Embodiments of the Buddha in Sarvâstivāda Doctrine: With Special Reference to the *Mahāvibhāṣā

Embodiments of the Buddha in Sarvâstivāda Doctrine:

With Special Reference to the *Mahāvibhāṣā

Michael RADICH (Wellington)

Introduction

Past studies of ideas about the embodiments of the Buddha have typically held that "Schools" ideas, among which those of the Sarvâstivāda are considered particularly important, predate Mahāyāna materials on the whole; and that where ideas are found in both "Schools" and Mahāyāna texts, the direction of borrowing and influence must be from the Schools to the Mahāyāna. Scholars have also thought that Sarvâstivāda in particular, or Mainstream schools more generally, propounded a doctrine holding that the Buddha is embodied in two main forms, a dharmakāya and a rūpakāya. This two-body doctrine is sometimes even imagined to reach back to the Nikāyas2 and Āgamas. The Mahāyāna, especially the Prajñāpāramitā, is then supposed to have passively accepted this doctrine. In the process, Mahāyāna texts are thought to have reinterpreted an originally pluralistic notion of dharmakāya (dharma = plural qualities of the Buddha) and replaced it with a more monistic, quasi-metaphysical concept.3

1 This paper is based upon work found in my PhD dissertation. See Michael Radich, "The Somatics of Liberation: Ideas about Embodiment in Buddhism from Its Origins to the Fifth Century C.E." (PhD diss.,

Harvard University, 2007), Chapter 4.4. I am grateful to the International Research Institute for Advanced Buddhology at Soka University in Tokyo for the opportunity to present the work contained here on December 1, 2009, and to Profs. Jan Nattier, John McRae, Kanno Hiroshi and Karashima Seishi for useful comments and corrections on that occasion that helped me improve the draft. All remaining errors are of course my responsibility.

2 There is no concept of *rūpakāya in the Pāli canon, though scholars sometimes assume there must be. 3 For expressions of this received view and its component parts, see for example Takeuchi Shōkō 武内紹晃, "Buddakan no hensen 仏陀観の変遷," in Daijō bukkyō vol. I, ed. Hirakawa Akira (Tokyo: Shunjusha, 1983), 160-162; Ruben L. F. Habito ルーベン・Ｌ・Ｆ・アビト, "Busshin ron no tenkai – sanshinsetsu no seiritsu wo megutte 仏身論の展開－三身説の成立をめぐって," Shūkyō kenkyū 宗教研究 52, no. 2

(1978): 4-5; John J. Makransky, Buddhahood Embodied: Sources of Controversy in India and Tibet (Albany: State University of New York Press, 1997), Ch. 2 and 3, where Sarvâstivāda precedence is implied by the order of presentation; Guang Xing, The Concept of the Buddha: Its Evolution from Early Buddhism to the Trikāya Theory (London: RoutledgeCurzon, 2005), Ch. 2. Nagao Gadjin, "On the Theory of the Buddha-Body (Buddha-kāya)," trans. Hirano Umeyo, The Eastern Buddhist, n.s. 6, no. 1 (1973): 25-53, does not mention Sarvâstivāda, but presents the two-body theory (within the "one-two-three" model, which Nagao still endorses) as an entirely Mahāyāna innovation, predominantly of the Prajñāpāramitā, 26; implying, primarily on the basis of the traditional ascription of the *Mahāprajñāpāramitôpadeśa 大智度論 T1509 to Nāgārjuna, that Nāgārjuna must have been instrumental in its elaboration; 30. Nagao's presentation, however, is very generalised and ahistorical. Habito seems similarly to regard the true twobody theory as a Prajñāpāramitā invention, implying that Sarvâstivāda merely prepares the ground; "Busshinron" 5. One significant departure from this general pattern is Louis de la Vallée Poussin, who presents Sarvâstivāda Abhidharma as propounding a theory of three bodies; "Notes sur les corps du Bouddha," Vijñaptimātratāsiddhi: La siddhi de Hiuan-Tsang (Paris: P. Geunther, 1928-29), 766-773. However, his claim that the Sarvâstivāda recognise nirmāṇakāya as a body of the Buddha is flawed; see

Appendix 2. For the view that a plural Mainstream dharmakāya was only gradually replaced by a Mahāyāna reinterpretation, see most influentially Paul Maxwell Harrison, "Is the Dharma-kāya the Real

In this paper, I will argue that this understanding is flawed for several reasons. First, while some Sarvâstivāda texts do posit two bodies, the ordinary body in this pair is not best characterised as rūpakāya. Rather, the most dominant Sarvâstivāda doctrine is that the Buddha has a "body of birth", which links their ideas closely with those of the *Mahāprajñāpāramitôpadeśa. Second, the dharmakāya in Sarvâstivāda doctrine, which has never been presented in its full range by previous scholars, actually centres on a notion that the Buddha is embodied in his gnosis (bodhi, jñāna), which is much more compatible with the Mahāyāna understanding than previous analysis would suggest. Third, our present record cannot prove that the doctrines in question chronologically precede important Mahāyāna body doctrines. Sarvâstivāda texts before the

• Mahāvibhāṣā ("Vibh") are in fact largely uninterested in the very problem of the Buddha's embodiment, and even the ideas found in Vibh may have remained in ferment much later than the advent of similar ideas in the Mahāyāna.

I will attempt to demonstrate these claims, first, by giving a more accurate description of actual Sarvâstivāda doctrine about Buddha bodies. I will examine in detail the actual doctrine of the Buddha's special bodies as found in the Vibh. I will then examine similar ideas as they are found (or not found) in earlier works in the Sarvâstivāda canon, and variations between body doctrine in the multiple extant translations of Vibh. On this basis, I will argue that it is highly unlikely any Sarvâstivāda ideas about Buddha-bodies predate similar Mahāyāna ideas. In closing I will briefly discuss the place of this revised picture of Sarvâstivāda in the larger history of ideas about the Buddha's bodies.

The *Mahāvibhāṣā

My main focus for the exploration of Sarvâstivāda body doctrine will be the massive Vibh. This is a commentary on the earlier Jñānaprasthāna, and is regarded as a seminal compendium of Sarvâstivāda doctrine. I have chosen this text as my focus for three main reasons. First, as opposed e.g. to a text like the Abhidharmakośa, its traditional date makes it still early enough to possibly count for the question of the chronological priority of Sarvâstivāda versus Mahāyāna body doctrine. Second, unlike earlier Sarvâstivāda texts, it contains substantial treatment of the doctrine of Buddha bodies, so that we have ample material for dicussion. Third, we are fortunate that it exists in several Chinese translations, comparison between which allows for detailed exploration of problems of dating.

Vibh exists in three Chinese translations. The two earlier translations are each partial: (1) by *Saṃghabhūti 僧伽跋澄, dating from 383 (*Vibhāṣāśāstra 鞞婆沙論, T1547); (2) by Buddhavarman 浮陀跋摩, dating from 437 (阿毘曇毘婆沙論, T1546); (3) by Xuanzang 玄奘 (602-664) (阿毘達磨大毘婆沙論 T1545). These three texts are best treated as plural texts in a genre of vibhāṣā (commentary), rather than as parallel translations of the same text.4 The *Saṃghabhūti version is much

'Phantom Body' of the Buddha?" Journal of the International Association of Buddhist Studies 15, no. 1 (1992): 44-94. 4 Charles Willemen, Bart Dessein and Collett Cox, Sarvāstivāda Buddhist Scholasticism (Leiden: Brill, 1998), "The Vibhāṣā Compendia", 229-239. There is evidence that more such texts existed and were lost, and that others in the genre were based upon other texts. smaller than the other two, at only fourteen fascicles. It is also very different in structure from the other two, and comments only on selected portions of the root text Jñānaprasthāna (JñP). The Buddhavarman translation was originally a hundred fascicles, but it was lost in political turmoil almost immediately upon completion in 439, and only partially restored. Our extant text is thus a partial text in sixty fascicles, covering only the first three of the eight chapters of JñP.

In what follows, I will take Xuanzang ("XZ") as my primary point of reference, but I will check whether each of the ideas treated is present in all three versions. Discrepancies between XZ and earlier versions will be noted in the footnotes, but also discussed in overview later in the paper (p. 147 ff.). Where an idea is missing from an earlier version but found in Xuanzang, we will obviously have reason to fear that it is of later origin. Vibh does not propound rūpakāya vs. a pluralistic dharmakāya

It is commonly said that the Sarvâstivāda ascribes two bodies to the Buddha, a rūpakāya and a dharmakāya. At least if we consider the problem in terms of Vibh, this claim is factually incorrect. First, rūpakāya is not, properly speaking, depicted as a body "of the Buddha" at all; it is rather used for the ordinary body of all sentient beings. Second, where the text does speak of the Buddha's fleshly body, it uses a different set of terms, which have been overlooked by previous scholarship in favour of the mistaken understanding of rūpakāya. Third, while the text does does teach a dharmakāya, the precise meaning of that term is somewhat different from the interpretations prior scholars have given it. I will address each of these points in turn. Rūpakāya is not really a body of the Buddha in Vibh

In Vibh, the Buddha has a rūpakāya only insofar as he shares dimensions of his being with ordinary sentient beings. Thus, in a significant sense, the Buddha can be said to "have" a rūpakāya only insofar as he is not a Buddha, that is, insofar as he is an ordinary sentient being. If anything, the rūpakāya is dissociated from the Buddha.

The text uses the term rūpakāya (seshen 色身 ) relatively frequently (38 times). In places, the context, in discussions about differences between rebirth in the three realms

(kāmadhātu, rūpadhātu, ārūpyadhātu) makes it clear that rūpakāya is the very general, ordinary body of incarnation in the kāmadhātu. Sometimes rūpakāya is even understood specifically as the unsatisfactory body of suffering, as for example when it is discussed in terms of the five upādānaskandhas (which a Buddha would not have). Other passages say sentient beings can either have (a) bodies of form (seshen) or (b) formless bodies (色無色身), depending upon the level of the cosmological hierarchy; this too shows seshen can be properties of any sentient being. In one passage, tiannü 天女 ("āpsaras") are even said to have seshen as a result of their magical powers. In all these passages, rūpakāya is clearly a body of ordinary sentient beings, in various senses. On the other hand, only once is seshen clearly predicated of the Buddha. The context is a discussion of whether or not it is possible to engage in a contemplation of impurity (不淨觀 , *aśubhabhāvaṇā14) taking the rūpa[-skandha] of the Buddha as the contemplative object. "Question: Is it possible to engage in contemplation of impurity taking the Buddha's body of physical form (佛色身, *buddhasya rūpakāya) as the meditative object?

"[Answer:] Some hold that this is impossible, because the Buddha's physical form (*buddha-rūpa, 佛色) is extremely subtle, exceedingly bright and clean, like pure light, and it is therefore impossible to be disgusted by it. Other masters, however, hold that it is possible for a Buddha to engage in contemplation of impurity taking himself as a meditative object, but that it is not possible for other people [to engage in such a contemplation with him as the object]. Others again hold that there are two kinds of contemplation of impurity: (1) of the conditioned nature (*pratītyasamutpannatva?) of physical form; and (2) of physical form as evil and ill-omened. [It is held that the contemplation of impurity] of the conditioned nature can take the Buddha as the meditative object, but that [the contemplation of impurity as] evil and ill-omened cannot take the Buddha as the meditative object. Finally, there are those who hold that there are [a different] two kinds of contemplation of impurity: (1) at the level of universal characteristics (sāmānyalakṣaṇa); and (2) at the level of particularity (svalakṣaṇa). [It is held that contemplation of impurity at] the level of universal characteristics can take the Buddha's body (佛身, *buddhakāya) as the

meditative object, while [the contemplation of impurity at] the level of particularity cannot take the Buddha (佛 ) as the meditative object."16 The use of the term rūpakāya here seems loose. In a parallel passage elsewhere in Xuanzang's translation ("XZ"), the term rūpakāya does not appear at all.17 Even in this passage, the text immediately goes on to speak of both "the visible form (rūpa) of the Buddha" (佛色 ) and "the body of the Buddha" (佛身 ) separately, so that that the association between the two members of the compound (se 色 and shen 身) is at best loose. In this light, we might well translate

the opening question, "Is it possible . . . taking the material form or body of the Buddha (佛色身) as the meditative object?" The passage also refers loosely back to the notion in question as just "the Buddha". Further, the term rūpakāya does not appear in the same passage in Saṃghabhūti.18 All of this suggests that the passage was not composed with a formalised doctrine of rūpakāya as one of two particular embodiments of the Buddha in mind. Thus, in Vibh, the term rūpakāya is most consistently used as part of the attempt to conceptualise rebirth at different levels of the cosmological schema.

Rūpakāya is explicitly associated with suffering and ignorance, and with sentient beings other than Buddhas. Even if a Buddha can have a rūpakāya, then, he shares it with ordinary worldlings, and it is incidental to his status as the Buddha. Finally, the one passage that does predicate a rūpakāya of the Buddha does so only loosely, in passing as it were, and furnishes no reason to believe rūpakāya is a component of a developed doctrine of various Buddha bodies. Rūpakāya is thus not an important part of the Vibh's doctrine of Buddha bodies. We now turn, therefore, to what the text does say about the Buddha's bodies.

Terms for the Buddha's ordinary physical body When Vibh speaks of the ordinary physical body of the Buddha, in opposition to his dharmakāya, it most commonly either (1) uses no special term, but just calls it "the body"; or (2) refers to the "birth body" 生身, "the body born of father and mother" 父母生身, etc. The Buddha's physical body simply called "the body" etc.

particular, local aspect of that object with which cognitive engagement is achieved moment-to-moment through the course of the contemplation. 16 問有緣佛色身起不淨觀不。有作是說。無有能者。佛色微妙最極鮮潔如淨光明不可厭故。有餘師說。佛能自緣起不淨觀。餘無能者。或有說者。不淨觀有二種。一色緣起。二色過患。色緣起者。能緣佛身。色過患者。不能緣佛。復有說者。不淨觀有二種。一共相境。二自相境。共相境者能緣佛身。自相境者不能緣佛, T1545:27.207b02-10. This passage is already contained in the Saṃghabhūti version of the text, T1547:28.504c27-505a11; also in Buddhavarman, T1546:28.340b28-c06. However, there are some key differences between these earlier presentations and the version we find in Xuanzang. Most notable for our present purposes, of course, is the fact that they do not contain the term rūpakāya (indeed, as we shall see below, this term never appears in Saṃghabhūti). Saṃghabhūti also does not present the opinion that the Buddha can contemplate impurity with his own body as object, but not that of others; other opinions cited also differ.

17 See T1545:27.439a15-26, where the question is presented thus: 緣佛身色頗有能起不淨觀不. 18 See n. 16. It is common for Vibh to speak of the Buddha's ordinary body simply as "the body" (shen 身) or "the body of the Buddha" (foshen 佛身), without calling it any particular kind of body. For example, the text clearly recognises the body of marks of the mahāpuruṣa19 as a dimension of the overall range of the Buddha's embodiments. In speaking of this possibility, however, it never uses the term rūpakaya;20 nor does it usually use the special terms that we will examine below. It speaks simply of "the body" or "the Buddha's body".21 In one passage alone, it speaks of "the great body"; and in the narrative of Māra showing the body of a Buddha to Upagupta, it associates the body of marks with the "body of birth" (生身, see immediately below), in opposition to the dharmakāya. This is the case also when the text speaks in isolation of the light emanated by the

Buddha's body.

"Body" alone, or interchangeably "body of the Buddha", are the terms used when the text discusses the reason that it is an ānantarya sin to draw blood from the body of the Buddha. The text also speaks simply of "the body of the Buddha" in discussing the 19 Like other texts at this stage of doctrinal history, Vibh regularly groups the eighty minor marks, the golden complexion, and the halo of light with the thirty-two major marks.

20 Pace Guang, Concept of the Buddha, Ch. 2, which contends that in supposed Sarvâstivāda two-body theory, rūpakāya is primarily the body of marks. 21 Two passages are particularly telling. (1) Much of juan 177 discusses the marks, as part of a larger discussion of karma of body, speech and mind. (This entire discussion is missing from earlier versions of the text.) The term foshen 佛身 features in this discussion thirteen times, and 如來身 *tathāgatakāya once. By contrast, no term for a special Buddha body is ever used. (2) In a passage discussing the means by which it was possible for the Tathāgata to prophesy that Maitreya will become a Buddha in future, the text lays down the twofold stipulation that "the bodies of all Buddhas must have the unsurpassed thirty-two marks and the unsurpassed *āryadharma(s), viz. anuttarasaṃyaksaṃbodhi" 一切佛身必有無上三十二相及無上聖法。謂正等菩提, T1545:27.894c13-14. Again, no special term is used for the body of marks. See also 159c14-16, 361b24-25, 428c02-05, 590b02-04, 730a08-10. whereabouts or transformation of the Buddha's body when he miraculously vanished.27 The text uses the same term when contrasting the Buddha's physical strength with his

"power of thought" (yili 意力).28 "Body of the Buddha" (fo shenxing 佛身形) is also the term used when the Buddha becomes visibly physically weak near the time of his death (parinirvāṇa).29

In the majority of these examples, the terms "body" or "body of the Buddha" refer to the visible, physical body of the Buddha, during his last earthly incarnation or some related existence. Sometimes, however, the same term also refers clearly to what Vibh will also call dharmakāya.30 It seems likely that dharmakāya is also meant when the text says ordinary compassion is based upon the body of the (ordinary) sentient being, whereas the "great compassion" (mahākaruṇā) of the Buddha is based upon the "body of the Buddha".31 In sum, even where the text discusses aspects of the embodiment of the Buddha that make it special, it does not necessarily apply any special term, but rather, often speaks just of "the Buddha's body".

The Buddha's "body born of father and mother"/"body of birth" Sometimes, Vibh does specify the kind of body it means when it speaks of the Buddha's ordinary body. In such contexts, far more common that rūpakāya are the related terms "body born of father and mother" and "body of birth". The former term, which we will see is quite rare outside Vibh, is the more specific, and so we will look at it first.

The term "body born of father and mother" (fumushengshen 父母生身 , Skt.

grave sin to attack his body of birth? (The aśaikṣadharmas presumably are not harmed by such an attack.) The text answers in part that "the aśaikṣadharmas that comprise the Buddha function (轉 , */vṛt) based upon (依, *āśritya) the body of birth. It should be understood that if one harms the basis (所依), one also harms what is based upon it (能依 ), just as, if the bottle is broken, the milk [in it] will also be lost;" T1545:27.620c09-18. 27 T1545:27.66a08-15. Such contexts would be associated with nairmāṇikakāya in later texts that posited such a body. Here we find hua shen 化身 but as a verb-object clause; see below. The passage varies considerably in Buddhavarman, T1546:28.54a15-24.

28 T1545:27.156c16. Under "power of thought" the passage clearly discusses items associated with the dharmakāya (such as the eighteen āveṇikadharmas). 29 T1545:27.680a13-14.

30 For example, when discussing "Buddha" in the context of three gems, the text says, "'Buddha' here refers to the aśaikṣadharmas in the body of the Buddha (foshen). Taintless (i.e. without outflows, anāsrava) faith that takes that as its object (緣 ) is called 'realisation and purification by means of/with regard to buddha';" T1545:27.533b20-26. In the midst of further complications, the text here draws distinctions between the three gems in terms of their embodiment: (1) the aśaikṣadharmas in the body of the Buddha is what is referred to by the term buddha; (2) the term dharma refers to "the śaikṣa and aśaikṣa dharmas, i.e. the three anāsravêndriyas etc." (三無漏根等學無學法 ) in the body of the pratyekabuddha; or śaikṣa dharmas in the body of the bodhisattva; (3) the term saṃgha refers to the śaikṣa- and aśaikṣadharmas in the body of the śrāvaka, etc. Not just the Buddha, but adepts of all three vehicles can be characterised by their different modes of embodiment. That this passage appeals to the model of the three vehicles, with śrāvakas at the bottom, and to the bodhisattva ideal is one respect among several in which Vibh seems be tend towards some Mahāyāna ideas. Cf. also n. 36, 56.

