These books are, in the
traditional
order in which they are listed:
1.
1.
Abhidharmakosabhasyam-Vol-1-Vasubandhu-Poussin-Pruden-1991
Further, this mdtrkd is an exhaustive and thorough-going analysis (of the dharmas).
From the above, then, we can see that the Mdtikds (or Mdtrkds) performed an important function in the development of the corpus of Buddhist literature, as admitted in traditional Buddhist literature itself.
By itemizing the component parts of the Dharma and the Vinaya, the
Pruden xxxix
? xl The Abhidharma
Mdtrkd did play an important part in the later elaboration which is Abhidharma literature. However, the Abhidharmikas worked on these lists, minutely analyzed the items on these lists, and then proceeded to give long, exhaustive treatises on each item. So the Mdtrkd represents but one aspect of the whole picture of the growth of the Abhidharma.
If the Abhidhamma was merely speaking on the dhamma as some scholars maintain, and if the Mdtikd served as the nucleus of later Abhidhamma elaborations, why, when the literature was codified into a Pitaka, was it named the Abhidhamma Pitaka, and not the Mdtikd Pitaka? It appears then that the Mdtikds did not directly develop into the Abhidharma literature as we now have it.
As Kimura has shown, in the earliest Buddhist Sangha, abhidhamma-katha-- discussion on the dharma--had a direa relationship with the later development of the commentarial literature on the Dhamma, the Abhidhamma commentaries. In the process, there were two tendencies; one: to summarize and to itemize, and two: to analyze and elaborate. This latter tendency came to predominate, and the name for this--abhidhamma--came to be attached permanendy to this new corpus of literature.
10. Abhidharma in the Agamas: the Religion of the Agamas.
There are abhidharmic tendencies in the extant sutras, in Pali as well as in those preserved in Chinese translation. Let us say a few words first, however, about the religion of the early Buddhist canon, the religion of the Agamas.
According to de La Vallee Poussin, all the teachings of the Buddha were not publically given out. Instead, much of the philosophy and the more subtle forms of the teaching were embodied in texts which were reserved for the study of monks in their monasteries; and the Agamas (or Nikdyas), the earliest form of the Buddhist sermons which have been preserved for us, are such philosophical texts as were transmitted from one generation of monks to those of a subsequent generation. Such texts are then the "clericalized" texts, and in these texts we see only a small bit of the popular side of early Buddhism.
Such is the case, to be sure, in any religion, and this is especially so in the case of Indian religions. Any Indian religion has two sides to it: a clerical, well worked- over doctrine, and a popular aspect of the religion, which includes many elements brought over in the mind of the new converts to the religion. But the important point to remember is that the extant literature of any religion is the technical literature used in the monasteries. The real face of early Buddhism in all of its aspects cannot be gotten at only through its literature, but must also be obtained
? through archaeology, art and chronology (Le dogme et la philosophie du Bouddhisme, 1930, Chap. VII). Such a mass Buddhism was the Buddhism that preceeded the canon, "precanonical Buddhism" {Bouddhisme precanonique). Its contents were not only a darsana--a systematic school of Indian philosophy, a consistent world-view--but a faith concerned with spirits and the release of these spirits from the round of birth and death, having, according to scholars, little in common with the doctrines of anitya, andtman, and duhkha so often stressed in the Agamas (see de La Vallee Poussin's Nirvana, 1925, p. 85, 115, 131). The spread of Buddhism was dependent upon its moral teaching, the personality of its founder, its wisdom embodied in memorable sentences and couplets (the Dhammapada or Uddnavarga), coupled with popular animal tales (the Jdtakas) (see The Way to Nirvana, 1917, Chap. V). Buddhism was also closely related to ancient Indian nature worship, the worship of certain trees, and the veneration of snakes.
Buddhism also came to be changed, especially on its popular level, by virtue of the influence of non-Buddhist religions, through the conversion to Buddhism of many non-Buddhists who brought their own ideas into the company of older believers. Popular Buddhist religion absorbed much of the pan-Indian pantheon of deities. However formal and set its doctrines might have been, most of the believers of Buddhism were but "demi-civilize" or semi-civilized (Bouddhisme, 1909, p. 349 ff. ). Such is also the view of A. B. Keith (Buddhist Philosophy in India and Ceylon, 1923) and C. A. F. Rhys-Davids (Sakya or Buddhist Origins, 1931, p. 431 ff. ).
11. The Agamas and the Nikdyas.
This above view was strongly opposed by Stcherbatsky (The Conception of Buddhist Nirvana, 1927). De La Vallee Poussin would appear to oppose the popular elements of early Buddhism to the scholarly works of later Buddhism: such would imply that there is something essentially different between early Buddhism and Abhidharma Buddhism. Stcherbatsky held that the Buddha was a produa of the philosophical environment of his time, and that he obviously had a well-defined philosophy with its attendant metaphysic.
