The Meaning of Vijnapti in Vasubandhu's Concept of Mind

THE JOURNAL OF THE INTERNATIONAL ASSOCIATION OF BUDDHIST STUDIES EDITOR-IN-CHIEF

A.K. Narain

University of Wisconsin, Madison, USA

The Meaning of Vijnapti in Vasubandhu's Concept of Mind

by Bruce Cameron Hall

For the Mahayana it is determined that the whole of the three realms is vijnapti-on\y (vijnapti-mdtra), according to the sutra: "It is thought-only (citta-matra), You Sons of the Conqueror, that is the whole of the three realms." Thought (citta), mind (manas), awareness (vijndna), and vijnapti are synonyms. Here "thought" (citta) implies "[[[thought]] itself] along with its concomitants." The [[[word]]] "only" serves to rule out [external] referents (artha). So begins the Vimsatikd-vrtti (VV),1 Vasubandhu's auto-commentary on the Vimsatikd-kdrikd (VK). The first stanza of VK reads:2

This [[[universe]]] is certainly vijnapti-only, since there are appearances of non-existent [external] referents, as when someone with an eye disease sees a non-existent "knot of hair" and so on. /VK 1/ VK (with VV) and the TrimJikd-kdrikd (TK) together make up the Vijnaptimdtratd-siddhi, or "Establishing That There is Vijnapti- On\y"5 Clearly, vijnapti-mdtra is being equated here with

citta-matra ("mind-only" or "thought-only"), which is an alternate title for Vasubandhu's Vijnanavada philosophy. While the Trirhsika presents Vasubandhu's own doctrine in some detail, the Vimsatikd (VK and VV) is a polemical work in the form of a dialogue between Vasubandhu and an imaginary opponent. This opponent is a fellow Buddhist, but a realist or, one might say, a literalist.

The vijnapti-mdtra of the Vijnaptimdtratd-siddhi has been variously translated as "representation-only,"4 "ideation-only,"5

"perception-only,"b and so on. Although none of these glosses is completely satisfactory, the purpose of the present essay is not to suggest another English equivalent, but rather to analyze the term vijnapti, its usage, and the concept it designates. Such analysis may help clarify the general conception of mind in the so-called "Yogacara idealism" as presented by Vasubandhu. This

analysis amounts to a commentary on the first paragraph of VV.

The elements to be explicated in this commentary are: (1) the term vijnapti itself, (2) its equation with other terms for mind, (3) the significance of the "only" in "mind-only," and (4) how it is that the "whole of the three realms" can be identified as "mind-only."

In addition to the Vijnaptimdtratd-siddhi itself, reference will be made to four other works ascribed to Vasubandhu: Abhidharmakoia-kdrikd (AKK), Abhidharmarkosa-bhasya (AKB), Karmasiddhi-prakarana (KSP), and Pancaskandha-prakarana (PSP).

7 The controversy over whether these works were written by one or more Vasubandhus is here ignored: the author of the Vijnaptimdtratd-siddhi is evidently fluent in the Sarvastivadin Abhidharma, which is outlined in AKK and treated critically in AKB, KSP, and PSP.

/. The Term Itself Vijnapti is a technical term of the Sarvastivadin Abhidharma, which Vasubandhu has here appropriated and used in a special sense. An investigation of this term can illustrate in miniature the widespread appropriation and redefinition of the Abhidharma in Vijnanavada philosophy. An interpretation of such a technical term should consider its ordinary use, its etymology, and its technical use (both in specific contexts and also in

relation to a cluster of other technical terms). This ought to reveal Vasubandhu's precise intention in reapplying the term. In ordinary parlance vijnapti (Pali vinnatti) means "information" or the act of informing someone, that is "report" or "proclamation," especially a report to a superior, and hence, "request"

or "entreaty." Vijnapti is a noun of action derived from the causative stem (jnapaya- or jnapaya-) of the verb root jnd ("know") with the prefix vi-.H Etymologically the term vijnapti would mean the act of causing [someone] to know [something] distinctly, or

in a concrete sense, that which causes [one] to know distinctly. Another important term derived from the causative oi'jnd (with the prefix pra-) is prajnapti (Pali pannatti), which means "declaration" or manifestation in words, and hence, "verbal or conventional designation," or perhaps even "concept."9 Another parallel formation which would be well known to a Buddhist monk is the unprefixed form jnapti or jnapti (Pali natti), the technical term for a formal "motion" or "proposal" in a meeting of the monastic community, for example: the motion to ordain a new monk.10 These three terms share the sense of a public act of "making known."