31 T1545:27.160b14-15. This passage is simpler in Buddhavarman, and this particular sentence is absent; see n. 171. Cf. 428a20-22: "[Ordinary] compassion is perfected in the body of the śrāvaka, the pratyekabuddha, and the Buddha; great compassion is only perfected in the body of the Buddha." This passage is already in Buddhavarman, T1546:28.322a04-08, where the text speaks of the "great body" (dashen, *mahākāya), rather than the "body of the Buddha"; see above n. 22. Immediately before this, it is stated that the Buddha's compassion is called "great" (mahā-) because "all the limitless merits in the Buddha's body are great", 以佛身中一切功德皆是大, 428a06-07.

• mātṛpitṛkasaṃbhavakāya) can be traced back to the oldest layers of Buddhist literature, in Dīghanikāya 2, the Sāmaññaphala-sutta. At an advanced stage of practice, the monk prepares himself for the elaboration of a manomayakāya (in what we called the "snake slough" model) by coming to the realisation:

"This my body is material (rūpī), made up from the four great elements, born of mother

and father (mātāpettikasambhavo, more lit. "parentally engendered"), fed on rice and

gruel, impermanent, liable to be injured and abraded, broken and destroyed, and this is my consciousness which is bound to it and dependent on it." This passage does note that the dissatisfactory given body, which is about to be transcended, is composed of visible material form (rūpī). However, Vibh has picked up not that characterisation (as would be the case if it spoke of rūpakāya), but rather the notion of being born of parents. Under the aegis of this *mātṛpitṛkasaṃbhavakāya, further, Vibh makes a systematic distinction not found in its source in DN 2; the term has undergone creative elaboration. In discussing the Buddha as refuge, Vibh explicitly opposes the *mātṛpitṛkasaṃbhavakāya to the dharmakāya. One does not take refuge in the Buddha's "body comprised of head, neck, belly, back, hands, feet etc."; "this body born of father and mother is a defiled (sâsrava) dharma, and is not the object of refuge-taking". Here, the emphasis is on the imperfection and impurity of the *mātṛpitṛkasaṃbhavakāya, and so it seems closer to the body shared by the Buddha with ordinary sentient beings.

Elsewhere, however, the emphasis is on ways the Buddha's *mātṛpitṛkasaṃbhavakāya is extraordinary. For example, the Buddha tells Ānanda that he can travel in a manomayakāya to the Brahmā Heavens. However, Ānanda is unimpressed, saying that since Śrāvakas can also achieve this, it is a pretty mediocre feat (何其劣哉!) and isn't much to brag about (世尊何足自歎). Hasn't the Buddha got something unique about him? Can the Buddha make the same journey in his body born of father and mother 父母生身, comprised of the material great elements (麁大種, cf. DN 2 Pāli cātummahābhūtiko). The Buddha replies that indeed he can. Ānanda is at last genuinely impressed. Another passage uses the same term for the Buddha's body as it is possessed of prodigious physical strength, which allows him to win a contest against a gaggle of strongmen. In these passages, it is clear that the Buddha's *mātṛpitṛkasaṃbhavakāya is not a type of body shared with ordinary sentient beings, but a special body only pertaining to the Buddha.

In sum, in DN 2, the notion of *mātṛpitṛkasaṃbhavakāya was an epithet of the ordinary material body of any practitioner (not just the Buddha), implying the radically imperfect and dissatisfactory nature of that body. In Vibh it is not readily predicated of other beings. This special body has a number of remarkable properties: superhuman physical strength; the ability to travel to the Brahmā Heaven without elaborating a manomayakāya; being malleable like molten metal to the Buddha's will. Nonetheless, it is not entirely perfect, being impure, and therefore inappropriate as a refuge. There are relatively few passages in which the "body born of father and mother" is named so precisely as such. In far more passages, the ordinary body of a Buddha, in

brahmalokaṃ upasaṇkamitā)" The Buddha replies, "I recall, Ānanda, having gone to the brahmā world by spiritual power with a mind-made body." La Vallée Poussin tries somewhat unconvincingly to link this to Sanaṃkumāra's assumption of a different kind of body when he visits the Heaven of the Thirty-Three; "Notes" 764-

contrast with his dharmakāya, is referred to by the obviously related but less specific term shengshen 生身, "body of birth" (*sāṃbhavikakāya, *saṃbhavakāya). I take both terms to refer to the same conception of the Buddha's body.

In most passages where we find "body of birth", it means the ordinary living body of the Buddha, in explicit contrast to the dharmakāya. The Buddha can receive offerings of goods because "the body of birth must depend upon clothing, food etc.;" he cannot recieve offerings of dharma because his dharmakāya is already perfect. The text stipulates the Buddha's "body of birth" when it discusses the Buddha's "bodily strength" 身力, by contrast with his "mental powers" 意力. Similarly, causing a schism in the saṃgha is a terrible sin because it harms the Buddha's dharmakāya, which is contrasted with "a malicious attack that draws blood from the Buddha's body", where the body is sometimes called "body of birth". The term "body of birth" is also used for the physical body in contrast to the dharmakāya in discussing which sense-organ is superior. Again,

paranirmitavaśavartidevas have nothing of the sort. [Conversely,] there are other conditions in virtue of which the karma of the paranirmitavaśavartidevas is superior, because the bodies of those gods are pure and subtle like the flame of a lamp, whereas the "body of birth" of the bodhisattva still has various excreta (便利) and impurities." 有餘緣故菩薩業勝。以菩薩身是力無畏等無邊功德所依止故。他化自在天身無如是事。有餘緣故他化自在天業勝。以彼天身清淨微妙如燈焰故。菩薩生身猶有種種便利不淨, T1545:27.101b06-13.

The eighteen āveṇikadharmas are often identified with the dharmakāya (see below), so that the passage is proposing that the body of the bodhisattva is based upon the dharmakāya. The most plausible interpretation of this problem is just that boundaries between bodhisattva and Buddha are blurred. In application to the concept of "body of birth", at least, this is not too problematic. The Buddha is still a bodhisattva at the beginning of his last earthly life, i.e. at the moment he is "born", and the "body of birth" is the body he is then born into. The problem of how a bodhisattva can have a dharmakāya is less easy to evade. If it is possible to speak of the body of a bodhisattva (before his awakening under the bodhi tree) as "based upon" buddhadharmas or even the dharmakāya, this almost smacks of the ("Mahāsāṃghika") docetistic doctrine that Sarvâstivādins should be opposing. This problem need not detain us here, however. in commenting on the Mahāparinirvāṇa sūtra tradition that the Buddha exhibited his physical body to the assembled monks just before he entered parinirvāṇa,46 the text proposes that "see" (觀 avalokayata) is said with regard to the body of birth, where the virtually synonymous "behold" (察 vyavalokayata) is said of the dharmakāya.47 "Body of birth" is also the term for the ordinary body in a curious passage that contrasts the clothing required by beings in the rūpadhātu with "modesty" as the garb of the dharmakāya.48

The term "body of birth" is also consistently used (interchangeably with "the body of the Buddha") in discussing the purity or impurity of the Buddha's body. Against the claim of the "Mahāsāṃghikas" that the Buddha's body is untainted ("without outflows", anāsrava), the Vaibhāṣikas argue that this would have made it impossible for Anupamā 無比女人 to feel lust for his body,49 for Aṇgulīmālya to feel anger towards him, etc. Since it can be the objective support (緣, *ālambana) for such impure emotions, "the body of birth of the Buddha is certainly not taintless (anāsrava)" (my emphasis).50 Similarly, when the Mahāsāṃghika opponents cite a passage stating that the Tathāgata was not rendered impure by worldly dharmas,51 the Vaibhāṣika retort that the scripture refers differentially to the "body of birth" and the dharmakāya. The statement that the Tathāgata was born and dwelt in the world, etc., is made with reference to the "body of birth" 依生身說 , whereas the statement that he was not soiled by worldly dharmas is made with reference to his dharmakāya.52 In all these passages, the opposition between the "body of birth" and the dharmakāya is clear.

bodies. This passage already in Saṃghabhūti and Buddhavarman.

46 avalokayata bhikṣavas tathāgatasya kāyaṃ; vyavalokayata bhikṣavas tathāgatasya kāyaṃ. Ernst Waldschmidt, Das Mahāparinirvāṇasūtra Teil III (Berlin 1951), §42.10; cited in Gustav Roth, "The Physical Presence of the Buddha and Its Representation in Buddhist Literature," in Investigating Indian Art: Proceedings of a Symposium on the Development of Early Buddhist and Hindu Iconography Held at the Museum of Indian Art Berlin in May 1986 (Berlin: Museen für Indische Kunst, Staatliche Museen Preussischen Kulturbesitz, 1987), 293.

47 應觀我者。謂於生身。應察我者。謂於法身, T1545:27.957c25-958a14.

48 T1545:27.362b14-c14. In the rūpadhātu, beings are clothed at the very moment of birth, "because shame is greater in the rūpadhātu" (色界中慚愧增故 ). By contrast, "Modesty is the garb of the dharmakāya. Just as the dharmakāya is clad in this most excellent of robes, so, too, the body of birth [is clad] (?? 慚愧即是法身衣服。如彼法身具勝衣服生身亦爾)." Most creatures in the kāmadhātu, by contrast, are born naked, because "most [[[beings]] in] the kāmadhātu are without shame" (欲界中多無慚愧). In both of the earlier Vibh the term is also "body of birth", T1546:28.267c27, T1547:28.518b26-27. Exceptions in the kāmadhātu are bodhisattvas and the nun Śuklā (白淨苾芻尼, later in the same para. 白淨尼, 白淨比丘尼 in Saṃghabhūti, T1546:28.267c29); on Śuklā, see Avadānaśataka ascribed to Zhi Qian

(this ascription is considered incorrect by modern scholars; Jan Nattier, A Guide to the Earliest Chinese Buddhist Translations: Texts from the Eastern Han 東漢 and Three Kingdoms 三國 Periods, Bibliotheca Philologica et Philosophica Buddhica X [[[Tokyo]]: The International Research Institute for Advanced Buddhology, Soka University, 2008], 121-122), T200:4.239b17-c11; P. L. Vaidya,ed. ed., Avadāna-Śataka, Buddhist Sanskrit Texts No. 19 (Darbhanga: The Mithila Institute, 1958), 180-182; Stanley Frye, trans., Sutra of the Wise and the Foolish (mdo mdzangs blun) or Ocean of Narratives (üliger-ün dalai) (Dharamsala: Library of Tibetan Works and Archives, 1981, 2000), 96-98, "The Beggar Woman Gives her Clothing." 49 T1545:27.391c27-28, 於佛身無比女人不應起愛. For the story of Anupamā, see Andy Rotman, "The Erotics of Practice: Objects and Agency in Buddhist Avadāna Literature," Journal of the American Academy of Religion 71, no. 3 (2003): 567-570; "The Story of Makándika the Wanderer", in Joel Tatelman, ed. and trans., The Heavenly Exploits: Buddhist Biographies from the Divyāvadāna, Volume One, The Clay Sanskrit Library (New York/Bury St Edmunds, Suffolk: New York University Press/JJC Foundation, 2005), 310-415; Yijing's 義淨 (635-713) Mūlasarvâstivāda vinaya, T1442:23.886a19-c20. 50 既緣起愛及瞋慢癡故佛生身定非無漏, T1545:27.229a15-24, 392a04-11, 392b13-17. 51 T1545:27.229a18-19, 871c02-05, etc.

52 T1545:27.229a15-b02, 391c26.392a15, 871c17-20 etc. Note that this interpretation is also found in the Saṃghabhūti version of the text, T1547:28.463b06-10; and in Buddhavarman T1546:28.176a24-b13.

Elsewhere in the text, the contrast between the Buddha's "body of birth" and the dharmakāya is explicit ("the bodies of the Buddha are of two types"). Here, the text is discussing the "fruits of the monastic life", and trying to resolve a dilemma: scriptures sometimes say there are four such fruits 四果, i.e. the four grades of attainment from Stream-Enterer to Arhat; sometimes they say that the life of a monk is itself one of the fruits won by "going forth". "There are two broad senses in which it is shown (現 ) that the Buddha is 'beyond the world' (*lokôttara, 出世): (1) the conventional (worldly) sense (世俗, *vyāvahārika); and (2) the ultimate sense (勝義, *pāramārthika). The 'worldly sense' refers to his abandoning the dharma of the householder and taking up [that of] the non-householder, [that is,] shaving off his hair and beard and donning the monastic robe, and taking up the observance of pure precepts with a correct mind of faith. The 'ultimate sense' refers to his attainment of perfect awakening to the four noble truths. When he first went forth from domestic life, it already shown that the Buddha was 'beyond the world' in the conventional sense; when, having left home, he engaged in practice (?展轉修行 ), it showed further that he was beyond the world in the ultimate sense. Now, when we speak of 'urging [[[people]]] to go forth from domestic life', this means urging people to cultivate

(or 'imitate', 學) the Buddha's bodies (即是勸人學諸佛身); it is for this reason that the sūtras teach [as they do?]. To explain, the bodies of the Buddha are, broadly speaking, of two types: (1) the body of birth; (2) the body of dharma (諸佛身略有二種。一者生身。二者法身). If someone abandons the dharma of the householder and takes up [that of] the non-householder, shaving off hair and beard, donning the monastic robe, and taking up the observance of pure precepts with a correct mind of faith, this should be understood to be cultivating/imitating the Buddha's body of birth (學佛生身 ). If, on the other hand, someone engages in the cultivation of correct practice, and gives rise to perfect awakening to the four noble truths, this should be understood to be cultivating/imitating the Buddha's body of dharma (學佛法身)."

In this rich passage, the following points are of special interest. (1) The Buddha, qua Buddha (and not as an ordinary sentient being) does indeed have two kinds of body, but that they are his dharmakāya and his "body of birth" (not rūpakāya). (2) Both are extraordinary (*lokôttara); neither is just the ordinary given body possessed by other, imperfect sentient beings. (3) Strictly speaking, in this passage, the "body of birth" (or its "cultivation") is only spoken of as such from the moment of departure from the household life. (4) The entirety of the monastic life is conceived of very broadly as an endeavour to cultivate and eventually realise these two kinds of Buddha body. When the "body of birth" is discussed in greatest detail, the text connects it with the problem of sôpadhiśeṣa- and nirupadhiśeṣanirvāṇa (有/無餘依涅槃), "Nirvāṇa with and without a support". This Vibh passage is thus part of an extended history in which sôpadhiśeṣa- and nirupadhiśeṣanirvāṇa came gradually to have corporeal meaning and be bound up in the larger discourse about the corporeal implications of the attainment of buddhahood. Because of the length of the passage, I translate it in full in Appendix 1

(below p. 155); here I will only summarise the main points pertinent to our discussion.

This passage says that the difference between the two types of nirvāṇa stems from the fact that in sôpadhiśeṣanirvāṇadhātu, the life(-force) is not yet exhausted; therefore the physical form comprised of the four elements (=mahābhūtas; 大種造色) still continues; because of this, mind (citta) and associated dharmas (caitāsikadharmas) continue to arise in dependence upon that physical form. Nirvāṇa without remainder is converse in all respects. "Physical form" (se 色, rūpa) is identified first with the "body of the five senseorgans" (五根身, *pañcêndriyakāya). It is then identified with rūpakāya: "[The phrase] 'physical form comprised of the elements' refers in general to the rūpakāya" (大種造色者總顯色身).” Most importantly for our current purposes, however, the text also speaks in the same connection of the Arhat's remaining physical body as a "body of birth".

This mention of rūpakāya and "body of birth" in the same sense and context might lead us to suspect that the two bodies are to be identified. This might seem to lend limited support to the notion that the Buddha has a rūpakāya. Here too, rūpakāya and "body of birth" once more refer to a kind of embodiment shared by the Buddha with other beings – however, here, it is shared with other Arhats. This means that the "body of birth" is not entirely a unique property of the Buddha, as the dharmakāya is (see below); but even if it is shared, it is still, in distinction to the rūpakāya, a body which is possessed by special beings. This confirms once more that "body of birth" is the specific term used to refer to the ordinary physical body in the special situation where it is possessed by a perfected being.

Rūpakāya here is not strictly relevant to the Buddha in another sense. The careful technical treatment of the rūpakāya in this passage shows that it refers primarily to a body of physical form (rūpa) as the property of sentient beings in the realms of desire (kāmadhātu) or form (rūpadhātu). This is contrasted to the kinds of bodies possessed by beings in the realm of formlessness/the immaterial realm (ārūpyadhātu). In other words, rūpa- in rūpakāya is not understood in opposition to dharma, but in opposition to ārūpya. We must therefore reorient our understanding of rūpakāya from Buddhalogical to cosmological questions. In fact, the passage alternates carefully between rūpakāya and "body of birth". Rūpakāya is the term in all cases describing Abhidharmic analysis of the relationship between ordinary body as "support" and the operation of mind and associated dharmas (para. <2>);59 or where discussing the physical body as it does not pertain to the Arhat (as something that has ceased upon entry into the nirupadhiśeṣanirvāṇadhātu, paras <7>, <9>). By contrast, "body of birth" is the term for the physical body as it does still persist for some Arhats (e.g. as opposed to the kleśas, paras <5>, <10>). The text only uses the term rūpakāya to refer to the body as it persists in an Arhat in discussion of anomalous cases (para. <12>), where the issue is the technical distinction between bodies obtaining in different cosmological realms. Thus, "body of birth" is the term for the special case where a physical body endures for a being who has already attained nirvāṇa. On the basis of this passage, then, we would say that "body of birth" refers to a body of the Buddha, too, insofar as it is the physical body of an Arhat.

In sum, we have seen in this section that Vibh's "two-body" theory opposes to the dharmakāya the "body of birth" or the "body born of father and mother", and not the rūpakāya. The most significant ideas about this body are: it is the Buddha's last physical earthly body right through his lifetime (from his early bodhisattva years to the very scene of his parinirvāṇa and even the relics); it is the body identified as the "remainder" in the corporeal interpretation of sôpadhiśeṣanirvāṇadhātu; it is the object of offerings, and requires food and clothing; it is immensely strong; is the body from which drawing blood is an ānantarya sin; it is impure (tainted/'with outflows', sâsrava); it is the "support" of the dharmakāya, and it is possible to harm the dharmakāya by harming it; it is "based upon" a set of "limitless merits" identical with the āveṇikadharmas and therefore upon the dharmakāya; part of the practice of a monk is understood as the imitation of this body; and the Buddha alone is capable of travelling miraculously even to the Brahmā Heaven in this body. This body is partially mundane, being tainted or defiled, heir to harms and ills, dependent upon conditions of sustenance etc.; on the other hand, it is also certainly no ordinary body, existing in some ill-defined relation to the dharmakāya; having extraordinary strength; being capable of other miraculous feats; being a great field of merit as the object of offerings to the Buddha; and so on.