But even if the set passages and formulas were removed from the Aamas, de La Vallee Poussin and Rhys-Davids cannot say that Buddhism is merely a faith concerned with spirits and immortality. So-called popular Buddhism and "pure" clerical Buddhism cannot be so clearly distinguished one from the other. Yet the Agamas do not give a clear picture of early Buddhism, especially in its popular
Pruden xli
? xlii The Abhidharma
aspects: the Agamas are the traditions (dgama) of a scholarly elite, as de La Vallee Poussin maintains.
The Agamas, literally the "transmitted" (doctrines) were in a sense also Nikdyas (compilations): the transmitted doctrines were collected together, formal- ized, and their vocabulary became technical terms; these terms came to be analyzed through vibhangas (long, explanatory definitions) and niruktas (ety- mologies); these same terms were also organized on the basis of numerical categories or on the basis of similarities (samyukta) into mdtrkds. So the tendencies that led ultimately to a systematic Abhidharma literature led in this same process to the systematization of the Agamas (the traditions) into Nikdyas (formal literary compilations).
In summary, the Agamas are doctrinal compilations from an early stage of Buddhism, and their recensions (Nikdyas) are in a sense Abhidharmic compila- tions and, being largely abhidharmic in tendency, they led ultimately to the growth of the Abhidharma as a separate literary genre.
12. Abhidharmic Tendencies in Extant Agamas: Numerical Categories, Samyuktas, and Vibhangas,
Abhidharmic tendencies, tendencies that led eventually to the growth of a separate literature, can be seen early in some scriptures. The use of numerical categories is one such tendency.
A. The Sangtti-suttanta (the "recited" sutra, Dtgha Nikdya no. 33, vol. Ill, pp. 207-271) lists a variety of items in a list from one to ten (one item, two items, three items . . . ) Dtgha Nikdya no. 34, the Dasuttara-suttanta, lists items in a list from one to ten, but now analyzes them according to various other categories, marking then a further development along abhidharmic lines.
This same tendency is seen in the case of the Anguttara Nikdya (anguttara, "increasing by an item"), which classifies all of the suttas in its collection on the basis of numerical categories: thus all suttas dealing with any group of "four" items are collected together, followed by all suttas dealing with "five" of anything (up to eleven items). This scheme then forms the basis for this one collection of texts, or nikdya.
B. Many other suttas or agamas were joined to one another on the basis of their affinity in subject matter. To be sure, this is not sharply distinguishable from the above numerical classification, but now the items are more meaningfully arranged. Such texts are called samyuktas (Pali: samyuttas) or "conjoined" texts.
? Examples of this tendency are the Sal-dydtana vagga (vagga = section or chapter) in the Majjhima Nikdya, and the Kamma-samyutta vagga in the Majjhima: i. e. , all those texts dealing with the ayatanas were grouped together, as were all those texts dealing with karma.
This too became the guiding principle in the compilation and editing of the
Samyutta Nikdya and parts of the Khuddaka Nikdya.
C. The concept of an expanded commentary (a vibhanga) is best seen in
individual suttas in the Sarhyutta Nikdya and in the Majjhima Nikdya. Many such texts have the word vibhanga in their titles:
Samyutta Nikdya XII. 2: the Vibhangam (II, pp. 2-4), which is an expansion of XII. 1 (desand); XLV. 8, the Vibhango (V. pp. 8-10), which is a commentary on the Noble Eightfold Path; and LI. 20, the Vibhanga (V, pp. 276-281), which is a commentary on the siddhi, or supernormal powers, of a Buddha.
The Majjhima Nikdya has Majj. 135, the Cula-kamma-vibhanga sutta (cula = smaller, lesser) (III, pp. 202-206), which treats of Karma, and is followed by the Mahd-kamma vibhanga-sutta, Majj. 136 (pp. 207-215) which is an elaboration of certain of the former sutta's sections; Majj. 137, the Saldyatana- vibhanga-sutta on the six ayatanas (p. 215-222) and all of its following suttas are vibhangas: Majj. 138, 139, 140 (on the dhdtus), 141, 142. Synonymous with the vibhanga is the word vedalla, which also means "expanded": as the Cula-vedalla- sutta, and the Mahd-vedalla-sutta. The Buddha would give a short sermon, and one of his disciples, such as Katyayana, or the Buddha himself, would elaborate on it; or the disciples would discuss it among themselves, and in this way it would reach its present form.
The dgamas (in their Chinese translations) which have the characteristics of vibhangas can be rather closely identified with these same Suttas extant in Pali: they are largely the same text (especially Majj. 131 to 142, as above), a fact which does not hold for the other dgamas.
Thus abhidharmic tendencies are clearly seen in many texts in both Pali and Chinese, so far advanced in many cases that it is merely a short step to real Abhidhamma literature, as the Samgtti-sutta has led to the Samgtti-parydya. There is in fact very little internal change from abhidharmic dgamas to Abhidhamma works; indeed, greater internal changes have come about in later Abhidharma works at a susequent period, as we shall see below.