In the Sarvastivadin Abhidharma both vijnapti and its opposite, avijriapti, appear as technical terms, particularly in the discussion of karma. Here karma specifically means ethically significant action: acts or deeds.11 In chapter 4 of the Abhidharmakosa Vasubandhu defines karma as: "volition and that derived from it,"12 quoting a sutra: "there are two [kinds of] karmas: volition and the act subsequent to volition."13 This two-fold karma is expanded into three types: mental (corresponding to "volition"), and vocal or bodily (corresponding to "post-volitional" action). Vocal and bodily karma is further classified as including both vijnapti- and avijnapti-karma, "manifest" and "unmanifest" acts, that is, karma manifest or not manifest to some consciousness.

One should note that, from an Abhidharmic perspective, the common-sense notion of "an act" is analyzed into a succession of momentary dharmas. In the case of vocal and bodily acts the dharmas would be moments of sound or color-shape. These audible or visible forms are understood to be dharmas included in the "aggregate of material forms" (rupa-skandka), since, given the momentariness of phenomena, the Abhidharma allows no real distinction between "acts" and "things." So far the Abhidharmic analysis is clear. Difficulty arises in that the "unmanifest" act is also included as one of the 75 dharmas accepted by the Sarvastivadins, and this dharma, avijnapti-rupa, is also included in the aggregate of material forms.14 Contrary to what one might assume at first glance, this "unmanifest" karma is not some kind of private act not observed by others, since an act is vijnapti (for the Sarvastivadins) if it could be "manifest" to another consciousness. [For the Vi10

jnanavadins vijnapti means "manifest to any consciousness," including that of the agent, and in this sense mental karma would be entirely vijnapti, since consciousness is, by definition, selfmanifesting.] Instead, avijnapti-rupa is used to explain karmic

continuity in certain contexts. A "manifest" vocal or bodily act is karmic, in the sense of ethically significant, because of its dependence on volition. But, since it is also "dharmic," that is, a momentary event, how can one account for the connection between this momentary act and its future consequence? The explanation of karmic continuity is a general problem for the Sarvastivadins, and it is in this context that avijnapti-rupa is added to the list of dharmas.

The following sequence is postulated:

(1)[[[manifest]], mental] volition,

(2) [[[manifest]], material] vocal or bodily act,

(3) [unmanifest, material] avijnapti-rupa,

(4) [manifest,material] consequence [that is, a later, consequent manifestation of rupa].

Since both the preceding (vijnapti) act and the succeeding (vijnapti) consequence are "material," it follows that the intervening (avijnapti) dharmas, although imperceptible, are also "material"—that is, they belong to the rupa-skandha. This notion of avijnapti-rupa is filled with difficulties, and Vasubandhu presents it with considerable qualms in the Abhidharmakoia. In the Karmasiddhi-prakarana, the whole concept of vijnapti- I avijnapti- rupa is rigorously criticized and finally rejected, and all karma is reduced to volition.15 Given this, it is tempting to see the title Vijnaptimdtrata-siddhi as Vasubandhu's proclamation that he has solved this problem by eliminating the category of avijnapti. In any case, the Vijnaptimdtratd system refers the problem of karmic continuity to the concept ofalaya-vijnana, the "store-consciousness" which contains the "residue" (vdsand) of past acts and the "seeds" (bija) of future ones. The new meaning assigned to vijnapti can best be explained by considering next the other three terms equated with it in the opening passage of VV.

The Other Terms for "Mind"

The translation of citta, manas, and vijnana above as "thought," "mind," and "awareness" should be understood as merely tentative. In fact, much argument has been devoted to

the problem of translating these terms. Brian Galloway, for example, has argued that the "correct translation" is "mind" (citta), "consciousness" (manas), and "perception" {vijriana), and that the "formerly prevailing" translations are wrong.16 Unfortunately, "the correct translation" is not so easy to come by. Not

only do the Sanskrit terms have several meanings and various uses, but the suggested equivalents are all imprecise terms in English. Any translation will thus mean something different depending on what a vague word such as "consciousness" suggests to each translator or reader. Once again, what is needed is not simply another English equivalent, but an explanation of the actual usage of the Sanskrit term. Furthermore, one should not forget that, in any case, these three terms are here stated to be synonyms.

VV presupposes that: "citta, manas, and vijndna have a single meaning."17 All three are terms for mind, not as a substantial entity, but as a stream of momentary mental dharmas. All three signify the same dharma or dharmas. Why, then, does Vasubandhu use three different terms for the same reality? One might

answer that, as a member of the Buddhist scholastic tradition that employs these three terms, Vasubandhu is obliged to explain them. One might also suggest that using three words, and thus pointing at the same reality from several perspectives, provides a depth of description that a single word could not. The

Abhidharma literature consists largely of intersecting and crossreferenced lists of terms. The three main terms for mind appear traditionally in different lists, with a different connotation and context.