Dharmakāya in Vibh A dharmakāya is indeed ascribed to the Buddha in Vibh. However, the text's specific conception of the dharmakāya has often been partially misconstrued. Scholars have often understood this and other "Mainstream" texts to present dharmakāya pluralistically, as comprised by such collections as the Buddha's eighteen āveṇikadharmas ("unique qualities") or the anāsravaskandhas/aśaikṣadharmas ("qualities without contamination", "qualities [of one] beyond any further practice"). Such interpretations are accurate, but fail to delve beneath the surface of the doctrine to a deeper claim. Following the lead of la Vallée Poussin, Makransky has shown that the Sarvâstivādin conception of the dharmakāya hinges on the idea that the Buddha is embodied in bodhi. I will argue that overlooking this underlying understanding can lead to exaggeration of the extent to which the Sarvâstivāda doctrine of dharmakāya is pluralistic. Further, this Sarvâstivādin doctrine of dharmakāya as embodiment in bodhi is part of a broader pattern in early Common-Era doctrines, and the gulf between Sarvâstivāda and Mahāyāna doctrines is thus not as great as has sometimes been thought.

The eighteen āveṇikadharmas as dharmakāya

The claim is sometimes seen in the secondary literature,61 and not entirely without justification, that the Sarvâstivāda define the dharmakāya as the eighteen āveṇikadharmas or "unique qualities" of a Buddha: ten "powers" (bala), four "confidences", three foundations of mindfulness, and great compassion. This characterisation is partially correct. However, it is also limited, first, because it presents only part of the overall picture; and second, because it fails to penetrate to the underlying idea of embodiment in bodhi.

The definition of the dharmakāya as the eighteen āveṇikadharmas62 is usually taken from a passage in which the text is discussing respects in which all Buddhas are equal.63 Among other things, the Buddhas are said to be equal "with regard to the dharmakāya"

(法身等).64 This resonates with more radical Mahāyāna statements that the Buddhas are one in virtue of the dharmakāya.65 Vibh here stops short of asserting the absolute identity of the Buddhas, but it obviously tends in the same direction. The passage goes on:

". . . that is to say, as one Buddha accomplishes (成就, */sidh) limitless merits [of]/(such as? 等) the eighteen unique properties [of a Buddha] (āveṇikadharmas, 十八不共法), viz. the ten powers (bala), the four confidences (四無畏, caturvaiśāradya), great compassion, and the three foundations of mindfulness (smṛtyupasthāna), (and so on? 等), so too do the other Buddhas, and thus we say [they are all] equal."66 Despite some ambiguity,67 this passage clearly identifies dharmakāya with the eighteen āveṇikadharmas. Underlying this definition is the idea that the Buddha is embodied in his awakening. The whole list breaks down to various aspects of the Buddha's special gnosis.

61 E.g. Guang Xing, Concept of the Buddha Ch. 2, esp. 35-36, more generally 35-44. 62 In Vibh, āveṇikadharmas so named do not feature very largely in connections other than discussion of dharmakāya. The āveṇikadharmas are part of a definition of buddhahood without any mention of dharmakāya, T1545:27.735c16-18; they are part of a discussion of the difference between Buddhas, śrāvakas and pratyekabuddhas, 189b05-16. Individual items on the list are discussed under the rubric of "powers of mind" (意力; cf. AKBh introducing AK 31a, mānasaṃ balam), 156c16-160c22; cf. AK & Bh

7.28-31, VP V, 66-74. 63 As opposed, e.g., to their lifespan, etc., in which respect they are not equal. Other respects in which they are not equal, according to AK, are the families and castes into which they are born, the size of their physical body, the time for which their Dharma endures, etc.; VP V, 81. 64 This doctrine of the "equality" of the Buddhas with respect to body and āveṇikadharmas is found in the Chinese canon as early as Zhi Qian's Vimalakīrtinirdeśa, though not yet under the name of dharmakāya. Lamotte implies it is there derived from Sarvâstivāda; once more, it is the central thrust of this study to show that we cannot be confident this is the case. See Radich, "Somatics" §4.5.2, p. 615. The other respects in which Buddhas are equal are in their "equipment" (saṃbhāra), that is, in the merits they have accumulated in virtue of countless aeons of prior bodhisattva practice; and in their service for the needs of other beings. Cf. AK 7.34, sambhāradharmakāyābhyāṃ jagataś cārthacaryayā/ samatā sarvabuddhānāṃ nāyurjātipramāṇataḥ; VP V, 79-81. 65 E.g. in the Drumakinnararājaparipṛcchā, "knowing that all Buddhas are one Buddha, by virtue of [their] entry into the inconceivable dharmadhātu;" chos kyi dbyings bsam gyis mi khyab pa la zhugs pas/ sangs rgyas thams cad sangs rgyas gcig tu shes pa; Lokakṣema reads, "All Buddhas are nothing but one Buddha. For what reason? Because [their] penetration of the 'body' of dharmas (fashen) is incalculable"一切佛為一佛耳。何以故。法身所入不可計故, T624:15.358b05-06. See Harrison, "Real Phantom Body"

62; discussed in Radich, "Somatics" 767. 66 謂如一佛成就十力四無所畏大悲三念住。十八不共法等無邊功德。餘佛亦爾故名平等 , T1545:27.85a26-28. See also 624a13-15, where the same ideas are repeated almost verbatim. See also 131b14-21, which differs slightly. This passage is in Buddhavarman; see below, n. 163. Compare AKBh to 7.34, discussed below p. )169.

67 Stemming from the placement of 等 and the unclear notion of "merits of" the āveṇikadharmas. The wording of this list is somewhat garbled in Buddhavarman also; see n. 163.

The ten "powers" are all kinds of jñāna ("wisdom", "gnosis", "special knowledge"): (1) the power of special knowledge of the possible and the impossible (處非處智力 , sthānâsthānajñānabala); (2) the power of special knowledge of the arising of dharmas [as a result of] karma (二業法集智力, karmavipāka(/phala)jñānabala); (3) the power of special knowledge of dhyāna, vimokṣa, samādhi and samāpatti (靜慮解脫等持等至發起雜染清淨智力, dhyānavimokṣasamādhisamāpattijñānabala); (4) the power of special knowledge of [the dispositions of] various classes of beings (種種界智力, nānādhātujñānabala); (5) the power of special knowledge of the aspirations of various beings (種種勝解智力, nānâdhimuktijñānabala); (6) the power of special knowledge of superior and inferior [[[moral]]] faculties [of beings] (根勝劣智力, indriyaparâparajñānabala); (7) the power of special knowledge of the ways that led to all destinies of rebirth (遍趣行智力, sarvatragāminīpratipajjñānabala); (8) the power of special knowledge of the recollection of prior dwelling-places (宿住隨念智力, pūrvanivāsajñānabala); (9) the power of special knowledge of the births and deaths of beings (死生智力 , cyutyupapādajñānabala); (10) the power of special knowledge of the destruction of taints (outflows) (漏盡智力, āsravakṣayajñānabala). The four respects in which a Tathāgata is confident, that is, "without fear, dauntless" (Ch. wuwei 無畏 ) or "assured" (VP's translation of AK's Skt. vaiśāradya), are also very close to the Tathāgata's jñāna. They either directly comprise jñāna, or are its immediate derivatives: (1) confidence in perfect, complete awakening [to all dharmas Skt.] (正等覺無畏 , sarvadharmâbhisaṃbodhivaiśāradya); (2) confidence in having eternally destroyed all taints (outflows, āsrava) (漏永盡無畏, sarvâsravakṣayajñānavaiśāradya); (3) confidence in the exposition of dharmas obstructing [realisation of the truth and deliverance] (說障法無畏, antarāyikadharmavyākāraṇavaiśāradya); (4) confidence in the exposition of the paths leading to deliverance (說出道無畏 , nairyāṇikapratipadvyākaraṇavaiśāradya).

(1) is a kind of confidence in the unshakeable nature of anuttarasaṃyaksaṃbodhi. Whether or not that confidence is part and parcel of awakening or derivative of it, it is inextricably tied up with the Buddha's gnosis. (2) is confidence in or deriving from a kind of jñāna. The relation of (3) and (4) to gnosis is slightly more indirect. Strictly speaking, they are confidence in skills of exposition that derive from that gnosis. Nonetheless, they are still related to jñāna, insofar as the Dharma is discourse about the truth which is the content of the Buddha's gnosis. The close links between these four "confidences" and jñāna are further emphasised in discussion of the overlaps and identities between them and the ten "powers". The next category in the list of the eighteen āveṇikadharmas is of course the three

"foundations of mindfulness" (smṛtyupasthānas). Each of these items represents a different way that the Buddha maintains a state of "indifference, mindfulness and full consciousness" (大捨住念正知, upekṣas . . . tathāgato. . . smṛtaḥ saṃprajānan) regardless of circumstances. Mindfulness and consciousness are, once more, fundamentally mental, not to say gnostic, qualities. Vibh even says outright that these three smṛtyupasthāna are incorporated in the first of the ten "powers", i.e. a kind of jñāna. The last āveṇikadharma is "great compassion" (mahākaruṇā). The close links between this item and gnosis is made explicit by Vibh's own statement that it too is "incorporated" in the first of the ten powers.

In sum, all eighteen āveṇikadharmas are identified with kinds of jñāna, or perhaps as products of jñāna. The various kinds of jñāna are clearly aspects of the Buddha's special gnosis, in virtue of which he is a Buddha. Gnosis is thus the central focus of the whole rubric of the āveṇikadharmas, and therefore of this definition of dharmakāya. The five anāsravaskandhas and the "aśaikṣadharmas comprising bodhi" as dharmakāya Secondary scholarship also notes an alternative interpretation of the dharmakāya in Sarvâstivāda. On this reading, dharmakāya comprises the five anāsravaskandhas ("pure aggregates") of morality (śīla), concentration (samādhi), wisdom (prajñā), liberation

(vimukti) and insight into the special knowledge of liberation (vimuktijñānadarśana). This interpretation is also supported by the texts. One problem with previous studies, however, is that they have not recognised the coexistence of both the āveṇikadharma model and this model. It is also important to note the way this notion of dharmakāya, too, is ultimately reducible to the Buddha embodied in his gnosis. We already saw above that in discussing the Buddha-refuge, Vibh says that one properly takes refuge in the dharmakāya, which it glosses as "the aśaikṣadharmas of the

Buddha which comprise bodhi". Despite surface appearances, this definition is in fact the same as the definition in terms of anāsravaskandhas. First, then, we must demonstrate this equivalence by turning to the Abhidharmakośa, which has been the source of the definition in terms of anāsravaskandhas for secondary scholars.

In AKBh 4.32, we find a similar definition, but without explicit mention of aśaikṣadharmas. AKBh further specifies that bodhi is comprised of kṣayajñāna and anutpādajñāna (as elsewhere, as we shall see below), in addition to (one dimension or kind of) samyagdṛṣṭi. AKBh also adds that the Buddha is bodhi so defined "along with the dharmas that attend [those dharmas]" (saparivārāḥ). To understand this definition, we must therefore understand: (1) the definition of bodhi as specific kinds of jñāna; (2) the definition of aśaikṣadharmas, explicit in Vibh but not in AK & Bh; (3) AKBh's term "attendant [[[dharma]]]s " (parivāraḥ).

Bodhi is repeatedly defined throughout AK & Bh as kṣayajñāna and anutpādajñāna. This highly technical definition is intimately linked to AK's account of the moment of transition to arhatship and buddhahood, and characteristics of the states thereafter. Complete liberation occurs with the transition from vajrôpamasamādhi ("the concentration like adamant", the last meditative accomplishment immediately before full awakening) to the arising of kṣayajñāna.94 This moment is explicitly identified with arhatship itself,95 and it is also the moment when a Buddha becomes a Buddha.96 Where the Arhat is of the class called "unshakeable/ unmoveable" (akopyadharman),97 he enters immediately after the arising of kṣayajñāna into anutpādajñāna, that is, the knowledge that tainted (āsrava) dharmas will not ever arise again.98

In the context of the same transition to arhatship and buddhahood, the text also broaches the definition of the term aśaikṣa. This term is glossed as referring to dharmas which are both untainted ("without outflows", anāsrava) and unconditioned (asaṃskṛta); by contrast, śaikṣadharmas are untainted but conditioned (saṃskṛta).99 In fact, the text identifies this transition into arhatship with entry into the category of aśaikṣa. Thus, the category of aśaikṣa is closely connected to that of bodhi, as defined by kṣayajñāna and anutpādajñāna. Even though AK & Bh does not specify that the aśaikṣadharmas are part of the Buddha-refuge, it is clear from context that this doctrine must be implicit (because the aśaikṣadharmas are an inalienable definitional component of buddhahood as defined by bodhi, which the text does identify with the Buddha as refuge).

It is therefore relevant to ask what the aśaikṣadharmas are for AK & Bh. Analysis shows they are identical to the five anāsravaskandhas. AKBh counts ten aśaikṣadharmas. (1) "perfect liberation" (saṃyagvimukti) and (2) "special knowledge of perfect liberation" (saṃyagvimuktijñāna) are considered exclusively aśaikṣa.100 Saṃyagvimutkijñāna is also called saṃyagjñāna, when it refers to kṣayajñāna and anutpādajñāna (=bodhi).101 (3-10) The eight members of the eightfold noble path are also considered to have both śaikṣa and aśaikṣa variants. One qualified as aśaikṣa possesses the aśaikṣa variants.102 How does this tenfold rubric map onto the fivefold set of anāsravaskandha? To answer

and anutpādajñāna are also identified with two other "pure" jñānas: (1) dharmajñāna ("special knowledge of dharma") and (2) anvayajñāna ("consequent/subsequent special knowledge"). They have as their respective objects the four noble truths in the kāmadhātu, or the rūpa- and ārūpyadhātu respectively; AK & Bh 2-4c, VP IV, 4-5; see also AK & Bh 4d-6b, VP IV, 6-8.

94 AK 6.44d-45a, VP IV, 230; see also Bh to AK 7.8, VP V, 13-14; 7.25a and Bh, VP V, 57. 95 AK 6.45b, VP IV, 230-231. 96 AKBh introducing AK 28ab, VP V, 66. See also AKBh 2.44ab and Bh, VP I, 204-205. 97 Defined later in the same chapter as part of a sixfold classification of types of Arhat, AK 6.56 ff., VP IV, 251 ff. 98 AK & Bh 6.50ab, VP IV, 240; see also AK 25b and Bh, VP V, 57-58. Where an Arhat is not "unshakeable", by contrast, he enters thereafter into another moment of kṣayajñāna, or into the aśaikṣasaṃyagdṛṣṭi. The latter therefore belongs to Arhats of both classes; AK 6.50b-d and Bh, VP IV, 240. 99 AKBh to 6.45, VP IV, 231-232.

100 Bh to AK 6.75b-d, VP IV 295-296. The text here says that the one qualified as aśaikṣa is so-called in virtue of the fact that, because of complete liberation from the bonds of all defilements, he is infused with/ empowered by (prabhāvita) vimukti and vimuktijñānadarśana (vimuktitatpratyātmajñānābhyāṃ): aśaikṣas tu sarvakleśabandhanātyantanirmokṣād vimuktitatpratyātmajñānābhyāṃ prabhāvita iti tasyaiva tadvacanaṃ nyāyyam. On prabhāvita, see Radich, "Somatics" n. 1525.

101 AKBh to 6.76, VP IV, 298. 102 I.e. "aśaikṣa right view" (aśaikṣī saṃyagdṛṣṭi) through to "aśaikṣa right concentration" (aśaikṣī saṃyaksamādhi); Bh to AK 6.75b, VP IV, 295. this question, we must know what the text means by parivāraḥ, "attendant [[[dharma]]]s", when it defines the Buddha-refuge as bodhi "along with [the dharmas that] attend [it]". La Vallée Poussin and Makransky identify these "attendants" as the five anāsravaskandhas. This identification is justified, but matters are not as straightforward as these authors make them seem. As far as I can determine, the only support for this identification is AKBh 1.2, and even that is somewhat tangential.

At AK 1.2, the kārikā identifies "Abhidharma" itself, in the "ultimate" sense (paramārthika), as "immaculate prajñā, with its retinue" (prajñâmalā sânucārâbhidharmaḥ). To this, Bh adds,

"Here, 'prajñā' means 'discernment of dharmas' (dharmāṇāṃ pravicayaḥ). 'Amalā' ('immaculate') means anāsrava ('taintless, without outflows'). 'With its retinue' means 'with its attendants' (saparivārā). In this manner, we come to say that the fivefold collection of anāsravaskandhas is 'Abhidharma'." An identity between prajñā and bodhi is supported by AK 7.1b and Bh, which speaks of kṣayajñāna and anutpādajñāna as dhī. Dhī here seems to be used metri causa as a synonym for prajñā. If kṣayajnāna and anutpādajñāna equal bodhi, and kṣayajnāna and anutpādajñāna are dhī=prajñā, then bodhi is prajñā. Thus, (Abhidharma=) "prajñā with its retinue" would seem a close equivalent to (the Buddha-refuge=) "bodhi with its attendants". Via these terminological identifications, we can infer that in the AKBh 1.2 definition of Abhidharma, the "attendants" of bodhi would be the anāsravaskandhas. Thus, the Buddha-refuge comprises bodhi and the five anāsravaskandhas.

We can support this identification from another direction. We just saw that AKBh defines the aśaikṣadharmas in terms of a tenfold set: (1) saṃyagvimukti (2) saṃyagjñāna; (3-10) aśaikṣa versions of the members of the eightfold noble path. The eightfold noble path is often analysed as śīla, samādhi and prajñā, the first three members of the fivefold anāsravaskandha rubric. By the addition of vimukti=saṃyagvimukti and vimuktijñ āna=saṃyagjñāna, the aśaikṣadharma rubric is exactly coterminous with the anāsravaskandhas.

Thus, this AKBh 4.32 definition of the Buddha refuge in terms of bodhi and "attendants" varies only in wording from the Vibh definition in terms of (dharmakāya =) "the aśaikṣadharmas that comprise bodhi". This is supported by other passages in Vibh. Vibh consistently identifies Buddha with the aśaikṣadharmas, understood to comprise the Buddha's body (unqualified) or dharmakāya. The aśaikṣadharmas are also understood to be identical with the five *aśaikṣaskandhas, i.e. anāsravaskandhas. Like the definition of dharmakāya as the āveṇikadharmas, these definitions focus centrally on the Buddha's embodiment in his gnosis. Vimuktijñānadarśana = saṃyagjñāna is the fifth anāsravaskandha and the tenth aśaikṣadharma, i.e. the culmination of each rubric. However, we saw that AKBh defines this instance as kṣayajñāna and anutpādajñāna. These are precisely the two jñānas elsewhere identified with bodhi itself. Thus, if the Buddha-refuge is the aśaikṣadharmas comprising bodhi (Vibh), that same bodhi is already the tenth aśaikṣadharma; and if it is [[[bodhi]] defined as] kṣayajnāna, anutpādajnāna and its attendants = the anāsravaskandhas (AKBh), the anāsravaskandhas also already incorporate bodhi. In both cases, the Buddha as refuge is merely the aśaikṣadharmas-cum-anāsravaskandhas, with bodhi given pride of place therein.