13. Sarvdstivddin Agamas.
There are sectarian, Sarvastivadin dgamas, but there must have been some
Pruden xliii
? xliv The Abhidharma
chronological distance between the Sarvastivadin dgamas and those dgamas in their final shape (in the form that we have them today) before the split of the
2 Sarvastivadins from the rest of the Sangha.
Thus early dgamas, non-sectarian in content, led to the growth of sectarian, i. e. , Sarvastivadin dgamas, which in turn led to the growth of Sarvastivadin Abhidharma. So to know the Sarvastivadin Abhidharma, it is important to know the Sarvastivadin dgamas. This however is almost impossible.
In the Abhidharma literature, when for instance the four types of pratyayas (conditions) are mentioned (as in Kosa T. 29, p. 36b, and the Nydya-anusdra, T 29 p. 440a) the reference "as given in the Sutra" is given, and since this particular sutra can be fully reconstructed (as explained below), this one text can be claimed for the Sarvastivadins. However this specific sutra is today not found in any of the extant Pali Canon or in the Chinese translation of the Agamas.
Since the publication, in the latter half of the 18th century, of the scholar- monk Hodo's work, the Kusha-ron Keiko, it has come to be generally agreed upon by scholars that the Madhyama Agama and the Ksudraka Agama, as they presently exist in Chinese translation, are, if not Sarvastivadin in affiliation and editorship, at least very close to it. But as we know from the dgamas quoted in Sarvastivadin Abhidharma works, there was a difference between the Sar- vastivadin dgamas that exist in Chinese and those dgamas that are directly quoted by the Sarvastivadin Abhidharma literature. How great a distance there was, we have no idea.
It is clear further that the Sarvastivadin authors of their Abhidharma literature were clearly aware that there were sutras which claimed various sectarian affiliations and that there were differences among them. In the Koia, phrases like "sutras of the Mahisasakas" (p. 11. 12a; 12. 16a), "sutras of the Kas'yaplyas" (p. 23. 17a), "sutras of the Sthaviravadins" (p. 1919b), and sutras of other groups (4. 48a, 5a) occur very often. In one passage in the Kosabhdsyam where the question whether sukha (pleasure) is experienced by the mind, or by the mind and by the body (the former is a Sautrantika position, the latter, Sarvastivadin), the Sarvastivadins quote as proof of their position a sutra which the Sautrantikas criticize by saying that "all sutras hold that pleasure is experienced by the mind, whereas the sutras of the Sarvastivadins add the word 'body' in this passage. "
In another passage, the Nydya-anusdra (T. 29, p. 330a-b) says, "This is not the teaching of the Buddha, but of the sutras. We see differences in words and meanings in the scriptures of the different sectarian groups. Because the sutras have different meanings, the sectarian teachings are different. That is, the
? Sautrantikas recite The Sutra of the Seven States of Existence, and (on the basis of this) posit, in their Abhidharma, the existence of an antara-bhava (an inter- mediary existence), and so too posit a gradual manifestation of insight. There is also a text, The Basic Teachings of All the Schools {Sarva-darsana mula-parydya? ) which is not read by the Sarvastivadins. The work The Simile of Grasping with the Fist (the Hasta-dvala? ) collects together many scriptural quotations, but there are those groups, among all groups, who do not read this work. For although it collects together a number of scriptures which are unanimously read by all the seas, yet there are differences in their phraseology. "
So it would thus be worthwhile to attempt a reconstruction of the sectarian sutras of the Sarvastivadins.
14. Samathadeva's Commentary.
How can we know the Sarvastivadin dgamas, and especially those dgamas quoted by Vasubandhu in the Kosabhdsyam! One source is a commentary on the Kosabhdsyam by one Samathadeva, an Indian scholar-monk about whom nothing is known. Only one work bearing his name remains, preserved in the Tibetan Tanjur (Tohoku no. 4094; Peking no. 5595), entitled the Updyika-ndmd Abhi- dharmakosa-Tikd. Although entitled a ttkd (sub-commentary), the Updyika is not a commentary in the usual sense of that word; in fact, the Updyika is only about half the length of the Kosabhdsyam.
Rather, wherever there is a passage in the Kosabhdsyam that quotes from an agama, that passage is given by Samathadeva by the full quotation of the passage from out of the sutra text. Often the whole paragraph is given, or if not, he gives the title of the agama and the chapter or section title where the passage is to be found. Occasionally, if there is no passage to be found in relevant dgamas, related passages are given from dgamas which illustrate examples of usage. The Updyika is valuable, not so much for understanding the Kofabhasyam, but for the reconstruction of sectarian, Sarvastivadin dgamas.
It is premature to suppose that the dgamas quoted in the Updyika are the same as those seen by Vasubandhu. Still they are close enough to get a good idea of the influences of the Sarvastivadin dgamas on Sarvastivadin Abhidharma. Yet the difficulties inherent in the Updyika are still great.