Citta is perhaps the most basic term for "mind" or "thought." It is the term that signifies a single thought, or better, a single thought-moment. Citta is also used to designate a particular mind as opposed to other minds, though in this sense the proper

technical term is citta-samtdna, "thought-series," a synonym for vijndna-srotas, "stream of consciousness." Citta is the mental as contrasted with the material (rupa), and bare consciousness as

contrasted with mental states (caitasika, caitta, or citta-samprayukta- samskara). [In the present passage, however, it is stated that citta here means consciousness along with its "concomitants," the mental states; it is "mind" or "consciousness" in the most general sense to which the equation of citta, manas, vijndna,

and vijnapti refers.] In the Sarvastivadin Abhidharma, the older classification of the five "aggregates" (skandhas) is replaced (schematically at least) by the tabulation of the 75 dharmas under the five headings of "material form" (rupa), "consciousness" (citta), "dispositions conjoined with consciousness" (citta-sarhprayukta- samskdra), "dispositions disjoined from consciousness"

(citta-viprayukta-sarhskdra), and the "unconditioned" {asamskrta). In this context, the single dharma, citta, takes the place of the "aggregate of awareness" (vijnana-skandha). Manas, on the other hand, is the term for "mind" as the

sixth of. the six organs or faculties of perception (sensory or mental) in the list of twelve "sense-fields" (dyatana): the six "sense organs" {indriya) and six "sense objects" (vi$aya). Here, "mental" is contrasted not with "material" but rather with "sensory." Perception can also be analyzed into three aspects: the object of

cognition (dlambana), the organ of cognition (dsraya), and the corresponding act of awareness (vijndna). In this way, the twelve sense-fields become the eighteen dhdtus (elements of perception), with the addition of visual, auditory, olfactory, gustatory,

tactile, and mental vijndna (perception or awareness). "Mental perceptions" are strictly-mental perceptions—that is, perceptions of ideas, concepts, or mental objects, whether derived from previous sensory perception or not. Vasubandhu (in AKK-AKB

1.16-17) finds no "dharmic" distinction between "mind" (manas as an "organ") and "mental perception" (mano-vijndna). Although the logic of the scheme requires a corresponding "organ" (dsraya or indriya) for each object and perception, the "organ" for mental perception is simply previous moments of awareness,

which serve as a causal basis for the arisal of the present mental perception. Once again, mind is not a substantial or quasi-material entity, but a stream of causally related thought-moments. Vijndna may be translated as "awareness," "consciousness," "cognition," "perception," and so on. Vasubandhu gives the following definition (AKK 1.16): "Vijndna is 'respective' vijnapti." (AKB:) "The apprehension that is the vijnapti with respect to

the various sense objects is called the vijnana-skandha."18 Vijndna occurs as a term in a great number of Abhidharmic lists; for example, the fifth of the five skandhas and six of the eighteen dhdtus are called vijndna. The vijnana-skandha has already been equated with citta above. In AKB on AKK 1.16, it is also equated

with manas and with the last six of the eighteen dhatus. In this last case, vijnana quite clearly means "perception." To translate caksur-vijndna and srotra-vijndna, for example, as "eye-consciousness" and "ear-consciousness" (with the plural "consciousnesses")— rather than "visual perception" and "auditory perception"— is not simply pecular English; it might even suggest a

strangely animistic notion of consciousness. However, "perception" or even "cognition" does not quite fit some of the other uses of vijnana. Such, for example, is vijnana as the sixth of the six dhatus—a different set of dhatus—the other five being earth, air, water, fire, andafewa (here meaning "ether" rather than "space," as it usually does in Buddhism). There is

also the third of the twelve links in the chain of dependent arising (pratltya-samutpdda): a vijnana that arises in dependence on sarhskaras, and in dependence on which there arises "name and form" (ndma-rupa). This is related to the use of vijnana as a term for that which "passes over" in rebirth. The specifically

Yogacara term dlaya-vijndna ("store-consciousness") has also been mentioned above. In these instances, where vijnana suggests something prior to or more general than perception or cognition, "awareness" or "consciousness" would seem to be a better translation.

These terms, citta, manas, and vijnana, all illuminate a concept of mind as a stream of causally related thought-moments, each of which is a specific act of bare awareness. What is it that vijndpti, which Vasubandhu here gives as the fourth synonym,

adds to the picture? Vijndpti here signifies more than vijnaptikarma, but retains a sense of activity or function. Vijfiapti designates the basic phenomenon of conscious experience, without requiring its separation into object, subject, and act of cognition. What then is "vijnapti-or\\y"}

HI. The Significance of "Only"

The word mdtra means "measure" or "extent." It is frequently added as the second member of compounds (which may be understood as bahuvrihis based on appositional karmadhdrayas) in the form of "X-mdtra" meaning "having X as its full extent," hence "only X." Often the sense is pejorative: "mere X." It is

clear, however, that in the present context no pejorative sense is required. "Mind-only" means "consisting only of mind," "nothing but mind," "bare mind," "sheer mind," and so on.19 A mdtra compound, then, affirms one factor while denying all others that might apply. The passage itself makes clear what is denied here: objects as external substantial entities. Yet cittamatra, the term used in the sutra passage cited, would suffice to deny external objects. What is the point of specifically affirming vijnapti by using the compound vijnapti-mdtra?