Further, this whole discussion (according to AKBh) hinges on the transition from the "concentration that is like adamant" (vajrôpamasamādhi) to arhatship and buddhahood. eightfold noble path to these three aggregates in reply to a query from her former husband Visākha. Thus, śīlaskandha includes right speech (saṃyagvāc), right action (saṃyakkarmānta) and right livelihood (saṃyagājīva); samādhiskandha includes right effort (saṃyagvyāyāma), right mindfulness (saṃyaksmṛti) and right concentration (saṃyaksamādhi); and prajñāskandha includes right view (saṃyagdṛṣṭi) and right thought (saṃyaksaṃkalpa). See Bhikkhu Ñāṇamoli and Bhikkhu Bodhi, trans., The Middle Length Discourses of the Buddha: A New Translation of the Majjhima Nikāya (Boston: Wisdom Publications, The new Buddha acquires the two jñānas that comprise bodhi, and also, presumably, converts his attainments on the eightfold path into the aśaikṣa variants. These aśaikṣa variants of the members of the path are defined precisely by the fact and knowledge that "What is to be done is done" etc., i.e. that there is "no need for further training" (aśaikṣa). In all respects, this moment is of its essence a moment of gnostic attainment, and any "attendant" phenomena follow upon and are ancillary to that central gnostic fact. This is also shown by the very language of "attendants" (parivāras), "accompanying" etc. Vibh's main conception of dharmakāya as embodiment in bodhi: Summary

In conclusion, in Vibh we find two main definitions of dharmakāya – in terms of āveṇikadharmas, and in terms of aśaikṣadharmas. To these two definitions, it was necessary for us to add a third Sarvâsitvāda definition (perhaps later) from AKBh, in terms of the ānāsravadharmas. This diversity of definitions has been overlooked by previous scholars, who have tended to notice only one or another. The diversity is significant for at least three reasons. First, it shows that Sarvâstivāda Buddha-body doctrine is more complex than secondary scholarship has suggested. Second, as we will see below, that complexity may be a sign that Sarvâstivāda Buddha-body doctrine continued to develop until quite late, and this may have implications for relative chronological relations between Sarvâstivāda and Mahāyāna body doctrine. Third, this very flexibility in terms of surface expression, I believe, is another sign that the key issue is not this or that list of specific "dharmas", but a deeper underlying conception of the dharmakāya.

I have argued that the definitions of dharmakāya in terms of aśaikṣadharmas and anāsravadharmas turn out to be reducible to the same concept, expressed in different terms. I argued further that both this definition of dharmakāya as aśaikṣaskandha-cumanāsravaskandhas and the definition in terms of āveṇikadharmas can be shown to hinge on a more fundamental underlying notion of the Buddha as embodied in his gnosis (bodhi, jñāna etc.). Not unobserving this underlying conception, previous scholars have sometimes suggested that Mainstream models of dharmakāya (of which Sarvâstivāda is a paradigmatic case) are pluralistic, which they take as a key point of contrast to Mahāyāna models.113 This exaggerates the distance between the Sarvâstivāda model and Mahāyāna models, and attention to the underlying notion of the Buddha's embodiment in gnosis corrects this distortion, revealing that the Sarvâstivāda conception of dharmakāya is in fact quite close to the Mahāyāna conception in some respects.

Dharmakāya in Vibh: Complications

Thus far, this discussion of the Vibh doctrine of dharmakāya has focused on what I take to be the main characterisations of the term, in terms of an underlying embodiment of the Buddha in his gnosis, parsed variously in terms of several different lists of specific dharmas. Before closing, we should observe some complications to this tidy picture, lest we fall into an image of dharmakāya in Vibh that is unrealistically abstract and "bloodless".

above is that AKBh discusses the āveṇikadharmas immediately following this moment of transition. 113 I am thinking particularly of Harrison, "Phantom Body". In discussing its contrast with the Buddha's ordinary body above, we already observed in passing several features of Vibh's dharmakāya: The body of the bodhisattva is "based upon" the dharmakāya. Dharmakāya is the "body of the Buddha" upon which his mahākaruṇā is based. Dharmakāya is perfect and complete, and therefore the Buddha cannot receive offerings of dharma. Dharmakāya never wanes or deteriorates, and neither does its "power" or strength. The sin of creating a schism in the saṃgha is an attack on the dharmakāya. The ear is considered by some authorities superior to other organs in "guiding the dharmakāya". The Buddha as refuge is the dharmakāya, glossed as "the aśaikṣadharmas comprising bodhi". "Modesty is the garb of the dharmakāya." Where scripture says that the body of the Buddha is taintless ("without outflows", anāsrava), it means the dharmakāya. To engage in correct practice and awaken to the four noble truths is "to cultivate the dharmakāya of the Buddha".

Some of these passages seem to imply a very corporeal, organic understanding of the dharmakāya. Such an understanding is echoed elsewhere. In one curious passage, the possibility is envisaged that there might be a stage of realisation where the kuśalamūlas are small, which is equated with a state in which the dharmakāya is "not yet grown" (法身未長). In this condition, it is not possible to sever the kleśas, but it is also said that the practitioner is not "overcome" (摧伏) by the kleśas, because the fledgling dharmakāya is still "powerful" (有威勢). Such a dharmakāya is compared to a tiger cub, who, not yet being fully grown, cannot catch other beasts, but is equally not harmed by other beasts, because it is powerful or awe-inspiring (有威勢). This passage seems to suggest that a kind of nascent or embryonic dharmakāya might be possessed by a practitioner before they attain to the stage of buddhahood. Similarly, we already saw that the monastic vocation is urged in terms of the imitation of the Buddha's body of birth and dharmakāya. In this context, the dharmakāya, also, can be "cultivated", and this cultivation specifically comprises "the cultivation of correct practice" and "giving rise to perfect awakening". This would also seem to indicate that the dharmakāya can be created in a nascent form, and only eventually brought to full fruition, on an understanding of growth closely modelled on the processes of organic "bodies".

In another passage, we read (as the third of a series of partly nirukta-style glosses on the term upādāna, "attachment, grasping, clinging"): "'Upādāna' also means 'to harm' (傷害義). Just as, when one's body is repeatedly stabbed by sharp daggers, one's body will weaken and deteriorate, so too, where the *dharmakāya of a sentient being (有情) is repeatedly stabbed by the daggers of the poisons and the kleśas, that dharmakāya will weaken and deteriorate (法身便壞)."

This passage is quite strange. It says that the dharmakāya could "weaken and deteriorate", but we have seen elsewhere that when dharmakāya is predicated of the Buddha, it is said that this cannot happen. It also seems to attribute a dharmakāya to an ordinary "sentient being" (youqing有情). Once more, it seems that the text here envisages a nascent dharmakāya that grows gradually over time. A possibly related idea is the claim that it is the dharmakāya that is "sustained and nourished" by the "food" of the dhyānas.

The unusual and somewhat chaotic impression of dharmakāya that we receive from these passages is amplified by some mentions of dharmakāya found only in earlier versions of Vibh, but not in XZ. One passage in Saṃghabhūti compares the Buddha to a dragon-king. The dragon king "nurtures his body" in the sea, and then rises into the air to make it rain. Similarly, the Buddha nutures his dharmakāya through long ages of past Buddhas before making the rain of the Dharma fall. Once more, we are given the impression that the dharmakāya has a nascent form that grows gradually before the attainment of buddhahood. Elsewhere, Saṃghabhūti says that Brahmans of other schools have three truths, which comprise a "tripartite dharmakāya" comprising a śīlaskandha, a samādhiskandha, and a jñānaskandha. The ascription of this view to non-Buddhists is very unusual. Another passage unique to Saṃghabhūti129 describes the dharmakāya in very concrete terms that clearly equates it with the Dharma qua teaching, even going so far as to count texts and syllables. Of greatest interest here is the fact that this interpretation is so unusual among our materials, against some claims in secondary scholarship that Mainstream texts often present the dharmakāya in terms of the teaching (Dharma).

In sum, the notion of the Buddha's embodiment in his gnosis is central to the Vibh definition of dharmakāya, with that gnosis unpacked into lists of āveṇikadharmas and aśaikṣadharmas. Here, the dharmakāya appears quite abstract, and naturally enough, to properly belong only to a fully-fledged Buddha. Alongside this concept, however, we find other traces of a much more literalist, corporeal interpretation of dharmakāya, where dharmakāya seems oddly enough to be understood as something that grows from an embryonic form well before the attainment of buddhahood, and indeed, that can as such even be ascribed in some sense to quite ordinary sentient beings.

The disinterest in Buddha-bodies in earlier Sarvâstivāda Abhidharma texts On internal evidence, it is thought that seven works in the Sarvâstivāda canon, extant almost exclusively in Chinese, are earlier than Vibh. They are the six "limbs" of the Sarvâstivāda (more literally, "feet", Skt. *ṣaṭpādaśāstra, ṣaṭpādâbhidharma Ch. 六足論, 六足阿毘曇, 六分阿毘曇), and the Jñānaprasthāna, figured as the torso or trunk. Thus, in approximate chronological order, scholars distinguish three main strata in the literature as a whole. (1) Three texts are considered oldest:

Saṃgītiparyāya (阿毘達磨論, T1536); Dharmaskandha (法蘊足論, T1537); Prajñapti śāstra (阿毘達磨施設足論, T1538);

(2) They are followed by a second stratum comprising: Dhātukāya (阿毘達磨界身足論, T1540); Vijñānakāya (阿毘達磨識身足論, T1539); Prakaraṇapāda (阿毘達磨品類足論, T1542; 眾事分阿毘曇論, T1541); Jñānaprasthāna (阿毘達磨發智論, T1544, 阿毘曇八犍度論 T1543; "JñP"), the root text upon which Vibh is a commentary;

(3) These in turn are followed by Vibh. However, thus far in this study, we have taken Vibh as our main point of reference. The reason we have not examined these earlier texts instead is that those texts, even JñP, have little to say about the Buddha's special bodies. This suggests that at the stage represented by Vibh, Sarvâstivāda saw a sudden burst of interest in the problematic of the Buddha's embodiments. In what follows, I will demonstrate this by briefly surveying the evidence in this corpus for each of the major concepts we examined above for Vibh in turn.

Rūpakāya almost never refers to a body of the Buddha. It refers only to the case of the ordinary sentient beings in the kāma- or rūpadhātu in the Saṃgītiparyāya and the Dharmaskandha. The earliest text in which rūpakāya is used of a body of the Buddha is the Prajñapti śāstra (only once). It is entirely absent in the Dhātukāya, the Prakaraṇapāda, the Vijñānakāya, and JñP. The doctrine that the Buddha bears the thirtytwo marks of the mahāpuruṣa is almost entirely absent; it only appears twice in the Prakaraṇapāda. Not even the word "body" is mentioned. The "body of birth" (shengshen 生身 , *sāṃbhavikakāya) and "body born of father and mother" (fumushengshen 父母生身 , *mātṛpitṛsaṃbhavakāya) are entirely absent. The term dharmakāya is entirely unknown prior to Vibh, even in JñP.

Even aside from these terms, these earlier texts are uninterested in the Buddha's bodies in general. With few exceptions, we do not find even the general terms "body of the Buddha" or "body of the Tathāgata". The tangetially relevant anāsravaskandhas/ aśaikṣaskandhas do appear in isolated parts of this corpus, but are never discussed in terms of a corporeal conceit or related to the Buddha's embodiment. Sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu are simply not discussed in this corpus prior to JñP.

Significantly, even JñP is no more interested than its predecessors in the overt discussion of the problematic of the Buddha's embodiments. We only have a couple of scant signs of the explosion of interest to come in Vibh: the definition of the Buddha refuge in terms of the aśaikṣadharmas that comprise bodhi (without explicit mention of dharmakāya); the five pure aggregates, called aśaikṣaskandhas (without equating them with the dharmakāya or a corporeal conceit); the key passage relating sôpadhiśeṣaand nirupadhiśeṣanirvāṇadhātu to rūpakāya and the "body of birth".142 In the case of the last passage, in the earlier Saṃghadeva translation the discussion is not still framed in terms of the "body", but rather "the four gross [[[elements]]]" and "the five sense-organs". In XZ, the five sense-organs have become "the body of the five sense-organs" (*pañcêndriyakāya). Even in Xuanzang's JñP, however, neither of the terms rūpakāya and "body of birth" appear; this step is only taken in Vibh.

Comparison between different versions of Vibh Thus, the rich ideas in Vibh about the Buddha's embodiments do not represent a Sarvâstivāda heritage from time immemorial. Instead, they can be specifically dated to Vibh itself. Much therefore hinges on how we date these ideas as they appear in Vibh.

This question can be approached, if not solved, by comparison of XZ's Vibh with the earlier versions. In the course of the discussion above, I noted in the footnotes places where earlier translations diverge from XZ. Here I will summarise that information. Saṃghabhūti

Saṃghabhūti's Vibh never uses the term rūpakāya. Saṃghabhūti does mention the thirtytwo marks and other distinguishing features of the Buddha's physical body; reference is made, in some of these passages only, to his "body", but as in XZ, this body is never named as a specific type. Saṃghabhūti never mentions the "body born of father and mother". It does use "body of birth" (twice, opposed to dharmakāya as in XZ). In comparison to earlier Sarvâstivāda texts, Saṃghabhūti's discussion of dharmakāya is extensive. As we saw above (p. 144; see also Appendix 3), it even contains some passages on dharmakāya not found in XZ. Other passages too, speaking simply of "the Buddha's body" etc., represent a marked increase in interest in this problem: the passage explaining that the Buddha chose to be "womb-born" in order to leave relics; reference to a "great body" to explain reference to the Buddha's compassion a "great" (mahākaruṇā); discussion of whether it is possible to contemplate impurity using the body of the Buddha as an object. The five pure aggregates feature only once, as part of a discussion of the dharmakāya.152 Finally, the text knows the corporeal reading of sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu, though it does not relate this understanding to the Buddha's bodies.

In sum, Saṃghabhūti's version of Vibh already contains very much that is new in Sarvâstivāda. In particular, a host of new ideas about the dharmakāya have suddenly burst upon the scene. Buddhavarman

Buddhavarman's Vibh uses the term rūpakāya once in connection to the Buddha, in a context unique to this version of the text. It mentions the thirty-two marks of the Buddha in a few places, but in that connection, simply speaks of the "body". The term "body of birth" is used in the passage about sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu (see Appendix 1); when the Buddha bests the strongmen; and in other passages already in Saṃghabhūti. The "body born of father and mother" appears for the first time in Buddhavarman, in two passages only: the discussion of the Buddha's prodigious physical strength; and in discussion of the Buddha-refuge. Buddhavarman contains new dharmakāya passages after Saṃghabhūti: the passage on the Buddha's strength just mentioned also speaks of the powers of the dharmakāya; the nascent dharmakāya compared to a tiger cub; and definition of dharmakāya in terms of the eighteen āveṇikadharmas. Buddhavarman knows the anāsrava-/aśaikṣa-skandhas, and calls them "bodies" (shen 身) without associating them with the dharmakāya. It knows the corporeal reading of sôpadhiśeṣa- and anupadhiśeṣanirvāṇadhātu. Ideas only found in XZ

Thus, in sum, many ideas about Buddha-bodies are already present in Saṃghabhūti or  Buddhavarman, and thus date to the early fifth century at the latest. However, we must also consider passages that only appear in XZ. So far as I could determine, the following contents are unique to XZ: over a dozen mentions of the Buddha's body in discussion of the thirty-two marks; the discussion of marks and anuttarasaṃyaksaṃbodhi as conditions of Buddhahood in relation to Maitreya; Māra adopting a Buddha-body for Upagupta; how the Buddha won the hide-and-seek contest with the Brahma king; Cuṇḍa seeing that the Buddha's body is weak just before the parinirvāṇa; the Buddha's great compassion is based upon his body; the Buddha's funeral pyre was extinguished with milk because his body of birth was nourished on milk; the Buddha cannot receive dharma-offerings from others because his dharmakāya is already perfect; breaking the saṃgha is an ānantarya sin because it is an attack on the dharmakāya;174 exegesis of the two verbs of sight in the MPNS tradition of the Buddha's last display of his body in terms of "body of birth" and dharmakāya;175 two senses in which the Buddha is lokôttara explained in terms of "body of birth" and dharmakāya, and the life of the ordinary monk parsed in terms of "cultivation" of each of these Buddha-bodies; it is an ānantarya sin to harm the Buddha's body, even though he is not truly that body, because to harm the basis harms what is based upon it, viz. the dharmakāya.

These many differences between the two earlier texts and XZ are significant. We have to remember that we are contending with various complications of textual history, and arguments from silence are obviously dangerous when the early texts are partial. However, not all the differences between XZ and Buddhavarman can be explained by loss of part of Buddhavarman, for example. Four passages fall where Buddhavarman comments on the same parts of JñP as XZ. In addition, there are numerous differences in detail in the Appendix 1 sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu passage.180 We have also seen that some ideas are present in Saṃghabhūti and Buddhavarman but absent from XZ, which cannot be explained by the smaller scope or loss of parts of the earlier texts. While these discrepancies might reflect differences of recension behind the two translation, they could also result from ideas being added to the text between Buddharvaman and XZ. This possibility is supported by other evidence about the general nature and history of the texts.181 If this was the case between Buddhavarman and XZ, it could equally have been true before all of our Chinese translations. Thus, even ideas about Buddha- bodies shared by all three Vibh, may not necessarily date back early enough to predate relevant Mahāyāna ideas.

Sarvâstivāda body doctrine cannot be proven earlier than Mahāyāna body doctrine The received understanding holds that "Mainstream" Buddha-body doctrine, which is often taken to be represented by Sarvâstivāda doctrine, predates our earliest Mahāyāna sources, including Prajñāpāramitā.182 This paper shows that we have no evidence warranting this assumption. Here I will review the several reasons that we cannot be sure any of the Sarvâstivāda doctrine we have studied predates key Mahāyāna ideas. There seems to be a general tendency to presuppose that Mahāyāna texts were open to influence, while Mainstream texts were generally not. This may indeed have been the general tendency. But even the Āgamas contain numerous examples of apparent acceptance of apparently new ideas, such as dharmakāya, into contexts where they were previously unknown.183 Such cases make it impossible to assume in any specific case that Mainstream sources cannot be adopting innovative concepts from Mahāyāna sources.

Our starting point in dating any early Prajñāpāramitā ideas must be the Aṣṭasāhasrikā of Lokakṣema (道行般若經 T224; "Aṣṭa"), which serves as a terminus ad quem (Lokakṣema was active between 168 and 186 C.E.). It is often observed that the doctrines of this text must have already undergone some evolution. If, conservatively, the terminus ad quem for the composition of the underlying Indic text might be 125-150 C.E.,184 some

as "cultivation" of the "body of birth" and dharmakāya. In each case, the context is found in Buddhavarman: see n. 27, 31, 42, 55. 180 See n. 165.

181 The Pelliot fragment, Buddhavarman and XZ "correspond closely", but "some revealing differences" show that "both Chinese translations include additional material;" Willemen, Dessein and Cox 233-234. Consider, too, the size of Buddhavarman and XZ. Even before the loss of some text, Buddhvarman was only ever 100 juan, whereas XZ is 200. Willemen, Dessein and Cox think that texts of the Sarvâstivāda Abhidharma corpus in general were probably subject to ongoing interpolation and growth over a long period; 167. Frauwallner discusses aspects of the Dharmaskandha that seem to reflect such ongoing growth and modification; Studies in Abhidharma Literature 20-21. Willemen, Dessein and Cox give reasons that the Prajñapti śāstra probably underwent significant modification over a long history; 191-193: "Yamada Ryūjō contends that the Prajñapti śāstra, like other sūtra collections and Abhidharma texts, undoubtedly underwent expansion and restructuring over the centuries . . ." Willemen, Dessein and Cox also discuss categories apparently added to the Prakaraṇapāda between the translation of Guṇabhadra and Bodhiyaśas and that of Xuanzang; 312.

182 Arguments similar to the following apply for the relation between almost any extant "Schools" material and our earliest Mahāyāna sources. In relation to body doctrine, this is also particularly relevant with regard to the chronological relation between Mahāsāṃghika/Lokôttaravada and early Mahāyāna ideas. See Radich, "Somatics" Ch. 4.2, esp. §4.2.3. 183 See also, for example, the shift in the Chinese Āgamas towards a corporeal reading of sôpadhiśeṣaand nirupadhiśeṣanirvāṇadhātu, as detailed in Radich, "Somatics" Ch. 3.4.