For example, the Kosabhdsyam gives "in the ninth sutra of the Dirgha Agama . . . " (19. 17b), or "in the third sutra, included in the thirteenth sloka of the Sravaka-vydkhydna. . . " (2. 7b), or "in the second sutra in the Vibhanga- samgraha . . . " (10. 10b), or sometimes simply "srdvastyam niddnam . . . " ("in the
Pruden xlv
? xlvi The Abhidharma
episode at Sravasti"), or "Evam mayd srutam . . " ("Thus have I heard"). But since many of these works cannot be found in the extant Chinese or Pali editions of the Canon, comparison is impossible.
Sakurabe (in an article on pp. 155-161 of the Yamaguchi Hakase Kanreki- kinen Indogaku Bukkyogaku Ronso) traces a number of Sarvastivadin agamas, and found a remarkably close correlation between the agamas quoted in the Kosabhasyam (and the Updyika) and those texts traditionally held to be Sarvastivadin agamas in the Chinese Canon.
Several of Sakurabe's findings are: 1. that the Sarvastivadins had a fifth agama (as does the Pali Canon today), for the Updyika gives the sentence "in the Artha-vargiya sutra of the Ksudraka" (see Kosabhasyam 1. 10a), which Ksudraka (miscellany) is the same name given the fifth nikdya (the Khuddaka Nikdya) of the Pali Canon; 2. that the Sarvastivadin Ksudraka Agama has texts that a) are in the Chinese Tso O-han Ching (the Chinese translation of the Ksudraka Agama), b) which circulated separately, and c) which do not exist in the Pali Canon; and 3. that the Dtrgha Agama of the Kosabhasyam and Updyika is totally different from the Chinese edition of the Dtrgha Agama, the Ch'ang O-han Ching. His studies have shown that whereas the arrangement within the texts is often very close, texts not in the Pali can exist in both the Kosabhasyam (and Updyika) and in the Chinese Canon.
15. Sarvastivadin Abhidharma literature.
We can thus see that the early period of this literary genre went through three major states of development, as given above: 1. the early usage of the word abhidharma, 2. Abhidharmic elements in the Agamas and Nikdyas, which in turn formed the basis for 3. an independent, elaborated literature, a literature the vast bulk of which (with the exception of the Kosabhasyam and its commentary, the Vydkhyd) exists today only in Chinese and Tibetan translation.
16. Origin of the Abhidhamma
According to the Pali tradition, at night the Buddha would ascend to TavatimSa Heaven, and there he preached the Abhidhamma to his mother, Maya, and to the Devas residing in that Heaven. In the daytime he would return to the earth, where he preached this same Abhidhamma to his disciple Sariputta. Sariputta, through his supernormal powers of memory, memorized the totality of
? this teaching, and in turn recited it to Bhaddaji; Bhaddaji in turn recited it, in toto and without any error whatsoever, to his disciple, and in this way it was finally recited to Revata who, in turn, recited it publicly at the Third Council, held in Pataliputra in 251 B. C. some 235 years after the death of the Buddha.
In the Theravada tradition of Pali Buddhism there are some seven long books that go to make up the third Pitaka,the Abhidhamma Pitaka.
These books are, in the traditional order in which they are listed:
1. Dhamma-sangani, "Enumeration of the Dhammas" 2. Vibhanga, "The Book of Treatises"
3. Patthdna, "The Book of Origination"
4. Dhdtu-kathd, "Discussion of the Dhatus"
5. Puggala-pannatti, "Description of Individuals" 6. Yamaka, "The Book of Pairs"
7. Kathd-vatthu, "The Book of Controversy. "
Historians, however, place the Dhamma-sangani and the Vibhanga as the earliest of these works, followed by the Dhdtu-kathd, the Puggala-pannatti, the Kathd-vatthu, the Yamaka and the Patthdna. With the end of the composition of the Patthdna, tins Abhidhamma Pitaka is closed, and no subsequent Abhidhamma work in Pali is included within the Canon.
The Sarvastivadins claimed some six treatises (see below); these six works went to make up the Jndprasthdna, the Jndprasthdna gave rise to the Mahd- vibhdsd, and this work in turn gave rise to later compilations of doctrine.
There was no closed canon for the Sarvastivadins as far as the Abhidharma Pitaka was concerned, and the numerous references to "the seven Abhidharma books of the Sarvastivadins" must be understood in this context.
In the traditional view of the six smaller works that stand in relation to the Jndnaprasthdna, it is held that the Jndnaprasthdna is an earlier, more important work (the body), and that the other six works--shorter in length and dealing with only one or two topics--are its legs (Skt. , pdda), implying by this that they were
written subsequent to the Jndnaprasthdna in order to comment in greater detail on topics raised first in the Jndnaprasthdna. The contemporary scholarly opinion, however, is that these six works were the precursors of the largerJndnaprasthdna.
These six, the padasdstras, are:
la. Sangiti-parydya, by Sariputra (var. by Mahakausthila) composed approxi- mately 200 years after the Nirvana of the Buddha. The contents of this work closely resemble the Dasuttara-suttanta of the Dtgha Nikdya; exists in Chinese translation, T vol. 26, no. 1536.
lb. Dharma-skandha, by Maudgalyayana (var. by Sariputra). This work also
Pruden xlvii
? xlviii The Abhidharma
exists in Chinese translation, T. vol. 26, no. 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.
