In chapter six of his translation, de La Vallee Poussin published the
complete
text of all the kdrikds as then recently discovered by Sylvain Levi in Nepal, a total of some 210 slokas.
Abhidharmakosabhasyam-Vol-1-Vasubandhu-Poussin-Pruden-1991
1537.
2a. Vijndnakdya, by Devasarman, composed approximately one hundred years
after the death of the Buddha. Exists in Chinese translation, T. vol. 26, no. 1539. 2b. Dhdtukdya, by Vasumitra (var. by Purna). Composed approximately three hundred years after the death of the Buddha. Preserved in the Chinese Canon, T
vol. 26, no. 1540.
2c. Prakarana-pdda, by Vasumitra, composed some three hundred years after
the death of the Buddha. This work, which closely resembles the Pali Vibhanga- prakarana, exists in two Chinese translations, T. vol. 26, no. 1541 and no. 1542.
2d Prajnapti-fdstra, composed by Maudgalyayana. This work exists in one late, incomplete Chinese translation (T vol. 26, no. 1538), and in a Tibetan translation. According to Sakurabe, the bulk of Sarvastivadin Abhidharma literature is divided into three major periods: 1. the early period, the period of the six padasdstras, before the composition of the Jnanaprasthdna; 2. the period of the Jnanaprasthdna, its commentaries (the vibhdsds), to the composition of the
Amrta-rasa\ and 3. all works subsequent to the Amrta-rasa.
Abhidharmic tendencies found within the Sutras were extended and devel-
oped into such texts as the Sangiti-parydya (la) and the Dharrna-skandha (lb). There is a Sangiti-suttanta in the Dtgha Nikdya; the Sarvastivadin text
entitled the Sangiti-parydya is an extension of this sutra.
The Dharrna-skandha is not an extension or commentary on a sutra, but it
takes a topic from the sutra literature and reclassifies it. The topic is taken from one sutra preached by the Buddha at theJetavana-vihara, and quotations are taken from other texts to serve as a commentary to its subject matter.
The Dharrna-skandha is made up of 21 chapters, and all but two of them are taken verbatim from sutras. Of these two, one is a miscellaneous chapter, and one is on the indriyas.
The miscellaneous chapter in the Dharrna-skandha lists some 78 different types of defilements, a list found verbatim nowhere in the sutras. What the editor of this text obviously did was to bring together any and all descriptions of the defilements found scattered throughout the Canon, and collect them in this one work where they now form a total of some 78 defilements.
The chapter on the indriyas in the Dharrna-skandha gives 22 different types of indriyas; again nowhere in the sutras are the number of indriyas given as 22, but the editor collected all sutra references to the indriyas, and these came to 22.
According to Kimura, these sdstras were composed as a type of commentary to the Sutras, and in the words of Lamotte, these works are "tres proches des sutra cathechetiques," "close to those sutras which are catechetical in form. "
? According to Kimura, these two works have already left the sutra form (that is, they are not attributed to the Buddha) and are now in the form of an independent Abhidharma sastra. Nevertheless they are not totally outside of sutra influence, and so they still have the appearance of being an edition (Lamotte: recension) of a sutra.
Further, these works are not fully abhidharmic in their treatment of their subject matter; that is, there is no attempt to be inclusive in their range of topics; rather they are devoted to only one topic. Their sectarian tendencies are still quite small, and there is no attempt at polemics or defense of their specific doctrines. In this last respect, the Pali Vibhanga, the Dharma-skandha, and the Sariputra- Abhidharma areveryclosetooneanother. Andtoo,boththeSangiti-parydyaand the Dharma-skandha are attributed to direct disciples of the Buddha, to Sariputra and to Maudgalyayana respectively (with the variants, to Mahakausthila and to Sariputra, respectively).
The Prajnapti-idstra (2d) is included by Lamotte (Histoire, p. 206), Frauwallner and Ryujo Yamada in the earliest period of this Abhidharma literature. Indeed, the fact that it is not fully abhidharmic in its treatment of subject matter, that its sectarian or polemical tendency is small, and that it is attributed to a direct disciple of the Buddha (here Maudgalyayana) does apply to this text as well as to the Sangiti-parydya and the Dharma-skandha. Nevertheless this work is not as close as these other two works to their origins, that of the sutra form, but appears to mark a further step away from, or a development from its sutra prototype, and so is placed by Sakurabe in the second stage of the development of early Abhidharma literature.
Both Frauwallner and Sakurabe place the Sangiti-parydya as the earliest Abhidharma text, and the Dharma-skandha as being slightly later in time than the Sangiti-parydya. De La Vallee Poussin and Lamotte place these texts however at approximately the same period of composition.
17. The Second Period of Early Sarvastivddin Literature.
A slightly later period in the development of this Abhidharma literature saw an advance in terms of the texts' internal organization and their doctrinal development. Characteristics of the literature of this period are a) the numerical classification of items, b) the detailed commentary given to each item of the series, and c) a greater elaboration in the contents of these works than was seen in the earlier period of the literature. In this period of literature we have the growth of Sarvastivadin sectarian concepts and vocabulary, and by now the vocabulary comes
Pruden xlix
? 1 The Abhidharma
to be set.
The Sarvastivadin sectarian influence can be seen primarily in their method of
dividing, or classifying the dharmas: the dharmas are divided into defiled (afrava) and undefiled (anasrava) dharmas, or into the mind (citta) and its mental events (cetasika, caitasika). In this period too we have an elaboration of cause and effect relationships.
The Prakarana-pdda (2c), a work from this period, is noteworthy in three respects: a) in form it is the first purely sastra work of this literature; b) in doctrine it is the first purely Sarvastivadin sectarian work; and c) it is the first work to divide the dharmas into five major divisions: uncompounded dharmas (asamskrtd dharma) and compounded (samskrtd) dharmas. The compounded dharmas are made up of four groups: physical matter (rupa), the mind {citta), mental states (caitasikd dharma) and dharmas or elements that are neither mind nor matter. (Sakurabe, however, finds evidences of the fivefold division of the dharmas in both the Sangiti-parydya and the Dharma-skandha. )
The Prakarana-pddd is made up of eight chapters. Several chapters (nos. I, IV, and V) have had an independent translation into Chinese, which might point to the fact that they had an independent circulation in India itself.
According to the Prajna-pdramitd Upadesa (the Ta-chih-tu lun, traditionally attributed to Nagarjuna), the Prakarana-pddd comes from two different sources, Vasumitra and the Kasmirian Arhats, each writing four chapters apiece. Each of the eight chapter deals with a different subject: even though the chapter divisions in this work are meaningful divisions, each chapter is almost totally unrelated to the other chapters. The Prakarana-pddd is thus perhaps a collection of eight independent works, brought together by one or two persons who were the final editors of this work. Frauwallner terms the Prakarana-pddd "ein Sammelwerk. "
Chapter Four of the Prakarana-pddd is an elaboration of the mental states as first raised in the Dhdtukdya; however, they are revised and augmented here by the editor of the Prakarana-pddd.
Chapters Six and Eight have traceable origins in the Sangiti-parydya, but the Prakarana-pddd has added considerable new material to them. Chapter Seven is one chapter from the Dharma-skandha; here, however, the contents undergo a reclassification.
The remaining chapters, I, II, III, and V, are the Prakarana-pddd's independent contribution to the development of Sarvastivadin thought, for it is in these chapters that we find the fivefold division of the dharmas, the tenfold wisdoms, a new classification of the dyatanas, and an elaboration of some 98 types of mental laziness.
? 18. The Jndnaprasthdna.
The Jrlanaprasthdna is a major compendium of Sarvastivadin thought, and its bulk is considerably larger than all of the previous works. Its 44 chapters occupy (in Chinese translation) some 120 pages of the Taisho edition of the Canon.
The Jndnaprasthdna (The Foundation of Knowledge) also presents original contributions to Sarvastivadin doctrinal development: it applies dependent origination (j>ratityasamut-pdda)to biological life, and it elucidates some six different types of material causes {hetu).
Because of its central position in the development of Sarvastivadin Abhi- dharma thought, it is termed the mula-sdstra (Ch. pen-luri) and "the basic Abhidharma treatise" {gen-pen o-pi ta-mo).
Both the Vibhdsd and the Abhidharmakosa state that Katyayaniputra collected the teachings of the Buddha which had hithertofore been scattered throughout the Scriptures, and brought them together in one work, the Jndnaprasthdna.
The internal organization of the work leaves much to be desired: within each topic, all information relevant to this topic is indeed collected together in one spot, but the overall organization of the text is haphazard.
The Jndnaprasthdna exists in two very different Chinese translations. Tradi- tionally this was thought to be simply two rather differing editions of the same work, the one work (now termed the Asta-grantha, T no. 1544) being the basic text, and the other translation (now termed the Abhidharma-sdstra, T. no. 1543) being a variant of it. Modern Japanese scholars now hold, however, that these two works represent two different traditions of Sarvastivadin Abhidharma learning, one tradition being centered in Kasmlr (the Asta-grantha), another being centered in Gandhara (the Abhidharma-sdstra). This is also seen in the faa that the Mahdvibhdsd (ostensibly a commentary on the Jndnaprasthdna) is also preserved in two very different Chinese translations.
(Kasmlr)
3a Asta-grantha
trans. 383 (in 30 chiian) by Sanghadeva and
Chu Fo-nien
(Gandhara)
3a Abhidharma-sdstra
trans. 657-660 (in 20 chiian) by Hsiian-tsang
Pruden li
? lii The Abhidharma
4a Mahd-Vibhasd
by Vasumitra and the
500 arhats, compiled
400 years after the Parinirvana of the Buddha, trans. 656-659 (in 200 chilan) by Hsiian-tsang
19. The Vibhdsds
4b Vibhdsd
by Katyayaniputra
trans. 437-439 (in 14 chiian) by Buddhavarman
The Jndnaprasthdna was commented upon (in both Kasmir and Gandhara? ) and its commentaries (vibhdsd) were termed simply "The Commentary," or Vibhdsd. Since these works exist in two Chinese translations, the commentary from Kasmir is termed the Mahdvibhdsd (The Great Commentary, T. 1545), and the commentary from Gandhara is termed simply the Vibhdsd (T 1546).
The Mahdvibhdsd is a voluminous commentary upon the Asta-grantha. It is made up of some 43 long chapters and the whole work occupies some 200 fascicules (volumes) in its Chinese translation, or one whole volume, vol. 27, of the Taisho Tripitaka. Volume 27 is 1,004 pages in length, each page having at most some 1,392 Chinese characters!
The Gandharan Vibhdsd, however, is much shorter in length, occupying only 14 fascicules in its Chinese translation: it has only 16 chapters to its text.
Such a voluminous commentary demonstrates to us the importance of the Jndnaprasthdna to the Sarvastivadins, and secures for the Jndnaprasthdna a position as the authoritative text of the Sarvastivadins.
The Mahdvibhdsd is a detailed analysis of everything in the Jndnaprasthdna', in one place, the Mahdvibhdsd devotes some I6V2 pages to commenting on one passage in the Jndnaprasthdna of only 100 characters. The work also raises new issues, issues not raised previously in the Jndnaprasthdna. Much new doctrine is introduced, and in addition, the Mahdvibhdsd quotes differing opinions on topics from outside orthodox Sarvastivadin ranks. Also many non-Buddhist theories are quoted at length.
