, Tokyo 1933-1936;
photomechanical
reprint edition, Tokyo, 1971.
Abhidharmakosabhasyam-Vol-1-Vasubandhu-Poussin-Pruden-1991
The Kosa and its commentaries, Sanskrit, Tibetan, and Chinese sources.
a. Abhidharmakosavydkhyd, Bibliotheca Buddhica, Sphutdrtha Abhidharma- kocavydkhyd, the work of Yacpmitra, first Kocasthdna, edited by Prof. S. Levi and Prof. Th. Stcherbatsky, 1st fasc, Petrograd, 1918; 2nd fasc by Wogihara, Stcherbatsky and Obermiller, (part of the second chapter), Leningrad, 1931.
Text of the third chapter, kdrikds and vydkhyd, in Bouddhisme, Cos- mologie . . . L. de La Vallee Poussin [with the collaboration of Dr. P. Cordier], Brussels, 1914-1919.
b. Tibetan translation of the Abhidharmakocakdrikdh and of the Abhi- dharmakocabhdsya of Vasubandhu, edited by Th. I. Stcherbatsky, 1st fasc. 1917,2nd fasc. 1930.
3. Tibetan sources, Palmyr Cordier, Catalogue de fonds tibetain de la Biblioteque Nationale, third part, Paris 1914, p. 394 and 499:
a. Abhidharmakosakadrikd and Bhdsya of Vasubandhu, Mdo 63, fol. 1-27, and fol. 28---Mdo 64, fol. 109.
b. Sutrdnurupd noma abhidharmakosavrttih of Vinltabhadra, 64, fol. 109-304.
c. Sphutdrtha ndrna abhidharmakosavydkhyd of Ya^omitra, 65 and 66. This is the commentary preserved in Sanskrit.
d. Laksandnusdrini ndma abhidharmakosattkd of Purnavardhana, a student of Sthiramati and master of Jinamitra and Silendrabodhi, 67 and 68.
e. Updyikd ndma abhidharmakosattkd of Samathadeva, 69 and 60, fol. 1-144. f. Marmapradipo ndma abhidharmakosavrttih of Dignaga, 70, fol. 144-286.
g. Laksandnusdrini ndma abhidharmakosattkd, an abridged recension of the
"Brhattika," above item d, 70, fol. 286-316.
h. Sdrasamuccayo ndma abhidharmavataratikd, anonymous, 70, fol. 315-393. i. Abhidharmdvatdraprakarana, anonymous, 70, fol. 393-417.
j. Tattvdrtho ndma abhidharmakosabhdsyatikd of Sthiramati, 129 and 130.
4. Abhidharmakofasdstra, of Vasubandhu, trans, by Paramartha in the period
564-567, Taisho volume 29, number 1559, p. 161-309; trans, by Hsiian-tsang, 651-654, Taisho volume 29, number 1558, p. 1-160.
The references in our translation are to the edition of Kyokuga Saeki, the
? Kando Abidatsuma Kusharon (Kyoto, 1891), the pages of which correspond to those of the Ming edition, a remarkable work which notably contains, in addition to interesting notes of the editor, copious extracts 1. from the two major Chinese commentators, 2. from the Vibhdsd, 3. from the commentary of Samghabhadra, and 4. from the work of K'uei-chi on the Trimsikd.
5. Among the Chinese commentaries on the Kosa:
a. Shen-t'ai, the author of a Shu: the Chil-she lun shu, originally in twenty Chinese volumes, today only volumes 1, 2, 4, 5, 6, 7, and 17 are extant; Manji Zoku-zokyo-1. 83. 3-4.
b. P'u-kuang, the author of the thirty-volume Chil-she lun Chi; TD 41, number 1821.
c. Fa-pao, the author of a thirty-volume Chil-she lun Shu; TD 41, number 1822.
Two other disciples of Hsiian-tsang, Huai-su and K'uei-chi, have written commentaries on the Kosa which are lost. P'u-kuang has also written a short treatise on the teachings of the Ko/a.
d. Yuan-hui wrote a thirty-volume Shu on the Kdrikds of the Kosa, a work with a preface written by Chia-ts'eng and dated before 727; this work, the Chil-she lun sung Shu (var. Chil-she lun sung shih), is preserved in TD volume 41, number 1823. This Shu "was commented upon many times in China and very widely disseminated in Japan; it is from this intermediary text that Mahayanists in general draw their knowledge of the Kola. But from the point of view of Indology, it does not offer the same interest as the three preceding com- mentaries. "
Hsiian-tsang dictated his version of Samghabhadra to Yuan-yu. There are some fragments of a commentary written by him.
6. Gunamati and the Laksandnsdra.
Gunamati is known through his commentary on the Vydkhydyukti\ many fragments of this commentary are quoted in the Chos-'byun of Bu-ston, trans. Obermiller, 1931. It is mentioned four times by Yasomitra in his Abhidharmakofavydkhyd.
a. Introductory stanzas: Gunamati comments on the Kosa, as has Vasumitra; Yasomitra follows this commentary when it is correct.
b. "Gunamati and his disciple Vasumitra say that the word nutrias is declined in the fourth case. But when the word namas is not independent, we have the accusative. This is why this master (Vasubandhu), in the Vydkhydyukti, says, 'Saluting the Muni with my head' . . . " (Kosa, Vydkhyd, i p. 7).
c. Gunamati holds that the Kosa wrongly teaches that "Conditioned things,
Poussin 9
? 10 Introduction
with the exception of the Path, are sasrava (Kosa, i. 4b)," for all of the dharmas, without exception, can be taken as an object by the dsravas(Vydkhyd i, p. 13).
d. On the subject of the continuity of the mental series, "the master Gunamati, with his disciple the master Vasumitra, through affection for the doctrine of his own nikaya, instead of confining themselves to explaining the Kosa, refute it" (Kosa,iiL Ua-b, note).
N. Peri (Date, 41) recalls that Burnouf mentioned (Introduction, 566), according to Yasomitra, the commentary of Gunamati. He adds: "An author very rarely quoted. His Laksananusarasastra (Taisho 1641) forms part of the Canon, where it is classified among the HInayana works. It summarizes the ideas of the Kosa, and then presents his own opinions on several points. The Hsi-yu-chi, after having listed him among the celebrated monks of Nalanda (TD 51, p. 924a2), tells us that he left the monastery where he had been living in order to move to V alabhF(p. 936c2). "
Taisho 1641 is only an extract of the treatise of Gunamati, the chapter which examines the sixteen aspects of the truths (Kosa, vii. 13): "Do we have sixteen things or sixteen names? The masters of the Vibhasa say that sixteen names are posited because there are sixteen things. But the siltra-upadesa masters say that there are sixteen names, but only seven things; four things for the first Truth, one thing for each of the three others. In the beginning the Buddha promulgated the Upadesasutra. After the disappearance of the Buddha, Ananda, Katyayana, etc. , recited that which they had heard. In order to explain the meaning of the Sutra, as disciples do, they composed a sastra explaining the Sutra, which is thus called a siltra-upadesa. Then the Vibhasa extracted an upadesa from that which was to be found [in this upadesa']; since it only indirectly comes from the Sutra, it is not called a sutra-upadefa. "
Gunamati continues as in the Kosa, vii. 13a, "According to the first explana- tion, anitya, impermanent, because it arises dependent on causes (pratyayd- dhinatvdt)" And he comments, "Conditioned things, without force, do not arise in and of themselves . . . "
The first volume ends, "The thesis of Vasubandhu is similar to the meaning of the siitra-upadesa masters. "
The second begins, "The author says, 1 am now going to give the explanation of what I believe. Anitya, impermanent, because, having arising, it has extinction. Conditioned things, having arising, and extinction, are not permanent. Arising is existence . . . "
The treatise touches on diverse points of philosophy, the absence of dtman, etc. In this work we encounter some very interesting notes, for example (Taisho,
? page 168b9), "In the Hinayana, the pretas are superior to animals; in the Mahayana, the opposite. In fact, the pretas are enveloped in flames . . . "
It is curious that the title of the work of Gunamati, literally Laksandnusd- rasdstra, is exactly identical to that of the book attributed to Purnavardhana in the Tanjur. We have Gunamati, a teacher of Sthiramati, and Purnavarudhana, a
25
student of Sthiramati.