From the above, then, we can see that the Mdtikds (or Mdtrkds) performed an important function in the development of the corpus of Buddhist literature, as admitted in traditional Buddhist literature itself.
By itemizing the component parts of the Dharma and the Vinaya, the
Pruden xxxix
? xl The Abhidharma
Mdtrkd did play an important part in the later elaboration which is Abhidharma literature. However, the Abhidharmikas worked on these lists, minutely analyzed the items on these lists, and then proceeded to give long, exhaustive treatises on each item. So the Mdtrkd represents but one aspect of the whole picture of the growth of the Abhidharma.
If the Abhidhamma was merely speaking on the dhamma as some scholars maintain, and if the Mdtikd served as the nucleus of later Abhidhamma elaborations, why, when the literature was codified into a Pitaka, was it named the Abhidhamma Pitaka, and not the Mdtikd Pitaka? It appears then that the Mdtikds did not directly develop into the Abhidharma literature as we now have it.
As Kimura has shown, in the earliest Buddhist Sangha, abhidhamma-katha-- discussion on the dharma--had a direa relationship with the later development of the commentarial literature on the Dhamma, the Abhidhamma commentaries. In the process, there were two tendencies; one: to summarize and to itemize, and two: to analyze and elaborate. This latter tendency came to predominate, and the name for this--abhidhamma--came to be attached permanendy to this new corpus of literature.
10. Abhidharma in the Agamas: the Religion of the Agamas.
There are abhidharmic tendencies in the extant sutras, in Pali as well as in those preserved in Chinese translation. Let us say a few words first, however, about the religion of the early Buddhist canon, the religion of the Agamas.
According to de La Vallee Poussin, all the teachings of the Buddha were not publically given out. Instead, much of the philosophy and the more subtle forms of the teaching were embodied in texts which were reserved for the study of monks in their monasteries; and the Agamas (or Nikdyas), the earliest form of the Buddhist sermons which have been preserved for us, are such philosophical texts as were transmitted from one generation of monks to those of a subsequent generation. Such texts are then the "clericalized" texts, and in these texts we see only a small bit of the popular side of early Buddhism.
Such is the case, to be sure, in any religion, and this is especially so in the case of Indian religions. Any Indian religion has two sides to it: a clerical, well worked- over doctrine, and a popular aspect of the religion, which includes many elements brought over in the mind of the new converts to the religion. But the important point to remember is that the extant literature of any religion is the technical literature used in the monasteries. The real face of early Buddhism in all of its aspects cannot be gotten at only through its literature, but must also be obtained
? through archaeology, art and chronology (Le dogme et la philosophie du Bouddhisme, 1930, Chap. VII). Such a mass Buddhism was the Buddhism that preceeded the canon, "precanonical Buddhism" {Bouddhisme precanonique). Its contents were not only a darsana--a systematic school of Indian philosophy, a consistent world-view--but a faith concerned with spirits and the release of these spirits from the round of birth and death, having, according to scholars, little in common with the doctrines of anitya, andtman, and duhkha so often stressed in the Agamas (see de La Vallee Poussin's Nirvana, 1925, p. 85, 115, 131). The spread of Buddhism was dependent upon its moral teaching, the personality of its founder, its wisdom embodied in memorable sentences and couplets (the Dhammapada or Uddnavarga), coupled with popular animal tales (the Jdtakas) (see The Way to Nirvana, 1917, Chap. V). Buddhism was also closely related to ancient Indian nature worship, the worship of certain trees, and the veneration of snakes.
Buddhism also came to be changed, especially on its popular level, by virtue of the influence of non-Buddhist religions, through the conversion to Buddhism of many non-Buddhists who brought their own ideas into the company of older believers. Popular Buddhist religion absorbed much of the pan-Indian pantheon of deities. However formal and set its doctrines might have been, most of the believers of Buddhism were but "demi-civilize" or semi-civilized (Bouddhisme, 1909, p. 349 ff. ). Such is also the view of A. B. Keith (Buddhist Philosophy in India and Ceylon, 1923) and C. A. F. Rhys-Davids (Sakya or Buddhist Origins, 1931, p. 431 ff. ).
11. The Agamas and the Nikdyas.
This above view was strongly opposed by Stcherbatsky (The Conception of Buddhist Nirvana, 1927). De La Vallee Poussin would appear to oppose the popular elements of early Buddhism to the scholarly works of later Buddhism: such would imply that there is something essentially different between early Buddhism and Abhidharma Buddhism. Stcherbatsky held that the Buddha was a produa of the philosophical environment of his time, and that he obviously had a well-defined philosophy with its attendant metaphysic.
But even if the set passages and formulas were removed from the Aamas, de La Vallee Poussin and Rhys-Davids cannot say that Buddhism is merely a faith concerned with spirits and immortality. So-called popular Buddhism and "pure" clerical Buddhism cannot be so clearly distinguished one from the other. Yet the Agamas do not give a clear picture of early Buddhism, especially in its popular
Pruden xli
? xlii The Abhidharma
aspects: the Agamas are the traditions (dgama) of a scholarly elite, as de La Vallee Poussin maintains.
