Articles by alphabetic order
 Ā Ī Ñ Ś Ū Ö Ō
1 2 3 4 5 6 7 8 9 0

Abhidharma II

From Chinese Buddhist Encyclopedia
Jump to navigation Jump to search


As most traditional biographies of the Buddha have it, Śākyamuni spent the last 45 years of his ultimate existence preaching the law (dharma) he had discovered in his awakening. Buddhist accounts of the Buddha’s dispensation agree in regarding it as a therapeutic and pragmatic approach to salvation adapted to the language, the religious needs, the social and psychological profiles as well as the intellectual capacities of the audience. However, while such “skill in means” (upāyakauśalya) perfectly suited the needs of early Buddhism as a missionary religion, it made a consistent and unitary account of the Buddhist doctrine difficult to provide, due to the apparent contradictions and competing levels of truth involved by the rhetorical plasticity of the discourses (sūtra). In other words, the (as tradition would have it, only apparent) doctrinal diversity of the Buddhist sūtras made a situation­ independent description of the Buddhist doctrine, not conditioned by any particular audience, a desideratum. The Abhidharma may have been developed in part in a quest to meet this need, well before the sectarian fragmentation of the Buddhist community. Since the Buddha’s words were uttered in specific situations, but the truth is context ­independent, “Abhidharma texts were considered to be explicit in meaning (nītārtha) and the interpretations presented in them were accepted as the authoritative standard by which the sūtras, which were only of implicit meaning (neyārtha), were to be interpreted” (Cox, 1995, 14).

Origins and Early Development of the Abhidharma

From an early date, the Abhidharma likely served as a vehicle for factional identity, and gradually developed into the privileged expression of Buddhist sectarian self­assertion and intersectarian polemics. Many among the extant Abhidharma works, both canonical and commentarial, reflect a clear sense of sectarian identity and contentiousness. However, there are reasons to believe that these texts, while born in a common project, underwent a complex development both before and after the emergence of sects (Cox, 1998, 145; Anālayo, 2014, 13). According to C. Cox (1998, 142), early Abhidharma likely consisted not in a set of individual texts such as those we now possess, “but rather [in] a type of exegesis that gradually developed in tandem with distinctive content, and eventually resulted in an independent branch of inquiry and a concomitant and separate genre of texts.” Originally, teachings could have been

presented orally in an abbreviated form appropriate for oral transmission, together with an attendant elaborating commentary. At the extreme, the texts could be contracted to a “skeleton” or outline format for purposes of preservation in memory and expanded in oral recitation through the insertion of stock descriptive phrases and formulaic patterns. (Cox, 1998, 141; Anālayo, 2014, 36–37)

This interpretation comes very close to the hypothesis proposed recently by Anālayo (2014, 55–89), according to whom the early Abhidharma works “seem to have grown out of a common nucleus, which appears to have been mainly discourse quotations on central themes . . . combined with a commentarial exegesis.” According to most Western scholars, Abhidharma “evolved from the practice of formulating matrices, or categorizing lists (mātṛkā), of all topics of the teaching arranged according to both numeric and qualitative criteria” (Cox, 1995, 18–19n29; Anālayo, 2014, 22n26). As pointed out by Anālayo,

the term mātṛkā derives from mātṛ, “mother,” and conveys the sense of a succinct list or summary which can be expanded and serve as the skeleton for a detailed exposition. A textual mātṛkā is thus comparable to a “mother” in the sense that it can give birth to a full exposition of a particular topic.

(Anālayo, 2014, 20–21)

This includes, on the one hand, “lists containing fundamental concepts under which it was attempted to subsume all the various elements,” and on the other hand, “attribute­mātṛkās” reflecting a method “consisting of composing a list of attributes and discussing the nature of the relevant elements with the aid of this list” (Frauwallner, 1995, 4–5).

Examples of the first type of matrix include the five constituents or aggregates (skandha, i.e., corporeality [[[rūpa]]], feeling [[[vedanā]]], conception [[[saṃjñā]]], conditioning factors [[[saṃskāra]]], bare cognition [[[vijñāna]]]), whereas examples of the second type are supplied by dyads (but also triads, tetrads, etc.) such as corporeal/incorporeal (rūpin/arūpin), visible/invisible (sanidarśana/anidarśana), and conditioned/unconditioned (saṃskṛta/asaṃskṛta). An important taxonomic list or “ meta­list” (Ronkin, 2005, 27) consists of the 37 “limbs of awakening” (bodhipākṣikadharma). As Anālayo points out,

“[t]his basic list [covers] the mental qualities that tradition considers crucial for progress to awakening; and it is this mātṛkā which according to the Mahāparinirvāṇa-sūtra was taught by the Buddha just before his passing away, and the mātṛkā that the Pāsādika-sutta and the Sāmagāma- sutta . . . recommend for ensuring communal har­mony. (Anālayo, 2014, 50)

The limbs of awakening include:

1. f our applications of mindfulness (smṛtyupasthāna);

2. f our right exertions (samyak-prahāṇa/pradhāna);

3. four bases of supernatural ability (ṛddhipāda);

4. five faculties (indriya);

5. five powers (bala);

6. seven members of awakening (bodhyaṅga); and

7. the eightfold noble path (āryāṣṭāṅgamārga). These 37 items are variously described as a precondition for liberation from the fluxes (āsrava, negative pollutants of one’s mental stream), as the cultivation of the path (mārgabhāvanā), as the jewels (ratna) of the dharma, and as the quintessence of the Buddha’s teaching “about which no disagreement exists” (Gethin, 1992).

According to Anālayo, of decisive influence on the evolving Abhidharma must have been attempts made to be as comprehensive as possible, to supplement the directives given in the early discourses for progress on the path with a full picture of all aspects of the path in an attempt to provide a complete map of everything in some way related to the path . . . Equipped with a complete map of the doctrine, the disciples are fortified in their struggle for survival in competition with non­Buddhist teachers and in their attempts to maintain harmony within the Buddhist fold.

(Anālayo, 2014, 168) This drive towards giving a comprehensive inventory of all that there is may have been meant as a symbolic counterpart of the Buddha’s omniscience (Anālayo, 2014, 94, 109).

This pre­ or proto­Abhidharmic “taxonomic organization through lists or matrices” (Cox, 1995, 9) is visible in sūtras using lists to summarize the Buddha’s teaching, such as the Saṅgītisūtra and the Daśottarasūtra, both of which arrange their items numerically (from ones/monads to tens/ decads and ten­times­ten, respectively) and the Arthavistarasūtra, which organizes its elements thematically (Anālayo, 2014, 29–39; note that the three sūtras are ascribed to the Buddha’s great disciple and “Abhidharma specialist” Śāriputra; Migot, 1954, 519–532). In addition, numerical and topical arrangement was not only meant to organize individual sūtras, but also to “dictate the structure of entire collections of sūtras

(Cox, 1995, 9), the most famous examples being the Aṅguttaranikāya or Ekottarikāgama grouping, in which the sūtras are classified according to the increasing number of the doctrinal items they deal with, and the Saṃyuttanikāya or Saṃyuktāgama, in which the sūtras are organized according to their respective topics. Recent scholarship on the origins of Abhidharma insists, however, that the use of lists, which may have characterized the

Vinaya (monastic discipline) and especially the Prātimokṣa (disciplinary precepts) at least as much as the Abhidharma, cannot be considered specific to Abhidharma, and should rather be interpreted as a common feature of oral transmission (Cox, 1995, 8; Anālayo, 2014, 22–28, 83). In another reference to the wider cultural context, for E. Frauwallner (1995, 40), the Buddhist mātṛkās likely were an answer to the Brahmanical sūtras, concisely formulated rules such as those that can be found in the works of various philosophical schools or in grammatical works, and which, like the mātṛkās, require an explanation to be understood.

According to another hypothesis, the origin of Abhidharma is not simply to be found in this taxonomic approach, but “in dialogues concerning the doctrine (abhidharmakathā), or monastic discussion in catechetic style characterized by an exchange of questions and interpretative answers intended to clarify complex or obscure points of doctrine” (Cox, 1995, 8; Kimura, 1968, 27–43; Sakurabe, 1969, 13–29). This hypothesis of the origins of Abhidharma in a certain question and answer format is certainly not incompatible with the idea that it arose from matrices used to organize concepts, since

[t]he tendency toward organization represented by mātṛkā and that toward discursive explanation represented by abhidharmakathā together constitute the exegetical method characteristic of mature Abhidharma analysis, and both tendencies are found in incipient form in Abhidharma treatises from the earliest period onward. (Cox, 1995, 9)

We must remember, however, that whatever we understand to be the origin of Abhidharma, these origins are not the Abhidharma per se: the two main approaches that have been considered as explaining the coming into being of the Abhidharma – the use of mātṛkās and the question­and­answer format – are features that are not in themselves necessarily characteristic of Abhidharma thought, however much they may have contributed to its formulation. (Anālayo,

Meaning and Function of the Abhidharma

The meaning of the word abhidharma and its two components, the prefix abhi and the substantive dharma, seems to have changed according to scholarly milieu, time and place. The northern Indian scholiasts understand abhi in the sense of “with regard to” and dharma in the sense of “teaching,” the compound expression thus meaning “with

regard to the teaching,” that is, the study of the Buddha’s teaching. This usage seems to be well documented in early Pali literature as well as in most early discoursesunderstanding of the expression abhidharma (Anālayo, 2014, 70–79 and 70n49). The later Pali commentators seem to favor an interpretation of abhi as “highest,” “further,” and of dharma as “teaching,” that is already listed in the important commentary *

Abhidharma Mahāvibhāṣā (T. 1545 [XXVII] 4b13; Anālayo, 2014, 151n65). According to this second interpretation, abhidharma consists in the highest or supreme teaching, “that which goes beyond what is given in the Buddha’s discourses, in a sense somewhat reminiscent of the term ‘metaphysics’ ” (Ronkin, 2005, 26). Both interpretations are represented in the Chinese translations of the word: whereas early renderings such as wubifa

(無比法) reflect an interpretation of abhi as “highest” or “beyond compare,” translations such as duifa (對法) and xiangfa (向法) rely on an understanding of abhi in the sense of “directed toward” (Cox, 1995, 16n1, 16n7). The great commentator and Abhidharma master Vasubandhu (350–430 CE[?]) understands abhi in the sense of “directed toward,” “facing” (abhimukha), but dharma in the sense of either nirvāṇa (the supreme dharma) or the numerous individual factors of existence making up the world of ordinary experience (AKBh. 2,10–11; AKVy.