7. Sthiramati, a student of Gunamati, defended the Kosa against Samgha-
bhadra. "His commentary on the Kosa is mentioned many times by Shen-t'ai, P'u-kuang, and Fa-pao in their work on the same text. The precise manner in which they-quote it, in which they note and discuss its opinions, causes us to believe that Hsiian-tsang may have brought it to China, and perhaps they themselves had also read it" (N. Pe'ri, Date, 41). Sthiramati, the author of the Tsa-chi, is one of the great masters of the Vijnaptimitrata system.
There exists (Taisho 1561) a small treatise by Sthiramati (transcription and translation) entitled Kosatattvdrthatikd or Abhidharmakosasdstratattvdrathaptkd, which is doubtless an extract of a voluminous work of the same name and by the same author preserved in Tibetan (Cordier, 499).
We observed, at the beginning, the commentary on the seven points indicated in the introductory stanza of the Kosa.
On the wisdom of the Buddha, superior to that of the saints, the author quotes the Kalpanamanditika stanza (Huber, Sutrdlamkara), Kosa, i. l, vii. 30; and recalls the ignorance that Maudgalyayana had of the place where his mother had been reborn, Kosa, i. 1.
In order to demonstrate the thesis of the Kosa that sraddhendriya can be impure, ii. 9, the author quotes at length the sutra on the request of Brahma to the Buddha (setting into motion the Wheel of the Law), a sutra briefly indicated by V asubhandu.
The work ends with some remarks on the duration of life: The stanza says: "Among the Kurus life is always 1,000 years in length; half of this to the west and the east. In this continent, it is not set: atits end, some ten years; in the beginning, without measure" {Kosa, iii. 75-77), "There are, in fact, in this world, some beings who have extra meritorious actions and who make the resolution, 'May I have a long life! , without desiring more precisely, 'May I live one hundred years, ninety years, eighty years! ' Or rather some venerable persons, parents and friends, say, 'May you live long! ' without saying more precisely how long a time. If one makes similar vows, it is because the actions done by persons of this continent are associated with thoughts of desire. The Sutra says, 'Know, oh Bhiksus that the length of life was over 80,000 years under Vipasyin, 20,000 years under Kasyapa;
Ponssin 11
? 12 Introduction
the length of life is now 100 years; few will go beyond this, and many will have less. ' If the length of life is not set, why does the Blessed One express himself in this way? . . . " The treatise concludes with the well-known stanza: sucirna- brahmacarye'smin . . . (Kosa, vi. 60a).
8. Samghabhadra has written two works.
The first (TD 29, number 1562), the title of which is transcribed into Chinese as Abhidharmanydyanusdrasdstra--or perhaps better as Nydydnusdro ndma Abhidharmaidstram--is a commentary which reproduces without any changes the Kdrikds of the Abhidharmakosa. But this eighty-volume commentary criticizes the Kdrikds, which present the Vaibhasika doctrine by noting them with the word kila, which means "in the words of the School"; it refutes the Bhdsyam, the auto-commentary of Vasubandhu, when this work presents views opposed to those of the Vaibhasikas,and it corrects them when it attributes to the Vaibhasikas views which are not theirs.
The title of the second treatise (TD 29, number 1563) is not completely transcribed: Abhidharmasamaya-hsien-sdstra or Abhidharmasamaya-kuang-sastra. J. Takakusu proposes Abhidharmasamayapradipikds'dstra, which is not bad;
however pradipa, "lamp," is always teng, and we have for hsien the equivalents prakdsa and dyotana.
This is a forty-volume extract from the Nydydnusdra, from which all polemic
is excluded and which is thus a simple presentation of the system (samaya) of the
Abhidharma. It differs from the Nydydnusdra by the presence of a rather long
introduction, in seven stanzas and prose, and also by the manner in which it treats
the Kdrikds of Vasubandhu: these Kdrikds are either omitted (ii. 2-3) or corrected
( i l l , 14) when they express false doctrines or when they cast suspicion on true
26 doctrines by the addition of the word kila.
Samghabhadra is an innovator, and K'uei-chi distinguishes the earlier and the later Sarvastivadins, Siddhi, 45 (theory of atoms), 65 (laksanas of "conditioned things"), 71 (the viprayukta called ho-ho)y 147 (vedand? ), and 311 (divergent Sarvastivadins, on adhimoksa).
***
(Additions to the Bibliography, by Hubert Durt. )
The following titles are editions of texts and works related to the Abhi- dharmakosabhasyam, which have appeared in print since the first appearance of de La Valle'e Poussin's French translation (1923-1931).
? Sanskrit:
Gokhale, V. V. , The Text of the Abhidharmakosakdrikd of VasubandhuJournal
of the Bombay Branch, Royal Asiatic Society, n. s. , vol. 22,1946, p. 73-102. Pradhan, P. , Abhidharm-Koshabha/ya of Vasubandhu, Tibetan Sanskrit Works
Series, vol. vrn, K. P. Jayaswal Research Institute, Patna, 1967; 2nd edition, revised with introduction and indices, by Dr. Aruna Haldar, 1975.
Shastri, Swami Dwarikadas, Abhidharmakosa & Bhdsya of Acharya Vasu- bandhu with Sphutdrthd Commentary of Acdrya Yasomitra, Part I (I and II Kosasthdna), critically edited, Bauddha Bharati, Varanasi, 1970.
Wogihara, Unrai, Sphutdrthd Abhidharmakosavyakhya, 2 vols.
, Tokyo 1933-1936; photomechanical reprint edition, Tokyo, 1971.
Tibetan:
Otani University, The Tibetan Tripitaka, Peking Edition, vol. 115 (number
5590), to Vol. 119 (number 5597), Suzuki Research Foundation, Tokyo, 1962.
Qiinese:
Takakusu, J. K. Watanabe, The Tripitaka in Chinese, vol. 29 (numbers 1558 to
1563), The Taisho Issai-kyo kanko kwai, Tokyo 1926.