The Agamas, literally the "transmitted" (doctrines) were in a sense also Nikdyas (compilations): the transmitted doctrines were collected together, formal- ized, and their vocabulary became technical terms; these terms came to be analyzed through vibhangas (long, explanatory definitions) and niruktas (ety- mologies); these same terms were also organized on the basis of numerical categories or on the basis of similarities (samyukta) into mdtrkds. So the tendencies that led ultimately to a systematic Abhidharma literature led in this same process to the systematization of the Agamas (the traditions) into Nikdyas (formal literary compilations).
In summary, the Agamas are doctrinal compilations from an early stage of Buddhism, and their recensions (Nikdyas) are in a sense Abhidharmic compila- tions and, being largely abhidharmic in tendency, they led ultimately to the growth of the Abhidharma as a separate literary genre.
12. Abhidharmic Tendencies in Extant Agamas: Numerical Categories, Samyuktas, and Vibhangas,
Abhidharmic tendencies, tendencies that led eventually to the growth of a separate literature, can be seen early in some scriptures. The use of numerical categories is one such tendency.
A. The Sangtti-suttanta (the "recited" sutra, Dtgha Nikdya no. 33, vol. Ill, pp. 207-271) lists a variety of items in a list from one to ten (one item, two items, three items . . . ) Dtgha Nikdya no. 34, the Dasuttara-suttanta, lists items in a list from one to ten, but now analyzes them according to various other categories, marking then a further development along abhidharmic lines.
This same tendency is seen in the case of the Anguttara Nikdya (anguttara, "increasing by an item"), which classifies all of the suttas in its collection on the basis of numerical categories: thus all suttas dealing with any group of "four" items are collected together, followed by all suttas dealing with "five" of anything (up to eleven items). This scheme then forms the basis for this one collection of texts, or nikdya.
B. Many other suttas or agamas were joined to one another on the basis of their affinity in subject matter. To be sure, this is not sharply distinguishable from the above numerical classification, but now the items are more meaningfully arranged. Such texts are called samyuktas (Pali: samyuttas) or "conjoined" texts.
? Examples of this tendency are the Sal-dydtana vagga (vagga = section or chapter) in the Majjhima Nikdya, and the Kamma-samyutta vagga in the Majjhima: i. e. , all those texts dealing with the ayatanas were grouped together, as were all those texts dealing with karma.
This too became the guiding principle in the compilation and editing of the
Samyutta Nikdya and parts of the Khuddaka Nikdya.
C. The concept of an expanded commentary (a vibhanga) is best seen in
individual suttas in the Sarhyutta Nikdya and in the Majjhima Nikdya. Many such texts have the word vibhanga in their titles:
Samyutta Nikdya XII. 2: the Vibhangam (II, pp. 2-4), which is an expansion of XII. 1 (desand); XLV. 8, the Vibhango (V. pp. 8-10), which is a commentary on the Noble Eightfold Path; and LI. 20, the Vibhanga (V, pp. 276-281), which is a commentary on the siddhi, or supernormal powers, of a Buddha.
The Majjhima Nikdya has Majj. 135, the Cula-kamma-vibhanga sutta (cula = smaller, lesser) (III, pp. 202-206), which treats of Karma, and is followed by the Mahd-kamma vibhanga-sutta, Majj. 136 (pp. 207-215) which is an elaboration of certain of the former sutta's sections; Majj. 137, the Saldyatana- vibhanga-sutta on the six ayatanas (p. 215-222) and all of its following suttas are vibhangas: Majj. 138, 139, 140 (on the dhdtus), 141, 142. Synonymous with the vibhanga is the word vedalla, which also means "expanded": as the Cula-vedalla- sutta, and the Mahd-vedalla-sutta. The Buddha would give a short sermon, and one of his disciples, such as Katyayana, or the Buddha himself, would elaborate on it; or the disciples would discuss it among themselves, and in this way it would reach its present form.
The dgamas (in their Chinese translations) which have the characteristics of vibhangas can be rather closely identified with these same Suttas extant in Pali: they are largely the same text (especially Majj. 131 to 142, as above), a fact which does not hold for the other dgamas.
Thus abhidharmic tendencies are clearly seen in many texts in both Pali and Chinese, so far advanced in many cases that it is merely a short step to real Abhidhamma literature, as the Samgtti-sutta has led to the Samgtti-parydya. There is in fact very little internal change from abhidharmic dgamas to Abhidhamma works; indeed, greater internal changes have come about in later Abhidharma works at a susequent period, as we shall see below.