9,18ff.). The Vibhāṣā compendia, a series of fundamental Abhidharma commentaries, list 24 interpretations of the word (Watanabe, 1983, 23–24). These fall under three major categories: (1) the Abhidharma “enables one to discriminate and analyze the factors according to their generic and particular inherent characteristics”; (2) it “enables one to suppress non­Buddhist or false Buddhist doctrines and establish the true teaching” (see Anālayo, 2014, 124–125); (3) it “enables one to remove defilements and progress along the path” (T. 1545 [XXVII] 4a12ff.; T. 1546 [XXVIII] 3c4ff.; T. 1547 [XXVIII] 418a1ff.; Cox, 1995, 4).

These exegetical, expository(/apologetic) and soteriological meanings of abhidharma are well in tune with the six functions ascribed to the Abhidhar ma by the early * Āryavasumitrabodhisattvasaṅgīti (T. 1549 [XXVIII] 733a16ff.; Cox, 1995, 4). Considered in an exegetical perspective, the Abhidharma “enables one to discriminate the meaning or nature of dharmas as presented in the sūtras.” In its expository function, the Abhidharma “enables one to cultivate the four noble truths,” “enables one to realize the twelve members of dependent origination and dependently originated factors,” and “expounds the meaning of the eightfold noble path.” These three aspects have strong

soteriological overtones and obviously impinge upon the properly gnostic and cathartic, hence salvational, dimensions of the Abhidharma. Considered in this perspective, the Abhidharma can be defined as “that which analyzes and describes the causes of the various factors instrumental in the complete cessation of defilements” and “that through the cultivation of which one attains nirvāṇa.” The factor the cultivation of which brings about nirvāṇa is none other than insight or discernment (prajñā). This sixth function of the Abhidharma thus comes very close to a definition provided by the Vibhāṣā compendia and accepted by most later authorities, that is, Abhidharma as the controling faculty of pure (anāsrava) insight together with “that which furthers, is associated with, or contains this pure insight,” hence the Abhidharma treatises themselves (T. 1545 [XXVII] 2c23; T. 1546 [XXVIII] 2c27ff.;

T. 1547 [XXVIII] 417b3ff.; also AKBh. 2,3ff.; AKVy. 8,10ff.; T. 1562 [XXIX] 329a29ff.). These various functions must be considered as a whole, however, for through [the] exercise of completely describing the character of each factor in every instance of its occurrence, the factors of which experience is

composed can be seen as they actually are, the misconceptions obscuring our perception of experience can be discarded, the factors obstructing and ensnaring us can be abandoned, and the factors contributing toward liberation can be isolated and cultivated. (Cox, 1995, 12) The numerous functions of the Abhidharma are perhaps best described in a w ell­known passage from the canonical Sarvāstivāda Vijñānakāya (see below):

Abhidharma is the light of the true doctrine; without Abhidharma treatises, one would not be able to destroy the darkness surrounding what is to be known by knowledge. Abhidharma functions as the pure eye within the mind, as the basis of all knowledge; it is the sun illuminating the forest of things to be known, the sword that destroys heretical texts; it constitutes the authority

for those who open the eyes of sentient beings and is the womb of the Tathāgatas; it is the illumination in the three realms, the path of the eye of insight; it is the light of all factors and the ocean of the Buddha’s words; it is able to issue forth highest insight and to remove all doubts. (T. 1539 [XXVI] 531a13–17; trans. Willemen, Dessein & Cox, 1998, v)

Authenticating the Abhidharma

We do not know whether all groups once possessed their own Abhidharma corpus, but once the exegetical inquiry that was typical of the earliest, pre­sectarian Abhidharma had crystallized into sect­specific textual corpora, the authentication of these scriptures became an important task for those Buddhist denominations that did possess them (Lamotte, 1976, 197–210; Davidson, 1990, 303–305). The most radical attitude in this regard is the Theravāda view, according to which the newly awakened and omniscient Buddha grasped (adhigata) and collated (vicita) the seven Abhidhamma books under the Bodhi tree and subsequently taught them to his mother Māyā in the Trāyastriṃśa (Pal.

Tāvatiṃsa, “ Thirty­three”) heaven during a three­month rain retreat. Every evening, the Buddha descended to Lake Anavatapta (Pal. Anotattha) and repeated the day’s lesson to his disciple Sāriputta who, having “laid down the numerical series in order to make it easy to learn, remember, study and teach the Law,” taught it in turn to 500 disciples (As I.16; Buswell & Jaini, 1996, 80, 18–21). This Theravāda position was challenged by those who claimed that the Kathāvatthu, one of the seven treatises allegedly preached by the Buddha in the Trāyastriṃśa heaven, had been composed by the elder Moggaliputtatissa on the occasion of the Second Council at Pāṭaliputra (3rd cent. bce). As an answer, the orthodox Theravādins claimed that while preaching to his mother, the Buddha had contented himself with expounding a mātṛkā of the Kathāvatthu, foreseeing that Moggaliputtatissa would develop the treatise in full in Pāṭaliputra (As I.3–6; Lamotte, 1976, 201).

While admitting that their seven Abhidharma treatises ultimately went back to the Buddha, the Sarvāstivādins and especially the Vaibhāṣikas recognized that the treatises in question had been collected and organized by great disciples of the Buddha, or, alternatively, that they had been ascertained by later scholiasts such as Kātyāyanīputra thanks to their praṇidhijñāna, “a transtemporal intuition resulting from the power of their vows . . . by which they [were] able to perceive past and future events” (Cox, 1995, 17n19). Yaśomitra (AKVy. 12,2–7) offers

an analogy to the authenticity of the canonicity of the Abhidharma, referring to the Vaibhāṣika claim that the case is comparable to that of Dharmatrāta collecting the udānas (aphorisms) uttered by the blessed one to create the Udānavarga. Likewise, [persons] such as the elder Kātyāyanīputra gathered and fixed the Abhidharma that had been uttered here and there by the Blessed One in works such as the Jñānaprasthāna for the sake of the trainees, the Abhidharma which consists in an instruction on the characteristics of the factors. These disciples of the Buddha contented themselves, in this scenario, with adding summarizing stanzas (uddāna), dividing the teaching into

chapters (skandhaka), and so forth. This strategy of authentication likely was facilitated by the fact that the name of the scholiast Kātyāyanīputra (variously dated to the late 1st, to the 3rd, or to the 5th cent. after nirvāṇa by Buddhist authorities) was very similar to that of Kātyāyana, a prominent disciple of the Buddha considered to have played a significant role at the First Council at Rājagṛha (Kātyāyana was also famous for having picked out the quintessence of the dharma and submitted this work to the Buddha, who approved it and labelled it the “supreme law,”

abhidharma; Lamotte, 1976, 208; Przyluski, 1926, 201, 303). In this picture, the Sarvāstivādin “ six­membered Abhidharma” (ṣaṭpādābhidharma, i.e., the Jñānaprasthāna – the body [[[śarīra]]] – and its six ancillary treatises) was conceived of as the (editorial) work of great disciples of the Buddha or later authors: Mahāmaudgalyāyana, Śāriputra, Mahākauṣṭhila, Vasumitra, Devaśarman and Kātyāyana/Kātyāyanīputra (see below).

Moving yet a step further, the Sautrāntikas are said to have denied the Abhidharma any independent authority and regarded the sūtras alone as authoritative, whence their name (AKVy. 11,29–30). According to them, the Abhidharma treatises were nothing but human compositions. The early 5th­century Vaibhāṣika scholiast Saṅghabhadra records at least three reasons why his Sautrāntika opponent Vasubandhu (350–430 CE?) rejected the Buddha’s authority of the Abhidharma works (T. 1562 [XXIX] 329c6ff.; Honjō, 2010, 181–187): (1) these treatises are traditionally ascribed to disciples/authors such as Kātyāyana/Kātyāyanīputra, (2) the Buddha has advised Ānanda to rely

(pratisaraṇa) on the sūtras and not on the abhi dharma (T. 1562 [XXIX] 329c20–21), and (3) the Abhidharmas of the different Buddhist denominations exhibit contradictory positions on one and the same question (for Saṅghabhadra’s refutation of the views of Vasubandhu, see Cox, 1995, 6–7). The Dīpavaṃsa’s (V.37) Mahāsaṅgītikas also regarded the Abhidharma as not the word of the Buddha.