By way of illustration, at the start of the work the Jndnaprasthdna asks the question,"Whatisthehighestworldlydharma(laukikadharmd)V Theanswerin the Asta-grantha version of the Jndnaprasthdna is in some 70 Chinese characters, while the Mahdvibhasd's answer to this same question is in more than 8,000 Chinese characters, and takes up some two Chinese fascicules.
In the Mahdvibhdsd, in answer to the question, "Why, in the Scriptures, did
? the Blessed One first teach the highest worldly dharmaT, some 33 different answers are given, from this question's doctrinal implications for the concept of adherence {prdpti), to the various types of mindfulness (smrtyupasthdna), to the four different types of conditions (pratyaya), and gives, in all, the names of eight different masters and schools.
The Mahavibhasd spends many pages on important points and problems but, in the main, it follows the Jndnaprasthana's (the Asta-grantha's) internal organization.
20. Development of the Literature After the Vibhdsds.
Sakurabe places the Abhidharma-amrta-rasa (The Abhidharma Taste of the Deathless) between the Jndnaprasthana and the Mahavibhasd on the basis of its doctrines, for it introduces points not found in the Jndnaprasthana, points which, however, are found in the Mahavibhasd: the doctrine of the five types of results (phala), the theory of atoms (paramanu), etc.
This work is a simple summary of all Sarvastivadin doctrine. It is made up of 16 chapters, and so, according to Sakurabe, constitutes the first work in the third period of Sarvastivadin Abhidharma literature, the period after the composition of the Jndnaprasthana. In its internal organization it appears to be a precursor to Dharmajina's Abhidharmahrdaya, which was in turn the direa precursor to the Abhidharmakosabhdsam.
21. The Abhidharmahrdaya.
The Abhidharmahrdaya (Heart of the Abhidharma) is a work in seven integral chapters, the order of which almost perfectly parallels the later order of the chapters in the Abhidharmakolabhdsyam.
The chapters of the Abhidharmahrdaya are:
I. Dhdtu-nirdesa
II. Samskdra-nirdesa
in. Karma-nirdefa IV Anusaya-nirdesa
v. Pudgakt-mdrga-nirdesa
(Kosa chapters) I. Dhdtu-nirdesa
II. Indriya-nirdesa ffl. Loka-nirdesa TV Karma-nirdesa
V. Anusaya-nirdesa
VI. Pudgala-mdrga-nirdesa
Pruden liii
? liv The Abhidharma
VI. fndna-nirdesa VII. Jnana-nirdesa VII. Samdpatti-nirdesa VIII. Samdpatti-nirdesa
IX. Pudgala-pratisedha
The Abhidharmahrdaya concludes with a supplement (a Miscellanea) and an appendix (a Discussion), for a total of nine chapters.
The Abhidharmahrdaya was the first work to use kdrikds, or verses, followed by their prose commentary (bhdsyam). This was the first real innovation in the internal organization of an Abhidharmic text. The work is internally coherent from beginning to end, and it is not merely a summary or an elaboration of a previous text: it is a well thought-out presentation of doctrine. The author of this work, Dharmajina (or Dharmasresthi) clearly had a self-conscious awareness of this work as a whole, complete text.
This sevenfold chapter division of the Abhidharmahrdaya was adopted by later works, by the Abhidharmahrdaya of Upasanta (Taisho no. 1551), the Ksudraka-Abhidharmahrdaya (also called the Samyukta-Abhidharmahrdaya, Taisho no. 1552) of Dharmatrata, and, with some modifications, by Vasubandhu, in his Abhidharmakosabhdsyam.
22. The Abhidharmakosabhdsyam.
Even though Vasubandhu's Abhidharmakosabhdsyam is the outstanding Abhidharma text of Far Eastern Buddhism, it is not the purpose of this article to discuss the question of the authorship of the Kosabhdsyam, nor the circumstances surrounding its composition: these topics will be discussed in a later article. I should like rather to merely say a few words on the place of the Kosabhdsyam in the general course of development of the Sarvastivadin Abhidharma literature.
Vasubandhu changed the name of the second chapter (from Samskdra- nirdesa) to Indriya-nirdesa, and added another chapter, the third chapter, Loka- nirdesa, for a total of nine chapters. The former supplement and appendix material was incorporated into the body of the work, and Vasubandhu added a new appendix chapter, the Pudgala-pratisedha (Refutation of the Soul), to the end of the work as a ninth chapter.
Later post-Kosa works, and indeed even anti-Ko/rf works like Sanghabhadra's Nydya-anusdra and his Samaya-pradtpika, not only kept the kdrikd-bhdsyam style of composition, but Samghabhadra even adopted the Kosa's (Vasubandhu's) kdrikds verbatim, adding his own prose commentary, or Bhdsyam. Samghabhadra changed the chapter names, and he took the ninth chapter, the Pudgala-
? pratisedha, from the end of the work and put it at the beginning as a first chapter, there to serve as an introduction to what is the most essential feature of Buddhist thought, its doctrine of andtman.
Another work, the Abhidharma-dtpa (Lamp on the Abhidharma, or the Abhidharma-vrtti Marmadipa-ndma) was composed somewhat later than these above works. The author of this work (known only in Tibetan as Phyogs-kyi-glan- po) renamed the first chapter (the Dhdtu-nirdesa) the Skandha-dyatana-dhdtu- nirdesa, and the sixth chapter (the Pudgala-mdrga-nirdes'a) became simply the Mdraga-nirdesa. The author kept the kdrikd-bhasyam format, which was by now a distinctive feature of Sarvastivadin Abhidharma literature.
The kdrikd-bhasyam format has only one exception to it: the Abhidharma- avatdra. This work, whose full name is the Sdrasamuccaya-ndma Abhidharma- avatdra-tikd (Entry into the Abhidharma, being a Compendium of its Essentials) is a work roughly contemporary with the composition of the Kosa. Tradition names the author as one Parsva (or Skandati). This work does not have the kdrtkdbhdsyam format but is, rather, a short treatise completely in prose; moreover the work lacks chapter divisions. It classifies all Sarvastivadin doctrine on the basis of the five skandhas and the three uncompounded dharmas, an original departure from the division based on uncompounded and compounded dharmas (see above).
23. Sanskrit Remains of the Abhidharma.
Very little remains of the bulk of Abhidharma literature in its original Sanskrit or Indie languages, especially when compared with the remains of the various vinayas and sutras which have been uncovered. Thus the Abhidharma literature of the schools of Kasmfr and Gandhara--the Sarvastivadins and the Sautrantikas-- exist primarily in their Chinese and Tibetan translations, and almost not at all in their original Sanskrit.
A fragment thought to be of the Sangtti-paryaya was found on 31 July 1930 in Bamiyan. In the village of Akkan, in the foothills of the Himalayas, there is a 35-meter-high image of the Buddha, and to the east of this image is a cave. It was from the collapsed roof of this cave that one page of text, written in Guptan script, was found. This fragment was studied by Professor Sylvain Levi, and he discovered that it corresponded to a part of the Sangiti-parydya. The results of his study were published in the Journal Asiatique (1932), and were translated and reprinted in two Japanese journals within that same year.
Pruden lv
? lvi The Abhidharma
The passages in question were from that part of the Sangiti-paryaya which is in close agreement with a Digha Nikdya passage and, indeed, the rediscovered passage was so fragmentary that it could also be from the Vinaya or from the AngiMara Nikdyal But if it is indeed a seaion of the Sangiti-paryaya, there is then but one page from the early period of Sarvastivadin Abhidharma literature which has been preserved for us in its original language.
Furthermore, Bamiyan is 150 kilometers to the west of the city of Kabul, the present-day capital of Afghanistan. This area was the center--as Gandhara--of Sarvastivadin studies, a fact perhaps relevant to the identification of this fragment with the text of the Sangiti-paryadya.
24. The Sanskrit Kosabhdsyam.
Another piece of Sanskrit Abhidharma literature that has been found is the full text of the Abhidharmakos'abhdsyam.
The Abhidharmakos'abhdsyam is made up of two parts, the kdrikd or verse sections (the Kosa), and the auto-commentary to these verses (the bhdsya) by Vasubandhu.
The kdrikd seaion has traditionally been known. It has in faa a separate translation into Chinese, which points to its having had an independent circulation in India.
The prose or commentarial seaion, the bhdsyam, had long been lost, but in 1935 Rahula Samkrtyayana discovered a palm-leaf manusaipt of both the kdrikd and the bhdsyam of the Abhidharmakosa, that is, the full text of this work, in the Tibetan monastery of Ngor, a Sakyapa institution located some two days' ride south of Shigatse.
This palm-leaf manusaipt dates from the 12th or the 13th century. It is an incomplete text: in the sixth chapter, kdrikds nos. 53 to 68 are missing. Nevertheless, the manuscript has some 600 kdrikds, plus 13 from the last chapter.
The kdrikd seaion of this manuscript find was published in the Journal of the Bombay Branch of the Royal Asiatic Society, by V. V. Gokhale; but it was only recently (1967) that the prose seaion, the bhdsyam, was published together with these kdrikds (see below).
Preceeding the find by Sarhkrtyayana, however, much scholarly work had already been done on the text of the Abhidharmakos'abhdsyam.
? 25. Translations of the Abhidharmakosabhasyam.
A team of Japanese and French scholars had worked on the Kosa, based on the kdrikds as they had been preserved in the Sphuta-arthd Abhidharmakosa Vydkhya, a Sanskrit commentary on the Kosabhasyam by Yasomitra. In this Vydkhyd the kdrikds are quoted, as well as large parts of the prose text (the bhdsyam). Working with a Cambridge manuscript of Yasomitra's Vydkhya and with the Tibetan translation of the Vydkhya, Louis de La Vallee Poussin published a complete French translation of the Chinese text of the Abhidharma- kosabhasyam (i. e. , the Chinese text of Hsuan-tsang's translation) in six volumes in Brussels (1923-1931).
In chapter six of his translation, de La Vallee Poussin published the complete text of all the kdrikds as then recently discovered by Sylvain Levi in Nepal, a total of some 210 slokas.
Based on de La Vallee Poussin's work, Sarhkrtyayana published the kdrikds with his own Sanskrit commentary (1933).
In 1935 the Japanese scholar Yoshio Nishi published the Kusharon (the Abhidharmakosabhasyam in Hsiian-tsang's Chinese translation) in the Kokuyaku- issaikyo series, and in this work he included the Sanskrit text of the kdrikds. (The Kokuyaku-issaikyo series was an edition of important works from the Far Eastern Buddhist Canon, translated into Japanese with often valuable introductions and annotations to the texts). In 1936 Ryujo Yamada published the kdrikds of the first chapter of the Kosabhdsyam, the Dhdtu-nirdesa, with their Chinese and Tibetan Tibetan versions (in Japanese translation) in a leading Japanese cultural journal, Bunka (Culture).
More recently, Narendra Nath Law's edition of Yasomitra's commentary served as the basis for Aiyaswami Sastri's publishing all the kdrikds to the third chapter, the Loka-nirdesa, and his translation of them into English in the Indian Historical Quarterly, vol. XXIV (1953).