13. Sarvdstivddin Agamas.
There are sectarian, Sarvastivadin dgamas, but there must have been some
Pruden xliii
? xliv The Abhidharma
chronological distance between the Sarvastivadin dgamas and those dgamas in their final shape (in the form that we have them today) before the split of the
2 Sarvastivadins from the rest of the Sangha.
Thus early dgamas, non-sectarian in content, led to the growth of sectarian, i. e. , Sarvastivadin dgamas, which in turn led to the growth of Sarvastivadin Abhidharma. So to know the Sarvastivadin Abhidharma, it is important to know the Sarvastivadin dgamas. This however is almost impossible.
In the Abhidharma literature, when for instance the four types of pratyayas (conditions) are mentioned (as in Kosa T. 29, p. 36b, and the Nydya-anusdra, T 29 p. 440a) the reference "as given in the Sutra" is given, and since this particular sutra can be fully reconstructed (as explained below), this one text can be claimed for the Sarvastivadins. However this specific sutra is today not found in any of the extant Pali Canon or in the Chinese translation of the Agamas.
Since the publication, in the latter half of the 18th century, of the scholar- monk Hodo's work, the Kusha-ron Keiko, it has come to be generally agreed upon by scholars that the Madhyama Agama and the Ksudraka Agama, as they presently exist in Chinese translation, are, if not Sarvastivadin in affiliation and editorship, at least very close to it. But as we know from the dgamas quoted in Sarvastivadin Abhidharma works, there was a difference between the Sar- vastivadin dgamas that exist in Chinese and those dgamas that are directly quoted by the Sarvastivadin Abhidharma literature. How great a distance there was, we have no idea.
It is clear further that the Sarvastivadin authors of their Abhidharma literature were clearly aware that there were sutras which claimed various sectarian affiliations and that there were differences among them. In the Koia, phrases like "sutras of the Mahisasakas" (p. 11. 12a; 12. 16a), "sutras of the Kas'yaplyas" (p. 23. 17a), "sutras of the Sthaviravadins" (p. 1919b), and sutras of other groups (4. 48a, 5a) occur very often. In one passage in the Kosabhdsyam where the question whether sukha (pleasure) is experienced by the mind, or by the mind and by the body (the former is a Sautrantika position, the latter, Sarvastivadin), the Sarvastivadins quote as proof of their position a sutra which the Sautrantikas criticize by saying that "all sutras hold that pleasure is experienced by the mind, whereas the sutras of the Sarvastivadins add the word 'body' in this passage. "
In another passage, the Nydya-anusdra (T. 29, p. 330a-b) says, "This is not the teaching of the Buddha, but of the sutras. We see differences in words and meanings in the scriptures of the different sectarian groups. Because the sutras have different meanings, the sectarian teachings are different. That is, the
? Sautrantikas recite The Sutra of the Seven States of Existence, and (on the basis of this) posit, in their Abhidharma, the existence of an antara-bhava (an inter- mediary existence), and so too posit a gradual manifestation of insight. There is also a text, The Basic Teachings of All the Schools {Sarva-darsana mula-parydya? ) which is not read by the Sarvastivadins. The work The Simile of Grasping with the Fist (the Hasta-dvala? ) collects together many scriptural quotations, but there are those groups, among all groups, who do not read this work. For although it collects together a number of scriptures which are unanimously read by all the seas, yet there are differences in their phraseology. "
So it would thus be worthwhile to attempt a reconstruction of the sectarian sutras of the Sarvastivadins.
14. Samathadeva's Commentary.
How can we know the Sarvastivadin dgamas, and especially those dgamas quoted by Vasubandhu in the Kosabhdsyam! One source is a commentary on the Kosabhdsyam by one Samathadeva, an Indian scholar-monk about whom nothing is known. Only one work bearing his name remains, preserved in the Tibetan Tanjur (Tohoku no. 4094; Peking no. 5595), entitled the Updyika-ndmd Abhi- dharmakosa-Tikd. Although entitled a ttkd (sub-commentary), the Updyika is not a commentary in the usual sense of that word; in fact, the Updyika is only about half the length of the Kosabhdsyam.
Rather, wherever there is a passage in the Kosabhdsyam that quotes from an agama, that passage is given by Samathadeva by the full quotation of the passage from out of the sutra text. Often the whole paragraph is given, or if not, he gives the title of the agama and the chapter or section title where the passage is to be found. Occasionally, if there is no passage to be found in relevant dgamas, related passages are given from dgamas which illustrate examples of usage. The Updyika is valuable, not so much for understanding the Kofabhasyam, but for the reconstruction of sectarian, Sarvastivadin dgamas.
It is premature to suppose that the dgamas quoted in the Updyika are the same as those seen by Vasubandhu. Still they are close enough to get a good idea of the influences of the Sarvastivadin dgamas on Sarvastivadin Abhidharma. Yet the difficulties inherent in the Updyika are still great.
For example, the Kosabhdsyam gives "in the ninth sutra of the Dirgha Agama . . . " (19. 17b), or "in the third sutra, included in the thirteenth sloka of the Sravaka-vydkhydna. . . " (2. 7b), or "in the second sutra in the Vibhanga- samgraha . . . " (10. 10b), or sometimes simply "srdvastyam niddnam . . . " ("in the
Pruden xlv
? xlvi The Abhidharma
episode at Sravasti"), or "Evam mayd srutam . . " ("Thus have I heard"). But since many of these works cannot be found in the extant Chinese or Pali editions of the Canon, comparison is impossible.