Needless to say, one of the most decisive criteria for scriptural authorization was the recitation­ cum­compilation of a given text or group of texts by a great disciple of the Buddha during the First Council, held in Rājagṛha immediately after the death of the master. According to the Vinayas of the Theravādins, the Mahīśāsakas, and the Mahāsāṅghikas (resp. Vin. ii.287; T. 1421 [XXII] 191ac; T. 1425 [XXII] 491c16), the monks assembled in Rājagṛha contented themselves with reciting, after Ānanda and Upāli, the dharma and the Vinaya, “the law and the

discipline,” no mention being made of the Abhidharma as an independent third “basket” (Kimura, 1968, 27–28). According to Anālayo (2014, 18), “the accounts of the first saṅgīti [communal recitation] in these Vinayas seem to have been finalized at a time when the Abhidharma had not yet become a collection in its own right.” Other sources such as the Mūlasarvāstivāda Vinaya and the Aśokāvadāna ascribe the compilation of a mātṛkā to the convener of the first council, Mahākāśyapa (resp. T. 1451 [XXIV] 408b2–11; T. 2042 [L] 113c3–4; T. 2043 [L] 152a5). However, most

sources agree that Ānanda himself recited an Abhidharmapiṭaka of various length (including mātṛkās alone) in Rājagṛha. This is true of, among many others, the Vinaya of the Dharmaguptakas (T. 1428 [XXII] 968b25–26), the Vinayamātṛkā of the Haimavatas (T. 1463 [XXIV] 818a28–29), and late Pali commentaries (Sv I.17; As I.3; Smp I.18). According to the Sarvāstivāda Vinaya (T. 1435 [XXIII] 449a19), upon being asked by Mahākāśyapa, Ānanda revealed that the Buddha had taught the Abhidharma in Śrāvastī, and that “[t]he teaching given at that time was that breaches of the five precepts are conducive to rebirth in hell, whereas keeping the five precepts leads to a heavenly rebirth” (Anālayo, 2014, 19).

Extant Abhidharma Texts

Only the Theravāda has preserved a complete Abhidhammapiṭaka in an Indian language (Pali), while the seven canonical abhidharma treatises of the Sarvāstivādins have come down to us almost completely in Chinese translation. A very small portion of them is available in Tibetan, while a few scattered Sanskrit fragments have been preserved (Dietz, 2004). Only minute fragments

of canonical Abhidharma texts of sects other than Theravāda and Sarvāstivāda have been preserved: some materials presumably belonging to the Kāśyapīya tradition are attested (Cox, 2014), and fragments from the *Śāriputrābhidharma have been identified in the Schøyen collection (see below).

Theravāda canonical Abhidhamma

In style and language, the Theravāda Abhidhamma texts markedly differ from the Vinayapiṭaka and Suttapiṭaka (von Hinüber, 1996, § 137). Short questions and answers as well as many formulas often similar to those found in the Niddesa characterize the Abhidhamma literature. The Abhidhammapiṭaka and its parts are mentioned under this name for the first time in the Milindapañha (12,21–13,7). As there is no corresponding paragraph in the Chinese Nāgasenabhikṣusūtra, the Chinese version of the Milindapañha, the reference to this text may have been inserted only in Ceylon (von Hinüber, 1996, § 179). The Theravāda Abhidhammapiṭaka consists of the seven texts listed and briefly described below.


The Theravāda Abhidhammapiṭaka begins with the Dhammasaṅgaṇi (alternative title Dhammasaṅgaha; Collection of Dhammas; the form Abhidhammasaṅgaṇi found in some early manuscripts is a misnomer). The structure of the text is discussed at length by E. Frauwallner (1995, 53–86). The absence of the traditional introduction (nidāna) “Thus have I heard . . .” in a text considered the “word of the Buddha” (buddhavacana) created a problem for Buddhist exegetes. The issue is discussed in the commentary, where various attempts to add a nidāna are mentioned (As 30,16–31,16). Usually, the Buddha is reported to have recited the Abhidhamma in heaven to his deceased mother and to

the gods (see above), and later to have handed it down to Sāriputta (a list of the sequence of teachers and pupils from Sāriputta up to Mahinda is found in As 32,12–20). Certain manuscripts of the Dhammasaṅgaṇi even insert “At one time the Lord was residing among the Tāvatiṃsa gods . . .” in front of the mātikā with which the text begins. This abhidhammamātikā or “abhi dhamma list,” which also occurs in other Theravāda Abhidhamma texts, comprises the concepts kusala “salutary, (morally) good,” akusala “(morally)

bad,” and abyākataundetermined,” immediately followed by the suttantamātikā or “suttanta (sūtra) list,” which is largely based on concepts taken over from the Saṅgītisuttanta (i.e. suttanta no. 33 of the Dīghanikāya; Anālayo, 2014, 29–37). The text that follows (Dhammasaṅgaṇi §§ 1–1599) expands and explains the mātikās. It is divided into four sections (kaṇḍa): (1) Cittuppādakaṇḍa (§§ 1–582;

Section on the Arising of Thoughts), (2) Rūpakaṇḍa (§§ 584–980, with another mātikā, §§ 584–594; Section on Matter), (3) Nikkhepakaṇḍa (§§ 981–1367; Summary Section), and (4) Atthuddhārakaṇḍa (§§ 1368–1599; Commentary Section). The titles of the sections are slightly different in the commentary (As 6,13–7,9). As the mātikās located at the beginning of the text are discussed in sections 3 and 4 only, it is clear that sections 1 and 2 are later additions. The Dhammasaṅgaṇi is considered the youngest of the texts assembled in the Abhidhammapiṭaka (Frauwallner, 1995, 53).


Although the Vibhaṅga (Explanation; Frauwallner, 1995, 43–48; von Hinüber 1996, §§ 138–139) does not begin with a mātikā, it presupposes such a list, which can indeed be reconstructed from the text itself. Usually the starting point is old canonical lists such as the five khandhas (rūpakhandha, vedanākhandha, saññākhandha, saṅkhārakhandha, viññāṇakhandha, “the aggregates matter (body), feeling, perception, conditioning factors, consciousness”) or the twelve āyatanasspheres (of perception),” that is, eye/visible object, ear/sound, and so forth, up to mind/mental object and so forth. Thus the Vibhaṅga systematizes older material drawn from the Suttapiṭaka.

The text is divided into 18 chapters with chapters 1–6 and 7–15 forming two units expanding on two mātikās. The last three chapters were originally separate Abhidhamma treatises which were subsequently included into the Vibhaṅga. The arrangement of chapter 16 (Ñāṇavibhaṅga; Explanation of Knowledge) and chapter 17 (Khuddakavatthuvibhaṅga; Explan

ation on the Small Items) follows the same numerical principle as the Aṅguttaranikāya with groups of 1–10, 18, 36, 72, of which only the first refers to Abhidhamma matters. Chapter 18 (Dhammahadayavibhaṅga; Explanation of the Heart of the Teaching) also begins with a mātikā of its own comprising, again, the five khandhas, the twelve āyatanas, and so forth. This chapter could be identical with the Mahādhammahadaya, which,

according to the Atthasālinī, was considered as a possible candidate to replace the allegedly noncanonical Kathāvatthu (see above). The Vibhaṅga shows many similarities with the Dharmaskandha, thus connecting early Theravāda and Sarvāstivāda Abhidharma (Anālayo, 2014, 88; see also below). It is generally considered the oldest text in the Abhidhammapiṭaka.


The Dhātukathā (Discussion of the Elements; Frauwallner, 1995, 48f.; von Hinüber, 1996, § 140) is a brief text which starts from mātikās of concepts similar to those found in the Vibhaṅga. The Dhātukathā bears some similarities with the Sarvāstivāda Dhātukāya and both may ultimately go back to the same origin (see below). The treatise owes its title to the fact that it investigates the ways in which

different concepts are related to the dhātus or “elements.” These elements are systematically analyzed as to whether they are included or not (saṃgahīta/asaṃgahīta), associated or dissociated (sampayutta/vipayutta), in a fourfold way: “included with non­included,” “n on­included with included,” “included with included,” and “n on­included with non­included.” This in a way anticipates methods of analysis fully developed only in later Buddhist scholasticism.


The Puggalapaññatti (Concept of Person; Frauwallner, 1995, 49ff.; von Hinüber, 1996, § 141–143; Anālayo, 2014, 52ff.) begins with a mātikā partly based on the abhidhammamātikā (see above). Various types of puggalas or “individuals, persons” are listed in groups from 1 to 10. This makes the text formally close to the Aṅguttaranikāya and particularly to the Dasuttarasuttanta (DN no. 34). When the compiler of the Puggalapaññatti drew material from the Suttapiṭaka, particularly from the Aṅguttaranikāya, he effaced the original “remembered orality” (von Hinüber, 1996, § 55) by removing the address bhikkhave,

“o monks,” and so on, thus assimilating the text to the Abhidhamma style. Because he limited his effort to reassembling the materials, it is impossible to assign to the text a place in the history of Abhidhamma. In spite of a similarity in title, the Puggalapaññatti, which has no parallels outside Theravāda, is entirely different from the Sarvāstivāda Prajñaptiśāstra (see below).