The term vijnapti signifies a "phenomenon" of consciousness, 20 a "manifestation" to consciousness, or a "percept"21—so long as one bears in mind that these terms should not be taken in a naively realistic or a naively idealistic sense. The translation "perception" is not bad, especially considering the ambiguity of

the English word. "Perception" can denote a quality, a faculty, a process, or the apparently objective aspect of that process: its contents. However, it might be better to retain "perception" for translating pratyak$a or vijndna (at least in some of its senses). To translate vijnapti here by "representation" conveys its "public" aspect, but seems to imply representation of something,

presumably of an external object or referent, which suggests a "representational" theory of knowledge. On the contrary, the purpose of the argument throughout the Virhiatika is to show that the concept of vijnapti suffices to make sense of perception, and that the concept of an external referent (artha) is logically

superfluous. It is specifically stated in the initial quotation that "the [[[word]]] 'only' serves to rule out [external] referents." Clearly, then, when vijnapti is qualified as "vijnapti-only," it cannot be meant as a representation of anything else, especially not of an external object.

This would seem to imply that the theory of knowledge involved here, if not representational, is some sort of subjective or absolute idealism.22 This has, in fact, been the most common "outside" interpretation of Vijnanavada, not only by modern

writers, but by its ancient opponents, both Hindu and Buddhist. Any statement to the effect that the world is "mind-only" seems to imply that, given a set of material and mental factors, the former are denied and the latter are affirmed, or the former are reduced to the latter. Even the translation "ideation-only" for vijnapti-matra seems to suggest that matter is unreal while

consciousness alone is real. It is one thing to accuse the Vijnanavadins of falling into a reification of mind, and quite another to assume that such is their intention. The former position is certainly arguable, and was strongly argued by their Madhyamika rivals. If anything, it

is the concept of dlaya-vijndna, rather than that of vijnapti-matra, which most exposes the Vijnanavadins to the charge of turning consciousness into some kind of substance or self. Here I would argue that, as is so often the case in Buddhist philosophy, Vasubandhu is consciously navigating between two extremes, which

in this case may be called realism and idealism. In negative terms, vijnapti-matra rules out the realist extreme: substantial external objects of cognition are denied. However, vijnapti-matra has also a positive connotation, and the fact that Vasubandhu here affirms precisely vijnapti—rather than

vijndna or citta, which might more easily be misunderstood— seems to indicate an intent to avoid the idealist extreme as well. What is exclusively affirmed is not consciousness as an abiding entity, but the content of momentary acts of consciousness.

When this vijnapti is equated with citta, manas, and vijndna, it follows that mind itself is vijnapti-matra: it consists of nothing else than the contents of momentary mental acts. The intention here is not to reduce the material to the mental, but to deny the dichotomy, while affirming that the basic reality is more usefully discussed in the terms belonging to a correct understanding of the mental.

IV. The Whole of the Three Realms

The compound vijnapti-matra involves a denial and an affirmation. The extent of this denial/affirmation is truly universal. The term used in the passage is traidhdtukam, which may be understood, grammatically and contextually, as meaning idam traidhdtukam: "this [[[universe]]] consisting of three dhdtus"2 Here we meet a third, cosmological, sense of the term dhatu in the Abhidharma lexicon. There are three cosmological "realms." The "realm of desires" (kdma-dhdtu) is the world of "ordinary" experience, that is to say the world experienced by beings in hell, ghosts, animals, most humans, and the lower orders

among the divinities. The two higher "realms" may be entered through either meditation or apparitional birth. The "realm of forms" (rupa-dhdtu) corresponds to the refined experiences of those in the first four levels of meditation (the four dhyanas) and the analogous experiences of certain classes of gods. The "realm of formlessness" (drupya-dhdtu) consists of the experiences of those meditators and divinities abiding in the formless meditations. These are given as four: infinite space, infinite awareness (vijndna), nothingness, and neither conception nor non-conception (with sometimes a fifth added: the cessation of conception and feeling).

Taken together, these three "realms" comprise the whole cosmos. "The whole of the three realms" is synonymous with samsara and with "all conditioned (samskrta) dharmas." If Vasubandhu's statement that the whole of the three realms is nothing butvijnapti is taken in the "idealist" sense, then this implies that there is really nothing "out there"—the whole universe is in the mind. Among other consequences, this interpretation

inevitably raises the problem of solipsism: "the whole universe is in my mind." On the contrary, taking the interpretation argued in this paper, Vasubandhu's statement means that the whole universe is nothing but the contents of consciousness (that is, all the contents of everyone's consciousness). Does this amount to the truism that we cannot conceive of anything that we cannot conceive of? I would argue, instead, that the intention of the vijnapti-mdtra doctrine is not to draw boundaries around reality but rather to point at the nature of specific experiences. Vasubandhu himself states the purpose of this doctrine in