Sakurabe (in an article on pp. 155-161 of the Yamaguchi Hakase Kanreki- kinen Indogaku Bukkyogaku Ronso) traces a number of Sarvastivadin agamas, and found a remarkably close correlation between the agamas quoted in the Kosabhasyam (and the Updyika) and those texts traditionally held to be Sarvastivadin agamas in the Chinese Canon.
Several of Sakurabe's findings are: 1. that the Sarvastivadins had a fifth agama (as does the Pali Canon today), for the Updyika gives the sentence "in the Artha-vargiya sutra of the Ksudraka" (see Kosabhasyam 1. 10a), which Ksudraka (miscellany) is the same name given the fifth nikdya (the Khuddaka Nikdya) of the Pali Canon; 2. that the Sarvastivadin Ksudraka Agama has texts that a) are in the Chinese Tso O-han Ching (the Chinese translation of the Ksudraka Agama), b) which circulated separately, and c) which do not exist in the Pali Canon; and 3. that the Dtrgha Agama of the Kosabhasyam and Updyika is totally different from the Chinese edition of the Dtrgha Agama, the Ch'ang O-han Ching. His studies have shown that whereas the arrangement within the texts is often very close, texts not in the Pali can exist in both the Kosabhasyam (and Updyika) and in the Chinese Canon.
15. Sarvastivadin Abhidharma literature.
We can thus see that the early period of this literary genre went through three major states of development, as given above: 1. the early usage of the word abhidharma, 2. Abhidharmic elements in the Agamas and Nikdyas, which in turn formed the basis for 3. an independent, elaborated literature, a literature the vast bulk of which (with the exception of the Kosabhasyam and its commentary, the Vydkhyd) exists today only in Chinese and Tibetan translation.
16. Origin of the Abhidhamma
According to the Pali tradition, at night the Buddha would ascend to TavatimSa Heaven, and there he preached the Abhidhamma to his mother, Maya, and to the Devas residing in that Heaven. In the daytime he would return to the earth, where he preached this same Abhidhamma to his disciple Sariputta. Sariputta, through his supernormal powers of memory, memorized the totality of
? this teaching, and in turn recited it to Bhaddaji; Bhaddaji in turn recited it, in toto and without any error whatsoever, to his disciple, and in this way it was finally recited to Revata who, in turn, recited it publicly at the Third Council, held in Pataliputra in 251 B. C. some 235 years after the death of the Buddha.
In the Theravada tradition of Pali Buddhism there are some seven long books that go to make up the third Pitaka,the Abhidhamma Pitaka.
These books are, in the traditional order in which they are listed:
1. Dhamma-sangani, "Enumeration of the Dhammas" 2. Vibhanga, "The Book of Treatises"
3. Patthdna, "The Book of Origination"
4. Dhdtu-kathd, "Discussion of the Dhatus"
5. Puggala-pannatti, "Description of Individuals" 6. Yamaka, "The Book of Pairs"
7. Kathd-vatthu, "The Book of Controversy. "
Historians, however, place the Dhamma-sangani and the Vibhanga as the earliest of these works, followed by the Dhdtu-kathd, the Puggala-pannatti, the Kathd-vatthu, the Yamaka and the Patthdna. With the end of the composition of the Patthdna, tins Abhidhamma Pitaka is closed, and no subsequent Abhidhamma work in Pali is included within the Canon.
The Sarvastivadins claimed some six treatises (see below); these six works went to make up the Jndprasthdna, the Jndprasthdna gave rise to the Mahd- vibhdsd, and this work in turn gave rise to later compilations of doctrine.
There was no closed canon for the Sarvastivadins as far as the Abhidharma Pitaka was concerned, and the numerous references to "the seven Abhidharma books of the Sarvastivadins" must be understood in this context.
In the traditional view of the six smaller works that stand in relation to the Jndnaprasthdna, it is held that the Jndnaprasthdna is an earlier, more important work (the body), and that the other six works--shorter in length and dealing with only one or two topics--are its legs (Skt. , pdda), implying by this that they were
written subsequent to the Jndnaprasthdna in order to comment in greater detail on topics raised first in the Jndnaprasthdna. The contemporary scholarly opinion, however, is that these six works were the precursors of the largerJndnaprasthdna.
These six, the padasdstras, are:
la. Sangiti-parydya, by Sariputra (var. by Mahakausthila) composed approxi- mately 200 years after the Nirvana of the Buddha. The contents of this work closely resemble the Dasuttara-suttanta of the Dtgha Nikdya; exists in Chinese translation, T vol. 26, no. 1536.
lb. Dharma-skandha, by Maudgalyayana (var. by Sariputra). This work also
Pruden xlvii
? xlviii The Abhidharma
exists in Chinese translation, T. vol. 26, no. 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.