The Kathāvatthu (Text Dealing with Disputes; Frauwallner, 1995, 86ff.; von Hinüber, 1996, § 144–151; for a Japanese translation, see Satō, 1991) is the only part of the Abhidhammapiṭaka traditionally dated (to the time of Aśoka, 218 years after the nirvāṇa) and ascribed to an author, Moggaliputtatissa, who is reported to have explained and expanded a mātikā of the Kathāvatthu originally uttered by the Buddha in heaven (see above). This assumption was made to save the canonicity of the text, which was not beyond doubt (As 3,25–29; 6,1–12). In structure

and content, the Kathāvatthu, which consists of strictly formalized questions and answers, is quite different from the other Abhidhamma texts, because about 200 (according to tradition 500) controversial points of the Buddha’s teaching are discussed (McDermott, 1975; Ganeri, 2001). The different views are ascribed to specific Buddhist schools by the commentary, which may be separated from the oldest parts of the Kathāvatthu by more than half a millennium.

Although some parts of the Kathāvatthu seem very old, in particular the discussion on the puggala or “person,” which exhibits ancient linguistic features (Norman, 1979) and is close in content to the Sarvāstivāda Vijñānakāya (Bronkhorst, 1993), the specific structure of the text easily allows for additions, which are difficult to discern, as remarked long ago by La Vallée Poussin (1922, 520).


The Yamaka (Pairs; Frauwallner, 1995, 51–53; von Hinüber, 1996, § 152ff.) is a huge text which would cover about 2500 pages if it were printed in full. According to tradition, it is 25 times longer than the Majjhimanikāya. As suggested by E. Frauwallner, the title was chosen because two things form a pair when the second thing originates from the first. The fairly complicated structure of the Yamaka is explained in detail at

the beginning of the commentary. All pairs are discussed at great length with all imaginable combinations being enumerated, which inspired the following remark to Frauwallner (1995, 53): “This is a particularly glaring example of how an intrinsically interesting problem can be inflated to the point of inanity using the Abhidharma method.” Still, a later Theravāda Abhidhamma text contends: “The text is succinct to the extreme” (Mohavicchedanī 279,14).


The Paṭṭhāna (Basis [of All Other Abhidhamma Texts]; Frauwallner, 1995, 50ff.; von Hinüber, 1996, § 154 ff.), also known as Mahāpakaraṇa (Large Treatise), is by far the longest text in the Tipiṭaka. While explaining the title, the commentary even states that the actual length of the text is incalculable. All editions therefore abbreviate, that of the Pali Text Society into two volumes; to gain a better picture, it is imperative to use the f ive­volume Burmese edition (still abbreviated considerably!). The Paṭṭhāna begins with the section on Tika or triads followed by the section

on Duka or dyads (arranged in the wrong order in the Pali Text Society edition). The Paṭṭhāna, which tries to comprehensively explain causality and thus, in the traditional understanding, to facilitate the use of the suttantas for Abhidhamma specialists, does not provide any new insights but endlessly invents new possible combinations of concepts: “method has replaced genuine thought” (Frauwallner, 1995, 51).

One canonical Theravāda Abhidhamma text is found in the Tipiṭaka, but outside the Abhidhammapiṭaka: the Paṭisambhidāmagga (also called simply Paṭisambhidā; Path of Discrimination; Frauwallner, 1995, 87–89; von Hinüber, 1996, § 119–120). It is the only canonical Theravāda Abhidhamma text included in the Khuddakanikāya of the Suttapiṭaka, which seems to have been the only part of the Tipiṭaka still open for additions at the time when the Paṭisambhidāmagga was created. The commentary ascribes this work to Sāriputta. According to the Dīpavaṃsa (Chronicle of the Island [[[Ceylon]]]; V.37), the Paṭisambhidāmagga was rejected during the Second Council by the Mahāsāṅghikas, who would certainly never accept this typical Theravāda text; the Dīpavaṃsa’s assertion, moreover, is a glaring a

nachronism. The Paṭisambhidāmagga consists of various texts apparently put together more or less at random. Only the first part, the Ñāṇakathā (Discussion on Knowledge), which originally might have been a separate work, begins with a mātikā; the other parts quote and comment on canonical suttantas. This text, which may have been put together in the 2nd century ce, is perhaps the first attempt to create a systematic handbook of Abhi dhamma and is as such a forerunner of Upatissa’s Vimuttimagga (Path to Liberation) and Buddha ghosa’s Visuddhimagga (Path to Purity;

Frauwallner, 1995, 89–95; von Hinüber, 1996, § 245– 250). While the Visuddhimagga became the basic handbook for Theravāda (Mahāvihāra) orthodoxy, the Vimuttimagga enjoyed a wide international recognition, being still in use in 12th­century Bengal (Skilling, 1987, 7, 15).

Sarvāstivāda Canonical Abhidharma

Only a tiny percentage of the canonical Abhidharma texts of the Sarvāstivāda survive in an Indic language. With one exception (the Prajñapti), all are extant in Xuanzang’s (玄奘) Chinese translations; two (the Prakaraṇapāda and the Jñānaprasthāna) have been translated twice into Chinese. Three sections of the Prajñapti are available in Tibetan translation, and one section survives in an 1 1th­century Chinese translation by Dharmapāla (Fahu [法護]) and Wejing (惟淨).

The titles of the seven Sarvāstivāda canonical treatises, together with their Tibetan and Chinese equivalents, are as follows:

1. Saṅgītiparyāya; ’Gro ba’i rnam grangs; Jiyimen zulun (集異門足論; T. 1536; trans. Xuanzang);

2. Dharmaskandha; Chos kyi phung po; Fayun zulun (法蘊足論; T. 1537; trans. Xuanzang);

3. Prajñaptiśāstra; Btags pa; Shishe zulun (施設足論; T. 1538; trans. Dharmapāla);

4. Vijñānakāya; Rnam par shes pa’i tshogs; Shishen zulun (識身足論; T. 1539; trans. Xuanzang);

5. Dhātukāya; *Khams kyi tshogs; Jieshen zulun (界身足論; T. 1540; trans. Xuanzang);

6. 88Prakaraṇa]](pāda); Rab tu byed pa; Pinlei zulun (品類足論; T. 1542; trans. Xuanzang)/Zhong shifen apitan lun (衆事分阿毘曇論; T. 1541; trans. Guṇaprabha and Bodhiyaśas); and

7. Jñānaprasthāna/ Aṣṭaskandhaka; Ye shes la ’jug pa; Fazhi lun (發智論; T. 1544; trans. Xuanzang)/Ba jiandu lun (八犍度論; T. 1543; trans. Saṅghadeva, Zhu Fonian, and Dharmapriya)

Since the Jñānaprasthāna was recognized as the most important Abhidharma text, it came to be known as “the body­treatise (kāyaśāstra),” whereas the other canonical texts were referred to as its “six feet” (ṣaṭpāda), the seven canonical Sarvāstivāda treatises being referred to collectively as the “Abhi dharma with six feet” (ṣaṭpādābhidharma). This usage, which is not attested in the Vibhāṣā compendia, finds its first known occurrence in a postscript to the Chinese translation of the *Aṣṭaskandhaśāstra (T. 1543 [XXVI] 887a19–24) written in 379 or 390 ce, very close in time to Kumārajīva, whose translations of the Mahāprajñāpāramitopadeśa (T. 1509) and the *Tattvasiddhi (T. 1646) mention the ṣaṭpādābhidharma without stating which individual texts are meant (Watanabe, 1954, 36ff.; Willemen, Dessein & Cox, 1998, 160–162). Sthiramati’s commentary on the Abhidharmakośa, the Tattvārthā, however, does not regard the Dhātukāya as a canonical

Abhidharma text, and thus enumerates a total of only six treatises (D 4421, tho 324a2/ P 5875, tho 91a1). Pūrṇavardhana’s Lakṣaṇānusāriṇī (D 4092, cu 324a2/P 5594, ju 380b4) does likewise. This may explain the absence of citations of the Dhātukāya in important texts such as the Mahāvibhāṣā (Cox, 1998, 160n60; see below) and the Abhidharmakośa (Hirakawa, 1973). When cited in the Abhidharmakośa, the six treatises are called śāstras

(benlun [本論]). The chronology of the seven treatises is difficult to assess. According to Puguang (普光, T. 1821 [XLI] 8b24f.), the chronological sequence of the seven treatises is as follows: Saṅgītiparyāya, Dharmaskandha and Prajñaptiśāstra (written/edited by direct disciples of the Buddha), Vijñānakāya (1st cent. after nirvāṇa), Prakaraṇapāda and Dhātukāya (early 3rd cent. after nirvāṇa), and Jñānaprasthāna (late 3rd cent. after nirvāṇa; Cox, 1995, 47n55). E. Frauwallner (1995, 13–14) was inclined to accept this relative chronology.