Otani University, Index to the Taisho Tripitaka, no. 16, Bidon-bu III (vol. 29), Research Association for the Terminology of the Taisho Tripitaka, Tokyo, 1962.
Funabashi, Suisai and Issai Funabashi, Kando Abidatsuma Kusharon Sakuin, Kyoto, 1956. This index is based on the Chinese version of Kyokuga Saeki--the Qiinese version used by de La Valine Poussin--, the Kand6-bon Kusharon, in thirty volumesJCyoto, 1887.
Hi. The Date of Vasubandhu. The Former Vasubandhu.
We shall not undertake here a bibliography of Vasubandhu. But his treatise,
the Pratityasamutpddavydkhyd (Cordier, iii. 365), calls for the attention of the
reader of the Kosa. G. Tucci has published some fragments of this work (JRAS.
1930, 611-623) where the twelve links in the chain are explained in detail, with
numerous quotations from scriptures. G. Tucci also proposes to publish the
21
Trisvabhdvakarika and some parts of the commentary to the Madhydntavibhdga.
Concerning the "definition of pratyaksa by Vasubandhu," vdsubdndhava pratyaksalaksana, known through the Tdtaparyatikd, 99, and the Vddavidhi attributed to Vasubandhu, see the articles by G. Tucci, A. B. Keith, R Iyengar, JRAS. 1929, 473; Ind. Hist. Quarterly, 1928, 221; 1929, 81; Stcherbatski, Logic, ii. 161,382; G. Tucci, Maitreya [ndtha] et Asanga, 70-71, and finally Pramdna-
Poussin 13
? 14 Introduction
samuccaya, chap, i, by R. Iyengar, pp. 31-35. It appears that Dignaga denies the authorship of the Vddavidhi to Vasubandhu,in spite of universal opinion, and the Ttka quotes Kosa ii. 64, which contradicts the above-mentioned definition of pratyaksa. There are also numerous passages of the Vydkhyayukti in the Chos- 'byun of Bu-ston (above p. 16 ).
Wassiliew, Buddhismus, 235 (1860): "Life of Vasubandhu. "
Kern, Geschiedenis, trans. Huet, ii. 450.
S. Levi, JA. , 1890, 2. 252; Theatre indien, 1890, i. 165, ii. 35; "Donations
religieuses des rois de ValabhT' {Htes Etudes, vii, p. 97); "Date de Candragomin," BEFEO. ,1903,47;Sutrdkmkdra, trans,preface,2-3,1911.
Biihler, Alter der indischen Kunst-Poesie, p. 97,1890.
J. Takakusu, "Life of Vasubandhu," Voung-pao, 1904; "A Study of Para-
metria's Life of Vasubandhu and the date of Vasubandhu," JRAS. , 1905; "Sdmkhyakdrikd," BEFEO, 1904.
f
Wogihara, Asanga s Bodhisattvabhumi, 14, Strasbourg thesis, Leipzig, 1908.
Noel Peri, "A propos de la date of Vasubandhu," BEFEO. , 1911, 339-392.
Pathak, Bhandarkar, Indian Antiquare, 1911 - 1912 (V Smith, History, 3rd edition 328,4th edition, 346).
B. Shiiwo [Benkyo Shiio], Dr. Takakusu and Mr. Peri on the date of Vasubandhu (270-350), Tetsugaku Zasshi, Nov. -Dec. 1912.
Winternitz, ]eshichte, ii. 256 (1913), iii. 693 (1922).
H. . Ui, "On the Author of the Mahdydnasutrdlamkdra" Z. fur Ind. und Iranistik, vi. 1928,216-225.
A group of articles, many of which are summaries of articles written in Japanese, in Melanges Lanman (Indian Studies in Honor of Charles Rockwell
Lanman), 1929; J. Takakusu, Date of Vasubandhu, the Great Buddhist Phi- losopher, T Kimura, Date of Vasubandhu Seen from the Abhidharmakosa', G. Ono,DateofVasubandhuSeenfromtheHistoryofBuddhisticPhilosophy,H. Ui, Maitreya as an Historical Personage. Further, mention of the opinions of B. Shiiwo, S. Funabashi, E. Mayeda, S. Mochizuki.
"H. P. Sastri pointed out the historicity of Maitreyanatha from the colophon of the Abhisamaydlamkarakdrika, which is a commentary, from the Yogacara point of view, on the Panaivimsatisdhasnkd-prajfM-pdramftd-sutra by Maitreyanatha" Kimura, Origin of Mahdydna Buddhism, Calcutta, 1927, p. 170).
The date of Vasubandhu is bound to that of Asanga, his brother. Now some parts of the Yogasdstra, the work of Asanga, were translated into Chinese in 413-421, and in 431. However, the opinion is accepted among Japanese scholars that the works attributed to Asanga, writing under the inspiration of the future
? Buddha Maitreya, are in reality the works of a master Maitreya, an dcdrya, "an historical personage. " This thesis permits us to strip Asanga of one part of the library of which we thought he was the pious redactor, and to place him, along with his brother Vasubandhu, toward the middle or end of the 5th century, or--why not? --towards the 6th century. "If a scholar named Maitreya is found to be the author of those works hitherto attributed to Asanga, then the date of the latter ought to be shifted later, at least by one generation, if not more. The ground for an earlier date for Vasubandhu should give way altogether" (Takakusu, Melanges Lanman, 85).
H. Ui, in Philosophical Journal of the Imperial University, Tokyo, number
411, 1921, takes into account the arguments, developed afterwards in his Studies
of Indian Philosophy, i. 359, summarized in Melanges Lanman. These arguments
appear to be weak and, to my mind, non-existent (Note bouddhique, xvi, Maitreya
et Asanga, Ac. Royale de Belgique, January 1930). I do not think that they gain any
force from the observations of G. Tucci ("On some aspects of the doctrines of
Maitreya-[natha] and Asanga," Calcutta Lectures, 1930). The tradition of the
Vijnaptimatrata school establishes, as Tucci observes, the lineage Maitreya-
natha-Asanga-Vasubandhu, but Maitreyanatha is not the name of a man, but
rather "He who is protected by Maitreya"; ndtha is a synonym of buddha, or more 18
precisely of bhagavat. The commentary of the Abhisamaydlamkdra (p. 73 of the Tucci edition) gives to Maitreya the title of bhagavat in one place where he explains how "Asanga, in spite of his scriptural erudition and his insight (labdhddhigamo'pi, Kosa, viii. 39), did not understand the Prajndpdramita and lost heart. Then the Bhagavat Maitreya, for his sake (tarn uddisya) explained the Prajndpdramita and composed the treatise which is called the Abhisama- ydlamkdrakdrikd. " It is with the title of the Maitreyanatha that Santideva designates the saint who, in the Gandavyilha, explains to the pilgrim Sudhana the virtues of "the Bodhi mind" (Bodhicarydvatdra, i. 14, Rajendralal Mitra, Buddhist Nepalese Literature, 92). If the School holds as sacred, as dryddesand, the treatises of Asanga, it is because the Bhagavat Maitreya has revealed them. That the Tibeto-Chinese tradition varies in its attributions, sometimes naming as author a revealing deity, sometimes an inspired master, does not pose any difficulty.