VK-VV 7-10. His imaginary opponent poses the scriptural objection that, if consciousness alone exists, why does the Buddha speak of all twelve sense-fields (objects and organs of perception)? Vasubandhu replies that the Buddha taught with "a special intention," depending on the needs of a specific audience. Both the teaching of the twelve dyatanas and the teaching of vijnapti-mdtra involve such a special intention, being two stages in the teaching of "no-self." The purpose of the dyatana doctrine (and, by implication, of the whole Abhidharma analysis into dharmas) is to introduce pudgala-nairdtmya ("the fact that there is no self in persons").24 Analysis into dharmas can dissolve the substantiality of "things"

and "beings," yet dharmas too can be reified. The purpose of the vijnapti-matra doctrine is to introduce the second stage in understanding "no-self: dharma-nairdtmya ("the fact that there is no self in dharmas"). The opponent objects that this seems to deny the existence of dharmas altogether, which would mean that vijnapti-matra too is non-existent. Vasubandhu replies that what is denied is not the existence of dharmas as moments of experience, but rather "that mentally constructed self that is the intrinsic nature of dharmas imagined by naive people as object,

subject, and so on."25 This does not deny that the dharmas themselves, as perceived by the Buddha, exist. However, the dharmas as perceived by the Buddha are inexpressible (anabhildpya) by philosophy. Furthermore, Vasubandhu points

out that this teaching of dharma-nairatmya works only when vijnapti- matra itself is understood to be vijnapti-only. Clearly, no reification of consciousness is intended here. One further passage may clarify Vasubandhu's intention. To the repeated assertion that dreams prove the possibility of perception without external referents, the opponent objects: "Someone not yet awakened does not understand the non-existence of a sense object seen in a dream."26 Vasubandhu replies that, on the contrary, we are all asleep:27

In this way, the world—asleep with a sleep that is the residue of the imposition of false conceptualizing—is seeing, as in a dream, a non-existent [external] referent. Not having been awakened, it does not understand, as it is, the non-existence of that [referent]. But, when one has been awakened through obtaining the transcendental non-conceptual cognition that is the antidote to that [[[sleep]]], then—through the direct realization of the purified mundane cognition that is obtained after that—one understands, as it is, the non-existence of the [[[sense]]] object.

V, Conclusion

I find it misleading to call Vasubandhu's approach "absolute idealism." Instead, I would see Vasubandhu's argument in the Virhsatikd as one more attempt to find the Buddhist "middle way" between positive and negative extremes, in this case the

extremes of reification and reductionism. "Common sense" takes the objects of perception to be substantial external entities, that is, "things." Analytical concepts such as atoms or dharmas are powerful tools that can demolish such "things," but atoms or dharmas can themselves be reified. Vasubandhu's argument

denies the necessity to posit any entities external to perception itself, and rejects, successively, the reification of things, atoms, dharmas, and even vijnapti itself. In Vasubandhu's Vijnanavada, vijnaptis, in effect, take the place of dharmas in the Abhidharma: as conceptual devices to prevent the reification of objects. The doctrine of vijnapti-mdtra is not the metaphysical assertion of a transcendental reality consisting of "mind-only." It is a practical injunction to suspend judgment: "Stop at the bare percept; no need to posit any entity behind it."

Rather than asserting "mind-only" as the true nature of unconditioned reality, Vasubandhu presents "mind-only" as a description of our delusion: the dreams of this sleep from which the Buddha has awakened. It is, after all, sarhsara that is declared to be vijnapti-mdtra. Yet if "mind-only" is merely skepticism about reified external entities, how does it avoid the opposite extreme of reductionism? The world is neither completely real, nor completely unreal, but like a dream. A dream has its own presence

and continuity, but its objects lack the substantiality of external objects. Whether common-sense things or Abhidharmic dharmas, dream-objects are bare percepts. If the dream-world sarhsara is "mind-only" then freedom and the Buddhist path are possible—we can "change our minds." If the realms of meditation are "mind-only" then one can create a counter-dream within the dream of the world's delusion. Most important, one can awaken from a dream.

Is it then correct to call the Yogacara, as presented by Vasubandhu, "Buddhist idealism"? The term "idealism" designates a number of different philosophies. At the least one should be aware that the "idealism" that Vasubandhu attempts to assert and the "idealism" for which his opponents criticize him may

be different "idealisms" The argument over whether Vijnanavada is idealistic or realisitic bears a marked resemblance to the controversy as to whether Madhyamaka is nihilism or transcendental absolutism.