Saṅgītiparyāya or Abhidharmasaṅgītiparyā yapādaśāstra

The Saṅgītiparyāya, ascribed to Mahākauṣṭhila and Śāriputra in the Indian and Tibetan (Yaśomitra, Bu ston) and the Chinese traditions respectively (Cox, 1998, 177n114), is a commentary on the Saṅgītisūtra, a discourse that, in increasing numerical order (from one to ten), lists 205 major concepts, for example, the three kinds of training (śikṣā), the four noble truths (āryasatya), and the five aggregates (skandha; Cox, 1998, 179; for another, Gandhari commentary on the Saṅgītisūtra, see Cox, 2014, 36–39; on a Sanskrit folio of the Saṅgītiparyāya in the Schøyen collection, see Matsuda, 2002, 239n1). The frame story of the treatise

relates how disputes and schisms arose in the Jaina community after the Jina’s death due to differing interpretations of the doctrine, and how Śāriputra, in order to prevent similar disputes within the Buddhist community, recited a systematic collection of the Buddhist doctrinal concepts and how the Buddha approved of and endorsed Śāriputra’s recital. (Waldschmidt, 1955, 299, 309–314)

This treatise is generally regarded as the earliest Sarvāstivāda Abhidharma text; its historical importance lies in the fact that it marks “the onset of a commentarial genre that was to form the Abhidharma” (Cox, 1998, 178). Although the Saṅgītiparyāya shows nonsectarian features common to ancient Buddhism, one can find the concept of “cessation through non­intelligence” (apratisaṃkhyānirodha). This notion (in

Abhidharma technical vocabulary, this dharma) is characteristic of the Sarvāstivāda (a denomination that is mentioned at the end of each of the 20 chapters) in that it presupposes the theory that the past, the future, and the present exist. (Xuanzang’s Chinese translation is translated into Japanese by Watanabe, 1929, and into German by Stache­Rosen, 1968 together with the text’s extant Sanskrit fragments; newly found fragments of this text are included in Matsuda, 2006; English summaries of the work are available in Frauwallner, 1995, 14–15; Potter, 1996, 203–216; Willemen, Dessein & Cox, 1998, 67–68; Cox, 1998, 177–181.)

Dharmaskandha or Abhidharmadharmaskandhapādaśāstra

The Dharmaskandha, which, together with the Saṅgītiparyāya, belongs to the earliest stage of the Sarvāstivāda Abhidharma treatises, is considered in the Indian tradition to have been compiled by the Buddha’s direct disciple Śāriputra, while the Chinese tradition ascribes it to Maudgalyāyana (Cox, 1998, 181n126). The full text is extant only in Xuanzang’s Chinese translation. The treatise is also known through 22 Sanskrit folios from a Mūlasarvāstivāda recension found in Gilgit (Dietz, 1985), which cover about 20 percent of the text. This Sanskrit

recension, which seems to differ in structure from the Chinese version, has been edited by S. Dietz (1984) and supplemented by K. Matsuda (1986). The treatise consists of 21 chapters (Frauwallner, 1995, 16–17), at the beginning of each of which is quoted a sūtra, followed by explanations of the sūtra’s important concepts. E. Frauwallner (1995, 15–21) divides the treatise into three sections. The first section comprises 15 chapters providing detailed expositions on the “path to liberation” and various issues of religious praxis. The second section consists wholly of chapter 16, which deals with 78 defilements and is entitled kṣudravastuka (“miscellany”; zashi [雜事]). The third section includes the remaining five chapters dealing with basic doctrinal concepts such as the 22 faculties (indriya), 12 bases

(āyatana), 5 aggregates (skandha), 62 elements (dhātu), and dependent origination (pratītyasamutpāda). Similarities between the Dharmaskandha, the Pali Vibhaṅga, and the *Śāriputrābhidharmaśāstra (T. 1548) have been pointed out by various scholars (Fukuhara, 1965, 110–112, Frauwallner, 1995, 17–20). According to E. Frauwallner, the Sarvāstivāda Dharmaskandha and the Theravāda Vibhaṅga, “two versions of the same work,” go back to a common ancestor predating Aśoka’s missions; both of them “represent a step toward the composition of

truly analytical and scholastic treatises independent of the sūtras” (Cox, 1998, 187). Let it be noted that the Dharmaskandha is to be credited with the a ll­important distinction between the path of vision (darśanamārga) and the path of cultivation (bhāvanāmārga), which would structure all subsequent Sarvāstivāda, Sautrāntika, and Mahāyānist accounts of the path to liberation. (We owe an annotated Japanese translation to Watanabe, 1930a; English summaries and discussions include Frauwallner, 1995, 15–21; Potter, 1996, 179–187; Willemen, Dessein & Cox, 1998, 68–70; Cox, 1998, 181–189.)

Prajñaptiśāstra or Abhidharmaprajñaptipādaśāstra

The Prajñaptiśāstra, ascribed to Maudgalyāyana in India and to Mahākātyāyana in China (Cox, 1998, 189n143), is considered the latest of the three Sarvastivāda treatises belonging to this early stage of the overall development of Abhidharma literature (Sakurabe, 1969, 46). This treatise presents a sūtra quotation at the beginning of each of the chapters, followed by an account of various topics in question­and­answer form. Given the high number of its quotations in the Mahāvibhāṣā (145), the Prajñaptiśāstra likely “served as a handy sourcebook of succinct identifications,

definitions, and canonical references on a variety of topics” (Cox, 1998, 197). The Prajñaptiśāstra surviving in Tibetan translation possesses three sections: (1) Lokaprajñapti (Instruction on the World; ’Jig rten gdags/btags pa; D 4086/P 5587) dealing with Buddhist cosmology, (2) Kāraṇaprajñapti (Instruction about Causes; Rgyu gdags/btags pa; D 4087/P 5588) containing an account of the Bodhisattva and the wheel­turning king (cakravartin), and (3) Karmaprajñapti (Instruction on Karman; Las gdags/btags pa; D 4088/P 5589).

More than 20 percent of the Lokaprajñapti has been preserved in its original (Cox, 1998, 191) through Sanskrit fragments of Mūlasarvāstivāda provenance (Dietz, 1989). A great majority of the text of the Kāraṇaprajñapti has its counterpart in a seven­fascicle Chinese version (Shishe lun [施設論]; T. 1538). Louis de La Vallée Poussin (1918, 295–350) has provided an analysis and selected translations of the Tibetan Lokaprajñapti and Kāraṇaprajñapti. In 1922 Taiken Kimura conjectured, on the basis of a close analysis of the citations of the treatise in the Mahāvibhāṣā, that the Prajñaptiśāstra may originally have had more subdivisions than the three extant sections. These chapters would have dealt with defilements

(anuśaya), knowledge (jñāna), and meditation (samāpatti/samādhi). In the meantime, references to the titles *Anuśayaprajñapti (Phra rgyas btags) and *Nāmarūpaprajñapti (Ming dang gzugs kyi btags) in Śamathadeva’s Abhidharmakośopāyikā Ṭīkā have come to light (Honjō, 1998). These references support Taiken Kimura’s suggestion even though, as C. Cox rightly points out (1998, 192), the extent to which these eighth­century materials reflect the original structure of the treatise is difficult to assess. (In 1995 Takumi Fukuda discovered, in the library of Dōhō University, Nagoya, a draft of a full Japanese translation of the three Pra jñaptis, completed in 1934 by Sei Katō, a graduate of Ōtani University, and began to publish it [Fukuda, 1998–2012]; Aohara has begun to publish a Japanese translation of the Karmapraj ñapti [Aohara, 2012a, 2012b, 2013, 2014] in the light of Katō’s materials; for a Japanese translation of the Chinese version, see Watanabe, 1930b; Willemen, Dessein & Cox, 1998, 70–71, and Cox, 1998, 189–197, provide useful summaries of the text.)

Vijñānakāya or *Abhidharmavijñānakāyapādaśāstra

The Vijñānakāya is extant only in Xuanzang’s Chinese translation. Unlike the Saṅgītiparyāya, the Dharmaskandha, and the Prajñaptiśāstra, this treatise is not ascribed to a direct disciple of the Buddha, but to Devaśarman (reported to have lived 100 years after the parinirvāṇa and to have been active in Viśoka), and it is not a commentary on sūtra passages. These features suggest that it is much later than these three treatises. The Vijñānakāya is divided into 6 parts: (1) Maudgalyāyana (an opponent of unclear affiliation claiming that only the present exists), (2) person (pudgala), (3) causes and conditions (hetupratyaya), (4) condition as object­support

(ālambanapratyaya), (5) miscellany (saṃkīrṇa), (6) accompaniment (samanvāgama) or possession (prāpti). The first two chapters are polemical, while the other four are systematic, a feature that, together with differences in topics, suggests “that the Vijñānakāya is a composite text, perhaps compiled, rather than composed, by Devaśarman” (Cox, 1998, 198). The first chapter contains a controversy on the existence of past, future, and present. The second chapter deals with the existence of the person (pudgala) and exhibits close similarities with Kathāvatthu I.1 (i.e. §§ 1–69). Chapters three through six deal with “the nature, arising, operation, and transformation of states of

thought or consciousness (vijñāna)” (Cox, 1998, 201), a range of topics that accounts for the title of the work. The importance of this highly innovative treatise (Frauwallner, 1995, 30–31) is indicated by the fact that it is cited 39 times in the Mahāvibhāṣā. The Vijñānakāya indeed represents a major change, most conspicuously in the direction of a systematic theory of causality. (The first two chapters have been rendered into French by La Vallée Poussin, 1925; several parts of the second chapter have been translated into English by Fumimaro Watanabe [Watanabe, 1983,

154ff.]; here as elsewhere, we owe an annotated Japanese translation to Watanabe Baiyū, 1931; English summaries and discussions include Frauwallner, 1995, 28–31; Potter, 1998, 367–374; Willemen, Dessein & Cox, 1998, 72–73; Cox, 1998, 197–205.) Dhātukāya or

The Dhātukāya is extant only in Xuanzang’s Chinese translation. This treatise is ascribed to Pūrṇa in India and to Vasumitra in China (Cox, 1998, 206n193), an attribution that may be due to the close connection between the Dhātukāya and the Prakaraṇapāda, which was authored by Vasumitra perhaps in the 4th century after nirvāṇa. The Dhātukāya, a work “of the mātṛkā­type accompanied by an explanatory text” (Frauwallner, 1995, 21), discusses mind (citta) and mental factors (caitta) in question­and­answer form. This treatise is divided into two parts. In the first part (*mūlavastuvarga, “fundamental groups”), 14 groups comprising 91 mental factors are enumerated and defined, thus creating, and for the first time, a “system of psychology” (Frauwallner, 1995, 22). The names of these groups with Cox’s English equivalents are as follows (Cox, 1998, 208–209):

1. t en factors of great extension (mahābhūmikadharma);

2. en defiled factors of great extension (kleśamahābhūmikadharma);

3. ten factors whose extension is that of limited defilements (parīttakleśabhūmikadharma);

4. five defilements (kleśa);

5. five views (dṛṣṭi);

6. five factors (dharma);

7. five types of contact (saṃsparśa);

8. five controlling faculties (indriya);

9. six varieties of perceptual consciousness (vijñānakāya);

10. six varieties of contact (sparśakāya);

11. six varieties of feelings (vedanākāya);

12. six varieties of ideas (saṃjñākāya);

13. six varieties of volition (saṃcetanākāya); and

14. six varieties of craving (tṛṣṇākāya).