***
The biography of Vasubandhu (by Paramartha) is not without its difficulties. The Kosa excited the criticism of Samghabhadra who, in his large Nydydnusdra, brings up innumerable heresies of a Sautrantika character which mar the work of
Poussin 15
? 16 Introduction
Vasubandhu. We are told that Vasubandhu refused to enter into controversy: "I am now already old. You may do as you please" (Takakusu's version). But we are also assured that Vasubandhu was then converted to the Mahayana by his brother Asanga, that he decided to cut out his tongue in order to punish it for not confessing the Mahayana earlier, and, more wisely, that he wrote numerous treatises wherein the doctrines of the Mahayana were brilliantly elaborated.
Yasomitra, the commentator on the Ko/a, says that the expression purva- caryas, "former masters," of the Kosa, designates "Asanga, etc. " {dsangapra- bhrtayas). N. Peri thinks that Yasomitra means to designate the school of the Purvacaryas by their most illustrious name, and that the text does not imply that Asanga is in fact purva relative to Vasubandhu (see my Cosmologie bouddhique, p. ix).
The Kosa was only translated in 563, whereas the work of Dharmatrata, an
imperfect draft of the Kosa, was translated in 397-418, 426-431, and 433-442.
Takakusu observes, "If the Kosa had existed, why did so many translators linger
over the book of Dharmatrata? {Melanges Lanmari). And it is difficult to give a 29
pertinent answer to this question.
But it appears almost impossible to believe that Paramartha the biographer of
Vasubandhu and first translator of the Kosa, arriving in China in 548, erred when he made the author of the Kosa the contemporary and the brother of Asanga. It is a hopeless hypothesis to identify the brother and the convert of Asanga with the former, or earlier, Vasubandhu.
One should admit the existence and the "Abhidharmic" activity of an earlier Vasubandhu. The problem, which I have taken up in the preface to Cosmologie bouddhique (above, p. 6), has been taken up again by Taiken Kimura, "Examen lumineau de 1'Abhidharma" (contents in Eastern Buddhism, iii, p. 85), fifth part: "On the sources of the Kosa! * We can see a summary of his conclusions in Melanges Lanman. Subsequently, see Note Bouddhique xvii, Acad, de Belgique: "Vasubandhu l'ancien. "
Yasomitra, in three places {Kosa, i. 13, iii. 27, iv. 2-3), recognizes in a master refuted by Vasubandhu the author of the Kosa (and a disciple of Manoratha
30 according to Hsuan-tsang), a "Sthavira Vasubandhu, the teacher of Manoratha,"
an "earlier master Vasubandhu," vrddhdcdryavasubandhu. P'u-kuang (Kimura, Melanges Lanman, 91) confirms Yasomitra, and designates the master in question under the name of "the earlier Vasubandhu, a dissident Sarvastivadin master. "
On the other hand, the gloss of the initial five stanzas of the treatise of 31
Dharmatrata, the re-edition of the Abhidharmasara of DharmasVI, attributes an edition of the same book in 6,000 verses to Vasubandhu. These stanzas and this
? gloss are not very clear. Kimura has studied them (Melanges Lanman); I have amended his interpretation (Note bouddhique xvii).
Iv. The Seven Canonical Treatises of the Abhidharma.
The Sarvastivadins recognize the authority of seven Abhidharma treatises,
"the word of the Buddha. " Among them, the Abhidharmikas, "who only read the 32
Abhidharma with its six feet," are distinct from the Vaibhasikas "who read the Abhidharma. "
33
The Abhidharma "with its six feet" is the great treatise of Katyayaniputra,
entitled the Jndnaprasthdna, upon which the Vibhdsd is a long commentary, and six treatises the order and authorship of which vary somewhat according to our sources. Following the order of Abhidharmakofavydkhya, there are: the Prakara- napdda of Vasumitra, Vijndnakdya of DevaSarman, Dharmaskandha of Sariputra (or of Maudgalyayana, according to Chinese sources), Prajnaptisdstra of Maudgal- yayana, Dhdtukdya of Purna (or of Vasumitra, Chinese sources), and Sangiti-
parydya of Mahakausthila (or of aariputra, Chinese sources).
One should note that the Tibetans list the Dharmaskandha first, and the
Jndnaprasthdna only as sixth: "The Tibetans seem to regard the Dharmaskandha as the most important of all. " This is also the opinion of Ching-mai (664 A. D. ), the author of the Chinese colophon (Takakusu, 75,115).
Takakusu,in"OntintAbhidharmaLiterature"(JPTS,1905),bringstogethera number of details on these seven books which Burnouf was the first to list; he gives the contents of the chapters of each of them. The remarks which follow are an addition to this fine work.
34 a. Jndnaprasthdna.
1. According to Hsiian-tsang, Katyayaniputra composed this Sastra in the
35 monastery of Tamasavana 300 years after the Parinirvana (the fourth century).
However, the Vibhdsd (TD 26, p. 21c29), commenting on the Jndnaprasthdna (TD 26, p. 918) says, "When the Bhadanta composed the Jrldnaprasthdna, he lived in the East, and this is why he cites as an example the five rivers that are known in the East. " (Kosa, iii. 57).
2. We know through the quotations of Yasomitra that the chapters bore the name of skandhaka (Indriyaskandhaka, Samddhiskandhaka), and that the work which he is referring to was written in Sanskrit.
However, the first translation has for its title "Sastra in eight chien-tu"; Paramartha has "Sastra in eight ch'ien-tu. " We reminded of khanda, but
Poussin 17
? 18 , Introduction
Paramartha explains that ch'ien-tu is equivalent to ka-lan-ta, which is evidently grantha. S. Levi thinks that ch'ien-tu is the Prakrit gantho. Takakusu concludes, "All we can say is that the text brought by Sarhghadeva seems to have been in a dialect akin to Pali. . . But this supposition rests solely on the phonetic value of
36
Chinese ideographs. "
3. The Jndnaprasthdna, a very poorly composed work, begins with the study of
11 the laukikdgradharmas.
"What are the laukikdgradharmas} The mind and mental states which are immediately followed by entry into samyaktvanyama (see Kosa, vi. 26). There are those who say the five moral faculties {indriyas, faith, etc. ) which are immediately followed by entry into samyaktvanyama are called the laukikdgradharmas. " The text continues, "Why are this mind and these mental states so called . . . ? "
38
The Vibhasa, TD 27, p. 7cl, reproduces the two definitions of the
Jndnaprasthdna and explains: "Who are the persons who say that the laukikdgra- dharmas are the five faculties? The former Abhidharmikas. Why do they express themselves in this way? In order to refute another school: they do not intend to say that the laukikdgradharmas consist solely of the five faculties. But the Vibhajyavadins hold that the five faculties are exclusively pure (andsrava) (see Kosa, ii. 9) . . . In order to refute this doctrine, the former Abhidharmikas say that the lokottaradhannas consist of the five faculties. Now these dharmas are produced in the person of a prthagjana: thus it is proven that the five faculties can be impure. "
The interest of this commentary lies in the fart that it distinguishes the "former Abhidharmikas" from Katyayaniputra and from the Jndnaprasthdna.