Mistaking taxonomy for understanding is a fault not limited to modern writers on Buddhism. A similar excessive concern for and trust in doctrinal labels can be seen in ancient Indian philosophers and Tibetan scholastics, and even in the Abhidharma itself. Instead of seeking the correct label for Vasubandhu's philosophy, we would do better to try to understand

it in its own terms. The identification of one school with another (such as that of Vijnanavada with some Western form of idealism) is not only likely to be misleading; it is all too often the point at which the argument stops. A more fruitful approach to comparative philosophy would begin by tentatively accepting several comparable philosophies as coherent systems in their own terms, and would proceed to apply their several viewpoints to specific problems of philosophy.

NOTES

1. All translations are my own unless otherwise noted. For the sources of Sanskrit texts here translated see below, notes 3 and 7. The passage from VV translated here is missing from the Sanskrit MS, and was reconstructed by Sylvain Levi as follows (1925, p.3):

Mahdydne traidhdtuham vyavasthdpyate I cittamatram bho jinaputrd yad uta traidhdtukam iti sutrdl I citiarh mano vijndnarh vijnaplis ceti parydydh I cittam atra sasarhprayogam abhipretam I rndtram ity arthaprati$edhdrtham I.

2. The first kdrikd of VK is still quoted or translated by some writers as it was first reconstructed and printed in Levi 1925. As emended in Levi 1932 (on the basis of new MSS of VK) the stanza reads: vijnaptimdtram evedam asadarthdvabhdsandt I yadvat taimirikasydsatkesotidrakddidarsanam II.

3. Sources used here for the Vijnaptimdtrata-siddhi are: Louis de La Vallee Poussin, "Vasubandhu, Virhgakakarikaprakaraiia: Traite des vingt slokas, avec le commentaire de l'auteur," Museon (New Series) 13 (1912): 53-90. [Romanized Tibetan text of VV, with French translation.] Sylvain Levi, Vijnaptimdtratdsiddhi: Deux trates de Vasubandhu, Vimsatikd et Trimsikd, Bibliotheque de I'Ecole des Hautes Etudes (sciences historiques et philologiques) fascicule 245 (Paris: Librairie Ancienne Honore Champion, 1925). [[[Sanskrit]] text of VK, VV, TK, and TB (Sthiramati's commentary on TK).]

Sylvain Levi, Mate'riaux pour Vetude du systeme Vijnaptimdtra, Bibliotheque de I'Ecole des Hautes Etudes (sciences historiques et philologiques) fascicule 260 (Paris: Librairie Ancienne Honore Champion, 1932). [Includes Sanskrit emendations and French translation for Levi 1925.] Clarence H. Hamilton, Wei Shih Er Shih Lun: The Treatise in Twenty Stanzas

on Representation-only, by Vasubandhu (Translatedfrom the Chinese Version ofHsuan Tsang, TripHaka Master oftheTang Dynasty), American Oriental Series, 13 (New Haven: American Oriental Society, 1938). [Hsiian-tsang's Chinese text of VV with English translation.]

Sitamsu Sekhar Bagchi, "Vijnaptimatratasiddhi," Nava-Nalanda- Mahavihara Research Publication 1 (1957): 367-389 ( + Sanskrit pages 1-12). [[[Sanskrit]] text of VK-VV (Levi 1925 without emendations), with English translation of VK-VV embedded in Bagchi's interpretation.]

Wing-tsit Chan, "The Thirty Verses on the Mind-Only Doctrine," in Sarvepatli Radhakrishnan and Charles A. Moore editors, A Source Book in Indian Philosophy (Princeton: Princeton University Press, 1957), pp. 333-337. [English translation of TK from Hsuan-tsang's Chinese version, along with (pp. 328-333) a partial reprint of Hamilton 1938.]

Thomas A. Kochumuttom, A Buddhist Doctrine of Experience: A New Translation and Interpretation of the Works of Vasubandhu the Yogacarin (Delhi: Motilal Banarsidass, 1982). [Includes English translations of VK-VV and TK. Kochumuttom seems to have depended entirely for the Sanskrit text of VK,

VV, TK, and TB on an extremely unreliable edition: Svami Mahesvarananda, Acdrya vasubandhu pranita I vijnapti matratdsiddhih I pancds'atikd I savrttikd trirhsatikd kdrikd I acdrya sthiramati pranltam trimsikd bhds,yanca [sic\] (VaranasI: GItadharma Karyalaya, 1962). Some of the new departures in Kochumuttom's translation seem to be based on Mahesvarananda's misprints. The misprint on Mahesvarananda's title page has apparently misled Kochumuttom into consistently calling TK the "Trimsatika."}

Stefan Anacker, Seven Works of Vasubandhu: The Buddhist Psychological Doctor, Religion of Asia Series, 4 (Delhi: Motilal Banarsidass, 1984). [Includes English translations of Vddavidhi, Pancaskandhaprakarana (PSP), Karmasiddhiprakarana (KSP), VK-VV, TK, Madhyantavibhdgabhds.ya (MVB), and Trisvabhavanirdesa (TSN), and reprints Sanskrit editions of VV, TK, MVB, and TSN.]