In the second part (*vibhajyavarga, “analysis,” in 16 sections), these mental factors are analyzed according to two kinds of relations, that is, simultaneous mental association (saṃprayoga) and includedness (saṃgraha). The Dhātukāya was largely incorporated into the Prakaraṇapāda (Frauwallner, 1995, 25–27). This may explain why this treatise was not cited in the Mahāvibhāṣā or Abhidharmakośa and was not considered canonical by some Indian masters (see above, and Cox, 1998, 207). Whatever the case may be, “[t]he importance of the Dhātukāya

lies in its exclusive focus on mental operations and in its efforts to develop a more abstract classification of, at least, the soteriologically significant mental phenomena” (Cox, 1998, 211). (There is an annotated English translation by Ganguly, 1994, with an introduction and a reproduction of the Taishō edition; Watanabe, 1932a, has provided an annotated Japanese translation of the Dhātukāya; English summaries and discussions of the work are available in Frauwallner, 1995, 21–28; Potter, 1996, 344–358; Willemen, Dessein & Cox, 1998, 71–72; Cox, 1998,

206–212.) Prakaraṇapāda or Abhidharmaprakaraṇapādaśāstra

The Prakaraṇapāda has survived only in two Chinese translations: Zhong shifen apitan lun (衆事分阿毘曇論; T. 1541), translated by Guṇabhadra and Bodhiyaśas in 435–443 ce, and Apidamo pinlei zu lun (阿毘達磨品類足論; T. 1542), translated by Xuanzang in 660 ce. Sanskrit fragments from a commentary on the Prakaraṇapāda have been discovered in Turfan, which, besides testifying to the fact that the

Vibhāṣās were not the only available commentaries on “canonicalAbhidharma works, seem to reflect a divergent textual structure (Imanishi, 1975). The first chapter, entitled Pañcavastuka, may largely borrow from an independent text, the *Pañcavastuka, translated into Chinese twice: the

Apitan wufaxing jing (阿毘曇五法行經; T. 1557) by An Shigao (安世高), and the Sapoduo zung wushi lun (薩婆多宗五事論; T. 1556) by Facheng (法成). This indicates that some chapters of this treatise may at first have been composed separately (Cox, 1998, 218). The Prakaraṇapāda is attributed to Vasumitra in both India and China, but the Mahāprajñāpāramitopadeśa (T. 1509 [XXV] 70a16f.) attributes only four chapters to Vasumitra (chs. 4–7 according to Frauwallner’s hypothesis), the other four being ascribed to arhats from Kashmir (Cox, 1998, 212). It

is considered to be the latest and in many ways the most significant of the “six feet” treatises (see above). The work is quoted no less than 106 times in the Mahāvibhāṣā, which at times and quite surprisingly rejects the interpretations of the Jñānaprasthāna and chooses those of the Prakaraṇapāda instead (Cox, 1998, 214). The wide reception of the Prakaraṇapāda may be due to its “function as a sourcebook or p roto­compendium uniting significant textual materials in a single text” (Cox, 1998, 215).

The treatise consists of 8 largely independent chapters:

1. Pañcavastuka (Five Groups; very close to the Pañcavastuka);

2. Jñāna (Knowledge; very close to the Vijñānakāya);

3. Āyatana (Sense Spheres; very close to the Pañcaskandhaka);

4. Saptavastuka (Seven Groups; very close to the Dhātukāya);

5. Anuśaya (Contaminants);

6. Saṃgrahādi (Includedness Etc.);

7. Sahasraparipṛcchā (Thousand Question); and

8. Nirvedha (Penetration).

One of the most interesting features of the Prakaraṇapāda is its first chapter on the pañcavastuka, “the fivefold classification of factors that would form the foundation of later Sarvāstivāda classifications of factors” (Cox, 1998, 215). This “method of encompassing and organizing all possible phenomena” (Cox, 1998, 220) distributes all factors into corporeality (rūpa), mind (citta), mind concomitants (caitta), factors dissociated from the mind (cittaviprayukta), and unconditioned factors (asaṃskṛta). (There is an annotated Japanese translation of the Prakaraṇapāda by Watanabe, 1932b; English summaries include Frauwallner, 1995, 32–36; Potter, 1996, 375–379; Willemen, Dessein & Cox, 1998, 73; Cox, 1998, 212–221.) J

ñānaprasthāna or Abhidharmajñānaprasthānaśāstra

As mentioned above, the Jñānaprasthāna is considered the most important of the seven canonical Sarvāstivāda Abhidharma treatises. This is reflected in the fact that a vast commentary on this treatise, the Mahāvibhāṣā, was compiled, comprising 200 fascicles in Xuanzang’s Chinese translation (T. 1545) and generally regarded as “the definitive statement of Sarvāstivāda Abhidharma exegesis” (Cox, 1998, 229). The Jñānaprasthāna is ascribed to Kātyāyanīputra, whose dates are variously given as 100, 300 or even 500 after nirvāṇa (Cox, 1998, 221; see also above). Two Chinese translations are extant, one by Saṅghadeva, Zhu Fonian (竺佛念) and

Dharmapriya in 30 fascicles (T. 1543; 383 ce), entitled *Aṣṭagrantha or *Aṣṭaskandha(ka), the other by Xuanzang in 20 fascicles (T. 1544; 657–660 ce), entitled Jñānaprasthāna. The two translations differ in both structure and doctrine and most likely represent distinct Sarvāstivāda lineages. Sanskrit fragments have been discovered in Bamiyan and Kučā (Lévi, 1932, Demiéville, 1961). The Jñānaprasthāna consists of 8 chapters (*skandhakas) entitled:

1. *Saṃkīrṇa (Miscellany);

2. *Saṃyojana (Fetters);

3. Jñāna (Knowledge);

4. Karman (Action);

5. Mahābhūta (Fundamental Material Elements);

6. Indriya (Controlling Faculties); 7. Samādhi (Concentration); and 8. Dṛṣṭi (Views).

Explaining the structure of these eight chapters in terms of mutual dependence, the Mahāvibhāṣā (T. 1545 [XXVII] 7a25ff.) states that [t]he first chapter is explained as presenting the factor of enlightenment, which occurs through the abandonment of defilements (Chapter 2). The abandonment of defilements depends on knowledge (Chapter 3), and a person who has given up the effects of action (Chapter 4) gives rise to that knowledge that abandons defilements. These varieties of action are dependent upon the four fundamental material elements (Chapter 5), and most prominent among the varieties of derived elements are the controlling faculties (Chapter 6). The purification of the controlling faculties takes place through concentration (Chapter 7), and to practice concentration, one must have surmounted false views (Chapter 8).(trans. Cox, 1998, 227–228)

(The *Aṣṭaskandhaka [T. 1543] has been translated into Japanese with annotations by Nishi Giyū and Sakamoto Yukio, 1934; Xuanzang’s Chinese version of the Jñānaprasthāna [T. 1544] has been translated with copious notes and tables by Sakurabe Hajime and Kaji Yōichi, 1996–2000; parts of the work have been translated into French by La Vallée Poussin, 1930; 1936–1937); Hurvitz, 1977, has dedicated a study to the Jñānaprasthāna ’s path to liberation; English summaries include Potter, 1996, 417–449; Willemen, Dessein & Cox, 1998, 73–79; Cox, 1998, 221–229; the first two chapters have been retranslated from the Chinese of T. 1544 into Sanskrit by Śānti Bhikṣu Śastri, 1955.)