184 See Egil Fronsdal, "The Dawn of the Bodhisattva Path: Studies in a Religious Ideal of Ancient Indian Buddhists with Particular Reference on the Earliest Extant Perfection of Wisdom Sutra" (PhD dissertation,

form of Prajñāpāramitā literature must date back to at least the late first century. Little can be said for certain beyond this about the Prajñāpāramitā,185 but it will suffice for present purposes. Central also for the study of early Mahāyāna Buddha-body doctrine is the Lokânuvartaṇā sūtra, which was also translated by Lokakṣema, sometime between 168 and 186 (內藏百寶經 T807).186 Even on the traditional dates, therefore, we have no grounds for confidence that any version of Vibh is earlier than early Mahāyāna. Vibh was traditionally held to date from around or after the time of the Kuṣāna King Kaniṣka,187 i.e. the second century C.E. at the earliest. Even this traditional date would not make it antedate Lokakṣema. It is also difficult to show on internal grounds that the Prajñāpāramitā is indeed, as is commonly claimed, reacting to Sarvâstivādin ideas, as has been shown by Onishi Yoshinori.188

A further problem is that the traditional date for Vibh is highly unreliable and cannot be followed. The modern scholarly consensus is that a relative chronology of the texts can be determined with some degree of reliability, but the absolute date of each text and the overall corpus cannot, beyond the very late terminus ad quem of Chinese translations.189 Even early Chinese translations collectively date between 383 and ca. 443, already very late for purposes of comparison with the earliest known Mahāyāna ideas about Buddhabodies. In addition, Vibh may have grown and changed over time. We therefore cannot be confident that any particular passage belonged to the text from its very inception, and was not added later on.

We have seen in this paper that very little of Vibh's doctrine of Buddha bodies is found in the earlier Sarvâstivāda Abhidharma, even in JñP. We thus have a situation in which the apparently open Vibh, still in the process of ongoing formation, evinces a sudden interest in the new topic of Buddha-bodies. Further, details of that view seem to be part of more general trends apparently close in time to the first translations of Vibh into Chinese: the development of a complex corporeal interpretation of sôpadhiśeṣa- and

Stanford University, 1998), 105. 185 Some scholars see Aṣṭa itself as the oldest Prajñāpāramitā text; others consider the Vajracchedikā ("Diamond") possibly the oldest. A range of dates have been proposed for the origins of this literature, from the second century B.C.E. to the second century C.E. See Hanayama Shōyū, "A Summary of Various Research on the Prajñāpāramitā Literature by Japanese Scholars," Acta Asiatica 10 (1966): 16-93;

Nakamura Hajime, "A Critical Survey of Mahāyāna and Esoteric Buddhism Chiefly Based upon Japanese Sources," Acta Asiatica 6 (1964): 57-88; Gregory Schopen, "The Phrase sa pṛthivīpradeśaś caityabhūto bhavet in the Vajracchedikā: Notes on the Cult of the Book in Mahāyāna," in Fragments and Figments of Mahāyāna Buddhism in India: More Collected Papers (Honolulu: University of Hawaiˋi Press, 2005), 31-32, 55 n. 16, 17; Edward Conze, The Prajñāpāramitā Literature (New Delhi: Munshiram Manoharlal, 2000), 1.

186 Paul Maxwell Harrison, "Sanskrit Fragments of a Lokottaravādin Tradition," in Indological and Buddhist Studies: Volume in Honour of Professor J. W. de Jong on his Sixtieth Birthday, ed. L. A. Hercus et al. (Canberra: Faculty of Asian Studies, 1982), 211-234. 187 See e.g. Paul Demiéville, "L'origine des sectes bouddhiques d'apres Paramartha," Mélanges chinois et bouddhiques 1 (1931-32): 24, fn. "b". Vibh mentions Kaniṣka: 昔健馱羅國迦膩色迦王, T1545:27.593a15; cited in Onishi Yoshinori, "Is the Aṣṭasāhasrikā Prajñāpāramitā Sūtra Really Arguing against the Sarvāstivādins?" Buddhist Studies Review 16, no. 2 (1999): 169. Of course, if it was possible for material to be added over time, other parts of the text might predate Kaniṣka, even if this sentence did not.

188 The only "meagre" signs of such a possibility are that it does criticise the idea of real existence in three times, and that it may owe something to Sarvâstivāda in its notion of merit. Onishi 179, 180. The relation is also complicated by Vibh as cited by Akanuma Chizen: "'Prajñā' here refers to what is called vaipulya" (此中般若說名方廣, T1545:27.660a29). On the basis of this passage, Akanuma argues that Prajñāpāramitā in fact precedes Vibh; cited in Hanayama, "Various Research" 55. 189 See Willemen, Dessein and Cox, 166-167.

nirupadhiśeṣanirvāṇadhātu; or the term "body of birth", very rare outside Vibh and MPPU. For these reasons, we cannot be sure that the ideas about Buddha bodies that are found in Sarvâstivāda literature predate Mahāyāna ideas on the same topic. Conclusions and Implications In this study, I have attempted to point out several misconceptions in received opinion about Sarvâstivāda Buddha-body doctrine and its place in the history of ideas about

Buddha-bodies more generally, and to sketch a more accurate picture to put in its place. First, it is true, in a sense, that Sarvâstivādins taught a "two-body" theory. However, our understanding of both bodies requires fundamental revision. In literature early enough to matter, the rūpakāya is not ascribed to the Buddha as one of his two special bodies. Rūpakāya is ascribed to the Buddha only very rarely in this literature, and far more frequently to ordinary sentient beings. Rather, the Sarvâstivāda (as represented by the Vibhāṣā) said that the Buddha had a "body of birth" or "body born of father and mother".

The Sarvâstivāda does indeed teach a dharmakāya, though only beginning with Vibh. However, previous scholars have tended to describe the Sarvâstivāda Abhidharma dharmakāya as either (a) the eighteen āveṇikadharmas (Guang); or (b) the aśaikṣadharmas, with emphasis on kṣayajñāna, anutpādajñāna and the five anāsravadharmas (la Vallée Poussin, Makransky). In fact, both of these characterisations are propounded in different places in Vibh. I argued further that these definitions are related, and both can be reduced by analysis to an underlying conception of the Buddha embodied in his gnosis (bodhi, jñāna), which has been overlooked by in previous scholarship due to excessive attention to the surface enumeration of categories.

Thus, even though Sarvâstivāda is often characterised as the locus classicus of a supposed Mainstream rūpakāya-dharmakāya model, Sarvâstivāda texts show that their model was different. However, it is certainly the case that a rūpakāya-dharmakāya model exists. Where, then, does it come from? Proper treatment of this problem is beyond the scope of this paper. However, I have already argued elsewhere that based on the evidence of the Chinese translation record, it seems that the opposition of rūpakāya to dharmakāya spread suddenly sometime before the late third century at the latest – but overwhelmingly in texts of Mahāyāna and not Mainstream provenance. Beyond the Chinese canon, the opposition between rūpakāya and dharmakāya may already be present in the works of Nāgārjuna.192 While this question certainly requires further research, these findings at least suggest that the rūpakāya-dharmakāya opposition may have originated in the very heart of the Mahāyāna itself. If this is true, our picture of a "Mainstream" rūpakāya-dharmakāya model is doubly wrong. These findings require a fundamental revision in our idea of the relation between

Sarvâstivāda and other ideas about the Buddha's bodies. If Sarvâstivāda does not teach a rūpakāya-dharmakāya model, it is impossible that such a model was originally derived from the Sarvâstivāda, and was passively accepted therefrom by the Mahāyāna. More than this, however, we have seen that it is highly unlikely that any Sarvâstivāda model of Buddha-bodies – even the actual "body of birth"-dharmakāya model – preexisted or influenced the Mahāyāna. Before Vibh, Sarvâstivāda texts show almost no inkling of the very idea that the Buddha has special bodies, nor any interest in the nature of those bodies. Vibh speculation about the Buddha's bodies thus represents a sudden flourishing of interest in the topic, not the transmission of an old heritage from the earlier roots of the school. Moreover, it can be shown even after Vibh, major Sarvâstivāda and other Abhidharma texts (*Abhidarmahṛdaya, *Saṃyukâbhidharmahṛdaya, *Satyasiddhi etc.) show relatively little interest in the Buddha's bodies (see Appendix 5).

Even on traditional dating, there is no reason to think Vibh predates early interest in the Buddha's body in the Mahāyāna, as reflected in Lokakṣema. In addition, there are reasons to think that at least the exposition of Buddha-body doctrine in Vibh continued to evolve not only until the time of the first two translations in 383 and 437, but even thereafter, between them and XZ. Ideas about Buddha-bodies in our extant versions of Vibh could thus be substantially later than early Mahāyāna ideas. This means that the direction of any influence could easily be the opposite to what scholars have usually supposed, i.e. from Mahāyāna to Sarvâstivāda.193

context of the idea of seeing the Buddha, T458:14.437a19-24. (3) Dharmakāya is first opposed to rūpakāya in Dharmarakṣa, where this opposition is suddenly very frequent: *Lokadharaparipṛcchāsūtra[?], T481:14.628b22-27; Vimaladattāparipṛcchāsūtra, T338:

12.92c20-29; Aśokadattavyākaraṇa, T337:12.86b28-c04; Du shi pin jing 度世品經 T292:10.617b13; Daśabhūmika T285:10.479a08-12, Johannes Rahder, ed., Daśabhūmikasūtra (Leuven: J.-B. Istas, 1926) 55, Honda Megumu, "Annotated Translation of the Daśabhūmika-sūtra," in Studies in South, East, and Central Asia: Presented as a Memorial Volume to the Late Professor Raghu Vira by Members of the Permanent International Altaicist Conference, ed. Denis Sinor (New Delhi: International Academy of Indian Culture, 1968), 200; Daśabhūmika T285:10.491a09-22, Rahder 86, Honda 264; Tathāgatôtpattisaṃbhavanirdeśa T291:10.612c25-613a11. Note the number of proto-Avataṃsaka texts here (Du shi pin jing, Daśabhūmika, Tathāgatôtpattisaṃbhavanirdeśa).

In my dissertation, I mistakenly stated that the first opposition between dharmakāya and rūpakāya was in Mokṣala's Pañcaviṃśatisāhasrikā prajñāpāramitā. This error was based upon a confusion of T221 (Mokṣala) and T223 (Kumārajīva). I am grateful to Jan Nattier for saving me from repeating this embarrassing mistake. 192 In Ratnâvalī: (1) 3.10; gang tshe sangs rgyas gzugs sku yi/ rgyu yang de ltar 'jig rten bzhin/ gzhal med de tshe chos sku yi/ rgyu lta ji ltar gzhal du yod, Michael Hahn, Nāgārjuna's Ratnāvalī: Vol. 1, The Basic Texts (Sanskrit, Tibetan, Chinese) (Bonn: Indica et Tibetica Verlag, 1982) I, 74; Jeffrey Hopkins and Lati Rimpoche with Anne Klein, The Precious Garland and the Song of the Four Mindfulnesses (London: George Allen and Unwin Ltd., 1975), 48. No Skt. extant. 諸佛色身因/尚如世無量/況佛法身因/而當有邊際, T1656:32.498a15-16. (2) 3.12, sangs rgyas rnams kyi gzugs sku ni/ bsod nams tshogs las byung ba ste/ chos kyi sku ni mdor bsdu na/ rgyal po ye shes tshogs las 'khrungs/, Hahn I, 74; Hopkins et al., Precious Garland 49. No Skt. extant. 諸佛有色身/皆從福行起/ 大王佛法身/由智慧行成, T1656:32.498a19-20. For other possibly early Indic references, see Radich, "Somatics" 999-1000. 193 Cf. also n. 30, 36, 56. Finally, the results of this study urge further re-investigation of the relation between all

Mahāyāna ideas of the Buddha's bodies and similar Mainstream ideas, especially regarding dharmakāya. Paul Harrison has influentially argued that the early Mahāyāna doctrine of dharmakāya should be read in line with a pluralistic "Mainstream" view (in which it means the Buddha is embodied in plural dharmas). To the extent that Sarvâstivāda represents the Mainstream, our results suggest that there may have been no "Mainstream" view for Mahāyāna texts to inherit and continue. In addition, I have argued that careful study of Vibh shows that its notion of dharmakāya hinges on the idea that the Buddha is embodied in his gnosis (bodhi, jñāna), expressed in various rubrics. On this analysis, the gap between Sarvâstivāda and Mahāyāna doctrines of dharmakāya is less than it appears from Harrison's emphasis on the plurality of "dharmas" in the supposed "Mainstream" dharmakāya. Thus, not only in chronology, but also in content, my findings are compatible with a general working hypothesis that the very idea of special Buddha bodies, and especially the naming of those bodies as various –kāya, may have first arisen in the Mahāyāna, and only then spread to Mainstream texts. Further exploration of that hypothesis must await future work.

In closing, we might also ask, if the Sarvâstivāda were indeed adopting and adapting Mahāyāna ideas, rather than the other way around, what patterns might be discerned in the way the Sarvâstivādins work to modify body discourse. At the present stage of my research, I can only offer some preliminary observations on this issue. However, it seems on the evidence surveyed here that if Sarvâstivāda were appropriating ideas of Mahāyāna origin, they were, on the one hand, careful to retain a common basis with the ideas they were reacting against, in the basic reading of dharmakāya to mean that the Buddha is embodied in his bodhi or jñāna. On the other hand, however, they seem at the same time to have been grounding these ideas more firmly in more traditional “dharmas”, and that in two ways. First, they attempted to ground these ideas in the authority of the Āgamas, for example in the notion of the "body of birth", which I have suggested is founded upon the Sāmaññaphala /Śrāmaṇyaphala. Second, they attempted to tie the notion of dharmakāya, somewhat general in Mahāyāna, to specific Abhidharmic rubrics of multiple, named buddhadharmas. This second step does have the effect of making the dharmakāya look more pluralistic on the surface, since Abhidharma itself is what we might call a “pluralist realism”; but we should not forget the underlying identity between Mahāyāna and Sarvâstivāda visions, in the basic notion of embodiment in gnosis. This relation between Sarvâstivāda concepts and their possible Mahāyāna origins is another area that calls for further investigation.

Appendix 1: Vibh on the "body of birth" and sôpadhiśeṣa/ nirupadhiśeṣanirvāṇadhātu The most detailed Vibh passage on the "body of birth",197 as discussed above,198 is part of an extended history in which sôpadhiśeṣa- and nirupadhiśeṣanirvāṇa came gradually to have corporeal meaning.199 This Vibh passage may be one of the earliest extended treatments of the "corporeal" reading of sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu. In light of this historical interest, I translate it here in full. Text from JñP (i.e. the root-text upon which Vibh is commenting) is given in bold. Passages to do with embodiment are underlined. "Paragraphs" are introduced and numbered for convenience of discussion: "[JñP:] It says in the sūtras, 'There are two kinds of nirvāṇadhātu, namely the nirvāṇadhātu with a remaining support, and the nirvāṇadhātu without a remaining support,' etc. etc.

<1> "[Vibh:] Why does [the root text] engage in this disputation? In order to trace out the detailed ramifications of the message of the sūtra 為廣分別契經義. . . . Although the sūtras say this, they do not elaborate on the distinction in detail 不廣辨 [to say] what is the nirvāṇadhātu with a remainder of support, and what is the nirvāṇadhātu without a remainder of support. This is the basis for the present disputation, that those points not distinguished in the sūtra should now be distinguished . . . .

197 T1545:27.167b14-168b20. The passage is found in Buddhavarman but not Saṃghabhūti; see above n. 156. 198 See above p. 132-133.

199 I argue this in detail in Radich, "Somatics" Ch. 3.4. Stereotypically, on the basis of expositions in texts at least as late as the Common Era, sôpadhiśeṣa- and nirupadhiśeṣanirvāṇa are taken to refer respectively to a moment of awakening during the aspirant's final lifetime, after which the aspirant continues to have a physical body (the "support"); and the final parinirvāṇa (physical death) at the end of that lifetime, at which the physical body "drops away", as it were (the awakened person is now "without support", i.e. no longer has a body). However, this reading is the result of a long development. In the Pāli canon, saupādisesa originally referred to liberation "with a vestige of clinging", pertaining to a category of people who were destined to be liberated, but who still had to undergo further rebirths before they finally realised that destiny. This category was opposed to those liberated "here in this visible

world", i.e. in the here and now (diṭṭhe dhammeva). Over time, saupādisesa came to be opposed to anupādisesa, "without a vestige of clinging". The opposition between the two came to refer to a twofold typology of different kinds of "Nibbāna element/realm" (nibbānadhātu). This development perhaps begins with Itivuttaka 44, which, though it is usually taken as the locus classicus for these concepts, is in fact highly anomalous in the Pāli canon in precisely this respect. In the Chinese translation record, the new interpretation of the two terms can be shown to coexist, alongside remnants of the older one, down at least to the turn of the fifth century, sometimes in the same texts. The new reading creeps in alongside the old in Āgamas, for example, where comparative textual evidence suggests it was not originally present. A further development occurred when saupādisesa- and anupādisesanibbānadhātu (Skt. sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu) came to refer to different bodily concomitants to enlightenment in the same person. The terms thus finally took on their more

familiar corporeal significance. Here, upadhi-śeṣa refers to the physical body of a person as a "remnant" of that which is clung to, viz. the upādānaskandhas. Upadhi itself was correspondingly reinterpreted to mean "basis, ground, substratum", reflected in the ("New") Chinese translation 新譯 of this element of the term as yi 依 "basis, ground, support" and not yu 餘 "excess, remainder". It might be in principle conceivable to read Pāli upādi "clinging" to refer to the object of clinging, but this was not the reading in the Pāli canon. The overall trajectory of these developments seems to suggest that corporeal readings of sôpadhiśeṣa- and nirupadhiśeṣanirvāṇa are unlikely to date back much before the Common Era. The history of these concepts thus forms part of a larger pattern I believe we can observe in our evidence, where the idea of buddhahood as being embodied in various special forms tout court also does not emerge much before the start of the Common Era.

"[JñP:] What, then, is the nirvāṇadhātu with a remainder of support? If an Arhat, a person in whom taints (諸漏, āsravas) are eternally exhausted, nonetheless still has life 壽命 , the 'series (相續 saṃtāna, saṃtati) of physical form constituted by the gross elements 大種造色 is not yet severed, and, based upon (*āśritya) that body of the five senses 依五根身 , the series of mind [still] functions, because there is a remaining support. This eternal cessation of the fetters (諸結 , saṃyojanāni), [but with continued] apprehension (得獲 , *pratyanubhāvanā) and experience (觸證 , *pratisaṃveda),202 is called 'the nirvāṇadhātu with a remainder of support'

(sôpadhiśeṣanirvāṇadhātu).203 <2> "[Vibh:] Here, 'life' refers to the 'faculty of life[-force]' (命根, jīvitêndriya).204 Why does the text not mention the mutual identity [of sentient beings] ([眾 ]同分 ,

• [[[sattva]]-]sabhāga[tā])?205 Because of simple authorial intent, etc.206 Alternatively, if the author should have mentioned it but did not, we should know that there is a further intent here. Further, then, he only speaks of [[[Wikipedia:faculty|faculty]] of life-force] out of [the two categories at issue] because the faculty of life[-force] and the mutual identity [of sentient beings] are both alike effects of the karma that determines existence per se (牽引業 , ākṣepakakarma),207 but of these, the faculty of life[-force] is entirely 一向 an effect of a

202 Cf. Itivuttaka 44, "the five faculties still persist, as a result of whose non-disappearance he receives that which is charming and otherwise, experiences pleasure and pain" (manāpāmanāpaṃ paccanubhoti sukhadukkhaṃ paṭisaṃvedeti), contrasted to a state in which "all that is sensed right here, not being rejoiced in, will become cold" (idheva sabbavedayitāni anabhinanditāni sīti bhavissanti); PTS 38, Peter Masefield, trans., The Itivuttaka (Oxford: Pali Text Society, 2001), 35.