26. Commentaries on the Abhidharmakosabhasyam.
There are altogether some seven Indian commentaries to the Abhidharma- kosabhasyam preserved in Tibetan, Chinese, and Uighur translations. The only one whose complete Sanskrit text has been preserved is Yasomitra's Vydkhya (which also exists in Tibetan translation). Manuscripts of the Vydkhayd exist in libraries in Paris, Cambridge, Leningrad, and Calcutta, and partial editions of this text are preserved in the libraries of Tokyo University and Kyoto University, Japan. The Paris manuscript, the best edition of this Vydkhya, is preserved in the collection of the Societe Asiatique; this text was reproduced by the Japanese scholars Bun'yu Nanjo and Kenju Sasahara, and deposited in the Otani
Pruden lvii
? lviii The Abhidharma
University Library, Kyoto.
In 1912, international efforts were begun to publish this work under the
leadership of Sylvain Levi. Levi, Stcherbatsky, and Unrai Ogiwara began the publication of this work in Bibliographie Bouddhique, getting as far as the middle of the second chapter (1918,1931). De La Vallee Poussin independently published the text and French translation of the third chapter, the Loka-nirdefa (1914-1918).
In Japan an association was formed to aid in the publication of the Vyakhya, an association headed by Ogiwara. This edition of the Vyakhya was to be based primarily on the Calcutta manuscript, with reference to the Paris manuscript. It was then that the whole text of the Vyakhya was finally published in Roman script in Tokyo (1932-1936). This work was recently reprinted (1971) in Tokyo, and is still readily available. In this work Ogiwara compared the text of the Abhidharmakosabhdsyam with its Chinese and Tibetan translations. The text of the Kosabhdsyam is italicized in the body of Yasomkra's work, and all of the works quoted in both Vasubandhu and Yasomitra are checked out in the footnotes.
More recently, Narendra Nath Law has published Yasomkra's Vyakhya as far as the fourth chapter, Karma-nirdesa, in Devanagari script, based on the Cambridge manuscript edition of this text. Law's work was published in the Calcutta Oriental Series, no. 31 (1949-1955).
27. The Tibetan Kosabhdsyam.
All of the work described above is based almost exclusively on the Sanskrit editions of the Kosabhdsyam and the Vyakhya, and on its Chinese translations. Nevertheless, the Tibetan translation of the Kosabhdsyam has also received some attention from Western and Japanese scholars.
Stcherbatsky published the Tibetan text of the first chapter, the Dhdtu- nirdesa, in Bibliographie Bouddhique, XX, Part T(1918), and, in Part II (1930) of this same series, continued the publication of the text up to the 46th kdrikd of the second chapter.
In Japan, Shunga Teramoto published the Tibetan text of the first chapter in Kyoto (1936), and the Department of Buddhist Studies (Bukkyogaku kenkyu- shitsu) of Kyoto University published the Tibetan text of the fkst chapter of the Kosabhdsyam along with the Vyakhya of Yasomitra: they have now gotten as far as the sixth chapter of the work.
? 28. Translations of the Kosabhdsyam and the Vyakhyd.
At the present time there exists a number of translations of the Kosabhdsyam and of the Vyakhyd.
A complete French translation of the Kosabhdsyam was carried out by de La Vallee Poussin. This translation is primarily based on the Chinese translation of Hsuan-tsang, but frequent reference is made to the Sanskrit text of Vasubandhu (as preserved in Yasomitra), the Chinese translation of Paramartha, and the Tibetan. This work was published from 1921 to 1931 (see above).
Stcherbatsky "translated" the ninth chapter, the Pudgala-pratisedha, from the Tibetan into English, under the title "The Soul Theory of the Buddhists" (1920). This translation was first published in the Bulletin de VAcademie des Russie, but it has been recently reprinted in India. This work is actually a very loose paraphrase of the ninth chapter.
Yasomitra's Vyakhyd has also undergone a number of partial translations. De La Vallee Poussin translated the Nyakhyd's commentary on the third chapter of the Kosabhdsyam into French (1914-1919), and the combined efforts of Ogiwara, Susumu Yamaguchi, Gadjin Nagao, and Issai Funabashi have translated the Vyakhyd into modern Japanese up to the second chapter of the Kosabhdsyam. In addition, Yamaguchi and Funabashi have published a Japanese translation of the
Vyakhyd commentary on the third chapter, the Loka-nirdesa (1955). In this work, each sentence of the Sanskrit is compared with its Tibetan translation, Yasomitra's commentary is added, and illustrative material from Sthiramati and other Indian masters is added. Working in this same format, Funabashi translated parts of the fourth chapter, Karma-nirdesa, in 1956.
More recently, Sakurabe has translated the first and the second chapters of the Kosabhdsyam into Japanese (1969), based on the full Sanskrit text edition of Pradhan (Patna, 1967).
29. Indexes to the Kosabhdsyam
The first index to the Abhidharmakosabhdsyam was an index based on the Chinese translation of this work. This index was called the Kando-Kusharon- sakuin. This index lists all of the Chinese words of the Kusharon (the Kosabhdsyam) in the order of their Japanese reading. The text used as the basis for this index was the Kando-bon, or Kando edition of this text. The word kando literally means that the annotation or commentary {-do) to the text was placed at
Pruden lix
? lx The Abhidharma
the top or crown (kan-) of the page. When a text is termed the Kando edition, this means that the editing of the work was done by one eminent scholar monk of the late 19th and the early 20th century, Kyokuga Saeki. Saeki's edition (i. e, the Kando edition) of the Kusharon was the edition of this text used by de La Vallee Poussin in his French translation, and most of de La Vallee Poussin's annotation is taken directly from the work of Saeki.
With the publication in 1946 of the Sanskrit kdrtkas by V. V. Gokhale (see above), and especially with the publication, in Devanagari script, of the full text of the Abhidharmakosa-bhdsyam (Sarhkrtyayana's manuscript find) by P. Pradhan in 1967, it now became possible to compile an index to the Sanskrit text of the Kosabhdsyam. This was done in the Kusharon-sakuin, compiled by Professor Akira Hirakawa of Tokyo University. The English title of this index is "Index to the Abhidharmakosa-bhasyam, Part One", and it was published in Tokyo in 1973. The lead words in this index are given in Sanskrit, with their Tibetan and Chinese translations. The Chinese words are given as they appear in Hsiian-tsang's translation of the Kosabhdsyam, with the variants of Paramartha given when needed. In this index, the first and all subsequent occurrences of the Sanskrit lead words are given (as found in Pradhan's edition of the text), followed by the use of each word in a compound, then by its Tibetan and Chinese translation. Part One of the Kusharon-sakuin is prefaced by a long English essay by Professor Hirakawa dealing with a number of topics raised by the Kosabhdsyam: the date of Vasubandhu, the relation of the Kosabhdsyam to the Yogacara tradition of Indian Mahayana Buddhism, the relation of the Sautrantikas and Mahayana Buddhism, and a review of the internal structure and the contents of* the Kosabhdsyam.
Part Two of this Index was published in 1977 and in this index the lead entries are given in Chinese, with their Sanskrit equivalents; the occurrence of the Chinese words in both the translations of Hsuan-tsang and Paramartha are shown, as well as the location of their Sanskrit originals in Pradhan's edition of the text.
Part Three was published in 1978 and is a Tibetan-Sanskrit index to the Kosabhdsyam. The references to the Tibetan Koiabhasyam are taken from the Peking edition (vol. 115) of the Tibetan Canon, with occasional readings adopted from the Derge edition of the Canon. Part Three also includes a complete page concordance from the Pradhan edition of the Kosabhdsyam (published in the Bauddha Bharati Series, vols. V, VI, VII, IX), to the TaishO editions of the text (the translations of Hsuan-tsang and Paramartha), to the Kando edition of Kyokuga Saeki (see above), and to both the Peking and the Derge edition of the Tibetan Kosabhdsyam.
? Part Three also contains an Addenda seaion with a supplement to the Sanskrit of Part One, and a valuable supplement to the corrigenda of Pradhan's text; and a 53-page corrigenda to Parts One and Two of this index concludes this work.
Pruden Ixi
? lxii The Abhidharma
1. The Theravadin Tradition of Hinayana Buddhism--the religion of Ceylon, Burma, Thailand, Laos, and Cambodia--recognizes some seven works as comprising the totality of their Abhidharma Pitaka: the Sarvastivadins of Kasmir and Gandhara also have an Abhidharma Pitaka, but the contents of this corpus are not limited to seven and include*a larger number of works: nor it appears, was it ever a closed system like the Theravadins'.
2. On the split, see the Prefatory Notes of CA. F. Rhys-Davids, in her Points of Controversy. According to Vasumitra, the original Sangha split into two, the Mahasanghikas and the Sthaviravadins (Pali: Theravadins), and the Sthaviravadins then split into two: the Haimavata (the "snow dwellers", the present-day Theravadins) and the Sarvastivadins. This last split occurred around 250 B. C.
? A bhidharmakosabhasyam of Vasubandhu
Louis de La Vallee Poussin
? INTRODUCTION
l h e earliest literature of Buddhism is divided into two parts or "baskets": The
and commentary on this discipline; and the Dharma, later termed Sutra, a collection of discourses which explain the Dharma, that is to say, everything that directly or indirectly concerns the path to salvation--a little moral law (powerless though it is to definitively deliver one from suffering), and above all the Eightfold Path, the methods of contemplation and of meditation which lead to the definitive deliverance from suffering, that is to Nirvana. This is the essential thing, for "the
4 sole taste of the Good Law is the taste of deliverance.
The Sutra or Dharma cannot be practiced exclusively. One effectively combats desire and hatred (lobha, dvesa) only by destroying ignorance (mohd)\ the moral law presupposes samyagdrsti or correct view with respect to the existence and retribution of action. Even more so, the elimination of the defilements and their most minute traces, indispensable to liberation from the round of rebirth, presupposes penetrating illumination into the nature of things, their accidental and transitory character. The sutras always contained, we can believe, much
3
psychology and ontology. When catechesis developed, numerous discourses of
Vinaya, the rules and regulations of the monastic discipline, including a history 2
the Buddha were edited, which contained enumerations, filled with glosses, of 4
technical terms. These are what the early tradition calls mdtfkas or indices. The 5
Anguttara and Digha 33-4, where these categories are arranged according to the
increasing number of their terms, have preserved for us an early type of this
literature. [One of the most notable matrkds is the Sangitisuttanta. The Pali
Canon has made a sutra of this text and places it in the Digha. Under the name of
Sangitiparyaya, this matrka takes its place among the seven canonical Abhi- 6
One school, more famous than the others, and which was perhaps the first to constitute standardized baskets of Vinaya and of Sutra, was the school of the Pali language, also the first to compile a third basket. The first catechism had been incorporated into the Sutra. The name Abhidharma was given to the new, more systematic, catechisms. It was a name which designated a special manner of presenting the Dharma and the authenticity (if not historical, at least doarinal) of these texts came to be affirmed and they were grouped into a "basket" placed on the same level as the baskets of Vinaya and Sutra. [For a discussion of the
dharma texts of the Sarvastivada. ]
1
? 2 Introduction
authenticity of the Abhidhammapipaka, see for example Atthasalini, p. 35. ] "Let it, then, be clearly understood," says Mrs. Rhys Davids, "that our present knowledge of such philosophy as is revealed in the Buddhist Pali canon would be practically undiminished if the whole of the Abhidhammapitaka were non-existent . . . The burden, then, of Abhidhamma is not any positive contribution to the philosophy of early Buddhism (? ), but analytic and logical and methodological elaboration of what is already given . . . The chief methods of that (=Abhidhamma) training were: first, the definition and determination of all names or terms entering into the Buddhist scheme of culture; secondly, the enunciation of all doctrines, theoretical and practical, as formulas, with coordination of all such as were logically interrelated; and finally, practice in reducing all possible heterodox
7
positions to an absurdity . . . " Nevertheless, the word Abhidhamma takes on a
higher scope, which we can understand by example. The prohibition against drinking alcohol is a precept of the Vinaya; but to examine the transgression of alcohol as a transgression of nature or a transgression of disobedience is to bring pure theory to play upon the Vinaya, to "refined the Vinaya, and this is what is called Abhivinaya. In the same way, the Abhidharma did not remain a stranger to scientific research and philosophy; it concerns itself with questions whose relationship with the Dharma properly so-called are quite loose. This tendency is very much accentuated in the latest of the Paji Abhidhamma treatises, the
8
Kathdvatthu, which tradition dates from the Council of Asoka. This work is an
account of heresies, and fixes their positions very clearly with respect to a mass of purely speculative points: in this work one can verify the long work of exegesis of which the Sutra had been the object.