Abhidharma Canonical Texts of Other Sects

The extent to which Buddhist denominations other than the Theravādins and the Sarvāstivādins possessed canonical Abhidharma works, and a fortiori what these Abhidharmapiṭakas looked like, is very difficut to assess. As A. Bareau (1947–1950) has pointed out, there is no compelling reason to believe that every denomination had developed an Abhidharmapiṭaka of its own. Whereas certain groups that branched off from a mother sect likely kept the latter’s Abhidharmapiṭaka with slight adaptations and emendations where necessary, other groups lacking an

independent third basket may have borrowed one from a geographically close sect. According to Paramārtha, the Sāṃmitīyas, the Dharmottarīyas, the Bhadrayānīyas and the Ṣaṇṇagarikas, the four subsects that seceded from the Vātsīputrīya sect, kept the latter’s “*Śāriputrābhidharma” (not to be confounded with the homonym Dharmaguptaka work described below; the Vātsīputrīya *Śāriputrābhidharma was otherwise known as the *Dharmalakṣaṇābhidharma, in nine sections), although they deemed it insufficient, and completed it here and there with their own interpretations of the sūtras (Demiéville, 1931–1932, 58). In a similar way, in Sri Lanka and a Theravāda context, the Abhayagirivāsins and the

Jetavanīyas apparently kept the Abhidharmapiṭaka of the Mahāvihāravāsins (Bareau, 1947–1950, 3). When it alludes to abhidharma, the Vinaya of the Mahīśāsakas may well refer only to the discipline of dogmatic exegetics rather than to a separate basket (T. 1421 [XXII] 132b; T. 1422 [XXII] 204a). The Vinaya of the Mahāsāṅghikas, on the contrary, seems to point to the existence of a separate Abhidharmapiṭaka (T. 1425 [XXII] 295a, etc.; see Bareau, 1947–1950, 9n6), and the same applies to the Haimavatas, whose Vinayamātṛkā (T. 1463 [XXIV] 818a28f.) suggests that their Abhidharma was structurally very close, if not identical, to the Dharmaguptaka *Śāriputrābhidharma, the only extant non­Theravādin and non­Sarvāstivādin Abhidharma treatise.

The *Śāriputrābhidharmaśāstra has come down to us in a Chinese translation by Dharmayaśas and Dharmagupta in 30 fascicles, the Shelifu apitan lun

(舍利弗阿毘曇論; T. 1548; Frauwallner, 1955, 221n1). This treatise is generally attributed to the Dharmaguptaka school (Bareau, 1950; Mizuno, 1996–1997, 319–340) in spite of the fact that the Chinese tradition regards it a Sāṃmitīya text (see above for a homonym Vātsīputrīya/Sāṃmitīya Abhidharma treatise). Sanskrit fragments from the *Śāriputrābhidharma or a very similar text have been recently identified in the Schøyen collection by K. Matsuda (2002). The treatise consists of 5 chapters whose titles, in A. Bareau’s Sanskrit restoration, are as follows:

1. Sapraśnaka (With Questions);

2. Apraśnaka (Without Questions);

3. Saṃgraha (Includedness);

4. Saṃprayoga (Simultaneous Mental Association); and

5. Prasthāna (Base).

Resemblances with the Sarvāstivāda Dharmaskandha and the Pali Vibhaṅga and Puggalapaññatti have long been recognized (Kimura, 1922, 73–118). E. Frauwallner has analyzed the *Śāriputrābhidharma in terms of structure and content and as to the way in which its mātṛkās developed from those of earlier texts. According to him, this treatise of interesting idiosyncratic character “helps us to avoid the one­sided judgements that exclusive observation of the Abhidharma of the other schools might otherwise lead us into making” (Frauwallner, 1995, 116). (An annotated Japanese translation was published by Watanabe, 1934; English summaries and discussions include Frauwallner 1995, 97–117; Potter, 1996, 317–325.)

Bibliography Primary Sources

Abhidharmakośabhāṣya: Abhidharmakośabhāṣyam of Vasubandhu, ed. P. Pradhan, Patna, 1967; rev. ed. 1975. Abhidharmakośavyākhyā: Sphuṭârthā Abhidharmakośavyākhyā, the Work of Yaśomitra, ed. U. Wogihara, Tokyo, 1936; repr. 1989. Atthasālinī: The Atthasālinī, Buddhaghosa’s Commentary on the Dhammasaṅgaṇi, ed. E. Müller, London, 1897. Dīpavaṃsa: The Dîpavaṃsa: An Ancient Buddhist Historical Record, ed. H. Oldenberg, London, 1879.

Milindapañha: The Milindapañho, Being Dialogues between King Milinda and the Buddhist Sage Nāgasena, the Pali Text, ed. V. Trenckner, London, 1880; repr. 1986. Mohavicchedanī: Mohavicchedani: Abhidhammamātikatthavaṇṇanā, ed. A.P. Buddhadatta & A.K. Warder, London, 1961. Samantapāsādikā: SamantapāsādikāBuddhaghosa’s Commentary on the Vinayapiṭaka, 7 vols., ed. J. Takakusu & M. Nagai, London, 1924–1947. Sumaṅgalavilāsinī: Sumaṅgalavilāsinī (Dīghanikāya-aṭṭhakathā), ed. T.W. Rhys Davids & J.E. Carpenter, 3 vols., London, 1929–1932. Secondary Literature Anālayo, The Dawn of Abhidharma, Hamburg, 2014.

Anālayo, “Teaching the Abhidharma in the Heaven of the Thirty­three, The Buddha and his Mother,” JOCBS 2, 2012, 9–35. Aohara Norisato (青原令知編), “Gōsesetsu wayaku kenkyū (1)” (『業施設』和訳研究 (1); An Annotated Japanese

Translation of the Karmaprajñapti [[[chapter]]] 1), RDR 479, 2012a, 9–33; chapter 2: BK 68, 2012b, 1–27; chapter 3: BK 69, 2013, 75–113; chapter 4: RDR 483, 2014, 26–51. Bareau, A., “Les origines du Śāriputrābhidharmaśāstra,” MUSEON 63, 1950, 69–95. Bareau, A., “Les sectes bouddhiques du Petit Véhicule et leurs Abhidharmapiṭaka,” BÉFEO 44/1, 1947–1950, 1–11. Bronkhorst, J., “Kathāvatthu and Vijñānakāya,” in: Premier Colloque Étienne Lamotte (Bruxelles et Liège, 24–27 septembre 1989), Louvain­la­Neuve, 1993, 57–61.

Buswell, R.E., & P.S. Jaini, “The Development of Abhidharma Philosophy,” in: K.H. Potter et al., eds., Encyclopedia of Indian Philosophies, vol. VII: Abhidharma Buddhism to 150 A.D., Delhi, 1996, 73–119. Cox, C.,“Gāndhārī Kharoṣṭhī Manuscripts: Exegetical Texts,” in: P. Harrison & J.­U. Hartmann, eds., From Birch Bark to Digital Data: Recent Advances in Buddhist Manuscript Research, Vienna, 2014, 35–49.

Cox, C., “Kaśmīra: Vaibhāṣika Orthodoxy,” in: C. Willemen, B. Dessein & C. Cox, 1998, 138–254. Cox, C., Disputed Dharmas, Early Buddhist Theories of Existence, An Annotated Translation of the Section on Factors Dissociated from Thought from Saṅghabhadra’s Nyāyānusāra, Tokyo, 1995. Cox, C., “The Unbroken Treatise: Scripture and Argument in

Early Buddhist Scholasticism,” in: M.A. Williams, C. Cox & M.S. Jaffee, eds., Innovation in Religious Traditions: Essays in the Interpretation of Religious Change, Berlin, 1992, 143–189. Davidson, R.M., “An Introduction to the Standards of Scriptural Authenticity in Indian Buddhism,” in: R.E. Buswell, ed., Chinese Buddhist Apocrypha, Honolulu, 1990. Demiéville, P., “Un fragment sanskrit de l’Abhidharma des Sarvāstivādin,” JA 249, 1961, 461–465. Demiéville, P., “L’origine des sectes bouddhiques d’après Paramārtha,” MCB 1, 1931–1932, 14–64; repr. in: P. Demiéville, Choix d’études bouddhiques (1929–1970), Leiden, 1973, 81–64.

Dietz, S., “The Sanskrit Abhidharma Fragments from the Turfan Oasis,” in: D. Durkin­Meisterernst, S.­C. Raschmann, J. Wilkens, M. Yaldiz & P. Zieme, eds., Turfan Revisited – The First Century of Research into the Arts and Cultures of the Silk Road, Berlin, 2004, 59–67.

Dietz, S., “A Brief Survey on the Sanskrit Fragments of the Lokaprajñaptiśāstra,” AMOUSBCRI 7, 1989, 79–86. Dietz, S., “Untersuchungen zur Schulzugehörigkeit der in Ujjain liegenden G ilgit­Fragmente,” in: H. Bechert, ed., Zur Schulzugehörigkeit von Werken der Hīnayāna-Literatur, Göttingen, 1985, 163–179.

Dietz, S., Fragmente des Dharmaskandha. Ein AbhidharmaText in Sanskrit aus Gilgit, Göttingen, 1984. Frauwallner, E., Studies in Abhidharma Literature and the Origins of Buddhist Philosophical Systems, Albany, 1995.

Fukuda Takumi (福田琢), “Katō Sei ikō Zōbun wayaku In sesetsu (1)” (加藤清遺稿蔵文和訳『因施設』; Japanese Translation of the Kāraṇaprajnapti [part] 1: From Sei Katō’s Literary Remains), DGBBKH 43, 2007, 88–59; part 2: DGBBKH 48, 2012, 164–136. Fukuda Takumi, “Katō Kiyoshi (sic) ikō Zōbun wayaku Seken sesetsu” (加藤清遺稿蔵文和訳『世間施設』(1); Japanese Translation of Lokaprajñapti [part] 1: From Kiyoshi (sic) Kato’s Literary Remains”), DGBBKH 34, 1998, 140–99; part 2: DGBBKH 35, 1999, 88–72; part 3: DGBBKH 36, 2000, 128– 91; part 4: DDR 84, 2001, 86–63; part 5: DDR 85–86, 2002, 310–273; part 6: DDR 89, 2004, 146–130. Fukuhara Ryōgon (福原亮嚴), Ubu-abidatsumaronsho no hattatsu (有部阿毘達磨論書の発達; The Development of the Abhidharmaśāstras of the Sarvāstivāda School), Kyoto, 1965.