4. E.g., Hamilton (1938).

5. E.g., Chan (1957).

6. E.g., Anacker (1984).

7. For AKK-AKB I have used the Sanskrit edition of Prahlad Pradhan, Abhidharmakosabhdsyam of Vasubandhu, Tibetan Sanskrit Works Series, 8 (Patna: K.P. Jayaswal Research Institute, 1967; reprinted 1975). I have also consulted the French translation of Louis de LaVallee Poussin, L'Abhidharmakosa de Vasubandhu, 6 volumes, Melanges chinois et bouddhiques, 16 (Bruxelles: Institut Beige des Hautes Etudes Chinoises, 1971 reprint; lsted. Paris: PaulGeuthner, 1923-31).

KSP is lost in Sanskrit, but survives in one Tibetan translation (Tohoku 4062 = Otani 5563) and two Chinese translations (Taisho 1608, 1609). There is an English translation (from Tibetan) in Anacker (1984), and a French translation (with the Chinese and Tibetan texts) in Etienne Lamotte, "Le Traite de l'acte de Vasubandhu, Karmasiddhiprakarana," Melanges chinois et bouddhiques 4 (1935-36) (Bruxelles: Institut Beige des Hautes Etudes

PSP is also lost in Sanskrii, surviving in one 'Tibetan translation (Tohoku 4059 = Otani 5560) and one Chinese translation (Taisho 1612). There is an English translation (IVom Tibetan) in Anacker (1984), and a French translation (with Chinese and Tibetan texts) in Jean Dantinne, U Traite des cing aggregals (Paficaskandhaprakararia de Vasubandhu), Publications de l'lnstitut Beige des Hautes Etudes Bouddhiques, Serie "Etudes et Texts," 7 (Bruxelles: Institut Beige des Hautes Etudes Bouddhiques, 1980). PSP is discussed in V.V.

Gokhale, "The Pancaskandhaka by Vasubandhu and its Commentary by Sthiramati," Annals of the Bhandarhar Research Institute (Poona) 18.3 (1937): 276-286. There is a discussion and a Sanskrit retranslation of PSP in Shanti Bhikshu Shastri, "Pancaskandhaprakarana of Vasubandhu," Indian Historical Quarterly 32 (1956): 368-385. Another work by Shanti Bhikshu Shastri, which

I have seen cited but have not been able to find, is Pancaskarui/iaprakararia of Vasubandhu: A Restitution into Sanskrit from the Tibetan Version together with an Introduction, English translation, Notes, a Tibetan-Sanskrit vocabulary and an Index of important Sanskrit Words (Kelaniya, 1969).

8. Compare Franklin Edgerton, Buddhist Hybrid Sanskrit Grammar and Dictionary, Volume II: Dictionary (New Haven: Yale University Press, 1953: reprints Delhi: Motilal Banarsidass, 1970, etc.), s.v. vijnapli.

9. Compare Edgerton (1953), s.v. prajfiapli.

10. Compare Edgerton (1953), s.v. jnapli, and the documents dealing with upasampada ("ordination") in the collection of official acts of a Sangha known as the Karmavdcand (Pali Kammavdca).

11. The classic sources for the discussion of karma are AKK-AKB chapter 4, and KSP.

12. AKK 4.1b: cetand tatkrtam ca tat.

13. AKB on AKK 4.1b: dve karmani cetand karma celayitvd ca.

14. Avijnapti-rupa is discussed in detail at AKK-AKB 1.11 and 4.1-22, and in KSP. See also Thomas Lee Dowling, Vasubandhu on the "Avijnapti-rupa": A Study in Fifth-Century Abhidharma Buddhism (Ph.D. Dissertation, Columbia University: 1976).

15. AKB on AKK 4.4 presents the controversy between the Sautrantika position that avijrmpti does not exist as a substantial entity (dravya) and the Sarvastivadin or Vaibhasika position that it does. For KSP see especially Anacker's (1984) translation and notes.

16. See two articles by Brian Galloway, "Vijnana, Samjna, and Manas," The Middle Way vol. 53, no. 2 (1978): 72-75, and "A Yogacara Analysis of the Mind, Based on the Vijnana Section of Vasubandhu's Pancaskandhaprakarana with Gunamati's Commentary "Journal of the International Association of Buddhist Studies vol. 3, no. 2 (1980): 7-20. Galloway's very interesting argument is

unfortunately weakened by the stridency of his tone, which is not justified by the limited scope of his evidence. The ambiguity of the Sanskrit and English terms involved is not exhausted by considering the Pancaskandliaprakarana and the Oxford English Dictionary. For example, Galloway's argument for translating vijnana as "perception" rather than "consciousness" relies very heavily on the PSP and other contexts in which vijnana clearly does mean "perception," but rather ignores other uses of the term vijnana (discussed in the present

essay) for which "consciousness" is at least an arguable translation. Particularly helpful discussions of these terms are given by Th. Stcherbatsky in The Central Conception of Buddhism and the Meaning of the Word "Dharma" (London: Royal Asiatic Society, 1923; reprints Delhi: Motilal Banarsidass, 1970, etc.), and by Shwe Zan Aung in his introduction to the translation of Anuruddha's AhhidhammaUhasangaha in S.Z. Aung and C.A.F. Rhys Davids,

The Compendium of Philosophy, Pali Text Society, Translation Series, 2 (London: Luzac 8c Co., 1910; reprints by Pali Text Society, London).