The Paji Abhidhamma does not form part of the ancient patrimony common
to all the sects--which is not to say that it is Singhalese! Whereas all the
soundings carried out in the canonical literature of purely Indian Buddhism
reveals to us some Vinayas and sutras that have developed out of the Pali
literature, or which have a close connection with this literature, no one has yet
9 discovered the presence of any "prototypes" of the Pali Abhidhammas.
In any case, according to the tradition itself, the Kathavatthu belongs in its own right to a certain philosophic school, the Vibhajjavadins, "the followers of
distinction. "
10 11 To the old question, discussed in the sutras, "Does all exist? ",
these philosophers answered by distinguishing (vibhajya): "The present, and the
past which has not yet brought forth its result exist; the future and the past which
u
have brought forth their result do not exist.
To this school there is opposed--from ancient times, we may believe--the
school of "all exists," Sarvastivada, (Sarvastivadinas, Sabbatthivadino). This
? school--which also formed a sect, which had a special Vinaya and its own canon, 13
and which was Sanskritized --"carved out" the Dharma. In addition to "casuists," vinayadharas, they had "philosophers," dbhidhdrmikas. [Their Devasarman, the proponent of the existence of the past and the future, was opposed to Mu-lien or
Moggaliputta. ] A long work, with regard to which we are little informed,
14
led to
the redaction of numerous works among which are the seven books of the
Abhidharma, Treatises (sdstra) or Works (prakarana), the Jnanaprasthdna and its
six "feet" (pdda), the Dharmaskandha, etc. There were philosophies which came
15
out of this first level of wisdom literature.
and, towards the end of the first century of our era (Council of Kaniska), commentary was written on the Jnanaprasthdna: the Vibhdsd, a collective work which gives its name to all the masters who adopted it. The Vaibhasikas are the philosophers who refer to the Vibhdsd (Watters, i. 276). The center of the school appears to have been KaSmlr, even though there were Sarvastivadins outside of Kasmlr,--Bahirdesakas, "masters from foreign lands"; Pascattyas, "Westerners [relative to Kasmlr]"; Aparantakas, "masters from the western borders"--and some Kasmiris who were not Vaibhasikas.
The Sarvastivadins and the Vaibhasikas believed that the Abhidharmas were
the word of the Buddha. But there were masters who did not recognize the
authenticity of these books. When they were obliged to observe that there is no
"basket of the Abhidharma" outside of the Abhidharmas of the Sarvastivadins but
that each one of them knew that the word of the Buddha was embraced within
three "baskets," they answered that the Buddha taught the Abhidharma in the
Siltra itself--which is quite true. They recognized only the authority of the Sutra,
17
and took the name of Sautrantikas.
But we should not be mistaken with respect to their attitude. Even though
formally opposed to some of the theses of the Vibhdsd and of the Vaibhasikas, the Sautrantikas had a modern enough speculation and perhaps a Buddhology. They did not systematically combat their opponents, who were, without doubt, their predecessors. They admitted everything from the system of the Vaibhasikas which they had no formal reason to deny.
Such is, at least--to speak with greater prudence,--the attitude of our author,
18 V asubandhu.
His work, the Abhidharmakosa, a collection of approximately six hundred verses, describes itself as "a presentation of the Abhidharma as taught by the Vaibhasikas of Kasmlr. " This is not to say that Vasubandhu is a Vaibhasika; neither is he a Sarvastivadin. He has evident sympathies for the Sautrantikas, and utilizes the opinions of the "early masters"--namely "the Yogacarins, the chief
But the speculative work continued
Poussin 3
16
a
? 4 Introduction
among them being Asariga"--but without doubt, in his own mind, the system of
the Vaibhasikas is indispensable: the Vaibhasikas are "the School. " One does not
find anywhere else a body of doctrine as organized or as complete as theirs.
Nevertheless they are sometimes in error, and on important points too.
Vasubandhu completes his collection of technical verses, an impartial presentation
of the Vaibhasika system, with a prose commentary, the Abhidharmakosa-
bhdsyam, wherein his personal opinions, objections, and the opinions of diverse
schools and masters are found presented among numerous theses rejected by the
19
School. We know that Vasubandhu was, in his turn, combated and refuted by
orthodox Vaibhasikas.
But it matters little to us whether he is always right! The essential thing, for
us as for the masters who followed him, is that his book and his bhdsyam are truly a treasure (kosa).
###
From the point of view of dogmatics the Abhidharmakosa, with the Bhdsyam, is perhaps the most instructive book of early Buddhism (the Hinayana). It renders a great service in the study of canonical philosophy and in the study of scholasticism properly so-called.
It would be very wrong to say that we do not know the philosophy of canonical Buddhism: we know its essentials, its principle teachings, its major affiliations, and many of its details. But the history of this philosophy, its origins and development, is less clear: even though we can imagine that Buddhism, like the Buddha himself, took many steps at its birth, and these in all directions. But it is fair to say (and encouraging to repeat) that if the history of the canonical philosophy has not yet been done, the image that scholars such as Rhys Davids and Oldenberg have given of this philosophy either remains definitive or calls for but light retouching. We may believe, however, that we do not fully know any part, because we so imperfectly know the scholasticism which certainly enriched it and perhaps deformed it, but which certainly unfolded within it; which moreover should be, by its methods and its tendencies, completely parallel to the early speculation from whence the canonical philosophy itself arose. This philosophy is made up of the earlier strata of a speculation which continues within scholasticism proper, Pali as well as Sanskrit.
The impression of ignorance is very strong when we attempt to read early works such as the Dhammasangani or the Kathdvatthu; or when, with some rigor, we attempt to determine the sense of the sutras themselves, word for word
? (avayavdrtha). How many terms the exact significance of which escape us! It is easy and often correct to observe that these terms originally did not have a precise meaning; that the general orientation of Buddhist thought alone merits our interest; that, if we were to ignore precisely the four dhydnas and the four drupyasamdpattisdy vitarka and vicdra, rupa, the "fruits" and the "candidates for these fruits," we nevertheless have a sufficient idea of the major purport of and the methods leading to holiness within Buddhism; and that it is the candidates for these fruits who should preoccupy themselves with the details of the Eightfold Path rather than Western historians. Some think that scholasticism is not interesting; that, throughout Buddhist history, it remains alien to religion proper, as with the early doctrine. This is wrong: iti cen na sMravirodhatah, "If you think thus, no, for this is in contradiction with the Sutral" Buddhism was born complicated and verbose; its scholastic classifications are often pre-Buddhist; it is our good fortune to be able to examine them up close, in sources more ancient than Buddhaghosa; and the Abhidharmakosa bestows this good fortune upon us
in the measure in which we have the courage to be worthy of it.
An example of this is given by the Buddhists themselves. The Abhidharma-
kosa has had a great destiny: "This work . . . had an enormous influence. From
the time of its appearance, it became indespensable to all, friend and foe, we are
told; and there is reason to believe this, for the same fortune followed it
everywhere, first in China with Paramartha, and Hsiian-tsang and his disciples,
and then in Japan, where to this day specialized Buddhist studies begin with the
20 Kosas'astra. "
The author assures us that we will find in his book a correct summary of the
doctrine of the Vaibhasikas; but, however close may be his dependence on earlier
Abhidharma masters, we may believe that he improves upon what they have said.
When the Kos*a has been read, the earlier works of the Sarvastivadins, the
Abhidharmas and the Vibhdsa, undoubtedly lose part of their practical interest.
Though the Chinese have translated these works, the Tibetan Lotsavas did not
think it proper to put these works into Tibetan (with the sole exception of the
21
Prajnapti ), doubtless because the Abhidharmakosa, in accord with the resolution
of Vasubandhu, constitutes a veritable summa, embracing all problems--ontology, psychology, cosmology, discipline and the doctrine of aaion, the theory of results, mysticism and sanctity--and treating them with sobriety and in clear language, with all the method of which the Indians are capable. After Vasubandhu, the Northern Buddhists--whichever school they belonged to, and whether or not they adhered to the Mahayana--learned the elements of Buddhism from the Kosa. All schools, in fact, are in agreement with respect to a great number of
Poussm 5
? 6 Introduction
fundamental items, the same admitted by Pali orthodoxy, and the same, we may add, which are often subjacent to the sutras themselves. These items, which the Vaibhasikas have elucidated, are nowhere so wisely presented as in the Abhi- dharmakosa. This sufficiently explains the reputation of the author and the popularity of the book.
If Vasubandhu is an excellent professor of Buddhism, of Buddhism without
epithet of sect or school, he furthermore renders us a precious service by initiating
us into the systematic philosophy of these schools. He constructs before us the
spacious edifice of Vaibhasika dogma; he shows us its flaws; he explains what the
Sautrantika says, what the Vaibhasika answers, and what he himself thinks. Like
many philosophical treatises, and like the best of them, the Abhidharmakosa is a
creature of circumstances, written sub specie aeternitatis. We find in it many 22
proper names, and many allusions to contemporary debates. This is not a dull book.
We also find in it a great number of quotations which are shortened
of the earlier literature. Its quotations add to the numerous fragments of the Sanskrit canon which the sands of Turkestan have given us or which have been discovered under the modernist prose of the Divydvadana and the sutras of the Great Vehicle. These bear most often on texts of a doctrinal order, and we become clear with respect to the doctrinal, if not the historical, relationships of the canons.
***
For a long time the importance of the Abhidharma has been recognized by European scholars, initially by Burnouf. Let us see why the study of this work has been deferred for such a long period of time.
The work of Vasubandhu is made up of two distinct parts: the Abhidharma- kosa or the kdrikds, a collection of approximately six hundred verses; and their commentary or bhdsyam.
And of the vast exegetical literature that fills eight volumes of the Tibetan canon, the Nepalese scribes have preserved only a single document for us, a commentary on the Bhasya by Yasomitra, the Abhidharmakosavydkhyd, which bears the name of Sphutdrthd, "of clear meaning. "
This commentary by Yasomitra is not a complete commentary. It occasionally quotes the stanzas of Vasubandhu, and it elucidates such and such a passage of the Bhdsyam, indicating the passage in question by the first words of that passage,
elsewhere. Because of this, the Kosabhdsyam is a precious testament for the study 23
? following the general usage of commentators.