203 These two sentences translated by la Vallée Poussin, AK IV, 211 n. 3, who, however, takes them from Vibh, and does not note that they go back to JñP. 204 Jīvitêndriya is a cittaviprayuktadharma (a dharma not associated with mind). Collett Cox translates "vitality", Disputed Dharmas: Early Buddhist Theories on Existence: An Annotated Translation of the Section on Factors Dissociated from Thought from Saṇghabhadra's Nyāyānusāra (Tokyo: The International Institute for Buddhist Studies, 1995), 289 ff. Note that this definition of sôpadhiśeṣanirvāṇa is echoed in the Viniścayasaṃgrahaṇī of YBh; see Sakuma Hidenori 佐久間秀範, "Yugashi-ji ron ni okeru tenne shisō 『瑜伽師地論』における転依思想," Indogaku bukkyōgaku kenkyū 39, no. 1 (1991): 436; Sakuma, Die Āśrayaparivṛtti-theorie in der Yogācārabhūmi, 2 vols., Alt- und Neu-Indische Studien 40 (Stuttgart: Franz Steiner Verlag, 1990) I, 168; Radich, "Somatics" n. 2528.

205 In the Sarvâstivāda Abhidharma, sabhāga[tā] is a cittaviprayuktadharma (a dharma not associated with mind) which accounts for the mutual identity (categorical stability) of sentient beings over multiple incarnations, such that they remain sentient beings always and do not become insentient things like stones. Collett Cox translates "the homogeneous character"; Disputed Dharmas 229 ff. Clearly, the commentator asks here why this dharma was not mentioned because it, in addition to the jīvitêndriya, is a condition of the ongoing possession of a living human body at issue in the category of sôpadhiśeṣa-. 206 是作論者意欲爾故乃至廣說. This appears to be an abbreviated way of stating a frequent comment on the particularities of the root text. In a fuller form this reads, "This is a matter of simple authorial intent. The author disputes this because he wishes to do so; [in so doing] he does not contravene the nature of the case (lit. 'the nature of [the] dharma[s]), and so what need is there for arguments in proof?" 彼作論者意欲爾故。隨彼意欲而作此論。不違法性何煩徵詰, T1545:27.5b21-22, c21-22, 23c25-27 etc. In other words, such a comment on a feature of the root text means that it is merely an inconsequential quirk of the root author's mode of expression.

207 Lit. "karma that throws, casts (ā/kṣip), lays down" etc., i.e. the consequential act (karma) that "projects" (VP) a sentient being into a specific existence characterised by certain general conditions, such as human existence, existence as a hell being, etc. ("lays down" the basic conditions of that existence), prior to any particular features of that existence (being this or that human, here or there, etc.). Opposed to manye 滿業 (prob. *paripūrakakarma, cf. AK 95b anekaṃ paripūrakam, "the act(s) that fill out [that destiny] are not singular"), which does lead to the particular conditions of the new existence. Vasubandhu compares these two types of karma respectively to a single line with which an artist delimits the border of a picture (ākṣepakakarma) and the details with which he fills in that border (*paripūrakakarma); AKBh to 4.95b, VP III, 199-200. Cox uses the excellent terms "skeletal" for ākṣepaka and "fleshing [out]" for paripūraka; Disputed Dharmas 110. These two types of karma are also called, in Ch. translations, 一業 "single act" (*ekakarma) and "various acts" 別業 (*anekakarma), e.g. Buddhavarman Vibh, T1546:28.82b16, cf. AK 4.95a-b, ekaṃ janmākṣipaty ekam anekaṃ paripūrakam, "a single [act] projects a single birth, but various [acts] fill it out", VP III, 198-199. Later Chinese commentators were to refer to the fruits of these two types of karma respectively as zongbao 總報 "general reward" and biebao 別報 "specific cause heterogeneous to itself (異熟 , vaipākika).208 It is on the basis of this remaining

physical form that mind etc. [i.e. and associated dharmas] continue to function.209

<3> Because what is referred to here as 'the gross elements' is the [[[Wikipedia:Absolute (philosophy)|ultimate]]] support (所依),210 it is mentioned first. 'Physical form comprised of the gross elements' arises

on that basis. On the basis of the physical form so comprised[, in turn], mind and dharmas associated with mind arise.211 [However,] because mind is the main focus, the text here speaks only in terms of the partial [designation] 'physical form comprised of

the four elements', in order to indicate the body of physical form (rūpakāya) in general,

and of 'the mental series based upon the body of the five senses' in order to indicate

mind and dharmas associated with mind; alongside these, there also still exist the [the other] dharmas dissociated from mind, such as birth (生 ), etc.,212 [but] because these

reward"; so esp. Tunnyun 遁倫 (d.u., fl. 8th century) T1828:42.621c02-05; also e.g. 371c05-13; Chengguan 澄觀 (738-839) T1736:36.465c22-24; Yijing/Huizhao, 下別釋緣起, T1788:39.328a20ff, etc. These latter terms show it is possible to think of these two kinds of karma as "generic karma" or "karma of species" and "specific karma".

208 Vipākaphala ~果 , "effect of a heterogeneous cause", is opposed to niṣyandaphala 異熟果 , a "homogeneous effect" or an effect of a cause homogeneous to itself. Elsewhere in both JñP (T1544:26.920c27-921a10) and Vibh, vaipākika is defined (from the side of cause, hetu) in terms immediately relevant for the present passage. The first portion of that passage runs:

"[JñP:] 'What is heterogeneous cause? It refers to what happens when mind and dharmas associated with mind (citta, caitāsikadharma) undergo (受) [the effects of] physical form (rūpa), mind, dharmas associated with mind, or dharmas dissociated from mind (cittaviprayuktadharmas) heterogeneous to themselves, such that the mind or dharmas associated with mind are effects heterogeneous to those

[[[rūpa]], citta, caitāsika, cittaviprayukta], [and takes them] for their heterogeneous causes.' "[Vibh:] Here, 'mind and dharmas associated with mind' refers to all bad and good (不善善, i.e. excluding the category of the karmically neutral) states of mind and dharmas associated with mind; this term also includes the rūpa that comes into being in association with them (彼隨轉色 ,

• tadanuvartakarūpa) and the [[[citta]]-]viprayukta, because they constitute a single effect along with mind and dharmas associated with mind. Thus, visible form as an effect of a cause heterogeneous to itself (異熟色, *vipākarūpa) refers to the eye and the other five organs (pañcêndriya), [and their objects, viz.] visible form, [[[sound]],] scent, taste and tangibles; 'mind' refers to the aggregate of consciousness (vijñānaskandha), i.e. to the six consciousness of eye, etc.; [and] 'dharmas associated with mind' refers to [the other] three aggregates (skandhas), i.e. sensation (vedanā), conception (saṃjñā) and 思 (?, sic), etc. [sic; by my count we run out of skandhas!]. '[[[Dharmas]]] dissociated from mind' (cittaviprayukta) refers to the aggregate of karmic intention (saṃskāraskandha), i.e. the faculty of life[-force] (jīvitêndriya), the mutual identity of sentient beings (sattvasabhāga), etc. Thus, it is apparent that heterogeneous cause and heterogeneous effect (vipākahetu, ~phala) are both coterminous with the five aggregates (skandhas);" T1545:27.96b12-21.

The implication for our present passage seems to be that jīvitêndriya alone is purely vaipākika, even though, as we see, vaipākika karmic causation is a very broad category, by means of which all five skandhas are brought into existence. Presumably, therefore, there is something special about vaipākikacausation that makes it the real reason that the person of the Arhat, comprising all five skandhas, continues to exist. Bearing in mind the fact that the Arhat is one in whom, by definition, the "outflows" (āsravas) have ceased, might we suppose that it is only in virtue of heterogeneous causation that the five skandhas, understood as impure by definition, can continue to exist for such a person?

209 This entire paragraph is absent in Buddhavarman. 210 This reading of upadhi as "support", meaning the body of the five skandhas, begins here to dovetail with the meaning of āśraya, also translated yi 依 . This equation was eventually worked out in the later theory of āśrayaparāvṛtti. On āśrayaparāvṛtti as a doctrine of (re-)embodiment, see Radich, "Somatics" 5.2.

211 This definition of the "support" and what is "supported by it" (所依, *能依) recalls the definition of āśraya and āśrita at AKBh to 3.41, Swami Dwarikadas Shastri, Abhidharmakośa & Bhāṣya of Acharya Vasubandhu with Sphutārthā Commentary of AcharyaYaśomitra (Varanasi: Bauddha Bharati, 1970), 496: āśrayo 'ti sêndriyaḥ kāyaḥ. tasya puṣṭaye kavaḍīkārāhārāḥ. āśritāś cittacaittās teṣāṃ puṣṭaye sparśaḥ; VP AKBh II, 126. It thus affords a link between the notion of "remainder" (upadhi, 餘依) and "support" (依), and thereby perhaps between sôpadhiśeṣa- and nirupadhiśeṣanirvāṇadhātu and āśrayaparāvṛtti. See further Radich, "Somatics" Ch. 5.2.

212 "Birth" is not among the list of cittaviprayuktadharmas in Saṃghabhadra's list, as studied by Cox. It appears that Vibh assumes a different set of cittaviprayuktadharmas, viz., in addition to jīvitêndriya and sattvasabhāga, "possession" (prāpti), birth, old age, subdurance, and impermanence (得生老住無常); see points are abstruse (難了知 ), and because they are included in the dharmas already previously listed, they are not indicated or mentioned separately. Thus, the eternal cessation of the fetters attained by one in whom the series of these various dharmas have not ceased is called 'the nirvāṇadhātu with a remainder of support'.213

<4> "There are also those who propound the following interpretation: 'Material form comprised of the gross elements' is the body, 'the five sense-organs' are the senseorgans, and 'the series of mind' (cittasaṃtāna) is 'mind' (覺 , *bodhi). The eternal cessation of the fetters where these series of body, organs and mind have not been severed214 is called 'the nirvāṇadhātu with a remainder of support'. As it says in the sūtras, '"Nirvāṇa with a remainder of support" refers to [a state in which] body, sense-

organs, and mind have not yet ceased.' <5> "Now, [JñP] says, '[[[mind]] still functions] because there is a remainder of support.' There are two kinds of 'support':215 (1) the support comprised by the kleśas

(*kleśôpadhi); and (2) the support comprised by the body of birth (*sambhavikakāyôpadhi).216 Thus, although the Arhat in the case discussed here is without the base comprised by the kleśas, he nonetheless does [still] have the base comprised by the body of birth. Or again, support is of two other kinds: (1) the defiled

support (染污依), and (2) the undefiled support. The Arhat in the case discussed here is without the defiled support, but nonetheless does still have the undefiled support.217 Therefore, the eternal cessation of the fetters attained in such a case is called 'the nirvāṇadhātu with a remainder of support'.

<6> "[As for] 'apprehension (得獲 , *pratyanubhāvanā) and experience (觸證 ,

• pratisaṃveda)': although these words differ, they both indicate the same meaning.218

"[JñP:] What, then, is the nirvāṇadhātu without a remainder of support? It is [the state of] the Arhat, in whom taints (諸漏, āsravas) are eternally exhausted, [when] life is extinguished, and the series of physical form constituted by the gross elements has been severed, such that the mind does not function any longer on the basis of the body of the five senses, because there is no remaining support. This eternal cessation of the fetters is called 'the nirvāṇadhātu with a remainder of support' (nirupadhiśeṣanirvāṇadhātu).

<7> "[Vibh:] Here, 'life is extinguished' indicates that the faculty of life[-force] (jīvitêndriya) and the mutual identity of sentient beings (sattvasabhāgatā) are extinguished. These are both effects of the karma that determines existence per se (ākṣepakakarma);219 although the text here happens to bring up explicitly only the faculty of life-force, it should be understood that it also refers to the mutual identity of sentient beings.220 <8> 'The series of physical form constituted by the gross elements has been

T1545:27.a28-b01, 100b17, 100c03-04. 213 This paragraph is paralleled in Buddhavarman by a much simpler discussion, the gist of which is basically the same. "Here, 'the four gross [[[elements]]]' means the four gross elements. 'The sense-organs' means the material form (rūpa) comprised [by those elements]. 'The series of mind' means mind and dharmas associated with mind. This is [what is meant by] 'the nirvāṇadhātu with a vestigial body';" T1546:28.126a23-25. 214 Following an obvious repunctuation of the Taishō thus (my punctuation in square brackets, original Taishō in round brackets): 大種造色是身[。]五根(。)是根[。]心相續是覺[。]此身根覺相續未斷 . . . . 215 In keeping with its translation in terms of nirvāṇadhātu with/out a vestigial body, Buddhavarman here reads, "There are two kinds of body" etc., T1546:28.126a28-29.

216 Buddhavarman here has only "the body of birth". 217 Basically the same in Buddhavarman; T1546:28.126a28-b02. 218 This paragraph is absent in Buddhavarman. 219 See above n. 207. 220 This paragraph is absent in Buddhavarman.

severed' indicates that the series of the rūpakāya in general has been severed. 'The mind does not function any longer on the basis of the body of the five senses' indicates that the mind and dharmas associated with mind do not continue [in series] any longer. The significance of the fact that the text does not speak of birth, etc. is as above. "There are also those who propound the following interpretation: 'Physical matter comprised of the gross elements' indicates the body; 'the body of the five senseorgans' indicates the sense-organs; 'the series of mind' indicates mind (覺, *bodhi). <9> "A case where (a) (i) the rūpakāya, (ii) mind and (iii) dharmas associated with mind, or (或?) (b) the series of (i) body, (ii) sense-organs and (iii) and mind thus have been severed, [so that] the fetters cease eternally, is called 'the nirvāṇadhātu without a remaining support'. This means that when the Arhat is about to [enter] parinirvāṇa, [a] 'wind' arises in the body, and throws it out of kilter (不調適); because it is out of kilter,

the inner fire grows weak; because the inner fire is weak, food is not digested; because food is not digested, he has no appetite; because he has no appetite, he no longer eats or drinks; because he does not eat or drink, the gross elements deteriorate; because the gross [[[elements]]] wane, the sense-organs of physical form comprised [of the elements] also deteriorate with them; because the sense-organs deteriorate, mind and dharmas associated with mind, having no support, do not continue [in series, 相續 ] any longer; because mind and dharmas associated with mind do not continue [in series], the faculty of life[-force], etc. is severed; and because the faculty of life[-force] etc. is severed, it is called 'entering nirvāṇa'. <10> "When the text says 'because there is no remaining support', [it means] that two kinds of support do not exist: (1) there is no support comprised of kleśas; (2) there is no support comprised by the body of birth. Or, also, that (1) there is no defiled support; and

(2) there is no undefiled support. Because there is no remaining support [in these senses], the eternal cessation of the fetters is called 'the nirvāṇadhātu with no remainder of support'.228   <11> "Why do the words 'apprehension and experience' not occur here? Because it is only on the basis of present apprehension that we can speak of 'apprehension' etc. But [in this case] present apprehension is severed, and so we do not speak of it. Also, because it is [only as] a conventional designation (施設, *prajñapti) based upon the person (補特伽羅, pudgala) that [there is] 'apprehension and experience'. But in this case, there is no pudgala, only the true nature of dharmas (dharmatā); for this reason, [also,] we do not

speak [of 'apprehension and experience']. <12> "Is it really true (頗 . . . . 耶 ) that there are Arhats who dwell neither in the nirvāṇadhātu with a remainder of support nor the nirvāṇadhātu without a remainder of support? Although there are in truth (理) no such cases, there would indeed be such cases according to what is said here. For according to what is said here, each of the three criteria must be met in order for us to speak of the nirvāṇadhātu with a remainder of support; and only where all three criteria are absent can we speak of the nirvāṇadhātu without a remainder of support. [Thus,] an Arhat born into the formless realm (無色界, ārūpyadhātu) has an immaterial body (無色身, *ārūpyakāya?, alternatively perhaps 'does

not have a rūpakāya'), and therefore does not dwell in the nirvāṇadhātu with a remainder of support; [and yet] his mind does function, and therefore he does not dwell in the nirvāṇadhātu without a remainder of support. [Again,] if an Arhat born into the rūpadhātu enters into the 'absorption of complete cessation' (滅盡定, nirodhasamāpatti), he does not dwell in the nirvāṇadhātu with a remainder of support, because his mind has ceased to function, but neither does he dwell in the nirvāṇadhātu without a remainder of support, since he still has a rūpakāya. Finally, if an Arhat is born into the kāmadhātu with

incomplete sense-organs (不具根), then, because he does not have all five sense-organs, he does not dwell in the nirvāṇadhātu with a remainder of support, and yet, because he still has a rūpakāya, he does not dwell in the nirvāṇadhātu without a remainder of

support.

<13> "Some say, therefore, that the text here ought to say, 'What is the nirvāṇadhātu with a remainder of support? It refers to a case where an Arhat still has life, [where] all the fetters have eternally ceased, [and where there is] apprehension and experience. What is the nirvāṇadhātu without a remainder of support? It is the eternal cessation of the fetters in an Arhat whose life is extinguished.' If we were to adopt this interpretation, then Arhats of the three realms (kāma-, rūpa- and ārūpyadhātu), regardless of whether they have bodies of material form (rūpakāya) or they have an immaterial body; regardless of whether they have mental function, or do not have mental function; regardless of whether they have all five organs, or whether they do not have all five organs – just so long as they have life, they would all be said to dwell in the nirvāṇadhātu with a remainder of support; and when their life is extinguished, they would all be said to dwell in the nirvāṇadhātu without a remainder of support. That the author should propose this interpretation, and nevertheless does not do so, is because the author of the original treatise wished to favour his throng of disciples and make it easy for them to understand, for which reason he [instead] advanced the interpretation we see here."

The Chinese canon down to 400 C.E. contains a surprising paucity of passages reflecting the corporeal interpretation of sôpadhiśeṣa and nirupadhiśeṣa. If we accept the traditional date of Vibh to the Kaniṣka era, therefore, the extended interest in that interpretation here is truly exceptional. We see here not just clear evidence of the corporeal interpretation per se, but detailed exposition of its precise workings (paras <2>, <3>, <5>), including a detailed explanation of the precise process of transition from sôpadhiśeṣa~ to nirupadhiśeṣa~, i.e. what it means to "die" when that death constitutes entry into parinirvāṇa; and also accounts of opposing views on the same doctrine (para. <4>, revisited in para. <8>).