17. AKK 2.34a-bl: citlarh mano 'tha vijnanam ekartham. Compare AKB and PSP. The equating of these three terms for mind is at least as early as Samyutta-nikaya II: 95 {Kindred Sayings II: 66).

18. AKK 1.16c: vijnanam prativijnaptih. AKB: vujayam vujayamprati vijnaptir upalabdhir vijndnaskandha ity ucyate. Compare the definition in PSP, which Shanti Bhikshu Shastri (1956: 381), Galloway (1980: 10), and Anacker (1984: 71) all take as: [[[vijnanam]]] dlambanavijnaptih.

19. I am indebted to Professor M. David Eckel, Harvard University, for stressing to me the non-negative connotation of -mdtra in such contexts.

20. I have in mind here the first definition of "phenomenon" given in the American Heritage Dictionary: "an occurrence or fact that can be perceived by the senses"—remembering that for Buddhists "mind" (manas) is, of course, one of the senses.

21. I am following the excellent definition of "percept" given in J.P. Chaplin, Dictionary of Psychology (new revised edition, New York: Dell Publishing, 1975), p. 376; "percept: 1. that which is perceived. 2. a perceptual act. The use of the term percept refers to the conscious experience and not to the physical object. Physical objects of perception are referred to as stimuli."

22. This is the view of Vijnanavada held by such writers as Stcherbatsky and Dasgupta, and presented in A.K. Chatterjee, The Yogacara Idealism (2nd ed., Delhi: Motilal Banarsidass, 1975). One classical source for this interpretation is the hostile and wildly inaccurate chapter on Buddhism in the late Vedantin work, SarvadarSana-samgraha. However, those modern writers who interpret Vijnanavada in this way usually consider it very similar to Vedanta. Kochumuttom (1982) strongly criticizes this idealistic interpretation, especially in chapter 6.

23. Kochumuttom (1982: 165-166) objects strenuously to the translation of traidhatukam here as "the three worlds." He declares that: "This translation ignores the fact that the term 'traidhdtuka' is an adjective meaning 'belonging to the three worlds', and that it is not a substantive meaning 'the three worlds.'" He goes on to say that such an adjective must modify an understood noun,

supplying "citta and caittas" as the understood noun. Accordingly (p. 260), he translates the passage: "those belonging to the three worlds are mere representations of consciousness," and the sutra quotation: "those belonging to the three worlds are mere mind"—"those" being cittas and caittas. As evidence for calling Vasubandhu's philosophy "pluralistic realism" this fails to convince. Kochumuttom confuses etymology with meaning and forgets the grammar of the sentence. It is true that traidhdtuka- is etymologically an adjective. It is a secondary derivative in -ka from the compound word tri-dhdtu, and means

"pertaining to the three realms." In the passage, however, we have traidhatukam, which seems in context to be a nominative singular neuter form. Theoretically it could modify cittam, and theoretically the passage could mean that the citta that pertains to the triple world is mere-citta or mere-vijnapti, but this is mere-tautology. In Sanskrit, secondary derivatives are often used as substantives. A good example is the very term caitta, "pertaining to citta." The interpretation that best fits this passage is [idarh] or [[[sarvam]]] traidhatukam— "all this [[[universe]]] that pertains to the three realms."

24. Nairatmya is often translated "selflessness." Although the translation is etymologically correct, the English word "selflessness" connotes unselfish behaviour, which may in fact be encouraged by the philosophical idea of nairatmya, but is not identical with it. [I am indebted to Professor Luis O. Gomez, University of Michigan, for pointing out to me the ambiguity of "selflessness."]

25. VV on VK 10d2: yo bdlair dharmdndm svabhdvo grdhyagrdhakddih parikalpitas. . . .

26. VK 17cd: svapnadrgvi^ayabhdvam ndprabuddho 'vagacchati.

27. VV on VK 17cd: evam vitathavikalpdbhydsavdsandnidrayaprasupto lokah svapna ivabhutam artharh paiyan na prabuddhas tadabhdvam yathdvan ndvagacchati I yadd tu latpratipak$alokottaranin>ikalpajndnaldbhdt prabuddho bhavati tadd mprMhahbdhcL<uMhalaukikajMnasarhmukhibhdvdd vi$aydbhdvarh yathdvad avagacchailti samanam etat I.