2a. Vijndnakdya, by Devasarman, composed approximately one hundred years
after the death of the Buddha. Exists in Chinese translation, T. vol. 26, no. 1539. 2b. Dhdtukdya, by Vasumitra (var. by Purna). Composed approximately three hundred years after the death of the Buddha. Preserved in the Chinese Canon, T
vol. 26, no. 1540.
2c. Prakarana-pdda, by Vasumitra, composed some three hundred years after
the death of the Buddha. This work, which closely resembles the Pali Vibhanga- prakarana, exists in two Chinese translations, T. vol. 26, no. 1541 and no. 1542.
2d Prajnapti-fdstra, composed by Maudgalyayana. This work exists in one late, incomplete Chinese translation (T vol. 26, no. 1538), and in a Tibetan translation. According to Sakurabe, the bulk of Sarvastivadin Abhidharma literature is divided into three major periods: 1. the early period, the period of the six padasdstras, before the composition of the Jnanaprasthdna; 2. the period of the Jnanaprasthdna, its commentaries (the vibhdsds), to the composition of the
Amrta-rasa\ and 3. all works subsequent to the Amrta-rasa.
Abhidharmic tendencies found within the Sutras were extended and devel-
oped into such texts as the Sangiti-parydya (la) and the Dharrna-skandha (lb). There is a Sangiti-suttanta in the Dtgha Nikdya; the Sarvastivadin text
entitled the Sangiti-parydya is an extension of this sutra.
The Dharrna-skandha is not an extension or commentary on a sutra, but it
takes a topic from the sutra literature and reclassifies it. The topic is taken from one sutra preached by the Buddha at theJetavana-vihara, and quotations are taken from other texts to serve as a commentary to its subject matter.
The Dharrna-skandha is made up of 21 chapters, and all but two of them are taken verbatim from sutras. Of these two, one is a miscellaneous chapter, and one is on the indriyas.
The miscellaneous chapter in the Dharrna-skandha lists some 78 different types of defilements, a list found verbatim nowhere in the sutras. What the editor of this text obviously did was to bring together any and all descriptions of the defilements found scattered throughout the Canon, and collect them in this one work where they now form a total of some 78 defilements.
The chapter on the indriyas in the Dharrna-skandha gives 22 different types of indriyas; again nowhere in the sutras are the number of indriyas given as 22, but the editor collected all sutra references to the indriyas, and these came to 22.
According to Kimura, these sdstras were composed as a type of commentary to the Sutras, and in the words of Lamotte, these works are "tres proches des sutra cathechetiques," "close to those sutras which are catechetical in form. "
? According to Kimura, these two works have already left the sutra form (that is, they are not attributed to the Buddha) and are now in the form of an independent Abhidharma sastra. Nevertheless they are not totally outside of sutra influence, and so they still have the appearance of being an edition (Lamotte: recension) of a sutra.
Further, these works are not fully abhidharmic in their treatment of their subject matter; that is, there is no attempt to be inclusive in their range of topics; rather they are devoted to only one topic. Their sectarian tendencies are still quite small, and there is no attempt at polemics or defense of their specific doctrines. In this last respect, the Pali Vibhanga, the Dharma-skandha, and the Sariputra- Abhidharma areveryclosetooneanother. Andtoo,boththeSangiti-parydyaand the Dharma-skandha are attributed to direct disciples of the Buddha, to Sariputra and to Maudgalyayana respectively (with the variants, to Mahakausthila and to Sariputra, respectively).
The Prajnapti-idstra (2d) is included by Lamotte (Histoire, p. 206), Frauwallner and Ryujo Yamada in the earliest period of this Abhidharma literature. Indeed, the fact that it is not fully abhidharmic in its treatment of subject matter, that its sectarian or polemical tendency is small, and that it is attributed to a direct disciple of the Buddha (here Maudgalyayana) does apply to this text as well as to the Sangiti-parydya and the Dharma-skandha. Nevertheless this work is not as close as these other two works to their origins, that of the sutra form, but appears to mark a further step away from, or a development from its sutra prototype, and so is placed by Sakurabe in the second stage of the development of early Abhidharma literature.
Both Frauwallner and Sakurabe place the Sangiti-parydya as the earliest Abhidharma text, and the Dharma-skandha as being slightly later in time than the Sangiti-parydya. De La Vallee Poussin and Lamotte place these texts however at approximately the same period of composition.
17. The Second Period of Early Sarvastivddin Literature.
A slightly later period in the development of this Abhidharma literature saw an advance in terms of the texts' internal organization and their doctrinal development. Characteristics of the literature of this period are a) the numerical classification of items, b) the detailed commentary given to each item of the series, and c) a greater elaboration in the contents of these works than was seen in the earlier period of the literature. In this period of literature we have the growth of Sarvastivadin sectarian concepts and vocabulary, and by now the vocabulary comes
Pruden xlix
? 1 The Abhidharma
to be set.
The Sarvastivadin sectarian influence can be seen primarily in their method of
dividing, or classifying the dharmas: the dharmas are divided into defiled (afrava) and undefiled (anasrava) dharmas, or into the mind (citta) and its mental events (cetasika, caitasika). In this period too we have an elaboration of cause and effect relationships.
The Prakarana-pdda (2c), a work from this period, is noteworthy in three respects: a) in form it is the first purely sastra work of this literature; b) in doctrine it is the first purely Sarvastivadin sectarian work; and c) it is the first work to divide the dharmas into five major divisions: uncompounded dharmas (asamskrtd dharma) and compounded (samskrtd) dharmas. The compounded dharmas are made up of four groups: physical matter (rupa), the mind {citta), mental states (caitasikd dharma) and dharmas or elements that are neither mind nor matter. (Sakurabe, however, finds evidences of the fivefold division of the dharmas in both the Sangiti-parydya and the Dharma-skandha. )
The Prakarana-pddd is made up of eight chapters. Several chapters (nos. I, IV, and V) have had an independent translation into Chinese, which might point to the fact that they had an independent circulation in India itself.
According to the Prajna-pdramitd Upadesa (the Ta-chih-tu lun, traditionally attributed to Nagarjuna), the Prakarana-pddd comes from two different sources, Vasumitra and the Kasmirian Arhats, each writing four chapters apiece. Each of the eight chapter deals with a different subject: even though the chapter divisions in this work are meaningful divisions, each chapter is almost totally unrelated to the other chapters. The Prakarana-pddd is thus perhaps a collection of eight independent works, brought together by one or two persons who were the final editors of this work. Frauwallner terms the Prakarana-pddd "ein Sammelwerk. "
Chapter Four of the Prakarana-pddd is an elaboration of the mental states as first raised in the Dhdtukdya; however, they are revised and augmented here by the editor of the Prakarana-pddd.
Chapters Six and Eight have traceable origins in the Sangiti-parydya, but the Prakarana-pddd has added considerable new material to them. Chapter Seven is one chapter from the Dharma-skandha; here, however, the contents undergo a reclassification.
The remaining chapters, I, II, III, and V, are the Prakarana-pddd's independent contribution to the development of Sarvastivadin thought, for it is in these chapters that we find the fivefold division of the dharmas, the tenfold wisdoms, a new classification of the dyatanas, and an elaboration of some 98 types of mental laziness.
? 18. The Jndnaprasthdna.
The Jrlanaprasthdna is a major compendium of Sarvastivadin thought, and its bulk is considerably larger than all of the previous works. Its 44 chapters occupy (in Chinese translation) some 120 pages of the Taisho edition of the Canon.
The Jndnaprasthdna (The Foundation of Knowledge) also presents original contributions to Sarvastivadin doctrinal development: it applies dependent origination (j>ratityasamut-pdda)to biological life, and it elucidates some six different types of material causes {hetu).
Because of its central position in the development of Sarvastivadin Abhi- dharma thought, it is termed the mula-sdstra (Ch. pen-luri) and "the basic Abhidharma treatise" {gen-pen o-pi ta-mo).
Both the Vibhdsd and the Abhidharmakosa state that Katyayaniputra collected the teachings of the Buddha which had hithertofore been scattered throughout the Scriptures, and brought them together in one work, the Jndnaprasthdna.
The internal organization of the work leaves much to be desired: within each topic, all information relevant to this topic is indeed collected together in one spot, but the overall organization of the text is haphazard.
The Jndnaprasthdna exists in two very different Chinese translations. Tradi- tionally this was thought to be simply two rather differing editions of the same work, the one work (now termed the Asta-grantha, T no. 1544) being the basic text, and the other translation (now termed the Abhidharma-sdstra, T. no. 1543) being a variant of it. Modern Japanese scholars now hold, however, that these two works represent two different traditions of Sarvastivadin Abhidharma learning, one tradition being centered in Kasmlr (the Asta-grantha), another being centered in Gandhara (the Abhidharma-sdstra). This is also seen in the faa that the Mahdvibhdsd (ostensibly a commentary on the Jndnaprasthdna) is also preserved in two very different Chinese translations.
(Kasmlr)
3a Asta-grantha
trans. 383 (in 30 chiian) by Sanghadeva and
Chu Fo-nien
(Gandhara)
3a Abhidharma-sdstra
trans. 657-660 (in 20 chiian) by Hsiian-tsang
Pruden li
? lii The Abhidharma
4a Mahd-Vibhasd
by Vasumitra and the
500 arhats, compiled
400 years after the Parinirvana of the Buddha, trans. 656-659 (in 200 chilan) by Hsiian-tsang
19. The Vibhdsds
4b Vibhdsd
by Katyayaniputra
trans. 437-439 (in 14 chiian) by Buddhavarman
The Jndnaprasthdna was commented upon (in both Kasmir and Gandhara? ) and its commentaries (vibhdsd) were termed simply "The Commentary," or Vibhdsd. Since these works exist in two Chinese translations, the commentary from Kasmir is termed the Mahdvibhdsd (The Great Commentary, T. 1545), and the commentary from Gandhara is termed simply the Vibhdsd (T 1546).
The Mahdvibhdsd is a voluminous commentary upon the Asta-grantha. It is made up of some 43 long chapters and the whole work occupies some 200 fascicules (volumes) in its Chinese translation, or one whole volume, vol. 27, of the Taisho Tripitaka. Volume 27 is 1,004 pages in length, each page having at most some 1,392 Chinese characters!
The Gandharan Vibhdsd, however, is much shorter in length, occupying only 14 fascicules in its Chinese translation: it has only 16 chapters to its text.
Such a voluminous commentary demonstrates to us the importance of the Jndnaprasthdna to the Sarvastivadins, and secures for the Jndnaprasthdna a position as the authoritative text of the Sarvastivadins.
The Mahdvibhdsd is a detailed analysis of everything in the Jndnaprasthdna', in one place, the Mahdvibhdsd devotes some I6V2 pages to commenting on one passage in the Jndnaprasthdna of only 100 characters. The work also raises new issues, issues not raised previously in the Jndnaprasthdna. Much new doctrine is introduced, and in addition, the Mahdvibhdsd quotes differing opinions on topics from outside orthodox Sarvastivadin ranks. Also many non-Buddhist theories are quoted at length.