4. One of the last stanzas of the last chapter is the sloka on the meaning of which, according to Vasumitra (Sectes, Masuda, p.
a. Abhidharmakosavydkhyd, Bibliotheca Buddhica, Sphutdrtha Abhidharma- kocavydkhyd, the work of Yacpmitra, first Kocasthdna, edited by Prof. S. Levi and Prof. Th. Stcherbatsky, 1st fasc, Petrograd, 1918; 2nd fasc by Wogihara, Stcherbatsky and Obermiller, (part of the second chapter), Leningrad, 1931.
Text of the third chapter, kdrikds and vydkhyd, in Bouddhisme, Cos- mologie . . . L. de La Vallee Poussin [with the collaboration of Dr. P. Cordier], Brussels, 1914-1919.
b. Tibetan translation of the Abhidharmakocakdrikdh and of the Abhi- dharmakocabhdsya of Vasubandhu, edited by Th. I. Stcherbatsky, 1st fasc. 1917,2nd fasc. 1930.
3. Tibetan sources, Palmyr Cordier, Catalogue de fonds tibetain de la Biblioteque Nationale, third part, Paris 1914, p. 394 and 499:
a. Abhidharmakosakadrikd and Bhdsya of Vasubandhu, Mdo 63, fol. 1-27, and fol. 28---Mdo 64, fol. 109.
b. Sutrdnurupd noma abhidharmakosavrttih of Vinltabhadra, 64, fol. 109-304.
c. Sphutdrtha ndrna abhidharmakosavydkhyd of Ya^omitra, 65 and 66. This is the commentary preserved in Sanskrit.
d. Laksandnusdrini ndma abhidharmakosattkd of Purnavardhana, a student of Sthiramati and master of Jinamitra and Silendrabodhi, 67 and 68.
e. Updyikd ndma abhidharmakosattkd of Samathadeva, 69 and 60, fol. 1-144. f. Marmapradipo ndma abhidharmakosavrttih of Dignaga, 70, fol. 144-286.
g. Laksandnusdrini ndma abhidharmakosattkd, an abridged recension of the
"Brhattika," above item d, 70, fol. 286-316.
h. Sdrasamuccayo ndma abhidharmavataratikd, anonymous, 70, fol. 315-393. i. Abhidharmdvatdraprakarana, anonymous, 70, fol. 393-417.
j. Tattvdrtho ndma abhidharmakosabhdsyatikd of Sthiramati, 129 and 130.
4. Abhidharmakofasdstra, of Vasubandhu, trans, by Paramartha in the period
564-567, Taisho volume 29, number 1559, p. 161-309; trans, by Hsiian-tsang, 651-654, Taisho volume 29, number 1558, p. 1-160.
The references in our translation are to the edition of Kyokuga Saeki, the
? Kando Abidatsuma Kusharon (Kyoto, 1891), the pages of which correspond to those of the Ming edition, a remarkable work which notably contains, in addition to interesting notes of the editor, copious extracts 1. from the two major Chinese commentators, 2. from the Vibhdsd, 3. from the commentary of Samghabhadra, and 4. from the work of K'uei-chi on the Trimsikd.
5. Among the Chinese commentaries on the Kosa:
a. Shen-t'ai, the author of a Shu: the Chil-she lun shu, originally in twenty Chinese volumes, today only volumes 1, 2, 4, 5, 6, 7, and 17 are extant; Manji Zoku-zokyo-1. 83. 3-4.
b. P'u-kuang, the author of the thirty-volume Chil-she lun Chi; TD 41, number 1821.
c. Fa-pao, the author of a thirty-volume Chil-she lun Shu; TD 41, number 1822.
Two other disciples of Hsiian-tsang, Huai-su and K'uei-chi, have written commentaries on the Kosa which are lost. P'u-kuang has also written a short treatise on the teachings of the Ko/a.
d. Yuan-hui wrote a thirty-volume Shu on the Kdrikds of the Kosa, a work with a preface written by Chia-ts'eng and dated before 727; this work, the Chil-she lun sung Shu (var. Chil-she lun sung shih), is preserved in TD volume 41, number 1823. This Shu "was commented upon many times in China and very widely disseminated in Japan; it is from this intermediary text that Mahayanists in general draw their knowledge of the Kola. But from the point of view of Indology, it does not offer the same interest as the three preceding com- mentaries. "
Hsiian-tsang dictated his version of Samghabhadra to Yuan-yu. There are some fragments of a commentary written by him.
6. Gunamati and the Laksandnsdra.
Gunamati is known through his commentary on the Vydkhydyukti\ many fragments of this commentary are quoted in the Chos-'byun of Bu-ston, trans. Obermiller, 1931. It is mentioned four times by Yasomitra in his Abhidharmakofavydkhyd.
a. Introductory stanzas: Gunamati comments on the Kosa, as has Vasumitra; Yasomitra follows this commentary when it is correct.
b. "Gunamati and his disciple Vasumitra say that the word nutrias is declined in the fourth case. But when the word namas is not independent, we have the accusative. This is why this master (Vasubandhu), in the Vydkhydyukti, says, 'Saluting the Muni with my head' . . . " (Kosa, Vydkhyd, i p. 7).
c. Gunamati holds that the Kosa wrongly teaches that "Conditioned things,
Poussin 9
? 10 Introduction
with the exception of the Path, are sasrava (Kosa, i. 4b)," for all of the dharmas, without exception, can be taken as an object by the dsravas(Vydkhyd i, p. 13).
d. On the subject of the continuity of the mental series, "the master Gunamati, with his disciple the master Vasumitra, through affection for the doctrine of his own nikaya, instead of confining themselves to explaining the Kosa, refute it" (Kosa,iiL Ua-b, note).
N. Peri (Date, 41) recalls that Burnouf mentioned (Introduction, 566), according to Yasomitra, the commentary of Gunamati. He adds: "An author very rarely quoted. His Laksananusarasastra (Taisho 1641) forms part of the Canon, where it is classified among the HInayana works. It summarizes the ideas of the Kosa, and then presents his own opinions on several points. The Hsi-yu-chi, after having listed him among the celebrated monks of Nalanda (TD 51, p. 924a2), tells us that he left the monastery where he had been living in order to move to V alabhF(p. 936c2). "
Taisho 1641 is only an extract of the treatise of Gunamati, the chapter which examines the sixteen aspects of the truths (Kosa, vii. 13): "Do we have sixteen things or sixteen names? The masters of the Vibhasa say that sixteen names are posited because there are sixteen things. But the siltra-upadesa masters say that there are sixteen names, but only seven things; four things for the first Truth, one thing for each of the three others. In the beginning the Buddha promulgated the Upadesasutra. After the disappearance of the Buddha, Ananda, Katyayana, etc. , recited that which they had heard. In order to explain the meaning of the Sutra, as disciples do, they composed a sastra explaining the Sutra, which is thus called a siltra-upadesa. Then the Vibhasa extracted an upadesa from that which was to be found [in this upadesa']; since it only indirectly comes from the Sutra, it is not called a sutra-upadefa. "
Gunamati continues as in the Kosa, vii. 13a, "According to the first explana- tion, anitya, impermanent, because it arises dependent on causes (pratyayd- dhinatvdt)" And he comments, "Conditioned things, without force, do not arise in and of themselves . . . "
The first volume ends, "The thesis of Vasubandhu is similar to the meaning of the siitra-upadesa masters. "
The second begins, "The author says, 1 am now going to give the explanation of what I believe. Anitya, impermanent, because, having arising, it has extinction. Conditioned things, having arising, and extinction, are not permanent. Arising is existence . . . "
The treatise touches on diverse points of philosophy, the absence of dtman, etc. In this work we encounter some very interesting notes, for example (Taisho,
? page 168b9), "In the Hinayana, the pretas are superior to animals; in the Mahayana, the opposite. In fact, the pretas are enveloped in flames . . . "
It is curious that the title of the work of Gunamati, literally Laksandnusd- rasdstra, is exactly identical to that of the book attributed to Purnavardhana in the Tanjur. We have Gunamati, a teacher of Sthiramati, and Purnavarudhana, a
25
student of Sthiramati.
