal-Holstein showed himself more tolerant of ambiguity. Rejecting the idea that the inclusion of a dhra with questionable transliterations calls into question the status of a work in toto, he wrote: Neither can the Emperors view be accepted in its entirety (the dhra proves the authenticity of the larger ragama as a whole), but we must admit that the ragama (or Sugatoa) dhra makes the thesis of the ultra-sceptics (the larger ragama is a Chinese forgery from beginning to end) equally untenable. Although Stal-Holstein evidently did not explicitly reject the dichotomy between authenticity and forgery entirely, he nevertheless seems to have traveled quite some distance down the path I am now exploring. To conclude our investigation so far, then, despite the considerable circumstantial evidence, beginning with stra catalogues which are unable to provide any details about the translation of the work, and including the irregular mode of its transmission apparently centrally, although not exclusively, as an intrusion within other works, suggestive of an apocryphal origin, we must conclude that the work nevertheless is, at least in part, genuinely and authentically Indian, even if we would question whether it should be called a stra as such (although Chinese tradition does so regard it). So how shall we classify it? Is it a translation, then? There are some indications casting doubt on the supposition that it is a strict translation of an Indian work, in the sense of preserving the formal features of an Indic original, or of representing a literal translation as such. In other words, there are reasons to doubt that there ever existed a text in India (or the Indic world?) having the contents, and in the shape, of the JSTJ as we now know it. This would seem to rule out calling it a translation as such.

Li Yishao , Foxue weishu bianlue , in Guoli zhongyang daxue wenyi congkan 1/2: 746 [not seen].



Jonathan A. Silk

Is it, then, a non-translation? The presence in the work of a narrative the Indic origins of which can be decisively proved, and which appears on the basis of presently available evidence to have been otherwise unavailable in China in the period to which the creation of the JSTJ belongs, strongly suggests that the JSTJ transmits genuinely Indian materials (directly, and not on the basis of some previous Sinitic tradition), even if their presentation may dier from the form(s) in which those materials were transmitted within the Indian world. (This, incidentally, leaves aside the whole issue of the linguistic form of the dhra itself, concerning which, at this point, as confessed above, I can conclude nothing.) In the absence of other evidence, it does certainly seem, then, that the JSTJ cannot be termed a non-translation in its entirety. Is the text, then, both a translation and a non-translation? Is it, in other words, meaningful to say that the JSTJ is, simultaneously, both a translation and a non-translation, containing both genuine Indic and non-Indic elements? Or does the least bit of non-Indic content, like a crumb of bread on a Passover plate, spread contagion through the whole? What do we say of a work that was put together somewhere out of imported parts is a Toyota assembled in Nebraska a Japanese or an American car? I am reminded here of what the author Paul Theroux said when asked by an interviewer who seemed keen on the genre as a type of history what he would call a work of historical narrative in which conversations to which the author did not have access are nevertheless recreated. Theroux replied: I call it ction. This is one kind of answer, of the one drop pollutes the pot kind, but it need not be ours. Taken to its logical extreme, this approach suggests that the mere act of translation denitively separates a work from its source, since any localization at all fundamentally cannot but alter a work in uncontrollable ways. This very real philosophical issue operates on a more basic level than the comparatively gross one I engage here, and is a problem for another day. Finally, should we settle for saying that such works are neither translation nor non-translation, that, after all, the dichotomy itself is misleadingly, or even impossibly, posed? Is the most fruitful approach at this point to decide that we may have been asking the

The Jifayue sheku tuoluoni jing


wrong question, trying to t a square peg in a round hole? What would it do to our appreciation of such scriptures were we to set aside the binary opposition between translation and non-translation altogether? Could this help us to see something we have previously been unable to visualize about the process of inculturation of Buddhism in China, about the creation of authority and the locus or loci of creativity, the nexus of which is centered precisely in this moment of intercultural scripture production? Despite my playful evocation of the tetralemma, then, in the end these four possibilities may each, in their own ways, be equally true. And this very fact, I believe, has the potential to tell us an enormous amount about scripture production in Chinese Buddhism, and about the process of the creation of Chinese Buddhism as a whole. I will return to this basic point in the conclusion below.

VI. The legacy of the Jifayue sheku tuoluoni jing

When I began this research, which grew out of my work on Mahdeva, I had no notion that I was doing more than following up a minor detail in an otherwise itself already suciently complicated quilt of stories and historical accounts. But as sometimes happens when one begins to tug absent-mindedly on a loose thread, unexpected things emerged; I discovered that this little dhra may have been more popular than I had supposed. We noticed above that it was known to a few Chinese authors. Aside from catalogue entries, the earliest such reference I know is that of Kuiji (632682), in his Dasheng fayuan yilin zhang , in which we nd the following:122
It says in the Mahsaniptastra: The Buddha said: Innite aeons ago I was still in the stage of being an ordinary person, named Zhetatuo. Living in the land of Jiatouluo, I engaged in sales and peddling, but was dishonest and lied, and did all sorts of evil deeds. Its impossible to fully detail them. At that time, stupidly insensitive, I killed my father and made love to my mother. Everyone in the land knew of this, and considered me not dierent from the six kinds of


T.1861 (XLV) 307a1522 (juan 3).