By way of illustration, at the start of the work the Jndnaprasthdna asks the question,"Whatisthehighestworldlydharma(laukikadharmd)V Theanswerin the Asta-grantha version of the Jndnaprasthdna is in some 70 Chinese characters, while the Mahdvibhasd's answer to this same question is in more than 8,000 Chinese characters, and takes up some two Chinese fascicules.
In the Mahdvibhdsd, in answer to the question, "Why, in the Scriptures, did
? the Blessed One first teach the highest worldly dharmaT, some 33 different answers are given, from this question's doctrinal implications for the concept of adherence {prdpti), to the various types of mindfulness (smrtyupasthdna), to the four different types of conditions (pratyaya), and gives, in all, the names of eight different masters and schools.
The Mahavibhasd spends many pages on important points and problems but, in the main, it follows the Jndnaprasthana's (the Asta-grantha's) internal organization.
20. Development of the Literature After the Vibhdsds.
Sakurabe places the Abhidharma-amrta-rasa (The Abhidharma Taste of the Deathless) between the Jndnaprasthana and the Mahavibhasd on the basis of its doctrines, for it introduces points not found in the Jndnaprasthana, points which, however, are found in the Mahavibhasd: the doctrine of the five types of results (phala), the theory of atoms (paramanu), etc.
This work is a simple summary of all Sarvastivadin doctrine. It is made up of 16 chapters, and so, according to Sakurabe, constitutes the first work in the third period of Sarvastivadin Abhidharma literature, the period after the composition of the Jndnaprasthana. In its internal organization it appears to be a precursor to Dharmajina's Abhidharmahrdaya, which was in turn the direa precursor to the Abhidharmakosabhdsam.
21. The Abhidharmahrdaya.
The Abhidharmahrdaya (Heart of the Abhidharma) is a work in seven integral chapters, the order of which almost perfectly parallels the later order of the chapters in the Abhidharmakolabhdsyam.
The chapters of the Abhidharmahrdaya are:
I. Dhdtu-nirdesa
II. Samskdra-nirdesa
in. Karma-nirdefa IV Anusaya-nirdesa
v. Pudgakt-mdrga-nirdesa
(Kosa chapters) I. Dhdtu-nirdesa
II. Indriya-nirdesa ffl. Loka-nirdesa TV Karma-nirdesa
V. Anusaya-nirdesa
VI. Pudgala-mdrga-nirdesa
Pruden liii
? liv The Abhidharma
VI. fndna-nirdesa VII. Jnana-nirdesa VII. Samdpatti-nirdesa VIII. Samdpatti-nirdesa
IX. Pudgala-pratisedha
The Abhidharmahrdaya concludes with a supplement (a Miscellanea) and an appendix (a Discussion), for a total of nine chapters.
The Abhidharmahrdaya was the first work to use kdrikds, or verses, followed by their prose commentary (bhdsyam). This was the first real innovation in the internal organization of an Abhidharmic text. The work is internally coherent from beginning to end, and it is not merely a summary or an elaboration of a previous text: it is a well thought-out presentation of doctrine. The author of this work, Dharmajina (or Dharmasresthi) clearly had a self-conscious awareness of this work as a whole, complete text.
This sevenfold chapter division of the Abhidharmahrdaya was adopted by later works, by the Abhidharmahrdaya of Upasanta (Taisho no. 1551), the Ksudraka-Abhidharmahrdaya (also called the Samyukta-Abhidharmahrdaya, Taisho no. 1552) of Dharmatrata, and, with some modifications, by Vasubandhu, in his Abhidharmakosabhdsyam.
22. The Abhidharmakosabhdsyam.
Even though Vasubandhu's Abhidharmakosabhdsyam is the outstanding Abhidharma text of Far Eastern Buddhism, it is not the purpose of this article to discuss the question of the authorship of the Kosabhdsyam, nor the circumstances surrounding its composition: these topics will be discussed in a later article. I should like rather to merely say a few words on the place of the Kosabhdsyam in the general course of development of the Sarvastivadin Abhidharma literature.
Vasubandhu changed the name of the second chapter (from Samskdra- nirdesa) to Indriya-nirdesa, and added another chapter, the third chapter, Loka- nirdesa, for a total of nine chapters. The former supplement and appendix material was incorporated into the body of the work, and Vasubandhu added a new appendix chapter, the Pudgala-pratisedha (Refutation of the Soul), to the end of the work as a ninth chapter.
Later post-Kosa works, and indeed even anti-Ko/rf works like Sanghabhadra's Nydya-anusdra and his Samaya-pradtpika, not only kept the kdrikd-bhdsyam style of composition, but Samghabhadra even adopted the Kosa's (Vasubandhu's) kdrikds verbatim, adding his own prose commentary, or Bhdsyam. Samghabhadra changed the chapter names, and he took the ninth chapter, the Pudgala-
? pratisedha, from the end of the work and put it at the beginning as a first chapter, there to serve as an introduction to what is the most essential feature of Buddhist thought, its doctrine of andtman.
Another work, the Abhidharma-dtpa (Lamp on the Abhidharma, or the Abhidharma-vrtti Marmadipa-ndma) was composed somewhat later than these above works. The author of this work (known only in Tibetan as Phyogs-kyi-glan- po) renamed the first chapter (the Dhdtu-nirdesa) the Skandha-dyatana-dhdtu- nirdesa, and the sixth chapter (the Pudgala-mdrga-nirdes'a) became simply the Mdraga-nirdesa. The author kept the kdrikd-bhasyam format, which was by now a distinctive feature of Sarvastivadin Abhidharma literature.
The kdrikd-bhasyam format has only one exception to it: the Abhidharma- avatdra. This work, whose full name is the Sdrasamuccaya-ndma Abhidharma- avatdra-tikd (Entry into the Abhidharma, being a Compendium of its Essentials) is a work roughly contemporary with the composition of the Kosa. Tradition names the author as one Parsva (or Skandati). This work does not have the kdrtkdbhdsyam format but is, rather, a short treatise completely in prose; moreover the work lacks chapter divisions. It classifies all Sarvastivadin doctrine on the basis of the five skandhas and the three uncompounded dharmas, an original departure from the division based on uncompounded and compounded dharmas (see above).
23. Sanskrit Remains of the Abhidharma.
Very little remains of the bulk of Abhidharma literature in its original Sanskrit or Indie languages, especially when compared with the remains of the various vinayas and sutras which have been uncovered. Thus the Abhidharma literature of the schools of Kasmfr and Gandhara--the Sarvastivadins and the Sautrantikas-- exist primarily in their Chinese and Tibetan translations, and almost not at all in their original Sanskrit.
A fragment thought to be of the Sangtti-paryaya was found on 31 July 1930 in Bamiyan. In the village of Akkan, in the foothills of the Himalayas, there is a 35-meter-high image of the Buddha, and to the east of this image is a cave. It was from the collapsed roof of this cave that one page of text, written in Guptan script, was found. This fragment was studied by Professor Sylvain Levi, and he discovered that it corresponded to a part of the Sangiti-parydya. The results of his study were published in the Journal Asiatique (1932), and were translated and reprinted in two Japanese journals within that same year.
Pruden lv
? lvi The Abhidharma
The passages in question were from that part of the Sangiti-paryaya which is in close agreement with a Digha Nikdya passage and, indeed, the rediscovered passage was so fragmentary that it could also be from the Vinaya or from the AngiMara Nikdyal But if it is indeed a seaion of the Sangiti-paryaya, there is then but one page from the early period of Sarvastivadin Abhidharma literature which has been preserved for us in its original language.
Furthermore, Bamiyan is 150 kilometers to the west of the city of Kabul, the present-day capital of Afghanistan. This area was the center--as Gandhara--of Sarvastivadin studies, a fact perhaps relevant to the identification of this fragment with the text of the Sangiti-paryadya.
24. The Sanskrit Kosabhdsyam.
Another piece of Sanskrit Abhidharma literature that has been found is the full text of the Abhidharmakos'abhdsyam.
The Abhidharmakos'abhdsyam is made up of two parts, the kdrikd or verse sections (the Kosa), and the auto-commentary to these verses (the bhdsya) by Vasubandhu.
The kdrikd seaion has traditionally been known. It has in faa a separate translation into Chinese, which points to its having had an independent circulation in India.
The prose or commentarial seaion, the bhdsyam, had long been lost, but in 1935 Rahula Samkrtyayana discovered a palm-leaf manusaipt of both the kdrikd and the bhdsyam of the Abhidharmakosa, that is, the full text of this work, in the Tibetan monastery of Ngor, a Sakyapa institution located some two days' ride south of Shigatse.
This palm-leaf manusaipt dates from the 12th or the 13th century. It is an incomplete text: in the sixth chapter, kdrikds nos. 53 to 68 are missing. Nevertheless, the manuscript has some 600 kdrikds, plus 13 from the last chapter.
The kdrikd seaion of this manuscript find was published in the Journal of the Bombay Branch of the Royal Asiatic Society, by V. V. Gokhale; but it was only recently (1967) that the prose seaion, the bhdsyam, was published together with these kdrikds (see below).
Preceeding the find by Sarhkrtyayana, however, much scholarly work had already been done on the text of the Abhidharmakos'abhdsyam.
? 25. Translations of the Abhidharmakosabhasyam.
A team of Japanese and French scholars had worked on the Kosa, based on the kdrikds as they had been preserved in the Sphuta-arthd Abhidharmakosa Vydkhya, a Sanskrit commentary on the Kosabhasyam by Yasomitra. In this Vydkhyd the kdrikds are quoted, as well as large parts of the prose text (the bhdsyam). Working with a Cambridge manuscript of Yasomitra's Vydkhya and with the Tibetan translation of the Vydkhya, Louis de La Vallee Poussin published a complete French translation of the Chinese text of the Abhidharma- kosabhasyam (i. e. , the Chinese text of Hsuan-tsang's translation) in six volumes in Brussels (1923-1931).
In chapter six of his translation, de La Vallee Poussin published the complete text of all the kdrikds as then recently discovered by Sylvain Levi in Nepal, a total of some 210 slokas.
Based on de La Vallee Poussin's work, Sarhkrtyayana published the kdrikds with his own Sanskrit commentary (1933).
In 1935 the Japanese scholar Yoshio Nishi published the Kusharon (the Abhidharmakosabhasyam in Hsiian-tsang's Chinese translation) in the Kokuyaku- issaikyo series, and in this work he included the Sanskrit text of the kdrikds. (The Kokuyaku-issaikyo series was an edition of important works from the Far Eastern Buddhist Canon, translated into Japanese with often valuable introductions and annotations to the texts). In 1936 Ryujo Yamada published the kdrikds of the first chapter of the Kosabhdsyam, the Dhdtu-nirdesa, with their Chinese and Tibetan Tibetan versions (in Japanese translation) in a leading Japanese cultural journal, Bunka (Culture).
More recently, Narendra Nath Law's edition of Yasomitra's commentary served as the basis for Aiyaswami Sastri's publishing all the kdrikds to the third chapter, the Loka-nirdesa, and his translation of them into English in the Indian Historical Quarterly, vol. XXIV (1953).