7. Sthiramati, a student of Gunamati, defended the Kosa against Samgha-
bhadra. "His commentary on the Kosa is mentioned many times by Shen-t'ai, P'u-kuang, and Fa-pao in their work on the same text. The precise manner in which they-quote it, in which they note and discuss its opinions, causes us to believe that Hsiian-tsang may have brought it to China, and perhaps they themselves had also read it" (N. Pe'ri, Date, 41). Sthiramati, the author of the Tsa-chi, is one of the great masters of the Vijnaptimitrata system.
There exists (Taisho 1561) a small treatise by Sthiramati (transcription and translation) entitled Kosatattvdrthatikd or Abhidharmakosasdstratattvdrathaptkd, which is doubtless an extract of a voluminous work of the same name and by the same author preserved in Tibetan (Cordier, 499).
We observed, at the beginning, the commentary on the seven points indicated in the introductory stanza of the Kosa.
On the wisdom of the Buddha, superior to that of the saints, the author quotes the Kalpanamanditika stanza (Huber, Sutrdlamkara), Kosa, i. l, vii. 30; and recalls the ignorance that Maudgalyayana had of the place where his mother had been reborn, Kosa, i. 1.
In order to demonstrate the thesis of the Kosa that sraddhendriya can be impure, ii. 9, the author quotes at length the sutra on the request of Brahma to the Buddha (setting into motion the Wheel of the Law), a sutra briefly indicated by V asubhandu.
The work ends with some remarks on the duration of life: The stanza says: "Among the Kurus life is always 1,000 years in length; half of this to the west and the east. In this continent, it is not set: atits end, some ten years; in the beginning, without measure" {Kosa, iii. 75-77), "There are, in fact, in this world, some beings who have extra meritorious actions and who make the resolution, 'May I have a long life! , without desiring more precisely, 'May I live one hundred years, ninety years, eighty years! ' Or rather some venerable persons, parents and friends, say, 'May you live long! ' without saying more precisely how long a time. If one makes similar vows, it is because the actions done by persons of this continent are associated with thoughts of desire. The Sutra says, 'Know, oh Bhiksus that the length of life was over 80,000 years under Vipasyin, 20,000 years under Kasyapa;
Ponssin 11
? 12 Introduction
the length of life is now 100 years; few will go beyond this, and many will have less. ' If the length of life is not set, why does the Blessed One express himself in this way? . . . " The treatise concludes with the well-known stanza: sucirna- brahmacarye'smin . . . (Kosa, vi. 60a).
8. Samghabhadra has written two works.
The first (TD 29, number 1562), the title of which is transcribed into Chinese as Abhidharmanydyanusdrasdstra--or perhaps better as Nydydnusdro ndma Abhidharmaidstram--is a commentary which reproduces without any changes the Kdrikds of the Abhidharmakosa. But this eighty-volume commentary criticizes the Kdrikds, which present the Vaibhasika doctrine by noting them with the word kila, which means "in the words of the School"; it refutes the Bhdsyam, the auto-commentary of Vasubandhu, when this work presents views opposed to those of the Vaibhasikas,and it corrects them when it attributes to the Vaibhasikas views which are not theirs.
The title of the second treatise (TD 29, number 1563) is not completely transcribed: Abhidharmasamaya-hsien-sdstra or Abhidharmasamaya-kuang-sastra. J. Takakusu proposes Abhidharmasamayapradipikds'dstra, which is not bad;
however pradipa, "lamp," is always teng, and we have for hsien the equivalents prakdsa and dyotana.
This is a forty-volume extract from the Nydydnusdra, from which all polemic
is excluded and which is thus a simple presentation of the system (samaya) of the
Abhidharma. It differs from the Nydydnusdra by the presence of a rather long
introduction, in seven stanzas and prose, and also by the manner in which it treats
the Kdrikds of Vasubandhu: these Kdrikds are either omitted (ii. 2-3) or corrected
( i l l , 14) when they express false doctrines or when they cast suspicion on true
26 doctrines by the addition of the word kila.
Samghabhadra is an innovator, and K'uei-chi distinguishes the earlier and the later Sarvastivadins, Siddhi, 45 (theory of atoms), 65 (laksanas of "conditioned things"), 71 (the viprayukta called ho-ho)y 147 (vedand? ), and 311 (divergent Sarvastivadins, on adhimoksa).
***
(Additions to the Bibliography, by Hubert Durt. )
The following titles are editions of texts and works related to the Abhi- dharmakosabhasyam, which have appeared in print since the first appearance of de La Valle'e Poussin's French translation (1923-1931).
? Sanskrit:
Gokhale, V. V. , The Text of the Abhidharmakosakdrikd of VasubandhuJournal
of the Bombay Branch, Royal Asiatic Society, n. s. , vol. 22,1946, p. 73-102. Pradhan, P. , Abhidharm-Koshabha/ya of Vasubandhu, Tibetan Sanskrit Works
Series, vol. vrn, K. P. Jayaswal Research Institute, Patna, 1967; 2nd edition, revised with introduction and indices, by Dr. Aruna Haldar, 1975.
Shastri, Swami Dwarikadas, Abhidharmakosa & Bhdsya of Acharya Vasu- bandhu with Sphutdrthd Commentary of Acdrya Yasomitra, Part I (I and II Kosasthdna), critically edited, Bauddha Bharati, Varanasi, 1970.
Wogihara, Unrai, Sphutdrthd Abhidharmakosavyakhya, 2 vols.
, Tokyo 1933-1936; photomechanical reprint edition, Tokyo, 1971.
Tibetan:
Otani University, The Tibetan Tripitaka, Peking Edition, vol. 115 (number
5590), to Vol. 119 (number 5597), Suzuki Research Foundation, Tokyo, 1962.
Qiinese:
Takakusu, J. K. Watanabe, The Tripitaka in Chinese, vol. 29 (numbers 1558 to
1563), The Taisho Issai-kyo kanko kwai, Tokyo 1926.
Otani University, Index to the Taisho Tripitaka, no. 16, Bidon-bu III (vol. 29), Research Association for the Terminology of the Taisho Tripitaka, Tokyo, 1962.