Jonathan A. Silk

beasts. The king of the land desired me to be killed. Fearful, I left and set out for other regions. I became a ramaa [and practiced?] for thirty-seven years. Because of the obstacle of the ve sins of immediate retribution, my mind could not nd peace. Later, in order to beg for alms I took a large bowl. Within the bowl there was this dhra called Collecting the Joy of the Teachings and Getting Rid of Suering and Trouble. For a year I recited it without ceasing. Then I was able to attain meditative concentration. 1


Here we have a relatively brief synopsis of the core narrative. At almost the same time, the Korean scholar ijk cited the text in his Posal kyebon-so at greater length, as follows:123
Therefore, the Jifayue jing says: Zhetatuo committed the ve sins of immediate retribution, and was made to suer by the king. At that time he was astonished and fearful, and became a ramaa. Living in another country, he practiced the ten good [precepts], did seated meditation and studied the Way. He wept night and day for thirty [sic!] years. Because of the obstacle of the ve sins of immediate retribution, his mind could not nd peace. In a mountain cave, he was always crying out Oh, how painful it is! Oh, how painful it is! With what mental [technique?] can I get rid of this pain? At one time, he was about to go out begging for alms. Sobbing with grief, he went down from the cave, toward a village. At that time he found a large bowl on the road. Within the bowl he saw there was the Dhra on Collecting the Joy of the Teachings and Getting Rid of Suering. Taking this scripture, without begging for alms he immediately returned to the cave joyously. He burned incense and oered worship, and piteously weeping and venerating it, he practiced the recitation of this scripture within the cave. After a year had passed, he was rst able to extin-

T.1814 (XL) 657a620 (juan 1). The text is also referred to slightly later in the commentary of Taehyn , Pmmanggyng kojkki T.1815 (XL) 716b10 (juan xia).


The Jifayue sheku tuoluoni jing


guish his transgressions. But because of the obstacle of his karma, he was not able to get it to enter his mind. At that time, he [ritually] washed and practiced for seven days. Like a beginning student, he worried and was not the least bit at ease. Practicing for seven days, his worry was unchanged. Mentally suering, he did not know what to think. Therefore he contemplated the written form of the dhra, and after many perturbations, his mind was suddenly settled. Then he was delighted. Like a person who nds a hundred thousand gold jn on the ground, that others did not know were there, inside he was endlessly joyous. Practicing for many years, he was able to y without obstacle and see the buddhas of the three worlds in the ten directions.

Some centuries later, the Khitan-Liao monk Feizhuo (d. 1063) in his Sanbao ganying yaolelu quoted or closely paraphrased as follows.124
Innite aeons long ago when kyamuni was an ordinary person he was named Zheta [sic!]. Living in the land of Jiatouluo, he engaged in sales and peddling, but was dishonest and lied, and did all sorts of evil deeds, killing his father and making love to his mother. After many years, all the people of the land all came to know of it, and they loudly proclaimed: Zheta killed his father and made love to his mother. When [he] thought about it, [he considered himself] no dierent from the beasts. Then at night he lept over the city walls, and ed and hastened toward a deep marsh. The king of the land, Pisheluo, announced to the people of the country: This fellow has committed sexual perversities and immoral acts. Whoever can lay hands on him will be handsomely rewarded. Then

T.2084 (LI) 839c723 (juan zhong).


Jonathan A. Silk

each and every person in the country eagerly desired to get hold of him. He left the country, and became a ramaa. In another country he practiced seated meditation and studied the Way. He wept day and night for thirty-seven years. Because of the obstacle of having committed the ve sins of immediate retribution, his mind [was never at rest].125 For thirty-seven years he lived in a cave in the mountains, raising his voice and crying out in distress. Sobbing with grief, he went down from the cave to beg for alms, and then, on the road, found a large bowl. Within it there was a stra box, but only one stra inside, the Mahsanipta Dhra Stra on Collecting the Joy of the Teachings and Getting Rid of Suering. This scripture can remove the great transgression produced by the ve sins of immediate retribution committed through hundreds of millions of aeons of rebirth. If there is a person who upholds, reads and recites it, he will never fall into the three unfavorable realms. Why? Because it was preached by the buddhas of the past when they attained nirva. At that time, having obtained this scripture, he did not go begging for alms, but immediately returned to the cave joyously. He burned incense and oered worship, and piteously weeping and venerating it, he practiced the recitation of this scripture within the cave. After a year had passed, he was rst able to [extinguish his transgressions].126 But because of the obstacle of his karma, he was not able to make it enter his mind. After passionately practicing for years, he was able to y without obstacle and see the buddhas of the ten directions. [text missing?]

Perhaps ve or six characters appear to have dropped from the text here. 126 Again, some text is evidently missing here.


The Jifayue sheku tuoluoni jing


[text missing?]

As is evident, Chinese, Khitan and Korean authors all noticed the work, or to be more cautious, all referred to its story. The same was true further east as well. The ninth-century Japanese Tendai scholar Annen notes the JSTJ in his catalogue Shoajari shingon mikky burui sroku as an example of a text which contains methods to ward o disease jobyh , apparently