26. Commentaries on the Abhidharmakosabhasyam.
There are altogether some seven Indian commentaries to the Abhidharma- kosabhasyam preserved in Tibetan, Chinese, and Uighur translations. The only one whose complete Sanskrit text has been preserved is Yasomitra's Vydkhya (which also exists in Tibetan translation). Manuscripts of the Vydkhayd exist in libraries in Paris, Cambridge, Leningrad, and Calcutta, and partial editions of this text are preserved in the libraries of Tokyo University and Kyoto University, Japan. The Paris manuscript, the best edition of this Vydkhya, is preserved in the collection of the Societe Asiatique; this text was reproduced by the Japanese scholars Bun'yu Nanjo and Kenju Sasahara, and deposited in the Otani
Pruden lvii
? lviii The Abhidharma
University Library, Kyoto.
In 1912, international efforts were begun to publish this work under the
leadership of Sylvain Levi. Levi, Stcherbatsky, and Unrai Ogiwara began the publication of this work in Bibliographie Bouddhique, getting as far as the middle of the second chapter (1918,1931). De La Vallee Poussin independently published the text and French translation of the third chapter, the Loka-nirdefa (1914-1918).
In Japan an association was formed to aid in the publication of the Vyakhya, an association headed by Ogiwara. This edition of the Vyakhya was to be based primarily on the Calcutta manuscript, with reference to the Paris manuscript. It was then that the whole text of the Vyakhya was finally published in Roman script in Tokyo (1932-1936). This work was recently reprinted (1971) in Tokyo, and is still readily available. In this work Ogiwara compared the text of the Abhidharmakosabhdsyam with its Chinese and Tibetan translations. The text of the Kosabhdsyam is italicized in the body of Yasomkra's work, and all of the works quoted in both Vasubandhu and Yasomitra are checked out in the footnotes.
More recently, Narendra Nath Law has published Yasomkra's Vyakhya as far as the fourth chapter, Karma-nirdesa, in Devanagari script, based on the Cambridge manuscript edition of this text. Law's work was published in the Calcutta Oriental Series, no. 31 (1949-1955).
27. The Tibetan Kosabhdsyam.
All of the work described above is based almost exclusively on the Sanskrit editions of the Kosabhdsyam and the Vyakhya, and on its Chinese translations. Nevertheless, the Tibetan translation of the Kosabhdsyam has also received some attention from Western and Japanese scholars.
Stcherbatsky published the Tibetan text of the first chapter, the Dhdtu- nirdesa, in Bibliographie Bouddhique, XX, Part T(1918), and, in Part II (1930) of this same series, continued the publication of the text up to the 46th kdrikd of the second chapter.
In Japan, Shunga Teramoto published the Tibetan text of the first chapter in Kyoto (1936), and the Department of Buddhist Studies (Bukkyogaku kenkyu- shitsu) of Kyoto University published the Tibetan text of the fkst chapter of the Kosabhdsyam along with the Vyakhya of Yasomitra: they have now gotten as far as the sixth chapter of the work.
? 28. Translations of the Kosabhdsyam and the Vyakhyd.
At the present time there exists a number of translations of the Kosabhdsyam and of the Vyakhyd.
A complete French translation of the Kosabhdsyam was carried out by de La Vallee Poussin. This translation is primarily based on the Chinese translation of Hsuan-tsang, but frequent reference is made to the Sanskrit text of Vasubandhu (as preserved in Yasomitra), the Chinese translation of Paramartha, and the Tibetan. This work was published from 1921 to 1931 (see above).
Stcherbatsky "translated" the ninth chapter, the Pudgala-pratisedha, from the Tibetan into English, under the title "The Soul Theory of the Buddhists" (1920). This translation was first published in the Bulletin de VAcademie des Russie, but it has been recently reprinted in India. This work is actually a very loose paraphrase of the ninth chapter.
Yasomitra's Vyakhyd has also undergone a number of partial translations. De La Vallee Poussin translated the Nyakhyd's commentary on the third chapter of the Kosabhdsyam into French (1914-1919), and the combined efforts of Ogiwara, Susumu Yamaguchi, Gadjin Nagao, and Issai Funabashi have translated the Vyakhyd into modern Japanese up to the second chapter of the Kosabhdsyam. In addition, Yamaguchi and Funabashi have published a Japanese translation of the
Vyakhyd commentary on the third chapter, the Loka-nirdesa (1955). In this work, each sentence of the Sanskrit is compared with its Tibetan translation, Yasomitra's commentary is added, and illustrative material from Sthiramati and other Indian masters is added. Working in this same format, Funabashi translated parts of the fourth chapter, Karma-nirdesa, in 1956.
More recently, Sakurabe has translated the first and the second chapters of the Kosabhdsyam into Japanese (1969), based on the full Sanskrit text edition of Pradhan (Patna, 1967).
29. Indexes to the Kosabhdsyam
The first index to the Abhidharmakosabhdsyam was an index based on the Chinese translation of this work. This index was called the Kando-Kusharon- sakuin. This index lists all of the Chinese words of the Kusharon (the Kosabhdsyam) in the order of their Japanese reading. The text used as the basis for this index was the Kando-bon, or Kando edition of this text. The word kando literally means that the annotation or commentary {-do) to the text was placed at
Pruden lix
? lx The Abhidharma
the top or crown (kan-) of the page. When a text is termed the Kando edition, this means that the editing of the work was done by one eminent scholar monk of the late 19th and the early 20th century, Kyokuga Saeki. Saeki's edition (i. e, the Kando edition) of the Kusharon was the edition of this text used by de La Vallee Poussin in his French translation, and most of de La Vallee Poussin's annotation is taken directly from the work of Saeki.
With the publication in 1946 of the Sanskrit kdrtkas by V. V. Gokhale (see above), and especially with the publication, in Devanagari script, of the full text of the Abhidharmakosa-bhdsyam (Sarhkrtyayana's manuscript find) by P. Pradhan in 1967, it now became possible to compile an index to the Sanskrit text of the Kosabhdsyam. This was done in the Kusharon-sakuin, compiled by Professor Akira Hirakawa of Tokyo University. The English title of this index is "Index to the Abhidharmakosa-bhasyam, Part One", and it was published in Tokyo in 1973. The lead words in this index are given in Sanskrit, with their Tibetan and Chinese translations. The Chinese words are given as they appear in Hsiian-tsang's translation of the Kosabhdsyam, with the variants of Paramartha given when needed. In this index, the first and all subsequent occurrences of the Sanskrit lead words are given (as found in Pradhan's edition of the text), followed by the use of each word in a compound, then by its Tibetan and Chinese translation. Part One of the Kusharon-sakuin is prefaced by a long English essay by Professor Hirakawa dealing with a number of topics raised by the Kosabhdsyam: the date of Vasubandhu, the relation of the Kosabhdsyam to the Yogacara tradition of Indian Mahayana Buddhism, the relation of the Sautrantikas and Mahayana Buddhism, and a review of the internal structure and the contents of* the Kosabhdsyam.
Part Two of this Index was published in 1977 and in this index the lead entries are given in Chinese, with their Sanskrit equivalents; the occurrence of the Chinese words in both the translations of Hsuan-tsang and Paramartha are shown, as well as the location of their Sanskrit originals in Pradhan's edition of the text.
Part Three was published in 1978 and is a Tibetan-Sanskrit index to the Kosabhdsyam. The references to the Tibetan Koiabhasyam are taken from the Peking edition (vol. 115) of the Tibetan Canon, with occasional readings adopted from the Derge edition of the Canon. Part Three also includes a complete page concordance from the Pradhan edition of the Kosabhdsyam (published in the Bauddha Bharati Series, vols. V, VI, VII, IX), to the TaishO editions of the text (the translations of Hsuan-tsang and Paramartha), to the Kando edition of Kyokuga Saeki (see above), and to both the Peking and the Derge edition of the Tibetan Kosabhdsyam.
? Part Three also contains an Addenda seaion with a supplement to the Sanskrit of Part One, and a valuable supplement to the corrigenda of Pradhan's text; and a 53-page corrigenda to Parts One and Two of this index concludes this work.
Pruden Ixi
? lxii The Abhidharma
1. The Theravadin Tradition of Hinayana Buddhism--the religion of Ceylon, Burma, Thailand, Laos, and Cambodia--recognizes some seven works as comprising the totality of their Abhidharma Pitaka: the Sarvastivadins of Kasmir and Gandhara also have an Abhidharma Pitaka, but the contents of this corpus are not limited to seven and include*a larger number of works: nor it appears, was it ever a closed system like the Theravadins'.
2. On the split, see the Prefatory Notes of CA. F. Rhys-Davids, in her Points of Controversy. According to Vasumitra, the original Sangha split into two, the Mahasanghikas and the Sthaviravadins (Pali: Theravadins), and the Sthaviravadins then split into two: the Haimavata (the "snow dwellers", the present-day Theravadins) and the Sarvastivadins. This last split occurred around 250 B. C.
? A bhidharmakosabhasyam of Vasubandhu
Louis de La Vallee Poussin
? INTRODUCTION
l h e earliest literature of Buddhism is divided into two parts or "baskets": The
and commentary on this discipline; and the Dharma, later termed Sutra, a collection of discourses which explain the Dharma, that is to say, everything that directly or indirectly concerns the path to salvation--a little moral law (powerless though it is to definitively deliver one from suffering), and above all the Eightfold Path, the methods of contemplation and of meditation which lead to the definitive deliverance from suffering, that is to Nirvana. This is the essential thing, for "the
4 sole taste of the Good Law is the taste of deliverance.
The Sutra or Dharma cannot be practiced exclusively. One effectively combats desire and hatred (lobha, dvesa) only by destroying ignorance (mohd)\ the moral law presupposes samyagdrsti or correct view with respect to the existence and retribution of action. Even more so, the elimination of the defilements and their most minute traces, indispensable to liberation from the round of rebirth, presupposes penetrating illumination into the nature of things, their accidental and transitory character. The sutras always contained, we can believe, much
3
psychology and ontology. When catechesis developed, numerous discourses of
Vinaya, the rules and regulations of the monastic discipline, including a history 2
the Buddha were edited, which contained enumerations, filled with glosses, of 4
technical terms. These are what the early tradition calls mdtfkas or indices. The 5
Anguttara and Digha 33-4, where these categories are arranged according to the
increasing number of their terms, have preserved for us an early type of this
literature. [One of the most notable matrkds is the Sangitisuttanta. The Pali
Canon has made a sutra of this text and places it in the Digha. Under the name of
Sangitiparyaya, this matrka takes its place among the seven canonical Abhi- 6
One school, more famous than the others, and which was perhaps the first to constitute standardized baskets of Vinaya and of Sutra, was the school of the Pali language, also the first to compile a third basket. The first catechism had been incorporated into the Sutra. The name Abhidharma was given to the new, more systematic, catechisms. It was a name which designated a special manner of presenting the Dharma and the authenticity (if not historical, at least doarinal) of these texts came to be affirmed and they were grouped into a "basket" placed on the same level as the baskets of Vinaya and Sutra. [For a discussion of the
dharma texts of the Sarvastivada. ]
1
? 2 Introduction
authenticity of the Abhidhammapipaka, see for example Atthasalini, p. 35. ] "Let it, then, be clearly understood," says Mrs. Rhys Davids, "that our present knowledge of such philosophy as is revealed in the Buddhist Pali canon would be practically undiminished if the whole of the Abhidhammapitaka were non-existent . . . The burden, then, of Abhidhamma is not any positive contribution to the philosophy of early Buddhism (? ), but analytic and logical and methodological elaboration of what is already given . . . The chief methods of that (=Abhidhamma) training were: first, the definition and determination of all names or terms entering into the Buddhist scheme of culture; secondly, the enunciation of all doctrines, theoretical and practical, as formulas, with coordination of all such as were logically interrelated; and finally, practice in reducing all possible heterodox
7
positions to an absurdity . . . " Nevertheless, the word Abhidhamma takes on a
higher scope, which we can understand by example. The prohibition against drinking alcohol is a precept of the Vinaya; but to examine the transgression of alcohol as a transgression of nature or a transgression of disobedience is to bring pure theory to play upon the Vinaya, to "refined the Vinaya, and this is what is called Abhivinaya. In the same way, the Abhidharma did not remain a stranger to scientific research and philosophy; it concerns itself with questions whose relationship with the Dharma properly so-called are quite loose. This tendency is very much accentuated in the latest of the Paji Abhidhamma treatises, the
8
Kathdvatthu, which tradition dates from the Council of Asoka. This work is an
account of heresies, and fixes their positions very clearly with respect to a mass of purely speculative points: in this work one can verify the long work of exegesis of which the Sutra had been the object.