Funabashi, Suisai and Issai Funabashi, Kando Abidatsuma Kusharon Sakuin, Kyoto, 1956. This index is based on the Chinese version of Kyokuga Saeki--the Qiinese version used by de La Valine Poussin--, the Kand6-bon Kusharon, in thirty volumesJCyoto, 1887.
Hi. The Date of Vasubandhu. The Former Vasubandhu.
We shall not undertake here a bibliography of Vasubandhu. But his treatise,
the Pratityasamutpddavydkhyd (Cordier, iii. 365), calls for the attention of the
reader of the Kosa. G. Tucci has published some fragments of this work (JRAS.
1930, 611-623) where the twelve links in the chain are explained in detail, with
numerous quotations from scriptures. G. Tucci also proposes to publish the
21
Trisvabhdvakarika and some parts of the commentary to the Madhydntavibhdga.
Concerning the "definition of pratyaksa by Vasubandhu," vdsubdndhava pratyaksalaksana, known through the Tdtaparyatikd, 99, and the Vddavidhi attributed to Vasubandhu, see the articles by G. Tucci, A. B. Keith, R Iyengar, JRAS. 1929, 473; Ind. Hist. Quarterly, 1928, 221; 1929, 81; Stcherbatski, Logic, ii. 161,382; G. Tucci, Maitreya [ndtha] et Asanga, 70-71, and finally Pramdna-
Poussin 13
? 14 Introduction
samuccaya, chap, i, by R. Iyengar, pp. 31-35. It appears that Dignaga denies the authorship of the Vddavidhi to Vasubandhu,in spite of universal opinion, and the Ttka quotes Kosa ii. 64, which contradicts the above-mentioned definition of pratyaksa. There are also numerous passages of the Vydkhyayukti in the Chos- 'byun of Bu-ston (above p. 16 ).
Wassiliew, Buddhismus, 235 (1860): "Life of Vasubandhu. "
Kern, Geschiedenis, trans. Huet, ii. 450.
S. Levi, JA. , 1890, 2. 252; Theatre indien, 1890, i. 165, ii. 35; "Donations
religieuses des rois de ValabhT' {Htes Etudes, vii, p. 97); "Date de Candragomin," BEFEO. ,1903,47;Sutrdkmkdra, trans,preface,2-3,1911.
Biihler, Alter der indischen Kunst-Poesie, p. 97,1890.
J. Takakusu, "Life of Vasubandhu," Voung-pao, 1904; "A Study of Para-
metria's Life of Vasubandhu and the date of Vasubandhu," JRAS. , 1905; "Sdmkhyakdrikd," BEFEO, 1904.
f
Wogihara, Asanga s Bodhisattvabhumi, 14, Strasbourg thesis, Leipzig, 1908.
Noel Peri, "A propos de la date of Vasubandhu," BEFEO. , 1911, 339-392.
Pathak, Bhandarkar, Indian Antiquare, 1911 - 1912 (V Smith, History, 3rd edition 328,4th edition, 346).
B. Shiiwo [Benkyo Shiio], Dr. Takakusu and Mr. Peri on the date of Vasubandhu (270-350), Tetsugaku Zasshi, Nov. -Dec. 1912.
Winternitz, ]eshichte, ii. 256 (1913), iii. 693 (1922).
H. . Ui, "On the Author of the Mahdydnasutrdlamkdra" Z. fur Ind. und Iranistik, vi. 1928,216-225.
A group of articles, many of which are summaries of articles written in Japanese, in Melanges Lanman (Indian Studies in Honor of Charles Rockwell
Lanman), 1929; J. Takakusu, Date of Vasubandhu, the Great Buddhist Phi- losopher, T Kimura, Date of Vasubandhu Seen from the Abhidharmakosa', G. Ono,DateofVasubandhuSeenfromtheHistoryofBuddhisticPhilosophy,H. Ui, Maitreya as an Historical Personage. Further, mention of the opinions of B. Shiiwo, S. Funabashi, E. Mayeda, S. Mochizuki.
"H. P. Sastri pointed out the historicity of Maitreyanatha from the colophon of the Abhisamaydlamkarakdrika, which is a commentary, from the Yogacara point of view, on the Panaivimsatisdhasnkd-prajfM-pdramftd-sutra by Maitreyanatha" Kimura, Origin of Mahdydna Buddhism, Calcutta, 1927, p. 170).
The date of Vasubandhu is bound to that of Asanga, his brother. Now some parts of the Yogasdstra, the work of Asanga, were translated into Chinese in 413-421, and in 431. However, the opinion is accepted among Japanese scholars that the works attributed to Asanga, writing under the inspiration of the future
? Buddha Maitreya, are in reality the works of a master Maitreya, an dcdrya, "an historical personage. " This thesis permits us to strip Asanga of one part of the library of which we thought he was the pious redactor, and to place him, along with his brother Vasubandhu, toward the middle or end of the 5th century, or--why not? --towards the 6th century. "If a scholar named Maitreya is found to be the author of those works hitherto attributed to Asanga, then the date of the latter ought to be shifted later, at least by one generation, if not more. The ground for an earlier date for Vasubandhu should give way altogether" (Takakusu, Melanges Lanman, 85).
H. Ui, in Philosophical Journal of the Imperial University, Tokyo, number
411, 1921, takes into account the arguments, developed afterwards in his Studies
of Indian Philosophy, i. 359, summarized in Melanges Lanman. These arguments
appear to be weak and, to my mind, non-existent (Note bouddhique, xvi, Maitreya
et Asanga, Ac. Royale de Belgique, January 1930). I do not think that they gain any
force from the observations of G. Tucci ("On some aspects of the doctrines of
Maitreya-[natha] and Asanga," Calcutta Lectures, 1930). The tradition of the
Vijnaptimatrata school establishes, as Tucci observes, the lineage Maitreya-
natha-Asanga-Vasubandhu, but Maitreyanatha is not the name of a man, but
rather "He who is protected by Maitreya"; ndtha is a synonym of buddha, or more 18
precisely of bhagavat. The commentary of the Abhisamaydlamkdra (p. 73 of the Tucci edition) gives to Maitreya the title of bhagavat in one place where he explains how "Asanga, in spite of his scriptural erudition and his insight (labdhddhigamo'pi, Kosa, viii. 39), did not understand the Prajndpdramita and lost heart. Then the Bhagavat Maitreya, for his sake (tarn uddisya) explained the Prajndpdramita and composed the treatise which is called the Abhisama- ydlamkdrakdrikd. " It is with the title of the Maitreyanatha that Santideva designates the saint who, in the Gandavyilha, explains to the pilgrim Sudhana the virtues of "the Bodhi mind" (Bodhicarydvatdra, i. 14, Rajendralal Mitra, Buddhist Nepalese Literature, 92). If the School holds as sacred, as dryddesand, the treatises of Asanga, it is because the Bhagavat Maitreya has revealed them. That the Tibeto-Chinese tradition varies in its attributions, sometimes naming as author a revealing deity, sometimes an inspired master, does not pose any difficulty.