Ganeri, J., “Argumentation, Dialogue and the Kathāvatthu,” JIP 29, 2001, 485–493. Ganguly, S., Treatise on Groups of Elements, Abhidharmadhātukāya-pādaśāstra, Delhi, 1994. Geiger, M., & W., Pāli Dhamma vornehmlich in der kanonischen Literatur, Munich, 1920; repr. in: H. Bechert, ed., Wilhelm Geiger, Kleine Schriften, Wiesbaden, 1973, 101–228.

Gethin, R., “The Mātikās: Memorization, Mindfulness, and the List,” in: J. Gyatso, ed., In the Mirror of Memory, Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, Albany, 1993, 149–172. Gethin, R., The Buddhist Path to Awakening: A Study of the Bodhi-Pakkhiyā Dhammā, Leiden, 1992. Hinüber, O. von, A Handbook of Pāli Literature, Berlin, 1996.

Hirakawa, A., “The Meaning of ‘Dharma’ and ‘Abhidharma’,” in: Indianisme et Bouddhisme: Mélanges offerts à Mgr Étienne Lamotte, Louvain­la­Neuve, 1980, 159–175. Hirakawa, A., Index to the Abhidharmakośabhāṣya – Part I: Sanskrit-Tibetan-Chinese, Tokyo, 1973. Honjō Yoshifumi (本庄良文), “Bibashashi no bussetsukan” (毘婆沙師の仏説観; Vaibhāṣika Discussions about the Words of the Buddha: Some Passages from the Mahāvibhāṣā and the Nyāyānusāra), IL 1, 2010, 173–193. Honjō Yoshifumi, “ Zuimin­sesetsu to Myōshiki­sesetsu” (『随眠施設』と『名色施設』; *Anuśayaprajñapti and

Imanishi, J., Fragmente des Abhidharmaprakaraṇabhāṣyam in Text und Übersetzung, Göttingen, 1975. Kimura Taiken (木村泰賢), Kimura Taiken zenshū (木村泰賢全集; Collected works of Taiken Kimura), 4 vols., Tokyo, 1967–1969. Kimura Taiken, Abidatsumaron no kenkyū (阿毘達磨論の研究; A Study of Abhidharma Treatises), Tokyo, 1922. La Vallée Poussin, L. de, “Documents d’Abhidharma: La controverse du temps; les deux, les quatre, les trois vérités,” MCB 5, 1936–1937, 1–187.

La Vallée Poussin, L. de, “Documents d’Abhidharma: Textes relatifs au nirvāṇa et aux asaṃskṛta en général I –II,” BÉFEO 30, 1930, 1–28, 247–298. La Vallée Poussin, L. de, “La controverse du temps et du Pudgala dans le Vijñānakāya,” in: Études asiatiques pu- bliées à l’occasion du 25e anniversaire de l’École française d’Extrême-Orient, vol. I, 1925, 343–376.

La Vallée Poussin, L. de, L’Abhidharmakośa de Vasubandhu, traduction et annotations, 6 vols., Paris, 1923–1933.

La Vallée Poussin, L. de, “Notes bouddhiques II. Le Vijñānakāya et le Kathāvatthu,” BCLBA 8/11, Bruxelles, 1922, 516–520. La Vallée Poussin, L. de, Vasubandhu et Yaçomitra, Troisième chapitre de l’Abhidharmakoça Kārikā, Bhāṣya et Vyākhyā, avec une analyse de la Lokaprajñapti et de la Kāraṇaprajñapti de Maudgalyāyana, London, 1918.

Lamotte, É., Histoire du bouddhisme indien, des origines à l’ère Śaka, Louvain­la­Neuve, 1976; ET; idem., trans. S. WebbBoin, The History of Indian Buddhism from the Origins to the Śaka Era (with addenda), Louvain­la­Neuve, 1988. Lévi, S., “Notes sur des manuscrits provenant de Bamiyan (Afghanistan) et de Gilgit (Cachemire),” JA 220, 1932, 1–45.

Matsuda Kazunobu (松田和信), “Shūimonsokuron no bāmiyan shahon dankan ni kansuru oboegaki” (集異門足論のバーミヤン写本断簡に関する覚え書き; A Note on the Fragments of the Bāmiyan Manuscripts of the Abhidharma- saṃgītiparyāyapādaśāstra), ITBK 21, 2006, 19–33. Matsuda Kazunobu, “Three Fragments Related to the

Śāriputra­Abhidharma,” in: J. Braarvig, ed., Buddhist Manuscripts, vol. 2, Oslo, 2002, 240–248. Matsuda Kazunobu, Newly Identified Sanskrit Fragments of the Dharmaskandha in the Gilgit Manuscripts, Kyoto, 1986. Matsuda Kazunobu, “Bonbundanpen Loka-prajñapti ni tsuite” (梵文断片 Loka-prajñapti について; On the Sanskrit Fragments of the Lokaprajñapti), Bg 14, 1982, 1–21. McDermott, J.P., “The Kathāvatthu Kamma Debates,” JAOS 95, 1975, 424–433. Migot, A., “Un grand disciple du Bouddha: Śāriputra: Son rôle dans l’histoire du bouddhisme et le développement de l’Abhidharma,” BÉFEO 46/2, 1954, 405–554.

Mizuno Kōgen (水野弘元), Mizuno Kōgen chosakusenshū (水野弘元著作選集; Collected Works of Kōgen Mizuno), 3 vols., Tokyo, 1996–1997.

Nishi Giyū (西義雄) & Sakamoto Yukio (坂本幸男), Abidonhakkendoron (阿毘曇八犍度論), KYIK 17–18, 1934. Norman, K.R., “Māgadhisms in the Kathāvatthu,” in: A.K. Narain, ed., Studies in Pāli and Buddhism: A Memorial Volume in Honor of Bhikkhu Jagdish Kashyap, Delhi, 1979, 279–287; repr. in: Collected Papers, vol. II, Oxford, 1991, 59–70. Potter, K.H., Encyclopedia of Indian Philosophies, vol. VII: Abhidharma Buddhism to 150 A.D., Delhi, 1996. Przyluski, J., Le Concile de Rājagṛha: Introduction à l’histoire des canons et des sectes bouddhiques, Paris, 1926. Ronkin, N., Early Buddhist Metaphysics, The Making of a Philosophical Tradition, London, 2005. Sakurabe Hajime (櫻部建), Kusharon no kenkyū: kai-konbon (倶舎論の研究 界・根品; Studies of the Abhidharmakośa, Chapters 1 and 2: Dhātunirdeśa and Indriyanirdeśa), Kyoto, 1969.

Sakurabe Hajime & Kaji Yōichi (加治洋一), Hocchiron (発智 論), SKYD Nyoraizō, Yuishiki­bu, 1996–2000. Śastri, Ś.B., Jñānaprasthāna-śāstra of Kātyāyanīputra, Santiniketan, 1955. Satō Mitsuo (佐藤密雄), Ronji, fu Kakuon-chū Shintei zōho

(論事附覚音註新訂増補; Kathāvatthu together with Buddha ghosa’s Commentary Revised and Enlarged), Tokyo, 1991. Skilling, P., “Dharma, Dhāraṇī, Abhidharma, Avadāna: What was taught in Trayastriṃśa?” ARIRIAB 11, 2008, 37–60. Skilling, P., “The Saṃskṛtāsaṃskṛta­viniścaya by Daśabalaśrīmitra,” BSR 4/1, 1987, 3–23.

Stache­Rosen, V., Dogmatische Begriffsreihen im älteren Buddhismus. Das Saṅgītisūtra und sein Kommentar Saṅgītiparyāya, 2 vols., Berlin, 1968.

Tsukamoto Keishō (塚本啓祥), Matsunaga Yūkei (松長有 慶) & Isoda Hirofumi (磯田熙文), Bongo butten no kenkyū III: Ronsho hen (梵語仏典の研究 論書篇; A Descriptive Bibliography of the Sanskrit Buddhist Literature), vol. III, Kyoto, 1990.

Waldschmidt, E., “Die Einleitung des Saṅgītisūtra,” ZDMG 105, 1955, 298–318. Watanabe Baiyū (渡辺楳雄), Ubu abidatsuma ron no kenkyū

(有部阿毘逹麿論の研究; Studies on Abhidharma Literature of Sarvāstivāda Buddhism), Tokyo, 1954. Watanabe Baiyū, Sharihotsuabidonron (舍利弗阿毘曇論),

KYIK 19, 1934. Watanabe Baiyū, Abidatsumakaishinsokuron (阿毘達磨界身足論), KYIK 5, 1932a. Watanabe Baiyū, Abidatsumahonruisokuron (阿毘達磨品類足論), KYIK 5, 1932b. Watanabe Baiyū, Abidatsumashikishinsokuron (阿毘達磨識身足論), KYIK 4, 1931.

Watanabe Baiyū, Abidatsumahōunsokuron (阿毘達磨法蘊足論), KYIK 3, 1930a. Watanabe Baiyū, Sesetsuron (施設足論), KYIK 3, 1930b, 331–439. Watanabe Baiyū, Abidatsumashūimonsokuron (阿毘達磨集異門足論), KYIK 1–2, 1929. Watanabe, F., Philosophy and Its Development in the Nikāyas and Abhidharma, Delhi, 1983. Willemen, C., B. Dessein & C. Cox, Sarvāstivāda Buddhist Scholasticism, Leiden, 1998. Vincent Eltschinger Yoshifumi Honjo