The Paji Abhidhamma does not form part of the ancient patrimony common
to all the sects--which is not to say that it is Singhalese! Whereas all the
soundings carried out in the canonical literature of purely Indian Buddhism
reveals to us some Vinayas and sutras that have developed out of the Pali
literature, or which have a close connection with this literature, no one has yet
9 discovered the presence of any "prototypes" of the Pali Abhidhammas.
In any case, according to the tradition itself, the Kathavatthu belongs in its own right to a certain philosophic school, the Vibhajjavadins, "the followers of
distinction. "
10 11 To the old question, discussed in the sutras, "Does all exist? ",
these philosophers answered by distinguishing (vibhajya): "The present, and the
past which has not yet brought forth its result exist; the future and the past which
u
have brought forth their result do not exist.
To this school there is opposed--from ancient times, we may believe--the
school of "all exists," Sarvastivada, (Sarvastivadinas, Sabbatthivadino). This
? school--which also formed a sect, which had a special Vinaya and its own canon, 13
and which was Sanskritized --"carved out" the Dharma. In addition to "casuists," vinayadharas, they had "philosophers," dbhidhdrmikas. [Their Devasarman, the proponent of the existence of the past and the future, was opposed to Mu-lien or
Moggaliputta. ] A long work, with regard to which we are little informed,
14
led to
the redaction of numerous works among which are the seven books of the
Abhidharma, Treatises (sdstra) or Works (prakarana), the Jnanaprasthdna and its
six "feet" (pdda), the Dharmaskandha, etc. There were philosophies which came
15
out of this first level of wisdom literature.
and, towards the end of the first century of our era (Council of Kaniska), commentary was written on the Jnanaprasthdna: the Vibhdsd, a collective work which gives its name to all the masters who adopted it. The Vaibhasikas are the philosophers who refer to the Vibhdsd (Watters, i. 276). The center of the school appears to have been KaSmlr, even though there were Sarvastivadins outside of Kasmlr,--Bahirdesakas, "masters from foreign lands"; Pascattyas, "Westerners [relative to Kasmlr]"; Aparantakas, "masters from the western borders"--and some Kasmiris who were not Vaibhasikas.
The Sarvastivadins and the Vaibhasikas believed that the Abhidharmas were
the word of the Buddha. But there were masters who did not recognize the
authenticity of these books. When they were obliged to observe that there is no
"basket of the Abhidharma" outside of the Abhidharmas of the Sarvastivadins but
that each one of them knew that the word of the Buddha was embraced within
three "baskets," they answered that the Buddha taught the Abhidharma in the
Siltra itself--which is quite true. They recognized only the authority of the Sutra,
17
and took the name of Sautrantikas.
But we should not be mistaken with respect to their attitude. Even though
formally opposed to some of the theses of the Vibhdsd and of the Vaibhasikas, the Sautrantikas had a modern enough speculation and perhaps a Buddhology. They did not systematically combat their opponents, who were, without doubt, their predecessors. They admitted everything from the system of the Vaibhasikas which they had no formal reason to deny.
Such is, at least--to speak with greater prudence,--the attitude of our author,
18 V asubandhu.
His work, the Abhidharmakosa, a collection of approximately six hundred verses, describes itself as "a presentation of the Abhidharma as taught by the Vaibhasikas of Kasmlr. " This is not to say that Vasubandhu is a Vaibhasika; neither is he a Sarvastivadin. He has evident sympathies for the Sautrantikas, and utilizes the opinions of the "early masters"--namely "the Yogacarins, the chief
But the speculative work continued
Poussin 3
16
a
? 4 Introduction
among them being Asariga"--but without doubt, in his own mind, the system of
the Vaibhasikas is indispensable: the Vaibhasikas are "the School. " One does not
find anywhere else a body of doctrine as organized or as complete as theirs.
Nevertheless they are sometimes in error, and on important points too.
Vasubandhu completes his collection of technical verses, an impartial presentation
of the Vaibhasika system, with a prose commentary, the Abhidharmakosa-
bhdsyam, wherein his personal opinions, objections, and the opinions of diverse
schools and masters are found presented among numerous theses rejected by the
19
School. We know that Vasubandhu was, in his turn, combated and refuted by
orthodox Vaibhasikas.
But it matters little to us whether he is always right! The essential thing, for
us as for the masters who followed him, is that his book and his bhdsyam are truly a treasure (kosa).
###
From the point of view of dogmatics the Abhidharmakosa, with the Bhdsyam, is perhaps the most instructive book of early Buddhism (the Hinayana). It renders a great service in the study of canonical philosophy and in the study of scholasticism properly so-called.
It would be very wrong to say that we do not know the philosophy of canonical Buddhism: we know its essentials, its principle teachings, its major affiliations, and many of its details. But the history of this philosophy, its origins and development, is less clear: even though we can imagine that Buddhism, like the Buddha himself, took many steps at its birth, and these in all directions. But it is fair to say (and encouraging to repeat) that if the history of the canonical philosophy has not yet been done, the image that scholars such as Rhys Davids and Oldenberg have given of this philosophy either remains definitive or calls for but light retouching. We may believe, however, that we do not fully know any part, because we so imperfectly know the scholasticism which certainly enriched it and perhaps deformed it, but which certainly unfolded within it; which moreover should be, by its methods and its tendencies, completely parallel to the early speculation from whence the canonical philosophy itself arose. This philosophy is made up of the earlier strata of a speculation which continues within scholasticism proper, Pali as well as Sanskrit.
The impression of ignorance is very strong when we attempt to read early works such as the Dhammasangani or the Kathdvatthu; or when, with some rigor, we attempt to determine the sense of the sutras themselves, word for word
? (avayavdrtha). How many terms the exact significance of which escape us! It is easy and often correct to observe that these terms originally did not have a precise meaning; that the general orientation of Buddhist thought alone merits our interest; that, if we were to ignore precisely the four dhydnas and the four drupyasamdpattisdy vitarka and vicdra, rupa, the "fruits" and the "candidates for these fruits," we nevertheless have a sufficient idea of the major purport of and the methods leading to holiness within Buddhism; and that it is the candidates for these fruits who should preoccupy themselves with the details of the Eightfold Path rather than Western historians. Some think that scholasticism is not interesting; that, throughout Buddhist history, it remains alien to religion proper, as with the early doctrine. This is wrong: iti cen na sMravirodhatah, "If you think thus, no, for this is in contradiction with the Sutral" Buddhism was born complicated and verbose; its scholastic classifications are often pre-Buddhist; it is our good fortune to be able to examine them up close, in sources more ancient than Buddhaghosa; and the Abhidharmakosa bestows this good fortune upon us
in the measure in which we have the courage to be worthy of it.
An example of this is given by the Buddhists themselves. The Abhidharma-
kosa has had a great destiny: "This work . . . had an enormous influence. From
the time of its appearance, it became indespensable to all, friend and foe, we are
told; and there is reason to believe this, for the same fortune followed it
everywhere, first in China with Paramartha, and Hsiian-tsang and his disciples,
and then in Japan, where to this day specialized Buddhist studies begin with the
20 Kosas'astra. "
The author assures us that we will find in his book a correct summary of the
doctrine of the Vaibhasikas; but, however close may be his dependence on earlier
Abhidharma masters, we may believe that he improves upon what they have said.
When the Kos*a has been read, the earlier works of the Sarvastivadins, the
Abhidharmas and the Vibhdsa, undoubtedly lose part of their practical interest.
Though the Chinese have translated these works, the Tibetan Lotsavas did not
think it proper to put these works into Tibetan (with the sole exception of the
21
Prajnapti ), doubtless because the Abhidharmakosa, in accord with the resolution
of Vasubandhu, constitutes a veritable summa, embracing all problems--ontology, psychology, cosmology, discipline and the doctrine of aaion, the theory of results, mysticism and sanctity--and treating them with sobriety and in clear language, with all the method of which the Indians are capable. After Vasubandhu, the Northern Buddhists--whichever school they belonged to, and whether or not they adhered to the Mahayana--learned the elements of Buddhism from the Kosa. All schools, in fact, are in agreement with respect to a great number of
Poussm 5
? 6 Introduction
fundamental items, the same admitted by Pali orthodoxy, and the same, we may add, which are often subjacent to the sutras themselves. These items, which the Vaibhasikas have elucidated, are nowhere so wisely presented as in the Abhi- dharmakosa. This sufficiently explains the reputation of the author and the popularity of the book.
If Vasubandhu is an excellent professor of Buddhism, of Buddhism without
epithet of sect or school, he furthermore renders us a precious service by initiating
us into the systematic philosophy of these schools. He constructs before us the
spacious edifice of Vaibhasika dogma; he shows us its flaws; he explains what the
Sautrantika says, what the Vaibhasika answers, and what he himself thinks. Like
many philosophical treatises, and like the best of them, the Abhidharmakosa is a
creature of circumstances, written sub specie aeternitatis. We find in it many 22
proper names, and many allusions to contemporary debates. This is not a dull book.
We also find in it a great number of quotations which are shortened
of the earlier literature. Its quotations add to the numerous fragments of the Sanskrit canon which the sands of Turkestan have given us or which have been discovered under the modernist prose of the Divydvadana and the sutras of the Great Vehicle. These bear most often on texts of a doctrinal order, and we become clear with respect to the doctrinal, if not the historical, relationships of the canons.
***
For a long time the importance of the Abhidharma has been recognized by European scholars, initially by Burnouf. Let us see why the study of this work has been deferred for such a long period of time.
The work of Vasubandhu is made up of two distinct parts: the Abhidharma- kosa or the kdrikds, a collection of approximately six hundred verses; and their commentary or bhdsyam.
And of the vast exegetical literature that fills eight volumes of the Tibetan canon, the Nepalese scribes have preserved only a single document for us, a commentary on the Bhasya by Yasomitra, the Abhidharmakosavydkhyd, which bears the name of Sphutdrthd, "of clear meaning. "
This commentary by Yasomitra is not a complete commentary. It occasionally quotes the stanzas of Vasubandhu, and it elucidates such and such a passage of the Bhdsyam, indicating the passage in question by the first words of that passage,
elsewhere. Because of this, the Kosabhdsyam is a precious testament for the study 23
? following the general usage of commentators.