The dicussion is further complicated by another part of this passage, omitted above to focus attention on the complex relation between sôpadhiśeṣanirvāṇadhātu and the "body of birth"/ rūpakāya. Here Vibh says JñP engages this topic of dispute to refute a range of incorrect interpretations of the sôpadhiśeṣa-/nirupadhiśeṣa- distinction. The range of opinion cited is simply astonishing:

"[The root text] also engages in this disputation to counter other schools, and demonstrate the correct interpretation. Thus, there are those who insist that the nirvāṇadhātu with a remainder of support is an essential entity (lit. "has own-nature", 自性 ), whereas the nirvāṇadhātu without a remainder of support is not an essential entity; [the root text engages in this disputation] to refute this dogma, and to demonstrate that both kinds of nirvāṇadhātu alike are not essential entities. Again, there are those who insist that the nirvāṇadhātu with a remainder of support is tainted (sâsrava 有漏, lit. 'has outflow'), whereas the nirvāṇadhātu without a remainder of support is taintless (anāsrava 無漏, 'is without outflow'); [the root text engages in this disputation] in order to refute this dogma, and to demonstrate that both kinds of nirvāṇadhātu alike are taintless. There are others again who insist that the nirvāṇadhātu with a remainder of support is conditioned (有為 saṃskṛta), and the nirvāṇadhātu without a remainder of support is unconditioned (無為, asaṃskṛta);238 [the root text

engages in this disputation] to refute this dogma, and to demonstrate that both kinds of nirvāṇadhātu alike are unconditioned. There are others again who insist that the nirvāṇadhātu with a remainder of support is (karmically) good (善, kuśala), and the nirvāṇadhātu without a remainder of support is (karmically) neutral (無記, avyākṛta); [the root text engages in this disputation] to refute this dogma, and to demonstrate that both kinds of nirvāṇadhātu alike are good. Or, again, there are others who insist that the nirvāṇadhātu with a remainder of support is of the path (道, mārga) but is not the fruit of the path (道果 , mārgaphala), while the nirvāṇadhātu without a remainder of support is the fruit of the path but not of the path; [the root text engages in this disputation] to refute this dogma, and to demonstrate that both kinds of nirvāṇadhātu alike are the fruit of the path. Still others insist that the nirvāṇadhātu with a remainder of support is included in the [four noble] truths (諦攝), while the nirvāṇadhātu without a remainder of support is not included in the truths; [the root text engages in this disputation] to refute this dogma, and to demonstrate that both kinds of nirvāṇadhātu alike are included in the truths. Still others

insist that the nirvāṇadhātu with a remainder of support is of the stage beyond training (無學 , aśaikṣa), whereas the nirvāṇadhātu without a remainder of support is neither of the stage of training nor of the stage beyond training (非學非無學); [the root text engages in this disputation] to refute this dogma, and to demonstrate that that both kinds of nirvāṇadhātu alike are neither of the stage of training nor of the stage beyond training. It is for the various reasons laid out here that the root text engages in this disputation." The bulk of this passage is already in Buddhavarman, with significant exceptions: "with remainder of support" is translated "with a vestigial body"; the term rūpakāya itself is absent; technical discussion of "life-force" (jīvitêndriya) and "mutual identify [of sentient beings]" (sattvasabhāgatā) is absent, etc. This means that most of the complexity and detail cannot date much later than the turn of the fifth century C.E. However, if we wish to adhere to the much earlier traditional dating of Vibh for this material, we must be ready to accept the implication that Vibh (and the JñP passage upon which it comments) is a

single large exception to the overall pattern of evidence about the gradual growth of the corporeal interpretation of sa- and nirupadhiśeṣa. The alternative, of course, is to consider the possibility that the Chinese translations contain material that developed later than the traditional dates. This possibility would be supported by the discrepancies between Buddhavarman and XZ, as discussed above. Appendix 2: *Nairmāṇikakāya is not a body of the Buddha in Vibh Louis de la Vallée Poussin claimed that Sarvâstivāda also recognises a *nirmāṇakāya (huashen, 化身). He seems to have based this claim entirely upon a single passage in Vibh, in which the text asks why the Buddha had not left behind such a body when he entered into parinirvāṇa, so that he might continue to teach the Dharma thereafter to the benefit of sentient beings. This claim might indeed seem to bear a close resemblance to the notion of the nairmāṇikakāya known from elsewhere. However, closer examination of other passages using the same term show that this resemblance is largely coincidental. The term huashen features

in the text in both nominal uses ("body produced by magic") and verbal ("to produce a body by magic"). This body is not special to the Buddha. Further, even where such a body is attributed to the Buddha (even hypothetically), it does not match the "classic" understanding of nairmāṇikakāya. It seems, therefore, that la Vallée Poussin's identification of a supposed Sarvâstivāda *nirmāṇakāya is part of an overall tendency to fit all materials to the Procrustean bed of a supposedly perennial threebody theory of the Buddha. Immediately after la Vallée Poussin's passage, the text contemplates the possibility that a Buddha might elaborate a *nirmāṇabuddha (化佛 , referred to immediately thereafter as

• nirmāṇakāya 化身), replete with all the marks, and make the magical body sit silent while he talks, and vice versa. The Sarvâstivāda doctors then consider in bizarre detail the mental multitasking this extraordinary feat would entail. Elsewhere, as we saw, the Buddha assures Ānanda that he can travel to the brahmāloka even in his "body of the mundane elements born of father and mother". Two speculative explanations of this power invoke the concept of nirmāṇakāya.247 The concept is also invoked in one of the theories to explain how the Buddha won his contest of hide-and-seek with Mahābrahman, though here the term is used in a verbal sense: "[The Buddha] transformed his body (化身) so that it could not be seen."248 As in la Vallée Poussin's example, although the text here contemplates the possibility of the Buddha (among others) himself elaborating a nirmāṇakāya, the emphasis is on hypothetical or counterfactual situations. Thus, the only situations in which the text predicates such bodies of the Buddha are in speculative, hypothetical doctrinal debate.

Elsewhere in the text, nirmāṇakāya are ascribed to beings other than the Buddha. The text discusses whether it is possible for a magical voice to speak apart from a magically elaborated body (nirmāṇakāya).249 The same term is also used to label the copy of the kāmadhātu body that an accomplished meditator can produce and project into the rūpadhātu.250 Nirmāṇakāya is also the term used for the bodies magically produced by Māra's three daughters when they tempted the bodhisattva under the bodhi tree.251 It is also used for the Buddha body magically produced by Māra for the benefit of Upagupta.252

In sum, the text does indeed recognise the possibility of the Buddha elaborating a nirmāṇakāya, but only because many spiritually powerful beings, including mahāśrāvakas, Māra and his daughters, and accomplished meditators can elaborate a nirmāṇakāya at will. Vibh never ascribes a nirmāṇakāya to the Buddha as part of a regular schema of embodiment. Indeed, where the Buddha is indeed ascribed with such a body, the text is always speculating or entertaining hypotheses. We are far from the conceptual world in which the Buddha regularly elaborates nirmāṇakāya that (appear to) perdure for the duration of an entire lifetime; where such bodies are elaborated primarily for the purpose of teaching sentient beings; and where his final earthly body is interpreted as such a nirmāṇakāya. Appendix 3: Dharmakāya passages in earlier Vibh translations not found in XZ

As I mentioned above (p. 144), Saṃghabhūti's Vibh contains an interesting unique passage unusually, among the materials surveyed here, implying the equation of the dharmakāya with the teaching, and even with the linguistic form of the texts (without obvious prejudice as to whether those texts are oral or written). The passage reads:253

"The scriptures of the Buddha speak of 'the eighty-thousand-fold dharmakāya'. What then is the enumeration that delimits the extent of this dharmakāya? One theory holds that each scripture (jing, sūtra) counts as one dharmakāya, and this is what is meant by an enumeration that delimits a single dharmakāya [i.e., this is the definition of 'dharmakāya' as a unit of measure]; in this manner, one arrives at eighty thousand in all. There is another theory that holds that [one?] foundation of mindfulness (意止, smṛtyupasthāna) counts as one dharmakāya, and so, also, for [each of] the [four] [correct] endeavours (prahāṇa), the [four] supernatural powers (ṛddhipāda), the [five] faculties (indriya), the

Abhidharma recognises a nirmāṇakāya of the Buddha, the text nowhere uses the term nirmāṇakāya (or

• nirmāṇabuddha). The current passage can be traced back at least to the Prajñapti śāstra,

T1538:26.526a08-20; speaking however of a "magical person" 化人, not a *nirmāṇakāya.

247 T1545:27.699b10-12. 248 化身令不顯現, T1545:27.66a06-13. 249 T1545:27.640b04-13. 250 T1545:27.696b24-697a06. 251 T1545:27.697c07-15. 252 T1545:27.698a09-22.

253 For this passage, see Guang Xing, Concept of the Buddha 36. My translation differs from Guang's on many points. [five] powers (bala), the [seven] members of enlightenment (bodhyaṇga) and the [eight members of the eightfold] path – each of these, too, counts as one dharmakāya, and thus (?) in all it comes to eighty thousand. "But the reckoners (算者) say: Eight syllables (字) make a 'foot' (句, certainly for pada/ pāda); thirty-two syllables make a śloka (首盧), and thus Five hundred thousand, and five thousand again – Five hundred and five thousand

Is the measure of one dharmakāya."

This notion of the dharmakāya as the teaching is significant primarily because it is so unique in the Sarvâstivāda corpus. Immediately following, the text explains two different theories for why the word kāya (身) is applied to the dharmakāya so defined. These explanations are also noteworthy, but for other reasons. In each case, it is because it is possible to subsume the dharma so labelled to a rubric of the five skandhas. (1) The Buddha's word (佛語 , *buddhavācana), considered from the perspective of its nature as a verbal teaching (教性, *deśanatvāt?) falls under the rūpaskandha, whereas considered from the perspective of its nature as ideas (名性, *nāmatvāt?) it falls under the saṃskāraskandha. (2) For the first time, we see an understanding of the dharmakāya as the five pure aggregates (anāsrava/aśaikṣaskandha ), which are listed and subsumed to the ordinary five skandhas, thus justifying the use of the term dharmakāya.

Appendix 4: The "body born of father and mother" in Kumārajīva, esp. MPPU The term "body born of father and mother" (var. 父母生身, 父母所生身, 父母所生其身, 父母所生之身) is relatively unusual in the Chinese canon. Apart from Vibh, the only text in which it 

appears more than once is Kumārajīva's *Mahāprajñāpāramitôpadeśa (MPPU). MPPU is thus the earliest translation in which the term appears.265 The term also occurs there seven times,266 more often than in any other text. In Vibh, as we have seen, the "body born of father and mother" occurs in opposition to the dharmakāya, as part of a broadly "two-body" theory. In MPPU it is part of an apparent nascent theory of more than two bodies, very different to "classic" Yogācāra trikāya theory. The "body born of father and mother"/"body of birth" is sometimes opposed to dharma[tā]kāya, and sometimes to an apparent prototype of the nairmāṇikakāya.267 The following two key MPPU passages define the "body born of father and mother" in contrast to each of these other bodies in turn:

"Further, the Buddha has two bodies: (1) the body of dharmatā (法性身, *dharmatākāya); and (2) the body born of father and mother 父母生身. The *dharmatākāya fills the space of the ten directions; it is infinite and boundless; it is handsome in visible form, and ornamented by the major and minor marks; [it is characterised by] infinite light and infinite sound,268 and the throngs of sentient beings that listen to [its] Dharma also fill space.269 It constantly emits all kinds of bodies, all kinds of words and names, all kinds of rebirth destinies, and all kinds of expedient means (upāya) in order to save sentient beings; it constantly saves all [[[beings]]], without a moment's rest. In this manner, the Buddha in his *dharmatākāya 法性身佛 saves sentient beings in the worlds of the ten directions. It is the Buddha in his body of birth 生身佛 [, on the other hand,] who undergoes the retribution for sins. The Buddha in his body of birth teaches the Dharma in sequence, as if it were a human dharma. Because there are these two kinds of Buddha, there is nothing wrong with the fact that the Buddha undergoes retribution for sins."270

"It is fitting to ask these questions271 of a man, but it is not fitting to ask them even of the gods; how much the more so of a Buddha? [Reply:] The Buddha has two kinds of body: (1) the body of transformation 變化身272 [created by] his supernormal powers (神通, abhijñā); (2) the body born of father and mother 父母生身. The body born of father and mother assumes the conditions of human [[[existence]]] 受人法, and so it is not like [that of] a god. For this reason, one should inquire of the Buddha in accordance with human custom."273

(523-600), T310:11.613b03-11; in Yijing義淨 (635-713), T1448:24.38b08-10; Bodhiruci (fl. 693-727), T310:11.323a01-08, 683c18-28, T761:17.631c02-08; in Prajñāruci (f. 516-543), T354:12.232b12-15; Devākara (fl. 676-688) repeats Jñānagupta verbatim T347:12.182a15-22; in *Pāramiti? 般剌蜜諦 (fl. 705) T945:19.119b03-10, 137c25-28; in Amoghavajra (705-774), T961:19.333c17, T1000:19.594b20, T1133:20.575b20-23 (and in parallels), T1665:32.574c21-22; in Śubhākarasiṃha (637-735), T906:18.913b09-10, 914a17-18.

265 MPPU is earlier than either Buddhavarman or XZ Vibh, which contain the term; and earlier than Buddhabhadra, who uses the term once. 266 T1509:25.90a17-20, esp. 121c26-122a05, esp. also 131c06-08, 620a27-b06, 624c15-18; in Buddhabhadra, T643:15.692c17-693a01. 267 This MPPU doctrine is of great interest, especially considering its date. To my knowledge, it has not been considered in its possible bearing upon the genesis of trikāya theory in Yogācāra. 268 These qualities, here ascribed to the dharma[tā]kāya, might broadly be described under the rubric of the "body of marks". Secondary literature to date would thus normally ascribe them to the *rūpakāya. 269 An interlinear note here reads "These sentient beings are also *dharmatākāya, and cannot be seen by people in saṃsāra."

270 T1509:25.121c26-122a05, Lamotte, Traité 513. 271 I.e. Ratnākara's questions to the Buddha about his health and welfare, which might imply the Buddha's vulnerability and imperfection. 272 Lamotte here reconstructs *nirmāṇakāya, which seems apt enough, but uncertain. 273 T1509:25.131c05-08, Lamotte, Traité 585. The opposition between the "body born of father and mother" and the body of transformation also features in the following question, but not its answer: 問曰。乃至不生一念瞋心者。為是變化身。為是父母生身。若是變化身則不足為奇。若是父母生身未斷結人。 Thus, in the earliest Chinese text that use "body born of father and mother", it is apparently part of an emergent theory more complex than "two-body" theory. MPPU also uses rūpakāya more than any of the Vibhāṣā texts. MPPU thus seems to be a veritable hotbed of doctrinal ferment regarding Buddha bodies, and promises to reward further study.

MPPU is thought to represent a hybrid of Sarvâstivāda scholasticism and Prajñāpāramitā doctrine. If it is indeed connected to Sarvâstivāda , this obviously opens up the question of the relation between its ideas and those of Vibh. Does the fact it speaks so much of the *mātṛpitṛkasaṃbhavakāya mean it orginated the idea, and did it then find its way into a stillaccreting Vibh? Or might Vibh rather have been one of its sources for the doctrine? Regardless of the direction of borrowing, or whether both got the concept from an unknown third source, we should note that the "body born of father and mother", though otherwise rare, is tightly concentrated in Vibh and MPPU. It may not be coincidence that these texts appear in the Chinese record within three decades of one another. Thereafter, the fad for this term seems rapidly to fade, making it something of a flash in the pan. This may indicate that at least Vibh passages containing *mātṛpitṛkasaṃbhavakāya date quite close to MPPU. Perhaps, then, Vibh was open to new ideas quite late in its history.

Apart from Vibh and MPPU, another early text in which a closely related concept appears is, very interestingly, the Mahāyāna Mahāparinirvāṇa sūtra (MPNS). The context is an explanation of the formulae "originally did exist, originally did not exist" etc. (本有 , 本無 ), themselves interesting for the evidence they provide of possible links between MPPU and parts of MPNS.274 MPNS reads: "We might say 'originally did exist'. [For instance,] I originally had a body [born

from] the union of father and mother [i.e. such a body did originally exist]. But on these grounds

[i.e. were this true], I would not now have a subtle dharma body of adamant. . . ;"275 The text adds a little further on: "We might say, 'originally did not exist': [for instance,] that I originally had a

body sustained by various kinds of food, but on these grounds [i.e. where the infinite body did no

originally exist], I would now not have an infinite body . . ."276 Given the rarity of the “body born of father and mother” in Chinese texts of this period, this evidence, in conjunction with the notion of benyou/benwu, may be one more piece of evidence linking MPNS to MPPU.277 The term “body of birth” also appears as part of a threefold buddhânusmṛti meditation recommended in the “Essentials of the Manual on Five Methods of Meditation” 五門禪經要用法 T619, a text ascribed to Dharmamitra (trans. c. 441 C.E.). This threefold meditation, which has been studied by Yamabe, instructs the practitioner to visualise in turn 1) a statue of the Buddha (in order to “take” the image firmly, i.e. become familiar enough with it to be able to go away from it to a quiet place and visualise a perfect eidetic copy of it without distraction); 2) the “body of birth”,278 identified with the thirty-two major and eighty minor marks; and then 3) the

云何能不生一念瞋心, T1509:25.624c15-18.

274 See Takeda Kohgaku 武田浩学, "Daichido ron wa 'Hon'u konmu' ge wo ikanaru kū rikai no reveru de ronjiteiru no ka: yonjikkanbon Dainehan gyō seiritsuron to no kanren ni okeru ichikōsatsu 『大智度論』は「本有今無」偈を如何なる空理解のレヴェルで論じているのか：四十巻本『大涅槃経』成立論との関連における一考察," Journal of the International College for Postgraduate Buddhist Studies 4 (2001): 203-261. 275 言本有者。我本有父母和合之身。是故現在無有金剛微妙法身。言本無者。我身本無三十二相八十種好。以本無有三十二相八十種好故。現在具有四百四病 T375:12.707b07-13; Yamamoto Kosho, trans.,

The Mahayana Mahaparinirvana-sutra (Ube: Karinbunko, 1973-1975), II, 413. 276 言本有者。我昔本有雜食之身。以食身故現在無有無邊之身, 707b27-28, Yamamoto II, 413. 277 Note also that the Mahāmegha sūtra,like MPNS translated by Dharmakṣema and closely related to MPNS for many other reasons, also opposes the “birth body” to the dharmakāya identified as vajrakāya: the Buddha is asked how aspirants can attain the "adamant dharma body" (金剛法身 *vajradharmakāya); the same list of questions also asks about the "true birth body and true dharma body" 真實生身, 真實法身 of the Tathāgata, his *vajrakāya 金剛之身 , and his "destructible gross body" 破壞雜身 ; T387:12.1081a19-20, 1081a22-25.

278 Yamabe incorrectly translates this (admittedly obscure) term “living body”. dharmakāya, identified with the eighteen āveṇikadharmas. Appendix 5: Buddha-bodies in other early Abhidharma and related texts

I have argued in this paper that Sarvâstivāda down to Vibh neither taught a rūpakāya-dharmakāya doctrine, nor did its actual body doctrine necessarily predate relevant Mahāyāna ideas. On this basis, I have also suggested it is unlikely that any putative Mainstream rūpakāya-dharmakāya model of the Buddha's embodiment preceded Mahāyāna. However, even if my evidence and interpretation of it here is correct, it is of course logically conceivable that the rūpakāyadharmakāya model was found somewhere in the Mainstream, but not in Sarvâstivāda. I cannot here consider this possibility exhaustively. In this Appendix, however, I briefly survey a number of additional Abhidharma texts from a similarly early period in the Chinese translation tradition for their ideas about Buddha-bodies. In most cases, opinions of modern scholars about relative dating place the original Indic works themselves approximately as late or later than Vibh. So far as I could determine, none of the ideas under study here appear in Ghoṣaka's *Abhidharmâmṛtarasa 阿毘曇甘露味論 T1553 (the Chinese translation of which may date, in its present form, between the mid-third and late fourth centuries); nor in Dharmaśrī's *Abhidharmahṛdaya (or * Abhidharmasāra) 阿毘曇心論 T1550. The only place that T1551 *Abhidharmasāraśāstra (*Abhidharmahṛdayaśāstra) mentions the "body of the Buddha" is in a discussion of the ānantarya sins (drawing the Buddha's blood); otherwise it too is silent on all the ideas studied here.