***
The biography of Vasubandhu (by Paramartha) is not without its difficulties. The Kosa excited the criticism of Samghabhadra who, in his large Nydydnusdra, brings up innumerable heresies of a Sautrantika character which mar the work of
Poussin 15
? 16 Introduction
Vasubandhu. We are told that Vasubandhu refused to enter into controversy: "I am now already old. You may do as you please" (Takakusu's version). But we are also assured that Vasubandhu was then converted to the Mahayana by his brother Asanga, that he decided to cut out his tongue in order to punish it for not confessing the Mahayana earlier, and, more wisely, that he wrote numerous treatises wherein the doctrines of the Mahayana were brilliantly elaborated.
Yasomitra, the commentator on the Ko/a, says that the expression purva- caryas, "former masters," of the Kosa, designates "Asanga, etc. " {dsangapra- bhrtayas). N. Peri thinks that Yasomitra means to designate the school of the Purvacaryas by their most illustrious name, and that the text does not imply that Asanga is in fact purva relative to Vasubandhu (see my Cosmologie bouddhique, p. ix).
The Kosa was only translated in 563, whereas the work of Dharmatrata, an
imperfect draft of the Kosa, was translated in 397-418, 426-431, and 433-442.
Takakusu observes, "If the Kosa had existed, why did so many translators linger
over the book of Dharmatrata? {Melanges Lanmari). And it is difficult to give a 29
pertinent answer to this question.
But it appears almost impossible to believe that Paramartha the biographer of
Vasubandhu and first translator of the Kosa, arriving in China in 548, erred when he made the author of the Kosa the contemporary and the brother of Asanga. It is a hopeless hypothesis to identify the brother and the convert of Asanga with the former, or earlier, Vasubandhu.
One should admit the existence and the "Abhidharmic" activity of an earlier Vasubandhu. The problem, which I have taken up in the preface to Cosmologie bouddhique (above, p. 6), has been taken up again by Taiken Kimura, "Examen lumineau de 1'Abhidharma" (contents in Eastern Buddhism, iii, p. 85), fifth part: "On the sources of the Kosa! * We can see a summary of his conclusions in Melanges Lanman. Subsequently, see Note Bouddhique xvii, Acad, de Belgique: "Vasubandhu l'ancien. "
Yasomitra, in three places {Kosa, i. 13, iii. 27, iv. 2-3), recognizes in a master refuted by Vasubandhu the author of the Kosa (and a disciple of Manoratha
30 according to Hsuan-tsang), a "Sthavira Vasubandhu, the teacher of Manoratha,"
an "earlier master Vasubandhu," vrddhdcdryavasubandhu. P'u-kuang (Kimura, Melanges Lanman, 91) confirms Yasomitra, and designates the master in question under the name of "the earlier Vasubandhu, a dissident Sarvastivadin master. "
On the other hand, the gloss of the initial five stanzas of the treatise of 31
Dharmatrata, the re-edition of the Abhidharmasara of DharmasVI, attributes an edition of the same book in 6,000 verses to Vasubandhu. These stanzas and this
? gloss are not very clear. Kimura has studied them (Melanges Lanman); I have amended his interpretation (Note bouddhique xvii).
Iv. The Seven Canonical Treatises of the Abhidharma.
The Sarvastivadins recognize the authority of seven Abhidharma treatises,
"the word of the Buddha. " Among them, the Abhidharmikas, "who only read the 32
Abhidharma with its six feet," are distinct from the Vaibhasikas "who read the Abhidharma. "
33
The Abhidharma "with its six feet" is the great treatise of Katyayaniputra,
entitled the Jndnaprasthdna, upon which the Vibhdsd is a long commentary, and six treatises the order and authorship of which vary somewhat according to our sources. Following the order of Abhidharmakofavydkhya, there are: the Prakara- napdda of Vasumitra, Vijndnakdya of DevaSarman, Dharmaskandha of Sariputra (or of Maudgalyayana, according to Chinese sources), Prajnaptisdstra of Maudgal- yayana, Dhdtukdya of Purna (or of Vasumitra, Chinese sources), and Sangiti-
parydya of Mahakausthila (or of aariputra, Chinese sources).
One should note that the Tibetans list the Dharmaskandha first, and the
Jndnaprasthdna only as sixth: "The Tibetans seem to regard the Dharmaskandha as the most important of all. " This is also the opinion of Ching-mai (664 A. D. ), the author of the Chinese colophon (Takakusu, 75,115).
Takakusu,in"OntintAbhidharmaLiterature"(JPTS,1905),bringstogethera number of details on these seven books which Burnouf was the first to list; he gives the contents of the chapters of each of them. The remarks which follow are an addition to this fine work.
34 a. Jndnaprasthdna.
1. According to Hsiian-tsang, Katyayaniputra composed this Sastra in the
35 monastery of Tamasavana 300 years after the Parinirvana (the fourth century).
However, the Vibhdsd (TD 26, p. 21c29), commenting on the Jndnaprasthdna (TD 26, p. 918) says, "When the Bhadanta composed the Jrldnaprasthdna, he lived in the East, and this is why he cites as an example the five rivers that are known in the East. " (Kosa, iii. 57).
2. We know through the quotations of Yasomitra that the chapters bore the name of skandhaka (Indriyaskandhaka, Samddhiskandhaka), and that the work which he is referring to was written in Sanskrit.
However, the first translation has for its title "Sastra in eight chien-tu"; Paramartha has "Sastra in eight ch'ien-tu. " We reminded of khanda, but
Poussin 17
? 18 , Introduction
Paramartha explains that ch'ien-tu is equivalent to ka-lan-ta, which is evidently grantha. S. Levi thinks that ch'ien-tu is the Prakrit gantho. Takakusu concludes, "All we can say is that the text brought by Sarhghadeva seems to have been in a dialect akin to Pali. . . But this supposition rests solely on the phonetic value of
36
Chinese ideographs. "
3. The Jndnaprasthdna, a very poorly composed work, begins with the study of
11 the laukikdgradharmas.
"What are the laukikdgradharmas} The mind and mental states which are immediately followed by entry into samyaktvanyama (see Kosa, vi. 26). There are those who say the five moral faculties {indriyas, faith, etc. ) which are immediately followed by entry into samyaktvanyama are called the laukikdgradharmas. " The text continues, "Why are this mind and these mental states so called . . . ? "
38
The Vibhasa, TD 27, p. 7cl, reproduces the two definitions of the
Jndnaprasthdna and explains: "Who are the persons who say that the laukikdgra- dharmas are the five faculties? The former Abhidharmikas. Why do they express themselves in this way? In order to refute another school: they do not intend to say that the laukikdgradharmas consist solely of the five faculties. But the Vibhajyavadins hold that the five faculties are exclusively pure (andsrava) (see Kosa, ii. 9) . . . In order to refute this doctrine, the former Abhidharmikas say that the lokottaradhannas consist of the five faculties. Now these dharmas are produced in the person of a prthagjana: thus it is proven that the five faculties can be impure. "
The interest of this commentary lies in the fart that it distinguishes the "former Abhidharmikas" from Katyayaniputra and from the Jndnaprasthdna.
4. One of the last stanzas of the last chapter is the sloka on the meaning of which, according to Vasumitra (Sectes, Masuda, p.