In the Saṃyuktâbhidharmahṛdaya 雜阿毘曇心論 T1552 we find a small selection of the Vibh ideas discussed, and very few new ideas. Lying in order to create a schism in the Saṃgha is the worst of the ānataryakarmas because it "proceeds in" 轉 the dharmakāya. Giving (dāna) to a  person who teaches the Dharma bears much fruit because it increases the dharmakāya. Giving that nourishes the "body of birth" is opposed to giving that nourishes the dharmakāya. Discussion of the Buddha-refuge proceeds in the now familiar terms of "body of birth" and dharmakāya, and the metaphor of bottle and milk is again used to explain the relation between the two.289 The text also contains the notion that the dharmakāya is clothed.290 The text mentions once the decoration of the body of a bodhisattva by the thirty-two marks. The "Buddha's body" in general is mentioned in the context of the ānantarya sin of drawing the Buddha's blood, and of Māra taking on a "Buddha-body".

The *Satyasiddhi (*Tattvasiddhi) 成實論 T1646, not a Sarvâstivāda text, is relevant as a (partly) Abhidharma text translated into Chinese in the same period (by Kumārajīva 鳩摩羅什 [344-413]). It knows neither the rūpakāya nor the "body of birth". It mentions nirupādhiśeṣanirvāṇa twice, but not in any elaborate connection to the body. Twice only, it speaks simply of the "Buddha's body". It only mentions dharmakāya once, where it appears to be identifying it with the five anāsravaskandhas. In one passage it presents a doctrine we have not seen elsewhere, claiming that some sentient beings attain the end of defilement in reliance on the Buddha's "true body", and some on the basis of a "body of transformation (化身 ,

• nairmāṇikakāya?).

Finally, we will also look briefly at the Abhidharmakośa and Bhāṣya. This text was not translated into Chinese until Paramārtha (499-569), but Vasubandhu is thought to have been alive around the period that concerns us here (turn of the fourth-fifth centuries). His text casts interesting light on the state of Sarvâstivāda Buddha-body doctrine even at this relatively late stage.

Perhaps the most significant thing about Buddha-body ideas in AKBh, for our purposes, is that we finally see some signs of the idea of rūpakāya opposed to dharmakāya. It is therefore possible that AKBh is the source of the characterisation of Sarvâstivāda against which I have been arguing. In the discussion of the Buddha-refuge as the aśaikṣadharmas, AKBh 4.32 opposes rūpakāya to dharmakāya, where we saw Vibh speaks of "body of birth": "As for the rūpakāya of the Buddha, it has undergone no modification in virtue of the acquisition of the qualities of a Buddha. Thus one does not take refuge in the rūpakāya of the Buddha which is, in fact, the rūpakāya of the Bodhisattva." Again, in AKBh to 7.34, alongside perfection (saṃpad) of the dharmakāya, perfection of the rūpakāya is also discussed. It entails acquisition of the major and minor marks, prodigious strength, and the body. We saw above that Vibh tends to discuss these things in terms of the "body of birth". However, to my knowledge these are the only two passages, even in AKBh, which might support the claim that Sarvâstivāda teaches a rūpakāyadharmakāya model.301

There is also some discussion of other related concepts in AKBh, but little is new. Like Vibh, AKBh to 7.34 says the dharmakāya is a respect in which all Buddhas are equal. However, AKBh here presents a much more complex picture than Vibh. AKBh also relates the tradition that deliberately acting against the Dharma and lying to create schism in the Saṃgha is the worst of the ānantarya sins because it harms the dharmakāya. AKBh 4.32 also discusses the Buddharefuge as the aśaikṣadharmas. The text does not, however, overtly identify these dharmas with the dharmakāya. The only new idea about the dharmakāya that I could find is that offerings to a preacher of the Dharma (dharmabhāṇaka) are "immeasurable", even though not given to an Ārya, for reasons connected (in most versions of the text) with the dharmakāya: he "edifies the immaculate dharmakāya" (Skt); he "makes sentient beings give rise to the pure dharmakāya" (XZ); he "produces the dharmakāya" (Paramārtha).

XZ's AKBh knows the term "body of birth" 生身, but the Skt. text does not. XZ uses the term in posing the question: Why should it be an ānantarya sin to physically harm the Buddha, when it only affects his "birth body"? Similarly, XZ again uses the term "body of birth" 生身 (once) when discussing the Buddha's titanic physical strength, but also speaks just of the "body". AKBh knows the doctrine of the thirty-two marks of the mahāpuruṣa, and connects these marks with the bodhisattva as well as the Buddha per se. However, it does not use any special body term in connection with this doctrine. Finally, AK 7.51 and Bh contains a discussion of the production of magical bodies, but the term nirmāṇakāya (or nairmāṇikakāya) is not used. In some of these passages, XZ has body terms where Skt. and other versions of the text do not.

301 In XZ but not Skt., rūpakāya is mentioned, as in several Vibh contexts, as the form in which ordinary sentient beings are born in the kāmadhātu; AKBh to 3.3, Pradhan 112,VP II, 5, T1558:29.41b03-05. Another apparent instance of rūpakāya 色身 in XZ turns out not to be paralleled in Skt. 言身異者。謂彼色身種種顯形狀貌異故 etc., Skt. nānātvena kāya eṣām iti nānātvakāyāḥ. anekavarṇaliṅgasaṃsthānatvāt; Pradhan 115, VP II, 17. It is difficult to tell if this is a chance byproduct of translation, or a sign that the notion of rūpakāya has been introduced; certainly we would imagine Chinese readers would have understood it to refer to rūpakāya.

This variation may indicate, once again, that this body talk remained volatile even at quite late stages of the development of the Sarvâstivāda. Thus, AKBh is the only Sarvâstivāda text that gives even slight support to the common claim that the Sarvâstivāda teaches a rūpakāya-dharmakāya model. However, Vasubandhu is obviously far too late to support any claims that this model preceded similar Mahāyāna ideas. Moreover, the complete absence of the "body of birth" from (Skt.) AKBh alone shows that it is somewhat distanced from the more elaborate Vibh understanding of the Buddha's bodies. We also see that on the whole, AKBh is still less interested in the problem of Buddha-bodies than Vibh (though it is also, of course, a much shorter text). It is thus unclear how much AKBh should be taken as representative of Sarvâstivāda in respect to Buddha-body doctrine. Indeed, given the larger pattern I argue elsewhere seems to observable in the rise of the rūpakāya-dharmakāya model, this may ironically be a sign of Mahāyāna influence upon AKBh. In sum, even when we look further afield, to later Sarvâstivāda texts also translated before approx. 440, to the seminal AKBh, and to the *Satyasiddhi and the *Abhidharmasāraśāstra, we find that Sarvâstivāda and some other Abhidharma texts were still remarkably uninterested in Buddha-bodies; and that the only very slender support for a rūpakāya-dharmakāya model is found in two passages in AKBh.

Abbreviations

AK(Bh) Abhidharmakośa(bhāṣya) Aṣṭa Aṣṭasāhasrikā prajñāpāramitā DDB Muller, Digital Dictionary of Buddhism DN Dīgha Nikāya JñP Jñānaprasthāna MN Majjhima Nikāya MPNS Mahāparinirvāṇa sūtra T374 MPPU *Mahāprajñāpāramitôpadeśa T1509 para. paragraph PTS The Pāli Text Society Skt. Sanskrit SN Saṃyutta Nikāya T Taishō shinshū daizōkyō 大正新脩大藏經 (CBETA version) Vibh *Mahāvibhāṣā VP Louis de la Vallée Poussin XZ Xuanzang 玄奘; Xuanzang's Vibh T1545

Bibliography

Bodhi, Bhikkhu, trans. The Connected Discourses of the Buddha: A Translation of the Saṃyutta Nikāya. Somerville MA: Wisdom Publications, 2000. Conze, Edward. The Prajñāpāramitā Literature. New Delhi: Munshiram Manoharlal, 2000. Cox, Collett. Disputed Dharmas: Early Buddhist Theories on Existence: An Annotated Translation of the Section on Factors Dissociated from Thought from Saṇghabhadra's Nyāyānusāra. Tokyo: The

International Institute for Buddhist Studies, 1995. Demiéville, Paul. "Busshin." S.v. Hôbôgirin : dictionnaire encyclopédique du bouddhisme d'après les sources chinoises et japonaises, ed. Sylvain Lévi, Takakusu Junjirō, Paul Demiéville, et. al. Tokyo: Maison franco-japonaise, 1929-. Demiéville, Paul. "L'origine des sectes bouddhiques d'apres Paramartha." Mélanges chinois et bouddhiques 1 (1931-32): 15-62. Frauwallner, Erich. Studies in Abhidharma Literature and the Origins of Buddhist Philosophical Systems. Translated by Sophie Francis Kidd. Albany: SUNY Press, 1995.

Fronsdal, Egil. "The Dawn of the Bodhisattva Path: Studies in a Religious Ideal of Ancient Indian Buddhists with Particular Reference on the Earliest Extant Perfection of Wisdom Sutra." PhD dissertation, Stanford University, 1998. Frye, Stanley, trans. Sutra of the Wise and the Foolish (mdo mdzangs blun) or Ocean of Narratives (üligerün dalai). Dharamsala: Library of Tibetan Works and Archives, 1981, 2000. Guang Xing. The Concept of the Buddha: Its Evolution from Early Buddhism to the Trikāya Theory. London: RoutledgeCurzon, 2005. Habito, Ruben L. F. ルーベン・Ｌ・Ｆ・アビト. "Busshin ron no tenkai – sanshinsetsu no seiritsu wo megutte 仏身論の展開 － 三身説の成立をめぐって." Shūkyō kenkyū 宗教研究 52, no. 2 (1978): 1-22. Hahn, Michael. Nāgārjuna's Ratnāvalī: Vol. 1, The Basic Texts (Sanskrit, Tibetan, Chinese). Bonn: Indica et Tibetica Verlag, 1982. Hanayama, Shōyū. "A Summary of Various Research on the Prajñāpāramitā Literature by Japanese Scholars." Acta Asiatica 10 (1966): 16-93.

Harrison, Paul Maxwell. "The Earliest Chinese Translations of Mahāyāna Buddhist Sūtras: Some Notes on the Works of Lokakṣema." Buddhist Studies Review 10, no. 2 (1993): 135-177. Harrison, Paul Maxwell. "Is the Dharma-kāya the Real 'Phantom Body' of the Buddha?" Journal of the International Association of Buddhist Studies 15, no. 1 (1992): 44-94. Harrison, Paul Maxwell. "Sanskrit Fragments of a Lokottaravādin Tradition." In Indological and Buddhist Studies: Volume in Honour of Professor J. W. de Jong on his Sixtieth Birthday, ed. L. A. Hercus et al., 211-234. Canberra: Faculty of Asian Studies, 1982.

Honda, Megumu. "Annotated Translation of the Daśabhūmika-sūtra." In Studies in South, East, and Central Asia: Presented as a Memorial Volume to the Late Professor Raghu Vira by Members of the Permanent International Altaicist Conference, ed. Denis Sinor, 115-276. Śata-Piṭaka Series, Indo-Asian Literatures, vol. 74. New Delhi: International Academy of Indian Culture, 1968. Hopkins, Jeffrey, and Lati Rimpoche with Anne Klein. The Precious Garland and the Song of the Four Mindfulnesses. London: George Allen and Unwin Ltd., 1975. Lamotte, Étienne, trans. L'Enseignement de Vimalakīrti (Vimalakīrtinirdeśa). Louvain: Publications universitaires, 1962. Lamotte, Étienne, trans. La Traité de la grande vertu de sagesse (Mahāprajñāpāramitāśāstra). Louvain: Institut orientaliste de Louvain, 1966-1980. la Vallée Poussin, Louis de. L'Abhidharmakośa de Vasubandhu. Brussels: Institut belge des hautes études chinoises, 1980. la Vallée Poussin, Louis de. "Documents d'Abhidharma 2. Doctrine des Refuges." Mélanges chinois et bouddhiques 1 (1931-32): 65-109. la Vallée Poussin, Louis de. "Notes sur les corps du Bouddha." In Vijñaptimātratāsiddhi: La siddhi de Hiuan-Tsang, 762-813. Paris: P. Geunther, 1928-29. Makransky, John J. Buddhahood Embodied: Sources of Controversy in India and Tibet. Albany: State University of New York Press, 1997. Muller, A. Charles, ed. Digital Dictionary of Buddhism. <http://buddhism-dict.net/ddb>.

Nagao, Gadjin. "On the Theory of the Buddha-Body (Buddha-kāya)." Translated by Hirano Umeyo. The Eastern Buddhist, n.s. 6, no. 1 (1973): 25-53. Nakamura, Hajime. "A Critical Survey of Mahāyāna and Esoteric Buddhism Chiefly Based upon Japanese Sources." Acta Asiatica 6 (1964): 57-88. Ñāṇamoli, Bhikkhu, and Bhikkhu Bodhi, trans. The Middle Length Discourses of the Buddha: A New Translation of the Majjhima Nikāya. Boston: Wisdom Publications, 1995. Nattier, Jan. A Guide to the Earliest Chinese Buddhist Translations: Texts from the Eastern Han 東漢 and Three Kingdoms 三國 Periods. Bibliotheca Philologica et Philosophica Buddhica X. Tokyo: The International Research Institute for Advanced Buddhology, Soka University, 2008. Onishi, Yoshinori. "Is the Aṣṭasāhasrikā Prajñāpāramitā Sūtra Really Arguing against the Sarvāstivādins?" Buddhist Studies Review 16, no. 2 (1999): 167-180. Pradhan, P., ed. Abhidharmakośabhāṣya of Vasubandhu. Patna: K. P. Jayaswal Research Institute, 1967. Radich, Michael. "The Somatics of Liberation: Ideas about Embodiment in Buddhism from Its Origins to the Fifth Century C.E." PhD diss., Harvard University, 2007.

Rahder, Johannes, ed. Daśabhūmikasūtra. Leuven: J.-B. Istas, 1926. Reynolds, Frank. "The Several Bodies of the Buddha: Reflections on a Neglected Aspect of Theravada Tradition." History of Religions 16 (1977): 374-389. Roth, Gustav. "The Physical Presence of the Buddha and Its Representation in Buddhist Literature." In Investigating Indian Art: Proceedings of a Symposium on the Development of Early Buddhist and Hindu Iconography Held at the Museum of Indian Art Berlin in May 1986, 291-312. Berlin: Museen für Indische Kunst, Staatliche Museen Preussischen Kulturbesitz, 1987. Rotman, Andy. "The Erotics of Practice: Objects and Agency in Buddhist Avadāna Literature." Journal of the American Academy of Religion 71, no. 3 (2003): 555-578. Sakuma Hidenori 佐久間秀範. Die Āśrayaparivṛtti-theorie in der Yogācārabhūmi. 2 vols. Alt- und Neu- Indische Studien 40. Stuttgart: Franz Steiner Verlag, 1990. Sakuma Hidenori 佐久間秀範. "Yugashi-ji ron ni okeru tenne shisō 『瑜伽師地論』における転依思想." Indogaku bukkyōgaku kenkyū 39, no. 1 (1991): 66-74. Schopen, Gregory. "The Phrase sa pṛthivīpradeśaś caityabhūto bhavet in the Vajracchedika: Notes on the Cult of the Book in Mahāyāna." In Fragments and Figments of Mahāyāna Buddhism in India: More Collected Papers, 25-62. Honolulu: University of Hawaiˋi Press, 2005. Senart, Émile. Le Mahâvastu: Texte sanscrit publié pour la première fois et accompagné d'introduction et d'un commentaire. Paris: Impr. Nationale, 1882-1887. Shastri, Swami Dwarikadas. Abhidharmakośa & Bhāṣya of Acharya Vasubandhu with Sphutārthā Commentary of Acharya Yaśomitra. Varanasi: Bauddha Bharati, 1970. Silk, Jonathan A. "Good and Evil in Indian Buddhism: The Five Sins of Immediate Retribution." Journal of Indian Philosophy 35, no. 3 (2007): 253-286. Strong, John. The Legend and Cult of Upagupta: Sanskrit Buddhism in North India and Southeast Asia. Princeton: Princeton University Press, 1992. Study Group on Buddhist Sanskrit Literature, The Institute for Comprehensive Studies of Buddhism, Taisho University 大正大学総合佛教研究所梵語佛典研究会, ed. Bonbun Yuimakyō : Potara kyū shozō shahon ni motozuku kōtei 梵文維摩經 : ポタラ宮所蔵写本に基づく校訂; Vimalakīrtinirdeśa: A Sanskrit Edition Based upon the Manuscript Newly Found at the Potala Palace. Tokyo: Taisho University Press, 2006. Takeda Kohgaku 武田浩学. "Daichido ron wa 'Hon'u konmu' ge wo ikanaru kū rikai no reveru de ronjiteiru no ka: yonjikkanbon Dainehan gyō seiritsuron to no kanren ni okeru ichikōsatsu 『大智度論』は「本有今無」偈を如何なる空理解のレヴェルで論じているのか：四十巻本『大涅槃経』成立論との関連における一考察 [A Significant Verse of the Mahaparinirvanasutra and its Interpretation in the Mahaprajnaparamitasastra]." Journal of the International College for Postgraduate Buddhist Studies 4 (2001): 203-261. Takeuchi, Shōkō 武内紹晃. "Buddakan no hensen 仏陀観の変遷." In Daijō bukkyō vol. I, ed. Hirakawa Akira, 155-181. Tokyo: Shunjusha, 1983. Tatelman, Joel, ed. and trans. The Heavenly Exploits: Buddhist Biographies from the Divyāvadāna, Volume One. The Clay Sanskrit Library. New York/Bury St Edmunds, Suffolk: New York University Press/JJC Foundation, 2005. Thurman, Robert A. F. The Holy Teaching of Vimalakīrti: A Mahāyāna Scripture. University Park: Pennsylvania State University Press, 2003. Vaidya, P. L. ed., Avadāna-Śataka. Buddhist Sanskrit Texts No. 19. Darbhanga: The Mithila Institute, 1958. Van Den Broeck, José. La saveur de l'immortel (A-p'i-t'an Gan Lu Wei Lun): La version chinoise de l'Amṛtarasa de Ghoṣaka (T1553). Louvain-La-Neuve: Institut orientaliste, 1977. Walshe, Maurice, trans. The Long Discourses of the Buddha: A Translation of the Dīgha Nikāya. Boston: Wisdom Publications, 1995. Watson, Burton, trans. The Vimalakirti Sutra. New York: Columbia University Press, 1997. Willemen, Charles, Bart Dessein and Collett Cox. Sarvāstivāda Buddhist Scholasticism. Leiden: Brill, 1998. Willemen, Charles. The Essence of Metaphysics (Abhidharmahṛdaya). Publications de l'Institut Belge des Hautes Études Bouddhiques, Serie 'Études des Textes' No. 4. Brussells: 1975. Yamabe Nobuyoshi. "The Paths of Sravakas and Bodhisattvas in Meditative Practices." Acta Asiatica 95 (2009): 47-75. Yamamoto, Kosho, trans. The Mahayana Mahaparinirvana-sutra: A Complete Translation from the Classical Chinese in 3 Volumes, Annotated and with Full Glossary, Index and Concordance by Kosho Yamamoto. Ube: Karinbunko, 1973-1975.